Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Mechanisms of joint destruction in rheumatoid arthritis — immune cell–fibroblast–bone interactions

Abstract

Rheumatoid arthritis (RA) is characterized by inflammation and destruction of bone and cartilage in affected joints. Autoimmune responses lead to increased osteoclastic bone resorption and impaired osteoblastic bone formation, the imbalance of which underlies bone loss in RA, which includes bone erosion, periarticular bone loss and systemic osteoporosis. The crucial role of osteoclasts in bone erosion has been demonstrated in basic studies as well as by the clinical efficacy of antibodies targeting RANKL, an important mediator of osteoclastogenesis. Synovial fibroblasts contribute to joint damage by stimulating both pro-inflammatory and tissue-destructive pathways. New technologies, such as single-cell RNA sequencing, have revealed the heterogeneity of synovial fibroblasts and of immune cells including T cells and macrophages. To understand the mechanisms of bone damage in RA, it is important to clarify how the immune system promotes the tissue-destructive properties of synovial fibroblasts and influences bone cells. The interaction between immune cells and fibroblasts underlies the imbalance between regulatory T cells and T helper 17 cells, which in turn exacerbates not only inflammation but also bone destruction, mainly by promoting RANKL expression on synovial fibroblasts. An improved understanding of the immune mechanisms underlying joint damage and the interplay between the immune system, synovial fibroblasts and bone will contribute to the identification of novel therapeutic targets in RA.

Key points

  • T helper 17 (TH17) cells and autoantibodies promote inflammation and tissue destruction in rheumatoid arthritis (RA) by activating other immune cells and synovial fibroblasts, leading to synovitis, bone erosion and cartilage damage.

  • Synovial fibroblasts in RA comprise pro-inflammatory and tissue-destructive subsets, the latter of which express RANKL and matrix metalloproteinases that are involved in osteoclastic bone resorption and cartilage degradation, respectively.

  • Bone lesions in RA are classified as bone erosion, periarticular bone loss and systemic osteoporosis, which are induced by distinct mechanisms.

  • The integration of data from single-cell RNA sequencing and biological studies provides a detailed depiction of the interplay among immune cells, fibroblasts and bone in RA pathogenesis.

  • Therapeutic strategies to modulate pathogenic synovial fibroblasts and to achieve a balance between regulatory T cells and TH17 cells and/or between bone resorption and repair will help achieve structural remission.

Introduction

Rheumatoid arthritis (RA) is an autoimmune disease caused by a complex interaction between genetic and environmental factors, and possibly consists of aetiologically heterogeneous subpopulations1,2. Autoimmunity is the first step in the pathogenesis of RA and a high serum concentration of autoantibodies, such as anti-citrullinated peptide antibodies (ACPAs), is a hallmark of RA, although some patients are seronegative1,2. Immune cells, such as activated T cells, B cells and macrophages, produce pro-inflammatory cytokines and stimulate synovial fibroblasts (tissue-resident mesenchymal cells in the joints) to polarize into pro-inflammatory and tissue-destructive subsets3,4,5. Tissue-destructive synovial fibroblasts express receptor activator of NF-κB ligand (RANKL) and induce osteoclasts to promote bone destruction, and express matrix metalloproteinases (MMPs) that accelerate cartilage degradation (described in detail in Box 1)3,4,5,6. Current therapies for RA are not effective in all patients and are associated with adverse effects such as infection; therefore, it is necessary to develop new therapies, ideally targeting joint-specific pathogenic molecules or cells. Moreover, as the existing therapies for RA are not capable of completely preventing structural damage or restoring joints, the time has come to thoroughly elucidate the immune mechanisms of joint destruction in RA in order to establish the scientific basis for novel therapeutic approaches. To that end, it is important to recognize the importance of the interactions among the immune system, fibroblasts and bone.

Extending the concept of immune cell–fibroblast and immune cell–bone interactions, we herein provide an overview of recent advances in these areas and explore the interactivity within the immune cell–fibroblast–bone triad7. We provide an overview of the relevant immune mechanisms and mechanisms of structural damage (including bone erosion and periarticular and systemic bone loss), summarize the interplay among immune cells, fibroblasts and bone in both active disease and remission, and provide a perspective on current and future strategies for the treatment of structural damage in RA.

Overview of joint structure and biology

The synovium consists of two layers, an intimal lining layer and a sublining layer. In a healthy joint, the lining layer, which consists of resident macrophages and lining synovial fibroblasts, forms a thin barrier at the interface between the sublining and the synovial fluid space. The sublining layer contains endothelial cells and sublining synovial fibroblasts. Synovial fibroblasts produce matrix proteins such as collagen to maintain the structure of the synovium, and lubricate and nourish cartilage surfaces by producing hyaluronic acid and other joint lubricants such as lubricin.

Bone homeostasis is maintained by a balance between bone resorption by osteoclasts and bone formation by osteoblasts3,8. Osteoclasts are the exclusive bone-resorbing cells and differentiate from bone marrow-derived monocyte–macrophage lineage cells. Osteoclasts resorb bone via decalcification and matrix degradation that are mediated by the secretion of hydrogen ions and matrix-degrading enzymes, respectively. Bone-forming osteoblasts, which produce bone matrix proteins and mediate mineralization, are of mesenchymal origin3,8 Some osteoblasts become embedded in the bone matrix where they differentiate into osteocytes, which are thought to orchestrate both osteoclastic bone resorption and osteoblastic bone formation in response to mechanical stress and hormonal cues3,9,10.

RANKL, a TNF family cytokine, and macrophage colony-stimulating factor (M-CSF) are essential molecules for osteoclastogenesis3,8,11. RANKL binds to its receptor RANK and activates downstream signalling pathways such as NF-κB and AP-1, leading to the autoamplification of NFATc1, the master regulator of osteoclastogenesis12. Osteoblasts and osteocytes express RANKL and stimulate osteoclastogenesis necessary for the renewal of bone under physiological conditions3,9,10. Osteoprotegerin, a decoy receptor for RANKL, inhibits the RANK–RANKL interaction13. M-CSF promotes proliferation of osteoclast precursors and activation and survival of osteoclasts3,8. Osteoblast differentiation is stimulated by osteogenic cytokines such as Wnt and bone morphogenetic protein (BMP). Sclerostin, an inhibitor of Wnt signalling, is mainly produced by osteocytes. Mechanical loading decreases sclerostin expression in osteocytes and promotes bone formation, whereas mechanical unloading increases RANKL expression in osteocytes and promotes osteoclastogenesis10,14,15. Activation of the immune system in autoimmune diseases disturbs bone homeostasis by acting directly on bone cells or by stimulating joint-resident cells such as fibroblasts, as discussed below.

Immune mechanisms in RA

In RA, the immune system stimulates synovial fibroblasts to exert inflammatory and tissue-destructive effects and exacerbate RA pathogenesis4. Fibroblasts are the most abundant mesenchymal stromal cells and serve as structural cells that define the architecture of organs; however, attention has increasingly been given to the role of fibroblasts in the pathogenesis of fibrosis, cancer and autoimmunity16,17. New technologies, including single-cell RNA sequencing (scRNA-seq) and mass cytometry, have revealed the heterogeneity of synovial fibroblasts and enabled the identification of functionally and phenotypically distinct pro-inflammatory and tissue-destructive subsets of these cells in RA4,5,18. Herein we describe the autoimmune responses and the subsequent activation of the distinct fibroblast subsets that mediate inflammation and structural damage in RA.

Immune response activation

Genome-wide association studies have revealed strong genetic associations between RA and the HLA regions, indicating the importance of antigen recognition in RA pathogenesis1,2. The citrullination of peptides is mediated by the peptidylarginine deiminases that are upregulated by smoking and periodontitis, suggesting a link between environmental risk factors and RA pathogenesis1,2. The combination of genetic and environmental factors contributes to the breakdown of self-tolerance and the onset of autoimmune arthritis.

The cascade of autoimmune responses starts with T cells recognizing self antigens presented by antigen-presenting cells, such as dendritic cells. CD4+ T cells differentiate into T helper (TH) cells, among which TH17 cells have a critical role in autoimmune inflammation3,7. T follicular helper cells (CXCR5+PD-1hi), which reside in lymph nodes, as well as newly identified T peripheral helper (TPH) cells (CXCR5PD-1hi) that reside in the inflamed synovium, help B cells to produce autoantibodies such as ACPAs and rheumatoid factor19. TH17-derived IL-17 and other cytokines (such as IL-21, IL-22 and TNF) mediate the proliferation of synovial fibroblasts as well as innate immune cells, including neutrophils and macrophages, and induce the expression of pro-inflammatory cytokines (such as TNF, IL-6 and IL-1) and chemokines (such as CCL20 and CCL2) by these cells3,7. TH17 cells also increase the pro-inflammatory activity of autoantibodies via the desialylation of autoantibodies in an IL-21-dependent and IL-22-dependent manner20. A decrease in IgG glycosylation in ACPA+ asymptomatic individuals not only parallels the clinical onset of RA, but is also associated with disease activity in ACPA+ patients with RA, suggesting the importance of TH17 cells and autoantibodies in the immune activation phase of RA20,21. A study published in 2021 showed that tissue-resident memory CD8+ T cells cause arthritis flares22. Immune complexes activate innate immune cells to further upregulate pro-inflammatory cytokines and chemokines3,7. Synovial fibroblasts amplify inflammation in response to these inflammatory mediators as well as mechanical strain23. According to a 2019 scRNA-seq study, IL-6 is produced mainly by synovial fibroblasts, and macrophages are the main producers of IL-1 and TNF; T cells and B cells have also been found to produce TNF in RA18 (Fig. 1). Another scRNA-seq study revealed that CCL13, CCL18 and MMP3 are upregulated in synovial myeloid cell subsets in ACPA RA as compared with ACPA+ RA, which could explain, at least in part, the difference in immune mechanisms between seronegative and seropositive RA24.

Fig. 1: Mechanism of structural damage in rheumatoid arthritis.
figure 1

a | Under physiological conditions, synovial fibroblasts lubricate and nourish the cartilage surface by producing hyaluronic acid and other joint lubricants such as lubricin. b | In rheumatoid arthritis, dendritic cells present autoantigens and produce cytokines that induce the differentiation of naive CD4+ T cells into T helper (TH) cells such as TH17 cells, T follicular helper (TFH) cells and T peripheral helper (TPH) cells. The conversion of FOXP3+ T cells to exFOXP3TH17 cells is promoted by IL-6 produced by synovial fibroblasts. IL-17 activates sublining synovial fibroblasts, macrophages and neutrophils, and induces the expression of pro-inflammatory cytokines and chemokines from these cells. TFH and TPH cells help B cells to produce autoantibodies and immune complexes. TH17 cells upregulate the activity of autoantibodies by regulating antibody glycosylation. Immune complexes activate innate immune cells to further upregulate pro-inflammatory cytokines and chemokines. Pro-inflammatory cytokines upregulate expression of receptor activator of NF-κB ligand (RANKL) in synovial fibroblasts. Pro-inflammatory cytokines and IgG immune complexes directly promote differentiation of osteoclasts through Fc receptors. IL-17 activates lining synovial fibroblasts to express RANKL and matrix metalloproteinases (MMPs), which induce osteoclastogenesis and cartilage degradation. Pro-inflammatory cytokines as well as Wnt inhibitors (such as DKK1 and sclerostin) inhibit osteoblastic bone formation. Regulatory T (Treg) cells inhibit osteoclastogenesis. c | Periarticular bone loss. Plasma cells potently induce periarticular bone loss by expressing RANKL, autoantibodies and pro-inflammatory cytokines. B cells inhibit osteoblastic bone formation via TNF, CCL3 and IL-6. d | Systemic bone loss. Pro-inflammatory cytokines and immune complexes that circulate from inflamed joints as well as glucocorticoid administration promote osteoclastogenesis and inhibit osteoblastogenesis. Reduced mechanical loading induces expression of sclerostin and RANKL by osteocytes, leading to inhibition of osteoblastogenesis and stimulation of osteoclastogenesis, respectively. PRIME cell, pre-inflammatory mesenchymal cell.

Fibroblast activation

Under arthritic conditions, synovial fibroblasts acquire an aggressive (activated, proliferative and invasive) phenotype and are important in the pathogenesis of RA17. This aggressive phenotype of RA synovial fibroblasts is seemingly induced by the inflammatory milieu in the synovium. Analysis of DNA promoter methylation in synovial fibroblasts revealed that the DNA methylation pattern in synovial fibroblasts from patients with very early RA is already different from that in synovial fibroblasts under healthy conditions, suggesting that epigenetic modification is not just a consequence of inflammation, but could also be a cause of disease initiation and progression25.

Metabolic changes, particularly an increase in glycolysis in synovial fibroblasts, are also linked with the aggressive phenotype of synovial fibroblasts in arthritis17. Glycolysis is the source of ATP under hypoxic conditions (which is considered to be a feature of the synovial microenvironment in RA), although this metabolic pathway is less efficient than oxidative phosphorylation. The expression of hypoxia-inducible factor 1α (HIF1α), an inducer of glycolysis, is linked to the aggressive features of synovial fibroblasts26. In addition, activation of intracellular complement C3 and C3a receptor on synovial fibroblasts after repeated inflammatory challenges induces metabolic reprogramming that promotes the activation of these cells and the priming of synovial tissue for inflammation27. Although these findings suggest that tissue priming occurs independently of adaptive immunity, the efficacy of abatacept (CTLA4-immunoglobulin (Ig), which prevents T cell costimulation) in RA suggests that activation of T cells might be necessary even in the chronic phase of the disease and that the reprogramming of synovial fibroblasts by repetitive priming might require T cell help. Considering that synovial fibroblasts are functionally heterogeneous, as we discuss below, it will be important to clarify how the polarization of fibroblasts into distinct synovial fibroblast subsets with inflammatory or tissue-destructive properties is determined.

Fibroblast heterogeneity

The phenotypic and functional heterogeneity of synovial fibroblasts is attracting increasing attention. Cadherin-11 is one of the most frequently investigated synovial fibroblast surface markers and has been linked to synovial fibroblast activation and inflammation. Cadherin-11 is expressed mainly in lining synovial fibroblasts, although it is also expressed in certain sublining synovial fibroblasts28. Cadherin-11 activates mitogen-activated protein kinases and NF-κB, which induce synovial fibroblasts to secrete pro-inflammatory cytokines such as IL-6 (ref.29). Cdh11-deficient mice have a hypoplastic synovial lining and exhibit reduced inflammation and cartilage erosion in the serum transfer-induced arthritis model28.

scRNA-seq studies published in the past few years have identified synovial fibroblast subsets with inflammatory and tissue-destructive properties both in mouse and human arthritis4,5,18. Application of scRNA-seq to RA synovial fibroblasts characterized by the expression of podoplanin (PDPN) identified three major subpopulations: CD34THY1, CD34THY1+ and CD34+ cells. Among these subpopulations, the CD34THY1+ cells, which localize to the perivascular zone in the inflamed synovium, are highly proliferative and secrete pro-inflammatory cytokines5. Subsequently, the integration of single-cell transcriptomics and mass cytometry revealed that sublining CD34CD90+HLA-DRhi synovial fibroblasts are expanded in the RA synovium and constitute a major source of IL-6 and CXCL12 (also known as stromal cell-derived factor 1)18. Moreover, a separate single-cell transcriptional analysis found two distinct subpopulations of fibroblast activation protein-α (FAPα)-expressing fibroblasts: inflammatory FAPα+THY1+ cells, located in the sublining, and tissue-destructive FAPα+THY1 cells, located in the synovial lining layer4. Notably, in mice with serum transfer-induced arthritis, adoptive transfer of the FAPα+THY1+ synovial fibroblasts subset induces inflammation, whereas transfer of the FAPα+THY1 subset induces joint destruction4.

What is the mechanism underlying the generation of inflammatory or tissue-destructive synovial fibroblasts? Activation of Notch signalling induces the production of pro-inflammatory cytokines from RA synovial fibroblasts. In a 2020 study, scRNA-seq revealed that the interaction of Notch ligands expressed by endothelial cells with the receptor NOTCH3 on sublining synovial fibroblasts activates Notch signalling and drives the polarization of THY1+ inflammatory synovial fibroblasts30. The genetic deletion of Notch3 or the blockade of NOTCH3 signalling attenuates the inflammation and bone destruction in serum transfer-induced arthritis30. TNF signalling is also important for the polarization of inflammatory synovial fibroblasts, as the activation of TNF signalling exclusively in fibroblasts induces arthritis in mice31,32. Moreover, the IL-6 family cytokine leukaemia inhibitory factor (LIF), which is upregulated in synovial fibroblasts under arthritic conditions, acts in an autocrine manner via LIF receptor to promote STAT4 activation, which increases production of important pro-inflammatory factors, including IL-6, leading to the polarization of inflammatory synovial fibroblasts33.

Although inflammation is known to also promote tissue-destructive synovial fibroblasts, the stimulatory factors, intracellular signal transduction pathways and transcriptional machinery underlying the generation of these cells remain to be elucidated. The most important feature of tissue-destructive synovial fibroblasts is the production of RANKL, but they also produce tissue-destructive factors such as MMPs, thus orchestrating a variety of mechanisms that are involved in bone and cartilage destruction4,5. Understanding how RANKL production is induced in synovial fibroblasts will be helpful for determining the main mechanisms driving these cells.

Whether these distinct populations of synovial fibroblasts are subsets with fixed phenotypes or whether they have phenotypic plasticity also remains unclear. Fate mapping and/or tracing experiments in vivo are necessary to address this question.

Structural damage in RA

Structural abnormalities in RA involve bone erosion as well as periarticular and systemic bone loss. Osteoclasts were first observed at the interface between the inflamed synovium and bone in the 1980s, although it was initially unclear why autoimmunity increased osteoclastic bone resorption34. The generation of osteoclasts by culturing synovial cells from patients with RA indicated that both osteoclast precursor cells and osteoclastogenesis-supporting fibroblasts were present in the RA synovium35. Moreover, synovial fibroblasts were found to express a high level of RANKL, a molecule that is important for osteoclast differentiation36,37. This finding suggested that the immune system induced osteoclastogenesis mainly by stimulating synovial fibroblasts, rather than by acting directly on osteoclast precursor cells, and indicated the importance of tissue-destructive fibroblasts in arthritis.

Bone erosion

How does inflammation cause bone erosion?

Protein-degrading enzymes were originally thought to sufficiently explain the bone erosion that occurs in arthritis; however, the importance of osteoclast-mediated bone resorption was suggested by the essential role of these cells in bone resorption in physiological bone remodelling and by the observation that osteoclasts are numerous at the synovium–bone interface in RA. Indeed, mice lacking osteoclasts are protected from bone erosion in TNF-transgenic (TNF-Tg) and serum transfer-induced models of arthritis, indicating the primacy of osteoclasts in bone erosion in RA38,39. The increased osteoclastogenesis observed in arthritis is attributed mainly to the increased expression of RANKL, as RANKL-deficient mice exhibit much less severe bone damage in arthritis than their matched littermates38.

Synovial fibroblasts, as well as T cells and B cells, express RANKL when activated36,37,40,41,42,43. A long-standing question was which cell type induces osteoclastogenesis in the inflamed synovium. In the collagen-induced arthritis (CIA) model, mice lacking RANKL expression in synovial fibroblasts, but not those with T cell-specific or B cell-specific RANKL deficiency, are protected from bone erosion, indicating that synovial fibroblasts are the primary RANKL-producing cells in the synovium in autoimmune arthritis6,44. These findings lend support to the concept of ‘tissue-destructive fibroblasts’6,44.

Pro-inflammatory cytokines such as TNF, IL-6 and IL-1, which are abundant in the synovium and synovial fluid in RA, promote RANKL expression by synovial fibroblasts. In addition, the immune system enhances osteoclastogenesis by activating osteoclast precursor cells in several ways. Pro-inflammatory cytokines act directly on osteoclast precursor cells to enhance signalling downstream of RANK as well as to increase the expression of co-stimulatory receptors for RANK45,46,47. Moreover, IgG immune complexes directly promote osteoclast differentiation through Fc receptors48 (Fig. 1). Antibodies have been shown to stimulate the production of IL-8 and TNF as well as to promote osteoclastogenesis49,50. Serum concentration of soluble RANKL is associated with disease activity in RA51. However, studies in mice selectively lacking soluble RANKL have demonstrated that soluble RANKL does not contribute to physiological bone remodelling or to a model of postmenopausal osteoporosis52. The question of how soluble and membrane-bound RANKL contribute to bone destruction in RA will be an interesting issue to explore in future investigations.

How are T cells involved in bone erosion in RA?

Activated T cells express not only RANKL but also effector cytokines with either stimulatory or inhibitory effects on osteoclastogenesis40,41. TH1 and TH2 cells inhibit osteoclastogenesis through the expression of IFNγ and IL-4, respectively. TH17 cells comprise an exclusively osteoclastogenic T cell subset that induces RANKL expression on synovial fibroblasts via production of IL-17, IL-21 and IL-22 (refs7,53,54). Pro-inflammatory cytokines from IL-17-activated innate immune cells further induce RANKL expression on synovial fibroblasts and act on osteoclast precursor cells to activate the downstream pathway of RANK3,7. Desialylation of IgG by TH17 cells also increases the osteoclastogenic capacity of immune complexes20,55 (Fig. 1).

Regulatory T (Treg) cells are pivotal in the suppression of immune responses and the prevention of autoimmunity56. FOXP3 functions as the master transcription factor for the development and function of Treg cells57. Humans and mice deficient in FOXP3 exhibit lethal autoimmune diseases owing to a lack of Treg cells56. Evidence from a combination of genome-wide association studies with epigenetic analysis or expression quantitative trait locus analysis suggests that Treg cells are strongly associated with RA58,59. Treg cells express high amounts of CTLA4 and IL-10, which act on osteoclast precursor cells and inhibit osteoclastogenesis60,61. Thus, Treg cells not only regulate inflammation in arthritis, but also directly inhibit bone destruction. Considering that Treg cells recognize autoantigens, the loss of FOXP3 expression in Treg cells could exacerbate autoimmune arthritis. Indeed, some FOXP3+ T cells are plastic, and they lose FOXP3 expression and convert to arthritogenic TH17 cells, which exacerbate autoimmune arthritis62. These TH17 cells of FOXP3+ T cell origin (called exFOXP3TH17 cells) induce osteoclastogenesis more efficiently than TH17 cells derived from naive CD4+ T cells. exFOXP3TH17 cells produce copious amounts of effector molecules such as IL-17, CCR6, CCL20 and RANKL, and induce RANKL expression on synovial fibroblasts62,63. Thus, exFOXP3TH17 cells are the osteoclastogenic T cell subset that most potently induces bone erosion.

How is osteoblastic bone formation impaired in RA?

In arthritic joints, osteoblast function is impaired, especially in the bone adjacent to the inflammatory synovium64. Pro-inflammatory cytokines inhibit osteoblastic bone formation via several mechanisms8.

TNF suppresses osteoblast differentiation by suppressing expression of the transcription factor RUNX2 and by upregulating inhibitors of Wnt signalling65. In the RA synovium, endogenous Wnt inhibitors such as Dickkopf-related protein 1 (DKK1), sclerostin, and Frizzled-related proteins are upregulated8; DKK1 and sclerostin inhibit Wnt signalling by binding to LRP5 and LRP6, which are receptors for canonical Wnt signalling, whereas Frizzled-related proteins bind directly to Wnt ligands. DKK1 is produced mainly by TNF-stimulated synovial fibroblasts. In patients with RA, variants of Dkk1 variants are associated with severe joint destruction66. Sclerostin is produced not only by osteocytes, but also by TNF-stimulated synovial fibroblasts67. TNF also inhibits BMP signalling by inducing the production of BMP3, an endogenous BMP inhibitor, by osteoblasts68.

IL-1 inhibits osteoblast differentiation, whereas IL-6 promotes osteoblast differentiation and bone formation under certain conditions69,70. Administration of IL-6 stimulates bone formation in mice via trans-signalling, but treatment with an antibody targeting IL-6 receptor has no negative effects on bone mass71, possibly because IL-6 blockade has a positive influence on bone mass by suppressing inflammation mediated by IL-6 and other cytokines such as TNF as well as by inhibiting osteoclastogenic bone resorption.

The role of IL-17 in osteoblast differentiation is controversial72. Spondyloarthritis, including psoriatic arthritis and ankylosing spondylitis, is characterized by inflammation and new bone formation in entheses, both of which are known to be reduced by IL-17 blockade73,74. Conversely, IL-17 deficiency promotes bone formation without influencing inflammation and bone erosion in the joints of mice with serum transfer-induced arthritis75. IL-17 inhibits calvarial osteoblast differentiation in vitro by inducing osteoblast expression of secreted Frizzled-related protein 1 (ref.75). Reportedly, IL-17 from γδ T cells promotes bone formation and facilitates bone fracture healing76. It is likely that the effect of IL-17 on bone formation is dependent on the type of osteoblast precursors and the microenvironments of the affected sites72.

Pro-inflammatory cytokines can modulate osteoblasts by regulating the expression of semaphorins, which are known to act as osteoimmune factors. Under physiological conditions, semaphorin 3A (Sema3A) promotes osteoblastic bone formation, whereas Sema4D suppresses it77,78. TNF and IL-6 promote the expression of ADAMTS-4, which cleaves cell-surface Sema4D to generate soluble Sema4D79. In RA, serum levels of Sema3A are negatively correlated with disease activity, whereas levels of Sema4D are positively correlated with disease activity, suggesting that altered expression of Sema3A and Sema4D under inflammatory conditions might lead to impaired osteoblastic bone formation in arthritis79,80.

Whereas immune regulation of osteoblasts has been extensively studied, the regulation of the immune response by osteoblasts in RA remains largely unclear. Within the past few years, it has been reported that osteoblasts produce PLEKHO1, a negative regulator of osteoblastic bone formation that promotes the production of pro-inflammatory cytokines in osteoblasts81. Osteoblast-specific inhibition of PLEKHO1 ameliorated inflammation and promoted bone formation in a mouse model of arthritis81. Elucidation of osteoblast–immune interactions will contribute to the development of future therapeutic strategies for restoring joint structure.

Periarticular bone loss

Periarticular bone loss in RA is an osteoporotic lesion observed in the bone adjacent to joints82. Periarticular bone loss has been attributed to joint inflammation but the precise mechanism remains unclear. Periarticular bone loss is already present in the pre-RA state in ACPA+ individuals83. Indeed, ACPAs were shown to induce osteoclastogenesis and periarticular bone loss in a model of antigen-induced arthritis84.

Plasma cells are specialized B lineage cells that produce antibodies and reside primarily in the bone marrow84,85. Under arthritic conditions, plasma cells accumulate in the bone marrow proximal to inflamed joints and express RANKL at high levels44. Plasma cells efficiently induce osteoclastogenesis in vitro in a RANKL-dependent manner44. The ability of plasma cells to induce osteoclasts is much greater than that of B cells. Mice deficient in RANKL in the B cell lineage are protected from periarticular bone loss, although not from bone erosion44. Plasma cells also produce antibodies and pro-inflammatory cytokines such as IL-6 (refs85,86). Taken together, these findings show that bone marrow plasma cells promote osteoclastogenesis and thereby periarticular bone loss by expressing RANKL, pro-inflammatory cytokines and autoantibodies (Fig. 1). The contribution of RANKL derived from osteoblasts or osteocytes to periarticular bone loss needs to be further explored.

Impaired osteoblastic bone formation results in periarticular bone loss because bone mass that is removed by osteoclasts is not fully replaced. Subchondral bone marrow B cells inhibit osteoblast function by expressing CCL3 and TNF87. Expression of sclerostin is increased and expression of RUNX2 is decreased in periarticular bone before the onset of adjuvant-induced arthritis, suggesting that osteocytes may contribute to the impaired bone formation in periarticular bone loss88.

The bone marrow in proximity to the inflamed joints where periarticular bone loss occurs might be called ‘draining’ bone marrow, by analogy with draining lymph nodes, which are essential for the initiation and progression of immune responses at inflammatory sites. The immune dysregulation that elicits periarticular bone loss can also trigger joint damage. Patients with RA develop cortical microchannels in the bare area of the joint, where bone is not covered by articular cartilage within the joint capsule at an early stage of the disease, suggesting that the microchannels might facilitate communication between the draining bone marrow and the synovium, leading to the clinical onset and progression of RA89.

Systemic bone loss

Systemic bone loss in RA is observed as widespread osteoporosis, typically in the vertebrae and femurs, which have an increased risk of fracture compared with those in individuals without RA8,90,91. In general, osteoporosis is caused by diverse factors such as ageing, menopause and vitamin D deficiency11. The incidence of osteoporosis in patients with RA is approximately two times higher than in the general population92. This increased incidence is possibly attributable to RA-specific factors such as activation of the immune system, glucocorticoid treatment and loss of mobility.

Inflammatory factors such as immune complexes and pro-inflammatory cytokines contribute to systemic bone loss, but additional factors, such as glucocorticoid treatment and immobility, are also important. Glucocorticoid treatment inhibits osteoblast differentiation, induces osteoblast apoptosis and activates osteoclast differentiation90,91. Pro-inflammatory cytokines, which are produced in the inflamed joints, can induce systemic bone loss by activating osteoclastic bone resorption and inhibiting osteoblastic bone formation in bones at distant sites90,91. Immune complexes circulating in the bloodstream stimulate osteoclastic bone resorption systemically48,49,50. As for the source of RANKL in systemic bone loss in RA, plasma cell-derived RANKL is reportedly dispensable, suggesting the importance of RANKL derived from osteoblasts and osteocytes44. As mentioned above, a reduction in mechanical loading increases RANKL expression in osteocytes, leading to enhanced osteoclastic bone resorption10,14; in addition, mechanical unloading increases sclerostin expression in osteocytes, thereby decreasing osteoblastic bone formation15. These mechanisms could explain the immobility-related bone loss in RA (Fig. 1).

Immune cell–fibroblast–bone interactions

Active disease

As tissue-destructive synovial fibroblasts are important mediators of joint destruction, it is important to clarify the immune mechanism(s) that promotes the expression of genes encoding tissue-destructive molecules (such as RANKL) in synovial fibroblasts. Early in vitro experiments showed that pro-inflammatory immune mediators, such as IL-17, IL-6, IL-1, TNF, oncostatin-M and prostaglandin E2, or a combination thereof, increase RANKL expression in synovial fibroblasts93. Subsequent studies have provided more information regarding the interactions among immune cells, fibroblasts and bone at the cellular and molecular level.

CD40L on activated T cells was shown in the early studies to induce the proliferation and expression of adhesion molecules, such as intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion molecule 1 (VCAM1), as well as pro-inflammatory cytokines such as IL-6 by synovial fibroblasts94. ICAM1 and VCAM1 expressed by synovial fibroblasts further promote the interaction between T cells and synovial fibroblasts and CD4+ T cell activation95,96. T cells and synovial fibroblasts activate each other by producing pro-inflammatory cytokines and chemokines. For instance, T cells induce the expression of pro-inflammatory cytokines such as IL-6 and IL-8 by synovial fibroblasts97, and IL-7 from synovial fibroblasts promotes a homeostatic proliferation of T cells under arthritic conditions98. CX3CL1 (also known as fractalkine) expressed by synovial fibroblasts promotes the recruitment and activation of CX3CR1-expressing T cells99. Considering the high level of expression of CX3CR1 on TPH cells, it is possible that TPH cells and synovial fibroblasts might interact with each other19. CXCL10 expressed by synovial fibroblasts enhances the recruitment of CXCR3-expressing T cells, including TH1 cells, and CCL20 from synovial fibroblasts promotes the recruitment of CCR6-expressing TH17 cells to inflammatory joints100,101. Of note, IL-6 produced by synovial fibroblasts promotes the differentiation of TH17 cells, and IL-17 together with IL-6 further enhances IL-6 production by synovial fibroblasts, forming a positive feedback loop referred to as the IL-6 amplifier102. IL-6 is also important for the pathogenic conversion of FOXP3+ T cells into exFOXP3TH17 cells. IL-6 produced by synovial fibroblasts determines the fate of plastic FOXP3+ T cells and thereby promotes an imbalance between Treg cells and TH17 cells62. In line with this concept, in a clinical study an increase in peripheral blood Treg cells was correlated with clinical response in patients with RA treated with IL-6 blockade103. TH17 cells promote the production of not only RANKL, but also many pro-inflammatory mediators including granulocyte–macrophage colony-stimulating factor (GM-CSF) and cadherin-11, by synovial fibroblasts104,105. Thus, synovial fibroblasts and T cells activate each other to amplify inflammation and bone erosion in arthritis7. RA synovial fibroblasts can reportedly act as antigen-presenting cells by internalizing neutrophil extracellular traps that contain citrullinated peptides and presenting them as antigens to T cells106,107. Considering that the proportion of HLAhiTHY1+ sublining synovial fibroblasts is increased in the RA synovium18, it is possible that these synovial fibroblasts may amplify inflammation as antigen-presenting cells.

RA synovial fibroblasts promote the survival of B cells by producing VCAM1 and CXCL12 (ref.108). In addition, synovial fibroblasts stimulated with Toll-like receptor 3 ligands promote the differentiation and activation of B cells by producing TNF ligand superfamily members 13 and 13B (also known as APRIL and BAFF, respectively) as well as IL-6, thereby promoting the production of antibodies, including ACPAs109. In turn, immune complexes exacerbate inflammation and osteoclastogenesis3,48,49,50. In a study published in 2020, longitudinal genomic analysis of blood from patients with RA revealed that just before a flare of RA the activation of B cells is followed by an expansion of circulating CD3CD45PDPN+ pre-inflammatory mesenchymal (PRIME) cells, which resemble pathogenic sublining inflammatory synovial fibroblasts110. Levels of PRIME cells then decrease in the blood just after symptom onset and are considered to expand in inflammatory synovium, suggesting that they might migrate from the blood to the synovium. This suggests a possible contribution of an interaction between B cells and synovial fibroblasts to the recurrence of RA symptoms110. It would be interesting to clarify where PRIME cells come from and how they are activated, in order to better understand the mechanism of RA flare.

As mentioned above, RA synovial fibroblasts promote the recruitment of monocytes into joints by secreting chemokines such as CCL2 and CXCL10. Mechanical strain-mediated exacerbation of arthritis is dependent on these chemokines, suggesting that the interaction between monocytes and mechanosensitive synovial fibroblasts might underlie the joint specificity of RA pathogenesis23. Obviously, the recruitment of RANK+ monocytes and macrophages enhances osteoclastic bone erosion through their interaction with RANKL+ synovial fibroblasts. A study published in 2019 identified a CX3CR1hiLy6CintF4/80+I-A/I-E+ macrophage subset, termed arthritis-associated osteoclastogenic macrophages (AtoMs), as the pathogenic osteoclast precursor population in arthritis111. Because CX3CL1 is highly produced by endothelial cells and synovial fibroblasts, the CX3CR1–CX3CL1 axis is important for the migration of pathogenic osteoclast precursors to the inflamed synovium. In addition, prostaglandins produced by RA synovial fibroblasts drive the polarization of proheparin-binding EGF-like growth factor (HBEGF)-expressing macrophages112; in turn, these HBEGF+ macrophages promote synovial fibroblast invasiveness.

Aside from their interaction with immune cells, synovial fibroblasts interact with mesenchymal cells such as endothelial cells and osteoblasts in arthritis. As mentioned above, the differentiation of inflammatory synovial fibroblasts requires Notch signalling triggered by endothelial cells30. Moreover, RA synovial fibroblasts suppress osteoblastic bone formation via the expression of DKK1 (ref.113). Although the effects of immune cells on synovial fibroblasts and bone cells have been extensively studied, the influence of bone cells on synovial fibroblasts and/or immune cells has not been fully clarified and needs to be explored in further studies (Fig. 2).

Fig. 2: Immune cell–fibroblast–bone interplay in rheumatoid arthritis.
figure 2

Overview of the interactions among immune cells, fibroblasts and bone in bone erosion and remission. a | Left: Immune cells and pro-inflammatory synovial fibroblasts interact and activate each other via the production of pro-inflammatory cytokines and chemokines. Tissue-destructive synovial fibroblasts induce osteoclastogenesis by expressing receptor activator of NF-κB ligand (RANKL) and inhibit osteoblastogenesis by expressing Wnt inhibitors. Immune cells can directly promote osteoclastogenesis and inhibit osteoblastogenesis via pro-inflammatory cytokines and immune complexes. Bone cells reportedly activate immune cells and synovial fibroblasts via pro-inflammatory cytokines, but the effects of bone cells on the other cells are not well investigated. Right: Immune cell subsets including regulatory T (Treg) cells, MerTK+CD206+ macrophages and IL-9-producing innate lymphoid cells (ILCs) are thought to be involved in structural remission. Regulatory subsets of synovial fibroblasts and bone cells remain to be identified. b | Details of the interplay within the immune cell–fibroblast–bone interplay. Integration of findings from single-cell RNA sequencing and biological studies enables us to depict the interactions among immune cells, fibroblasts and bone at the cellular and molecular levels. ACPA, anti-citrullinated peptide antibody; AtoM, arthritis-associated osteoclastogenic macrophage; GM-CSF, granulocyte–macrophage colony-stimulating factor; MMP, matrix metalloproteinase; RA, rheumatoid arthritis; SF, synovial fibroblast; TFH cell, T follicular helper cell; TH17 cell, T helper 17 cell; TPH cell, T peripheral helper cell.

Remission

In patients with clinical remission, joint structural damage typically does not proceed. However, certain patients can be in a state of clinical remission in terms of signs and symptoms of inflammatory joint disease, but can have subclinical synovitis detectable by ultrasonography, which is associated with a high risk of bone erosion114. Thus, complete resolution of joint inflammation could be important for the achievement of structural remission with no further bone loss. Conversely, however, treatment with TNF blockade has been reported to suppress joint destruction even in patients who experience no or little clinical improvement, suggesting that joint destruction sometimes proceeds independently of inflammation115. Identifying the specific mechanism by which joint destruction occurs would help in the development of a method for establishing structural remission.

As mentioned above, Treg cells have an important role in immune suppression56. Treg cells also regulate bone homeostasis by decreasing osteoclastogenesis and increasing osteoblastic bone formation via CTLA4, IL-10 and transforming growth factor-β61,116,117,118. Under physiological conditions, the adoptive transfer of Treg cells suppresses osteoclastogenesis and increases bone volume119. Moreover, Foxp3-Tg mice are protected from bone erosion in a model of TNF-Tg arthritis120. Impaired Treg cell function or the emergence of exFOXP3TH17 cells has a pathological role in both autoimmune inflammation and bone resorption62. IL-9-deficient mice exhibit delayed resolution of antigen-induced arthritis as well as impaired activation of Treg cells and impaired proliferation of type 2 innate lymphoid cells (ILC2s)121. Administration of IL-9 in a serum transfer-induced arthritis model led to resolution of inflammation and joint destruction. IL-9 induces proliferation of ILC2s, which activate Treg cells in a manner dependent on inducible T cell costimulator (ICOS) and TNF receptor superfamily member 18 (GITR), supporting the importance of the interaction between Treg cells and ILC2s in the resolution phase of arthritis121. These findings suggest that controlling Treg cells would be a powerful approach to achieving structural remission.

As for anti-inflammatory subsets of macrophages, a 2019 study identified a population of CX3CR1+ resident synovial macrophages that restrict inflammatory reactions by providing a tight-junction-mediated protective barrier for the joint122. Determining how CX3CR1+ resident synovial macrophages and lining synovial fibroblasts interact with each other under physiological and arthritic conditions would be of interest. Another protective macrophage subset was identified by scRNA-seq analysis of synovial tissue macrophages from patients with early active RA, treatment-refractory active RA or treatment-sensitive RA in remission. These MerTK+CD206+ macrophages, which are enriched in the synovium of patients with RA in a state of sustained remission, resolve inflammation and induce a ‘repair’ phenotype of synovial fibroblasts via the production of lipid mediators123. Having a low proportion of MerTK+ macrophages is associated with an increased risk of disease flare after treatment cessation. This approach using scRNA-seq analysis will be important for the further identification of regulatory cell subsets that are necessary for inhibiting structural damage in RA (Fig. 2).

Treating structural damage

RA treatment is generally focused on immunomodulatory therapy to address joint inflammation. DMARDs, which are widely used for RA treatment, are now classified into three groups: conventional synthetic DMARDs, such as methotrexate; biologic DMARDs (bDMARDs), including anti-TNF, anti-IL-6 and anti-CD20 agents and CTLA4-Ig; and targeted synthetic DMARDs (tsDMARDs), such as Janus kinase (JAK) inhibitors8,90.

bDMARDs and JAK inhibitors are effective in preventing both joint inflammation and bone erosion. However, the effects of DMARDs on periarticular and systemic bone loss in RA are limited or have been poorly investigated90,91. The anti-RANKL antibody denosumab decreases bone erosion in RA124,125,126 and is approved in Japan for the treatment of bone erosion of RA. Anti-RANKL antibodies and bisphosphonates are effective for treating systemic osteoporosis and in reducing the risk of fracture in patients with RA, although they do not exert effects on inflammation or cartilage degradation126.

Current therapies effectively inhibit the progression of bone destruction in the majority of patients with RA, but in certain cases the response to even multiple DMARDs is inadequate, and it is thus difficult to completely prevent bone destruction. Therefore, we urgently need to fully elucidate the cellular and molecular network underlying structural damage in RA. In this section we provide an overview of the effects of bDMARDs and JAK inhibitors on joint structure and discuss novel candidates for future therapies to treat structural damage.

Biologic DMARDs

TNF inhibitors, IL-6 inhibitors and CTLA4-Ig are widely used bDMARDs. These bDMARDs effectively inhibit inflammation, bone erosion and cartilage degradation by suppressing the local inflammation mediated by synovial fibroblasts and macrophages as well as by inhibiting RANKL induction and RANK signalling pathways3,8,11. One might consider that structural protection is achieved mainly by the inhibition of inflammation, but inflammation-independent effects of TNF blockade on bone could exist, given that bone erosion is ameliorated in certain cases without any improvement in inflammation115. CTLA4-Ig inhibits inflammation by binding to CD80/CD86 on dendritic cells and suppressing T cell activation, and directly inhibits osteoclast differentiation by inducing apoptosis of osteoclast precursor cells in a CD80/CD86-dependent manner118,127.

In patients with established RA, inhibitors of IL-17A or IL-23 are less effective than other bDMARDs128, even though it has been well documented that TH17 cells are critical to arthritis pathogenesis (both inflammation and bone damage) and IL-17-deficient mice have been shown to be resistant to inflammation and bone destruction in various mouse arthritis models7. This reduced efficacy could be attributable to the heterogeneity of RA, with TH17-dependent mouse models reflecting the disease of only some patients with RA. Alternatively, it is possible that TH17 cells are important only for the early phase of RA rather than the established phase20. In line with this idea, an anti-IL-17 antibody was shown to be effective in the early phase rather than the late phase in a TH17-dependent mouse model129. Notably, dual blockade of IL-17A and IL-17F with bimekizumab produced a favourable result in a clinical trial involving patients with RA who had an inadequate response to a TNF inhibitor, suggesting that IL-17 blockade remains a promising approach if IL-17 family cytokines are fully blocked130. The current therapies for RA target pro-inflammatory cytokines mainly produced by innate immune cells and synovial fibroblasts, and thus target bystander pathways rather than antigen-specific pathways. Understanding the autoimmune mechanisms in RA could lead to the establishment of new therapeutic strategies in the future.

JAK inhibitors

JAKs (including JAK1, JAK2, JAK3 and TYK2) are widely expressed in immune and stromal cells in joints and are involved in various cellular responses initiated by cytokines. JAKs phosphorylate signal transducer and activator of transcription proteins (STATs), which then translocate to the nucleus to regulate gene transcription. JAK inhibitors suppress joint inflammation to an extent similar to the suppression produced by bDMARDs8,90. Which type of cells and signalling pathways are the specific targets of JAK inhibitors in vivo remains unclear, as most immune and bone cells are influenced by cytokine signalling that utilizes JAK–STAT pathways. In vitro studies have shown that JAK inhibitors suppress the production of IFNγ and IL-17 as well as the proliferation of CD4+ T cells131. JAK inhibitors also inhibit the expression of CD80/CD86 as well as pro-inflammatory cytokines such as IL-6 and TNF in dendritic cells132.

Certain JAK inhibitors inhibit bone erosion in patients with RA more potently than TNF blockade, suggesting that some JAK inhibitors might protect against structural damage through distinct mechanisms133,134. Although JAK inhibitors have no direct effects on osteoclast precursors, they suppress osteoclastogenesis by inhibiting the expression of RANKL on osteoclast-supporting mesenchymal cells135,136. In vitro, JAK inhibitors promote osteoblastogenesis in part by increasing the expression of anabolic proteins such as Wnt1 and β-catenin in osteoblasts135. In addition, it seems that JAK inhibition reverses bone erosion in RA by promoting the restoration of bone mass135. Further studies are necessary to elucidate whether and how JAK inhibition regulates structural damage in vivo.

Emerging therapeutic targets

To reinstate the joint structure, it is necessary to determine how to enhance the osteoblastic bone formation under arthritic conditions137. Blockade of DKK1 and sclerostin, both of which inhibit Wnt signalling, have been shown to exert considerable effects on bone formation in arthritis67,113,138,139. Treatment with an anti-DKK1 antibody prevents bone damage and leads to bone formation in TNF-Tg arthritis113. An anti-sclerostin antibody blocks periarticular and systemic bone loss in TNF-Tg arthritis and CIA, although it does not affect joint inflammation138,139. Thus, blockade with Wnt inhibitors could serve as a treatment for reinstating the joint structure in RA. However, TNF-Tg arthritis is exacerbated in sclerostin-deficient mice, consistent with a role for sclerostin in attenuating TNF signalling67. Therefore, treatment with a sclerostin inhibitor needs to be carefully conducted with much attention given to potential adverse effects.

There are other candidate molecules that can increase bone formation under arthritic conditions. Sema3A exerts bone anabolic effects by increasing osteoblastic formation and inhibiting osteoclastogenesis77. Sema3A has also been identified as an immunosuppressive factor, and the administration of Sema3A not only ameliorated inflammation and bone erosion but also increased bone formation in a serum transfer-induced model of arthritis140. In addition, Sema4D is known to be an osteoimmune molecule that promotes inflammation and inhibits osteoblastic bone formation78,79. Administration of an anti-Sema4D antibody inhibits inflammation and bone erosion in CIA79. Moreover, Notch signalling is important for the polarization of inflammatory synovial fibroblasts as well as the inhibition of osteoblastic bone formation, and studies in mice have shown that Notch inhibition increases bone volume by enhancing osteoblastic bone formation141. Furthermore, the CX3CR1–CX3CL1 axis and the CXCL10–CXCR3 axis promote not only the migration of T cells and macrophages, but also the activation of synovial fibroblasts99,100,142,143. Blockade of CX3CL1 and CXCL10 as well as blockade of GM-CSF and M-CSF have been shown to inhibit inflammation and bone erosion in mouse models of arthritis and are now being investigated in a clinical trial100,144,145,146,147,148. Thus, therapeutic strategies targeting molecules involved in the immune cell–fibroblast–bone triad will be beneficial for both inhibition of inflammation and restoration of the joint structure in RA (Table 1).

Table 1 The effect of molecules on immune cells, synovial fibroblasts and bone

Current therapies are not universally effective in all patients because RA pathogenesis is heterogeneous. The lack of predictors of treatment success presents a problem in relation to the choice of the best therapy for each individual patient. It is thus important to establish therapeutic strategies that are based on patient subpopulations. It remains to be seen whether the analysis of cells and transcriptomes in synovial tissue samples can appropriately delineate disease subsets and provide better targets for therapeutics. Alternatively, targeting the pathogenic synovial fibroblasts common to all patients with RA is an attractive therapeutic strategy. In terms of targeting surface molecules expressed on synovial fibroblasts, administration of antibodies directed against cadherin-11 and depletion of FAPα+ synovial fibroblasts have been shown to be effective in mouse models of RA4,28. An anti-cadherin-11 antibody has been shown to be ineffective in clinical studies, while therapies targeting FAPα are still under clinical investigation149,150. At present, there are no therapies targeting synovial fibroblasts that inhibit both bone erosion and cartilage degradation. Further identification of the surface or intracellular proteins specifically expressed by tissue-destructive synovial fibroblasts will contribute to the development of agents designed to treat structural damage.

Conclusions

Synovial fibroblasts play an important part in exacerbating inflammation and joint damage in RA by enhancing osteoclastogenic bone erosion and cartilage destruction as well as inhibiting osteoblastic bone formation. Structural remission will be achieved by completely inhibiting inflammation in addition to inhibiting the specific pathways related to joint damage. To this end, it will be important to further elucidate the mechanisms of immune cell–fibroblast–bone interplay and their effects on joint destruction and the generation of pathogenic synovial fibroblasts. TH17 cells has been considered to play a key role in autoimmune inflammation and bone destruction3,12. The activation of immune cells, including induction of a Treg cell–TH17 cell imbalance, is important for the arthritogenic effects of synovial fibroblasts. Treg cells not only inhibit inflammation, but also inhibit osteoclastogenic bone resorption and promote osteoblastic bone formation. It would be interesting to investigate whether Treg cells modulate joint damage by regulating the function or polarization of synovial fibroblasts. Thus, in future studies more attention will need to paid to Treg cells and synovial fibroblasts as well as cells that promote bone formation.

Technological advances in the past several years have revealed the heterogeneity of cell subsets and enabled the identification of pathogenic and protective cell populations. From a therapeutic point of view, it is important to restore the joint structure by increasing osteoblastic bone formation and by targeting the pathogenic immune cell–fibroblast axis. Clarification of the interaction between immune cells and fibroblasts at the single-cell level will provide new insights into the pathogenesis of RA. In order to prove the pathological relevance of the findings obtained by scRNA-seq analysis, it will be necessary to perform loss-of-function analysis in vivo, such as cell-type-specific gene deletion. Clarifying how skeletal stem cells or the nervous system contribute to joint damage in RA will also be important. Integration of in silico and in vivo studies will provide a complete atlas of the immune cell–fibroblast–bone triad in RA, providing a molecular basis for the development of future therapeutic strategies aimed at providing protection against structural damage as well as restoration of damaged joints.

References

  1. Firestein, G. S. Evolving concepts of rheumatoid arthritis. Nature 423, 356–361 (2003).

    CAS  PubMed  Article  Google Scholar 

  2. McInnes, I. B. & Schett, G. The pathogenesis of rheumatoid arthritis. N. Engl. J. Med. 365, 2205–2219 (2011).

    CAS  PubMed  Article  Google Scholar 

  3. Takayanagi, H. Osteoimmunology: shared mechanisms and crosstalk between the immune and bone systems. Nat. Rev. Immunol. 7, 292–304 (2007).

    CAS  PubMed  Article  Google Scholar 

  4. Croft, A. P. et al. Distinct fibroblast subsets drive inflammation and damage in arthritis. Nature 570, 246–251 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  5. Mizoguchi, F. et al. Functionally distinct disease-associated fibroblast subsets in rheumatoid arthritis. Nat. Commun. 9, 789 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  6. Danks, L. et al. RANKL expressed on synovial fibroblasts is primarily responsible for bone erosions during joint inflammation. Ann. Rheum. Dis. 75, 1187–1195 (2016).

    CAS  PubMed  Article  Google Scholar 

  7. Komatsu, N. & Takayanagi, H. Inflammation and bone destruction in arthritis: synergistic activity of immune and mesenchymal cells in joints. Front. Immunol. 3, 77 (2012).

    PubMed  PubMed Central  Article  Google Scholar 

  8. Shim, J. H., Stavre, Z. & Gravallese, E. M. Bone loss in rheumatoid arthritis: basic mechanisms and clinical implications. Calcif. Tissue Int. 102, 533–546 (2018).

    CAS  PubMed  Article  Google Scholar 

  9. Nakashima, T. et al. Evidence for osteocyte regulation of bone homeostasis through RANKL expression. Nat. Med. 17, 1231–1234 (2011).

    CAS  PubMed  Article  Google Scholar 

  10. Xiong, J. et al. Matrix-embedded cells control osteoclast formation. Nat. Med. 17, 1235–1241 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  11. Okamoto, K. et al. Osteoimmunology: the conceptual framework unifying the immune and skeletal systems. Physiol. Rev. 97, 1295–1349 (2017).

    CAS  PubMed  Article  Google Scholar 

  12. Takayanagi, H. et al. Induction and activation of the transcription factor NFATc1 (NFAT2) integrate RANKL signaling in terminal differentiation of osteoclasts. Dev. Cell 3, 889–901 (2002).

    CAS  PubMed  Article  Google Scholar 

  13. Tsukasaki, M. et al. OPG production matters where it happened. Cell Rep. 32, 108124 (2020).

    CAS  PubMed  Article  Google Scholar 

  14. Robling, A. G. et al. Mechanical stimulation of bone in vivo reduces osteocyte expression of Sost/sclerostin. J. Biol. Chem. 283, 5866–5875 (2008).

    CAS  PubMed  Article  Google Scholar 

  15. Miyazaki, T. et al. Mechanical regulation of bone homeostasis through p130Cas-mediated alleviation of NF-κB activity. Sci. Adv. 5, eaau7802 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  16. Davidson, S. et al. Fibroblasts as immune regulators in infection, inflammation and cancer. Nat. Rev. Immunol. 21, 704–717 (2021).

    PubMed  Article  CAS  Google Scholar 

  17. Nygaard, G. & Firestein, G. S. Restoring synovial homeostasis in rheumatoid arthritis by targeting fibroblast-like synoviocytes. Nat. Rev. Rheumatol. 16, 316–333 (2020).

    PubMed  PubMed Central  Article  Google Scholar 

  18. Zhang, F. et al. Defining inflammatory cell states in rheumatoid arthritis joint synovial tissues by integrating single-cell transcriptomics and mass cytometry. Nat. Immunol. 20, 928–942 (2019).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  19. Rao, D. A. et al. Pathologically expanded peripheral T helper cell subset drives B cells in rheumatoid arthritis. Nature 542, 110–114 (2017).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  20. Pfeifle, R. et al. Regulation of autoantibody activity by the IL-23–TH17 axis determines the onset of autoimmune disease. Nat. Immunol. 18, 104–113 (2017).

    CAS  PubMed  Article  Google Scholar 

  21. Bondt, A. et al. ACPA IgG galactosylation associates with disease activity in pregnant patients with rheumatoid arthritis. Ann. Rheum. Dis. 77, 1130–1136 (2018).

    CAS  PubMed  Google Scholar 

  22. Chang, M. H. et al. Arthritis flares mediated by tissue-resident memory T cells in the joint. Cell Rep. 37, 109902 (2021).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  23. Cambre, I. et al. Mechanical strain determines the site-specific localization of inflammation and tissue damage in arthritis. Nat. Commun. 9, 4613 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  24. Wu, X. et al. Single-cell sequencing of immune cells from anticitrullinated peptide antibody positive and negative rheumatoid arthritis. Nat. Commun. 12, 4977 (2021).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  25. Karouzakis, E. et al. Analysis of early changes in DNA methylation in synovial fibroblasts of RA patients before diagnosis. Sci. Rep. 8, 7370 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  26. Hua, S. & Dias, T. H. Hypoxia-inducible factor (HIF) as a target for novel therapies in rheumatoid arthritis. Front. Pharmacol. 7, 184 (2016).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  27. Friscic, J. et al. The complement system drives local inflammatory tissue priming by metabolic reprogramming of synovial fibroblasts. Immunity 54, 1002–1021 (2021).

    CAS  PubMed  Article  Google Scholar 

  28. Lee, D. M. et al. Cadherin-11 in synovial lining formation and pathology in arthritis. Science 315, 1006–1010 (2007).

    CAS  PubMed  Article  Google Scholar 

  29. Chang, S. K. et al. Cadherin-11 regulates fibroblast inflammation. Proc. Natl Acad. Sci. USA 108, 8402–8407 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  30. Wei, K. et al. Notch signalling drives synovial fibroblast identity and arthritis pathology. Nature 582, 259–264 (2020).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  31. Armaka, M. et al. Mesenchymal cell targeting by TNF as a common pathogenic principle in chronic inflammatory joint and intestinal diseases. J. Exp. Med. 205, 331–337 (2008).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  32. Armaka, M., Ospelt, C., Pasparakis, M. & Kollias, G. The p55TNFR-IKK2-Ripk3 axis orchestrates arthritis by regulating death and inflammatory pathways in synovial fibroblasts. Nat. Commun. 9, 618 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  33. Nguyen, H. N. et al. Autocrine loop involving IL-6 family member LIF, LIF receptor, and STAT4 drives sustained fibroblast production of inflammatory mediators. Immunity 46, 220–232 (2017).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  34. Bromley, M. & Woolley, D. E. Chondroclasts and osteoclasts at subchondral sites of erosion in the rheumatoid joint. Arthritis Rheum. 27, 968–975 (1984).

    CAS  PubMed  Article  Google Scholar 

  35. Takayanagi, H. et al. Involvement of receptor activator of nuclear factor κB ligand/osteoclast differentiation factor in osteoclastogenesis from synoviocytes in rheumatoid arthritis. Arthritis Rheum. 43, 259–269 (2000).

    CAS  PubMed  Article  Google Scholar 

  36. Takayanagi, H. et al. A new mechanism of bone destruction in rheumatoid arthritis: synovial fibroblasts induce osteoclastogenesis. Biochem. Biophys. Res. Commun. 240, 279–286 (1997).

    CAS  PubMed  Article  Google Scholar 

  37. Gravallese, E. M. et al. Synovial tissue in rheumatoid arthritis is a source of osteoclast differentiation factor. Arthritis Rheum. 43, 250–258 (2000).

    CAS  PubMed  Article  Google Scholar 

  38. Pettit, A. R. et al. TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am. J. Pathol. 159, 1689–1699 (2001).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  39. Redlich, K. et al. Osteoclasts are essential for TNF-α-mediated joint destruction. J. Clin. Invest. 110, 1419–1427 (2002).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  40. Takayanagi, H. et al. T-cell-mediated regulation of osteoclastogenesis by signalling cross-talk between RANKL and IFN-γ. Nature 408, 600–605 (2000).

    CAS  PubMed  Article  Google Scholar 

  41. Kong, Y. Y. et al. Activated T cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 402, 304–309 (1999).

    CAS  PubMed  Article  Google Scholar 

  42. Meednu, N. et al. Production of RANKL by memory B cells: a link between B cells and bone erosion in rheumatoid arthritis. Arthritis Rheumatol. 68, 805–816 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  43. Ota, Y. et al. Generation mechanism of RANKL+ effector memory B cells: relevance to the pathogenesis of rheumatoid arthritis. Arthritis Res. Ther. 18, 67 (2016).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  44. Komatsu, N. et al. Plasma cells promote osteoclastogenesis and periarticular bone loss in autoimmune arthritis. J. Clin. Invest. 131, e143060 (2021).

    CAS  PubMed Central  Article  Google Scholar 

  45. Herman, S. et al. Induction of osteoclast-associated receptor, a key osteoclast costimulation molecule, in rheumatoid arthritis. Arthritis Rheum. 58, 3041–3050 (2008).

    CAS  PubMed  Article  Google Scholar 

  46. Lam, J. et al. TNF-α induces osteoclastogenesis by direct stimulation of macrophages exposed to permissive levels of RANK ligand. J. Clin. Invest. 106, 1481–1488 (2000).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  47. Ochi, S. et al. Pathological role of osteoclast costimulation in arthritis-induced bone loss. Proc. Natl Acad. Sci. USA 104, 11394–11399 (2007).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  48. Negishi-Koga, T. et al. Immune complexes regulate bone metabolism through FcRγ signalling. Nat. Commun. 6, 6637 (2015).

    CAS  PubMed  Article  Google Scholar 

  49. Krishnamurthy, A. et al. Identification of a novel chemokine-dependent molecular mechanism underlying rheumatoid arthritis-associated autoantibody-mediated bone loss. Ann. Rheum. Dis. 75, 721–729 (2016).

    CAS  PubMed  Article  Google Scholar 

  50. Harre, U. et al. Induction of osteoclastogenesis and bone loss by human autoantibodies against citrullinated vimentin. J. Clin. Invest. 122, 1791–1802 (2012).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  51. Burska, A. N. et al. Receptor activator of nuclear factor κ-B ligand (RANKL) serum levels are associated with progression to seropositive/negative rheumatoid arthritis. Clin. Exp. Rheumatol. 39, 456–462 (2021).

    PubMed  Article  Google Scholar 

  52. Asano, T. et al. Soluble RANKL is physiologically dispensable but accelerates tumour metastasis to bone. Nat. Metab. 1, 868–875 (2019).

    PubMed  Article  Google Scholar 

  53. Sato, K. et al. TH17 functions as an osteoclastogenic helper T cell subset that links T cell activation and bone destruction. J. Exp. Med. 203, 2673–2682 (2006).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  54. Kotake, S. et al. IL-17 in synovial fluids from patients with rheumatoid arthritis is a potent stimulator of osteoclastogenesis. J. Clin. Invest. 103, 1345–1352 (1999).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  55. Harre, U. et al. Glycosylation of immunoglobulin G determines osteoclast differentiation and bone loss. Nat. Commun. 6, 6651 (2015).

    CAS  PubMed  Article  Google Scholar 

  56. Sakaguchi, S., Yamaguchi, T., Nomura, T. & Ono, M. Regulatory T cells and immune tolerance. Cell 133, 775–787 (2008).

    CAS  PubMed  Article  Google Scholar 

  57. Hori, S., Nomura, T. & Sakaguchi, S. Control of regulatory T cell development by the transcription factor Foxp3. Science 299, 1057–1061 (2003).

    CAS  PubMed  Article  Google Scholar 

  58. Okada, Y. et al. Genetics of rheumatoid arthritis contributes to biology and drug discovery. Nature 506, 376–381 (2014).

    CAS  PubMed  Article  Google Scholar 

  59. Ota, M. et al. Dynamic landscape of immune cell-specific gene regulation in immune-mediated diseases. Cell 184, 3006–3021 (2021).

    CAS  PubMed  Article  Google Scholar 

  60. Zaiss, M. M. et al. Treg cells suppress osteoclast formation: a new link between the immune system and bone. Arthritis Rheum. 56, 4104–4112 (2007).

    CAS  PubMed  Article  Google Scholar 

  61. Komatsu, N. & Takayanagi, H. Regulatory T cells in arthritis. Prog. Mol. Biol. Transl. Sci. 136, 207–215 (2015).

    PubMed  Article  Google Scholar 

  62. Komatsu, N. et al. Pathogenic conversion of FOXP3+ T cells into TH17 cells in autoimmune arthritis. Nat. Med. 20, 62–68 (2014).

    CAS  PubMed  Article  Google Scholar 

  63. Kochi, Y. et al. A regulatory variant in CCR6 is associated with rheumatoid arthritis susceptibility. Nat. Genet. 42, 515–519 (2010).

    CAS  PubMed  Article  Google Scholar 

  64. Walsh, N. C. et al. Osteoblast function is compromised at sites of focal bone erosion in inflammatory arthritis. J. Bone Miner. Res. 24, 1572–1585 (2009).

    CAS  PubMed  Article  Google Scholar 

  65. Gilbert, L. et al. Expression of the osteoblast differentiation factor RUNX2 (Cbfa1/AML3/Pebp2αA) is inhibited by tumor necrosis factor-α. J. Biol. Chem. 277, 2695–2701 (2002).

    CAS  PubMed  Article  Google Scholar 

  66. de Rooy, D. P. et al. Genetic studies on components of the Wnt signalling pathway and the severity of joint destruction in rheumatoid arthritis. Ann. Rheum. Dis. 72, 769–775 (2013).

    PubMed  Article  CAS  Google Scholar 

  67. Wehmeyer, C. et al. Sclerostin inhibition promotes TNF-dependent inflammatory joint destruction. Sci. Transl. Med. 8, 330ra335 (2016).

    Article  CAS  Google Scholar 

  68. Matzelle, M. M. et al. Inflammation in arthritis induces expression of BMP3, an inhibitor of bone formation. Scand. J. Rheumatol. 45, 379–383 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  69. Stashenko, P., Dewhirst, F. E., Rooney, M. L., Desjardins, L. A. & Heeley, J. D. Interleukin-1β is a potent inhibitor of bone formation in vitro. J. Bone Miner. Res. 2, 559–565 (1987).

    CAS  PubMed  Article  Google Scholar 

  70. Bellido, T., Borba, V. Z., Roberson, P. & Manolagas, S. C. Activation of the Janus kinase/STAT (signal transducer and activator of transcription) signal transduction pathway by interleukin-6-type cytokines promotes osteoblast differentiation. Endocrinology 138, 3666–3676 (1997).

    CAS  PubMed  Article  Google Scholar 

  71. McGregor, N. E. et al. IL-6 exhibits both cis- and trans-signaling in osteocytes and osteoblasts, but only trans-signaling promotes bone formation and osteoclastogenesis. J. Biol. Chem. 294, 7850–7863 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  72. Gravallese, E. M. & Schett, G. Effects of the IL-23–IL-17 pathway on bone in spondyloarthritis. Nat. Rev. Rheumatol. 14, 631–640 (2018).

    CAS  PubMed  Article  Google Scholar 

  73. Kampylafka, E. et al. Resolution of synovitis and arrest of catabolic and anabolic bone changes in patients with psoriatic arthritis by IL-17A blockade with secukinumab: results from the prospective PSARTROS study. Arthritis Res. Ther. 20, 153 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  74. Sherlock, J. P. et al. IL-23 induces spondyloarthropathy by acting on ROR-γt+ CD3+CD4CD8 entheseal resident T cells. Nat. Med. 18, 1069–1076 (2016).

    Article  CAS  Google Scholar 

  75. Shaw, A. T., Maeda, Y. & Gravallese, E. M. IL-17A deficiency promotes periosteal bone formation in a model of inflammatory arthritis. Arthritis Res. Ther. 18, 104 (2016).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  76. Ono, T. et al. IL-17-producing γδ T cells enhance bone regeneration. Nat. Commun. 7, 10928 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  77. Hayashi, M. et al. Osteoprotection by semaphorin 3A. Nature 485, 69–74 (2012).

    CAS  PubMed  Article  Google Scholar 

  78. Negishi-Koga, T. et al. Suppression of bone formation by osteoclastic expression of semaphorin 4D. Nat. Med. 17, 1473–1480 (2011).

    CAS  PubMed  Article  Google Scholar 

  79. Yoshida, Y. et al. Semaphorin 4D contributes to rheumatoid arthritis by inducing inflammatory cytokine production: pathogenic and therapeutic implications. Arthritis Rheumatol. 67, 1481–1490 (2015).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  80. Takagawa, S. et al. Decreased semaphorin3A expression correlates with disease activity and histological features of rheumatoid arthritis. BMC Musculoskelet. Disord. 14, 40 (2013).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  81. He, X. et al. Osteoblastic PLEKHO1 contributes to joint inflammation in rheumatoid arthritis. EBioMedicine 41, 538–555 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  82. Goldring, S. R. Periarticular bone changes in rheumatoid arthritis: pathophysiological implications and clinical utility. Ann. Rheum. Dis. 68, 297–299 (2009).

    PubMed  Article  Google Scholar 

  83. Kleyer, A. et al. Bone loss before the clinical onset of rheumatoid arthritis in subjects with anticitrullinated protein antibodies. Ann. Rheum. Dis. 73, 854–860 (2014).

    PubMed  Article  Google Scholar 

  84. Engdahl, C. et al. Periarticular bone loss in arthritis is induced by autoantibodies against citrullinated vimentin. J. Bone Miner. Res. 32, 1681–1691 (2017).

    CAS  PubMed  Article  Google Scholar 

  85. Lightman, S. M., Utley, A. & Lee, K. P. Survival of long-lived plasma cells (LLPC): piecing together the puzzle. Front. Immunol. 10, 965 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  86. Pioli, P. D. Plasma cells, the next generation: beyond antibody secretion. Front. Immunol. 10, 2768 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  87. Sun, W. et al. B cells inhibit bone formation in rheumatoid arthritis by suppressing osteoblast differentiation. Nat. Commun. 9, 5127 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  88. Courbon, G. et al. Early sclerostin expression explains bone formation inhibition before arthritis onset in the rat adjuvant-induced arthritis model. Sci. Rep. 8, 3492 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  89. Werner, D. et al. Early changes of the cortical micro-channel system in the bare area of the joints of patients with rheumatoid arthritis. Arthritis Rheumatol. 69, 1580–1587 (2017).

    PubMed  Article  Google Scholar 

  90. Tanaka, Y. Managing osteoporosis and joint damage in patients with rheumatoid arthritis: an overview. J. Clin. Med. 10, 1241 (2021).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  91. Dubrovsky, A. M., Lim, M. J. & Lane, N. E. Osteoporosis in rheumatic diseases: anti-rheumatic drugs and the skeleton. Calcif. Tissue Int. 102, 607–618 (2018).

    CAS  PubMed  Article  Google Scholar 

  92. Haugeberg, G., Uhlig, T., Falch, J. A., Halse, J. I. & Kvien, T. K. Bone mineral density and frequency of osteoporosis in female patients with rheumatoid arthritis: results from 394 patients in the Oslo County Rheumatoid Arthritis register. Arthritis Rheum. 43, 522–530 (2000).

    CAS  PubMed  Article  Google Scholar 

  93. Kim, K. W., Kim, H. R., Kim, B. M., Cho, M. L. & Lee, S. H. TH17 cytokines regulate osteoclastogenesis in rheumatoid arthritis. Am. J. Pathol. 185, 3011–3024 (2015).

    CAS  PubMed  Article  Google Scholar 

  94. Yellin, M. J. et al. Ligation of CD40 on fibroblasts induces CD54 (ICAM-1) and CD106 (VCAM-1) up-regulation and IL-6 production and proliferation. J. Leukoc. Biol. 58, 209–216 (1995).

    CAS  PubMed  Article  Google Scholar 

  95. Van Seventer, G. A., Shimizu, Y., Horgan, K. J. & Shaw, S. The LFA-1 ligand ICAM-1 provides an important costimulatory signal for T cell receptor-mediated activation of resting T cells. J. Immunol. 144, 4579–4586 (1990).

    PubMed  Google Scholar 

  96. Damle, N. K. & Aruffo, A. Vascular cell adhesion molecule 1 induces T-cell antigen receptor-dependent activation of CD4+ T lymphocytes. Proc. Natl Acad. Sci. USA 88, 6403–6407 (1991).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  97. Yamamura, Y. et al. Effector function of resting T cells: activation of synovial fibroblasts. J. Immunol. 166, 2270–2275 (2001).

    CAS  PubMed  Article  Google Scholar 

  98. Sawa, S. et al. Autoimmune arthritis associated with mutated interleukin (IL)-6 receptor gp130 is driven by STAT3/IL-7-dependent homeostatic proliferation of CD4+ T cells. J. Exp. Med. 203, 1459–1470 (2006).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  99. Sawai, H. et al. T cell costimulation by fractalkine-expressing synoviocytes in rheumatoid arthritis. Arthritis Rheum. 52, 1392–1401 (2005).

    CAS  PubMed  Article  Google Scholar 

  100. Lee, J. H. et al. Pathogenic roles of CXCL10 signaling through CXCR3 and TLR4 in macrophages and T cells: relevance for arthritis. Arthritis Res. Ther. 19, 163 (2017).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  101. Hirota, K. et al. Preferential recruitment of CCR6-expressing TH17 cells to inflamed joints via CCL20 in rheumatoid arthritis and its animal model. J. Exp. Med. 204, 2803–2812 (2007).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  102. Ogura, H. et al. Interleukin-17 promotes autoimmunity by triggering a positive-feedback loop via interleukin-6 induction. Immunity 29, 628–636 (2008).

    CAS  PubMed  Article  Google Scholar 

  103. Kikuchi, J. et al. Peripheral blood CD4+CD25+CD127low regulatory T cells are significantly increased by tocilizumab treatment in patients with rheumatoid arthritis: increase in regulatory T cells correlates with clinical response. Arthritis Res. Ther. 17, 10 (2015).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  104. Hirota, K. et al. Autoimmune TH17 cells induced synovial stromal and innate lymphoid cell secretion of the cytokine GM-CSF to initiate and augment autoimmune arthritis. Immunity 48, 1220–1232 (2018).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  105. Park, Y. E. et al. IL-17 increases cadherin-11 expression in a model of autoimmune experimental arthritis and in rheumatoid arthritis. Immunol. Lett. 140, 97–103 (2011).

    CAS  PubMed  Article  Google Scholar 

  106. Tran, C. N. et al. Presentation of arthritogenic peptide to antigen-specific T cells by fibroblast-like synoviocytes. Arthritis Rheum. 56, 1497–1506 (2007).

    CAS  PubMed  Article  Google Scholar 

  107. Carmona-Rivera, C. et al. Synovial fibroblast–neutrophil interactions promote pathogenic adaptive immunity in rheumatoid arthritis. Sci. Immunol. 2, eaag3358 (2017).

    PubMed  PubMed Central  Article  Google Scholar 

  108. Burger, J. A., Zvaifler, N. J., Tsukada, N., Firestein, G. S. & Kipps, T. J. Fibroblast-like synoviocytes support B-cell pseudoemperipolesis via a stromal cell-derived factor-1- and CD106 (VCAM-1)-dependent mechanism. J. Clin. Invest. 107, 305–315 (2001).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  109. Bombardieri, M. et al. A BAFF/APRIL-dependent TLR3-stimulated pathway enhances the capacity of rheumatoid synovial fibroblasts to induce AID expression and Ig class-switching in B cells. Ann. Rheum. Dis. 70, 1857–1865 (2011).

    CAS  PubMed  Article  Google Scholar 

  110. Orange, D. E. et al. RNA identification of PRIME cells predicting rheumatoid arthritis flares. N. Engl. J. Med. 383, 218–228 (2020).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  111. Hasegawa, T. et al. Identification of a novel arthritis-associated osteoclast precursor macrophage regulated by FoxM1. Nat. Immunol. 20, 1631–1643 (2019).

    CAS  PubMed  Article  Google Scholar 

  112. Kuo, D. et al. HBEGF+ macrophages in rheumatoid arthritis induce fibroblast invasiveness. Sci. Transl. Med. 11, eaau8587 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  113. Diarra, D. et al. Dickkopf-1 is a master regulator of joint remodeling. Nat. Med. 13, 156–163 (2007).

    CAS  PubMed  Article  Google Scholar 

  114. Brown, A. K. et al. An explanation for the apparent dissociation between clinical remission and continued structural deterioration in rheumatoid arthritis. Arthritis Rheum. 58, 2958–2967 (2008).

    CAS  PubMed  Article  Google Scholar 

  115. Smolen, J. S. et al. Evidence of radiographic benefit of treatment with infliximab plus methotrexate in rheumatoid arthritis patients who had no clinical improvement: a detailed subanalysis of data from the Anti-Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy study. Arthritis Rheum. 52, 1020–1030 (2005).

    CAS  PubMed  Article  Google Scholar 

  116. Roser-Page, S., Vikulina, T., Zayzafoon, M. & Weitzmann, M. N. CTLA-4Ig-induced T cell anergy promotes Wnt-10b production and bone formation in a mouse model. Arthritis Rheumatol. 66, 990–999 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  117. Tyagi, A. M. et al. The microbial metabolite butyrate stimulates bone formation via T regulatory cell-mediated regulation of WNT10B expression. Immunity 49, 1116–1131 (2018).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  118. Bozec, A. et al. T cell costimulation molecules CD80/86 inhibit osteoclast differentiation by inducing the IDO/tryptophan pathway. Sci. Transl. Med. 6, 235ra60 (2014).

    PubMed  Article  CAS  Google Scholar 

  119. Zaiss, M. M. et al. Increased bone density and resistance to ovariectomy-induced bone loss in FoxP3-transgenic mice based on impaired osteoclast differentiation. Arthritis Rheum. 62, 2328–2338 (2010).

    CAS  PubMed  Article  Google Scholar 

  120. Zaiss, M. M. et al. Regulatory T cells protect from local and systemic bone destruction in arthritis. J. Immunol. 184, 7238–7246 (2010).

    CAS  PubMed  Article  Google Scholar 

  121. Rauber, S. et al. Resolution of inflammation by interleukin-9-producing type 2 innate lymphoid cells. Nat. Med. 23, 938–944 (2017).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  122. Culemann, S. et al. Locally renewing resident synovial macrophages provide a protective barrier for the joint. Nature 572, 670–675 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  123. Alivernini, S. et al. Distinct synovial tissue macrophage subsets regulate inflammation and remission in rheumatoid arthritis. Nat. Med. 26, 1295–1306 (2020).

    CAS  PubMed  Article  Google Scholar 

  124. Cohen, S. B. et al. Denosumab treatment effects on structural damage, bone mineral density, and bone turnover in rheumatoid arthritis: a twelve-month, multicenter, randomized, double-blind, placebo-controlled, phase II clinical trial. Arthritis Rheum. 58, 1299–1309 (2008).

    CAS  PubMed  Article  Google Scholar 

  125. Takeuchi, T. et al. Effect of denosumab on Japanese patients with rheumatoid arthritis: a dose-response study of AMG 162 (Denosumab) in patients with RheumatoId arthritis on methotrexate to Validate inhibitory effect on bone Erosion (DRIVE)–a 12-month, multicentre, randomised, double-blind, placebo-controlled, phase II clinical trial. Ann. Rheum. Dis. 75, 983–990 (2016).

    CAS  PubMed  Article  Google Scholar 

  126. Takeuchi, T. et al. Effects of the anti-RANKL antibody denosumab on joint structural damage in patients with rheumatoid arthritis treated with conventional synthetic disease-modifying antirheumatic drugs (DESIRABLE study): a randomised, double-blind, placebo-controlled phase 3 trial. Ann. Rheum. Dis. 78, 899–907 (2019).

    CAS  PubMed  Article  Google Scholar 

  127. Axmann, R. et al. CTLA-4 directly inhibits osteoclast formation. Ann. Rheum. Dis. 67, 1603–1609 (2008).

    CAS  PubMed  Article  Google Scholar 

  128. Blanco, F. J. et al. Secukinumab in active rheumatoid arthritis: a phase III randomized, double-blind, active comparator- and placebo-controlled study. Arthritis Rheumatol. 69, 1144–1153 (2017).

    CAS  PubMed  Article  Google Scholar 

  129. Lubberts, E. et al. Treatment with a neutralizing anti-murine interleukin-17 antibody after the onset of collagen-induced arthritis reduces joint inflammation, cartilage destruction, and bone erosion. Arthritis Rheum. 50, 650–659 (2004).

    CAS  PubMed  Article  Google Scholar 

  130. Glatt, S. et al. Efficacy and safety of bimekizumab as add-on therapy for rheumatoid arthritis in patients with inadequate response to certolizumab pegol: a proof-of-concept study. Ann. Rheum. Dis. 78, 1033–1040 (2019).

    CAS  PubMed  Article  Google Scholar 

  131. Maeshima, K. et al. The JAK inhibitor tofacitinib regulates synovitis through inhibition of interferon-γ and interleukin-17 production by human CD4+ T cells. Arthritis Rheum. 64, 1790–1798 (2012).

    CAS  PubMed  Article  Google Scholar 

  132. Kubo, S. et al. The JAK inhibitor, tofacitinib, reduces the T cell stimulatory capacity of human monocyte-derived dendritic cells. Ann. Rheum. Dis. 73, 2192–2198 (2014).

    CAS  PubMed  Article  Google Scholar 

  133. Combe, B. et al. Filgotinib versus placebo or adalimumab in patients with rheumatoid arthritis and inadequate response to methotrexate: a phase III randomised clinical trial. Ann. Rheum. Dis. 80, 848–858 (2021).

    CAS  PubMed  Article  Google Scholar 

  134. Traves, P. G. et al. JAK selectivity and the implications for clinical inhibition of pharmacodynamic cytokine signalling by filgotinib, upadacitinib, tofacitinib and baricitinib. Ann. Rheum. Dis. 80, 865–875 (2021).

    CAS  PubMed  Article  Google Scholar 

  135. Adam, S. et al. JAK inhibition increases bone mass in steady-state conditions and ameliorates pathological bone loss by stimulating osteoblast function. Sci. Transl. Med. 12, eaay4447 (2020).

    CAS  PubMed  Article  Google Scholar 

  136. Murakami, K. et al. A Jak1/2 inhibitor, baricitinib, inhibits osteoclastogenesis by suppressing RANKL expression in osteoblasts in vitro. PLoS ONE 12, e0181126 (2017).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  137. Matzelle, M. M. et al. Resolution of inflammation induces osteoblast function and regulates the Wnt signaling pathway. Arthritis Rheum. 64, 1540–1550 (2012).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  138. Chen, X. X. et al. Sclerostin inhibition reverses systemic, periarticular and local bone loss in arthritis. Ann. Rheum. Dis. 72, 1732–1736 (2013).

    CAS  PubMed  Article  Google Scholar 

  139. Marenzana, M., Vugler, A., Moore, A. & Robinson, M. Effect of sclerostin-neutralising antibody on periarticular and systemic bone in a murine model of rheumatoid arthritis: a microCT study. Arthritis Res. Ther. 15, R125 (2013).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  140. Teng, Y. et al. Adenovirus-mediated delivery of Sema3A alleviates rheumatoid arthritis in a serum-transfer induced mouse model. Oncotarget 8, 66270–66280 (2017).

    PubMed  PubMed Central  Article  Google Scholar 

  141. Zhang, H. et al. NOTCH inhibits osteoblast formation in inflammatory arthritis via noncanonical NF-κB. J. Clin. Invest. 124, 3200–3214 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  142. Sawai, H., Park, Y. W., He, X., Goronzy, J. J. & Weyand, C. M. Fractalkine mediates T cell-dependent proliferation of synovial fibroblasts in rheumatoid arthritis. Arthritis Rheum. 56, 3215–3225 (2007).

    CAS  PubMed  Article  Google Scholar 

  143. Laragione, T., Brenner, M., Sherry, B. & Gulko, P. S. CXCL10 and its receptor CXCR3 regulate synovial fibroblast invasion in rheumatoid arthritis. Arthritis Rheum. 63, 3274–3283 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  144. Nanki, T. et al. Inhibition of fractalkine ameliorates murine collagen-induced arthritis. J. Immunol. 173, 7010–7016 (2004).

    CAS  PubMed  Article  Google Scholar 

  145. Hamilton, J. A., Cook, A. D. & Tak, P. P. Anti-colony-stimulating factor therapies for inflammatory and autoimmune diseases. Nat. Rev. Drug Discov. 16, 53–70 (2016).

    PubMed  Article  CAS  Google Scholar 

  146. Tanaka, Y. et al. Efficacy and safety of E6011, an anti-fractalkine monoclonal antibody, in patients with active rheumatoid arthritis with inadequate response to methotrexate: results of a randomized, double-blind, placebo-controlled phase II study. Arthritis Rheumatol. 73, 587–595 (2021).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  147. Yellin, M. et al. A phase II, randomized, double-blind, placebo-controlled study evaluating the efficacy and safety of MDX-1100, a fully human anti-CXCL10 monoclonal antibody, in combination with methotrexate in patients with rheumatoid arthritis. Arthritis Rheum. 64, 1730–1739 (2012).

    CAS  PubMed  Article  Google Scholar 

  148. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/ct2/show/NCT04134728 (2022).

  149. Finch, R. et al. Results of a phase 2 study of RG6125, an anti-cadherin-11 monoclonal antibody in rheumatoid arthritis patients with an inadequate response to anti- TNFα therapy [abstract OP0224]. Ann. Rheum. Dis. 78, 189 (2019).

    Google Scholar 

  150. Dorst, D. N. et al. Targeting of fibroblast activation protein in rheumatoid arthritis patients: imaging and ex vivo photodynamic therapy. Rheumatology https://doi.org/10.1093/rheumatology/keab664 (2021).

    Article  PubMed  Google Scholar 

  151. Pap, T. & Korb-Pap, A. Cartilage damage in osteoarthritis and rheumatoid arthritis–two unequal siblings. Nat. Rev. Rheumatol. 11, 606–615 (2015).

    PubMed  Article  Google Scholar 

  152. Araki, Y. & Mimura, T. Matrix metalloproteinase gene activation resulting from disordred epigenetic mechanisms in rheumatoid arthritis. Int. J. Mol. Sci. 18, 905 (2017).

    PubMed Central  Article  CAS  Google Scholar 

  153. Posthumus, M. D. et al. Serum levels of matrix metalloproteinase-3 in relation to the development of radiological damage in patients with early rheumatoid arthritis. Rheumatology 38, 1081–1087 (1999).

    CAS  PubMed  Article  Google Scholar 

  154. Chang, S. H. et al. Excessive mechanical loading promotes osteoarthritis through the gremlin-1-NF-κB pathway. Nat. Commun. 10, 1442 (2019).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  155. Han, E. J. et al. GREM1 is a key regulator of synoviocyte hyperplasia and invasiveness. J. Rheumatol. 43, 474–485 (2016).

    PubMed  Article  Google Scholar 

Download references

Acknowledgements

The authors are grateful to all the laboratory members, especially K. Okamoto, M. Tsukasaki and R. Ling for thoughtful discussion.

Author information

Authors and Affiliations

Authors

Contributions

The authors contributed equally to all aspects of the article.

Corresponding author

Correspondence to Hiroshi Takayanagi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Reviews Rheumatology thanks M. Nakamura, J. Lorenzo and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Komatsu, N., Takayanagi, H. Mechanisms of joint destruction in rheumatoid arthritis — immune cell–fibroblast–bone interactions. Nat Rev Rheumatol 18, 415–429 (2022). https://doi.org/10.1038/s41584-022-00793-5

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41584-022-00793-5

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing