RANKL Biology

Abstract

Since the receptor activator of nuclear factor-kappa B ligand (RANKL), its cognate receptor receptor activator of nuclear factor-kappa B (RANK), and the decoy receptor osteoprotegerin (OPG) were discovered, a number of studies have uncovered the crucial role of the RANKL-RANK-OPG pathway in controlling the key aspect of bone homeostasis, the immune system, inflammation, cancer, and other systems under pathophysiological condition. These findings have expanded the understanding of the multifunctional biology of the RANKL-RANK-OPG pathway and led to the development of therapeutic potential targeting this pathway. The successful development and application of anti-RANKL antibody in treating diseases causing bone loss validates the utility of therapeutic approaches based on the modulation of this pathway. Moreover, recent studies have demonstrated the involvement of the RANKL-RANK pathway in osteoblast differentiation and bone formation, shedding light on the RANKL-RANK dual signaling in coupling bone resorption and bone formation. In this review, we will summarize the current understanding of the RANKL-RANK-OPG system in the context of the bone and the immune system as well as the impact of this pathway in disease conditions, including cancer development and metastasis.

Introduction

The receptor activator of nuclear factor-kappa B ligand (RANKL), its receptor, receptor activator of nuclear factor-kappa B (RANK), and the soluble decoy receptor for RANKL, osteoprotegerin (OPG), were initially discovered within the immune and the bone systems. RANKL and RANK were first discovered by two independent groups during the study of T-cell activation. In the first report, the group identified a new cytokine of the TNF family as a regulatory molecule of T-cell activation and named it TRANCE [1]. In the second report, the group cloned the receptor RANK and the ligand RANKL from dendritic cells and murine thymoma lines, respectively [2]. TRANCE and RANKL were later deemed identical. RANKL stimulation activates not only T cells but also dendritic cells and augments the ability of dendritic cells to stimulate T cells, suggesting RANKL’s role in regulating T cell-mediated immune responses [1, 2]. Additionally, independent research groups contemporaneously identified RANKL from bone marrow-derived stroma cell lines and/or mouse myelomonocytic cell lines as the long-sought osteoclast differentiation factor (ODF) [3, 4] and its receptor ODFR, which was identical to RANK from mouse macrophage-like cell lines [5, 6]. OPG was identified as the inhibitory factor for osteoclastogenesis by two independent research groups prior to the cloning of RANKL and RANK [7, 8]. Given the history of cloning, it was conceivable that RANKL, RANK, and OPG are involved in the bone and the immune systems. Indeed, research employing various genetically deficient mouse models established the critical roles of this triad in regulating bone metabolism and physiological immune processes as well as pathological conditions including osteoporosis, rheumatoid arthritis, and certain types of cancer. In this review, we will introduce the key functions of the RANKL-RANK pathway in maintaining bone homeostasis and immunity. We will also discuss the emerging findings of this triad in bone homeostasis. We will then elaborate on the function of RANKL based on observations from a fully humanized anti-RANKL treatment in pathological bone condition and tumor.

RANKL in bone metabolism

The RANKL-RANK-OPG axis in osteoclast and bone homeostasis

Bone is a hard tissue that supports the body against gravity, assists locomotion, protects vulnerable organs, stores minerals, and plays a crucial role in the production of hematopoietic cells. Bone quantity and quality are maintained by the balance between mainly two types of cells: mesenchymal cell-derived bone-forming osteoblasts/osteocytes and hematopoietic precursor cell-derived bone-resorbing osteoclasts [9–11]. Postnatal bone is renewed continuously by osteoclastic bone resorption and osteoblastic bone formation in response to a variety of stimuli; this process is called bone remodeling. Bone remodeling is thought to be dynamically coupled, and the coupling is the key to the preservation of bone architecture and strength. The RANKL-RANK-OPG axis provides critical signals that control intercellular communication between bone forming-osteoblasts and bone resorbing-osteoclasts.

In bone tissue, RANKL is produced by several types of cells, including osteoblasts, osteocytes, stromal cells in bone, and immune cells. RANKL is a type II homotrimeric membrane protein that is encoded by the tumor necrosis factor superfamily member 11 (TNFSF11). RANKL has three isoforms due to alternative splicing: RANKL1 is a full-length form, RANKL2 has a shorter intracellular domain, and RANKL3 lacks the transmembrane domain [12]. RANKL in soluble form is also produced by proteolytic ectodomain shedding mediate by metalloproteinases [13, 14]. RANK is a type I membrane functional receptor encoded by the tumor necrosis factor receptor superfamily member 11a (TNFRSF11A). It is mainly expressed by cells of hematopoietic origin, including osteoclasts and their precursors [15], and it is also detected in mesenchymal stem cells [16, 17]. The ligation of RANK expressed on osteoclast precursors with RANKL derived from osteocytes and osteoblasts promotes the trimerization and activation of RANK [18]. The intracellular domain of RANK has binding sites for the adaptor protein TNF receptor-associated factors (TRAFs) [19]. Upon the activation of RANK, several signaling adaptors including TRAF6 are recruited and activated [20]. The recruitment of adaptors converges on kinase activation and promotes nuclear translocation and the activation of nuclear factor of activated T cells 1 (NFATc1) [21], c-fos [22], and nuclear factor kappa B (NF-kB) [23], which are the master regulators of osteoclast-specific transcriptional programs. Co-stimulatory signals mediated by ITAM-motif-containing proteins, DNAX-activation protein 12 (DAP12), and Fc receptor gamma-chain (FcRγ), which associate with cell surface receptor OSCAR, PIR-A, or TREM-2, activate the Syk-PLCγ pathway and flux calcium. The calcium flux evokes Ca2+-dependent calcineurin signaling and enhances activities of NFATc1 and the transcription factor cAMP response element-binding protein (CREB), which induces the expression of osteoclast-specific genes [24]. These signaling pathways initiate osteoclast differentiation. Consequently, fully functional mature multinucleated osteoclasts are generated through mainly cell–cell fusion and repeated incomplete cytokinesis [25]. RANKL–RANK signaling also regulates osteoclast activation and survival. OPG acts as a decoy receptor for RANKL and interferes with the RANKL–RANK interaction. OPG is a unique TNFR superfamily member that exists only as a secreted molecule. OPG is encoded by the tumor necrosis factor receptor superfamily member 11b (TNFRSF11B) and forms a dimer via the large carboxy-terminal domain. OPG is expressed primarily by osteoblasts and bone marrow stromal cells, and it can be induced in B lymphocytes, DC, and follicular DCs [26]. The expression of OPG is regulated by estrogen, IL-4, transforming growth factor beta (TGF-β) or prostaglandin E2 (PGE2) [3, 8, 27]. OPG competitively binds to RANKL and prevents it from binding to RANK, thereby preventing osteoclast formation and blocking osteoclastic bone resorption [7, 28–32] (Figure 1). The relative levels of OPG and RANKL are precisely and strictly controlled to ensure healthy bone, and the dysregulation of the RANKL-RANK-OPG system at the skeletal level has been widely documented in the context of bone-related pathologies such as post-menopausal osteoporosis. Recently, it has been reported that locally produced OPG rather than circulating OPG is crucial for bone and immune homeostasis [33].

Figure 1.

The RANKL-RANK-OPG pathway in osteoclast and osteoblast differentiation. RANKL is mainly produced by osteoblasts and osteocytes, while RANK is mainly expressed by cells of a hematopoietic origin, including osteoclasts and their precursors. The ligation of RANK by RANKL leads to the activation of downstream signaling pathway and promotes nuclear translocation and activation of NFATc1 and NF-kB, causing osteoclast differentiation. OPG and LGR4, both of which interact with RANKL, are endogenous inhibitors of the RANKL-RANK pathway. RANKL has also been reported to function as a receptor of RANK, serving reserve signaling. The ligation of RANKL by RANK activates RANKL reverse signaling through a proline-rich motif. The soluble form of RANKL (sRANKL) can be expressed by alternative splicing or produced through proteolytic cleavage by matrix metalloproteinases. sRANKL is functional to activate RANK signaling. The expression of RANK and LGR4 has been reported in osteoblastic lineage cells, and sRANKL might activate osteoblast differentiation through these receptors as an autocrine-paracrine loop.

The RANKL-LGR4 negative regulation of osteoclast differentiation

Although RANK has long been considered to be the sole receptor for RANK, a leucine-rich repeat G protein-coupled receptor 4 (LGR4), also known as G-protein-coupled receptor 48 (GPR48), and R-spondin receptor was identified as an additional receptor for RANKL [34]. In the bone, LGR4 is detected in bone marrow mesenchymal stromal cells, osteoblasts, chondrocytes, and mature osteoclasts. LGR4 competes with RANK for RANKL binding. Physical binding of LGR4 with RANKL activates the GSK3-β signaling pathway that suppresses the expression of NFATc1 during osteoclastogenesis, negatively regulating osteoclast differentiation [34]. Indeed, osteoclast-specific LGR4 deficiency increases bone erosion along with the hyperactivation of osteoclasts, leading to osteoporosis [34]. Therefore, in addition to OPG, LGR4 is another endogenous inhibitor of RANKL–RANK signaling. LGR4 and RANKL interaction is inhibited by OPG, indicating that the binding affinity of RANKL with LGR4 is lower than that with OPG [34] (Figure 1).

RANKL in osteoblastic bone formation

Since the discovery of RANKL, extensive efforts have been made to study its function in the bone. The skeletal phenotype arising in mice and humans from RANKL gene ablation is more prominent in the osteoclast compartment, and RANKL has pronounced roles in osteoclast differentiation and function. Therefore, the cell-autonomous effect of RANKL in bone marrow mesenchymal stromal cells (BM-MSCs), which form the origin of osteoblastic cells, have long been overlooked. Recently, a study reported a defect in RANKL-deficient osteoblast lineage. BM-MSCs derived from RANKL-deficient mice had a partial osteogenic differentiation defect, which was recovered through the lentiviral restoration of the soluble form of RANKL [16]. This is the first report revealing the contribution of RANKL to the osteogenic differentiation of bone marrow mesenchymal stromal cells. The expression of RANK and LGR4 is detected in BM-MSCs, suggesting an autocrine-paracrine loop in which RANKL possibly exerts its function through interaction with either its receptor RANK or the receptor LGR4 [16]. Note that this finding might be relevant for RANKL-dependent autosomal recessive osteopetrosis (ARO), which is a rare genetic bone disease characterized by increased bone density due to a defect that is not intrinsic to the hematopoietic lineage cells [16].

Some of TNF superfamily members are known to act as bidirectional signaling molecules that generate intracellular reverse signaling [35]. It was demonstrated that RANKL performs reverse signaling from osteoclasts to osteoblasts and contributes to osteoblastic bone formation and the coupling of bone resorption and formation [36]. The first implication of the presence of RANKL reverse signaling was obtained from studies that used RANKL-binding peptides. A peptide WP9QY (W9), which is designed to mimic the critical binding site of TNF type I receptor [37], binds RANKL and inhibits RANKL-induced osteoclastogenesis without preventing the interaction between RANKL and RANK in vitro and osteoclast-mediated bone loss in vivo [38]. High frequent administration of W9 peptides resulted in increased bone formation in addition to decreased osteoclastogenesis in vivo [39]. The W9 peptide stimulated osteoblast activities in vitro, which was attenuated by RANKL knockdown. These observations suggested the bone anabolic effect of W9 through RANKL. Another RANKL-binding peptide OP3–4, which is designed an OPG-like peptidomimetic and shown to inhibit osteoclastogenesis by interfering with the RANKL–RANK complex [40], also functions as a systematic stimulator of bone formation in an inflammatory bone destruction model [41]. The bone anabolic effect of RANKL-binding peptides suggests an osteogenic signaling pathway that is mediated by RANKL, wherein RANKL on osteoblasts works as a receptor to stimulate osteoblast differentiation.

As a ligand for RANKL, RANKL-binding proteins such as RANK and OPG are candidate molecules. OPG does not show any osteoblast stimulatory activity, while vesicular RANK derived from mature osteoclasts stimulates osteoblasts and promotes osteogenic activity via mammalian target of rapamycin complex 1 (mTORC1) signaling and Runt-related transcription factor 2 (Runx2) activation [36]. The bone anabolic effect of vesicular RANK and RANKL-binding peptides is blunted in RANKL-deficient osteoblasts [36]. The RANKL intracellular proline-rich motif is shown to be important for RANKL reverse signaling [36]. Osteoblasts derived from the mutant mice with a point mutation of proline to alanine (Pro29Ala) in the RANKL intracellular region failed to be activated by stimulation with vesicular RANK [36]. In addition, the mutant mice failed to increase bone formation following osteoclastic bone resorption [36]. These experimental results demonstrate that RANK–RANKL reverse signaling contributes to the coupling of bone resorption and formation (Figure 1). Vesicular RANK is detected not only in extracellular vesicle from mature osteoclasts but also in osteoclast apoptotic bodies that are released as a product of apoptotic cell disassembly [42]. In vitro cultured osteoclasts stimulated with bisphosphonates such as alendronate release RANK-carrying apoptotic bodies that can stimulate the RANKL-expressed osteoblastic cell line and induce mineralization. Alendronate has been known to induce osteoclast apoptosis in vivo [43], implying that apoptotic bodies exert bone anabolic effects in vivo. There is a report implying the bone anabolic effect of alendronate [44]. Further studies are required to clarify the role of vesicular RANK in the contest of pathophysiological mechanisms.

Although OPG binds RANKL with much higher affinity than RANK and OP3–4 peptides [40, 45], it fails to stimulate osteoblastic bone formation activity. These findings indicate that not all RANKL-binding molecules can stimulate osteoblast differentiation, raising a question about how vesicle RANK ligation activates RANKL reverse signaling. Recently, a report proposed that RANKL accumulation and clustering at the cell surface of osteoblasts is the mechanism that triggers RANKL reverse signaling [46]. Indeed, the stimulation of RANKL-expressing osteoblast cell line with RANK-coated beads and/or with RANKL binding peptides but not with OPG increases the accumulation of RANKL on the cell surface [46]. A cell-free system using a high-speed atomic force microscope (HS-AFM) showed the easy clustering molecular features of RANKL; OPG clings to the RANKL and disturbs the clustering. By contrast, the RANKL-binding peptide does not interfere with the RANKL clustering [46]. Given that clustering of membrane-bound receptors leads to cell activation [18, 47–50], RANKL accumulation, followed by clustering, seems to be the key to inducing RANKL reverse signaling. Additionally, artificial cluster induction using IgM to promote the pentameric assembly of the RANKL-OPG-Fc complex enhances the cell surface accumulation of RANKL and osteoblast activation in terms of the expression of osteogenic genes [46]. These observations support the idea that clustering formation may be a pivotal switch for RANKL-mediated osteoblast activation. However, whether the clustering assumes induction of RANKL reverse signaling requires clarification.

A recent study showed opposing results of the bone anabolic effect of RANKL–RANK signaling. The RANKL stimulation inhibited osteogenesis by promoting β-catenin degradation and also inhibiting its synthesis [17]. Bone marrow mesenchymal stem cell (BMSCs)-specific RANK knockout mice show increased bone formation with increased osteoblast maturation. Furthermore, RANK deficiency in BMSCs ameliorates OVX-induced bone loss [17]. The biological relevance of RANKL–RANK signaling in bone anabolism in the physiological and pathological context should be carefully defined. Nevertheless, the possibility of an osteogenic function of RANKL is worth further investigation since it will provide molecular basis for the development of new therapeutic strategies for diseases with bone loss.

RANKL in the immune system

RANKL–RANK signaling is critical for the immune system in terms of the development of immune organs and immune cells as well as the proper regulation of their function. Several types of immune cells and stromal cells express RANKL, RANK, and/or OPG and transduce signals for the development and functioning of the immune system. The RANKL-RANK-OPG-dependent processes have important implications not only for development but also for the regulation of immune responses and tumor.

The development of immune organs

Thymus is the primary lymphoid organ responsible for the development of T cells. T-cell progenitors, which arise in the liver during embryonic life and in the bone marrow in postnatal life, enter thymus and undergo positive and negative selections for acquiring a T-cell repertoire that is capable of responding to a diverse array of foreign antigens but are tolerant to self-antigens [51–53]. Cortical thymic epithelial cells (cTECs) control the functionality of TCRs during positive selection, while mature medullary thymic epithelial cells (mTECs) mediate the central tolerance process during negative selection by presenting self-peptides, including tissue restricted antigens (TRAs), on their MHC molecules [54, 55]. TRA expression is controlled by a transcriptional coactivator autoimmune regulator (AIRE) [56, 57]. In the thymus, RANKL is mainly produced by CD4/CD8 single positive thymocytes, Vγ5⁺ γδT cells, invariant natural killer T (iNKT) cells, and lymphoid tissue inducer (LTi) cells [58, 59]. By contrast, RANK is mainly expressed by mTECs [60]. The RANKL–RANK interaction induces the expression of AIRE in mTECs [61, 62], resulting in controlling of TRA expression and formation of TRA-MHC complex on mTECs [63–65]. RANKL stimulation triggers mTEC differentiation, meanwhile up-regulating the expression of OPG via the transcription factor Spi-B [66]. The expression of OPG limits the development of mature mTEC. mTEC-specific OPG deficiency resulted in an increased number of mTECs and Aire⁺ mTECs [33].

RANKL contributes to the development and functioning of secondary lymphoid organs. RANKL- and RANK-deficient mice display a complete absence of peripheral lymph nodes and abnormal development of Peyer’s patchs, cryptopatches and spleen [67–70]. Lymph node organogenesis requires hematopoietic LTi cells and mesenchymal lymphoid tissue organizer (LTo) cells. During embryogenesis, LTi cells produced in the fetal liver migrate towards the lymph node anlage [71]. The clustering of LTi cells that express both RANK and RANKL stimulates the production of lymphotoxinα1β2 (LTα1β2) on LTi cells in an autocrine manner. The produced LTα1β2 engages the LTβ receptor (LTβR) on the surrounding LTo precursors [72, 73]. Signaling via the RANKL–RANK and LTα1β2-LTβR axes creates a positive feedback loop, which drives the development of mature LTo cells. Mature LTo cells further recruit a larger number of LTi cells by expressing RANKL and chemokines and are necessary for tissue organization [74].

The gastrointestinal tract is the largest pathogenic bacteria entry site. Microfold (M) cells are specialized intestinal epithelial-derived cells in the follicle-associated epithelial (FAE) cells covering the gut-associated lymphoid tissues (GALT), including the Peyer’s patches (PPs) [13]. M cells are essential for the transfer of antigens from orally acquired pathogens to antigen-presenting cells and, subsequently, for the induction of efficient immune responses [13]. RANKL is necessary and sufficient for M cell development [67, 75]. Stromal cells in the subepithelial dome of the FAE provide RANKL to RANK-expressing stem cells in order to induce M cell differentiation, and RANKL-deficiency results in the loss of M cells.

RANKL–RANK signaling in B cell physiology is still a matter of debate. RANKL- and/or RANK-deficient mice show reduced peripheral B cell number associated with the absence of peripheral lymph nodes [68–70], although this B cell lymphopenia could reflect the dramatic reduction of the B cell niche caused by the fibrosis of bone marrow cavity due to severe osteopetrosis. Additionally, B cell-specific RANK-deficient mice show no alteration in B cell development [76]. A recent report describes involvement of RANKL–RANK signaling in the frequency of regulatory B cells and the production of anti-inflammatory cytokine IL-10 by B cells, suggesting a role of RANKL–RANK signaling in the modulation of B cell-mediated inflammation [77].

Involvement in autoimmunity and inflammation

Dysregulated RANKL–RANK signaling has been shown to be involved in autoimmunity. Although the specific role of RANKL–RANK signaling in autoimmune disease is still not fully understood, considerable amount of evidence suggests that the RANKL–RANK signaling affects T cells, B cells and more cell types. Rheumatoid arthritis (RA) is one of the most common autoimmune diseases, which is characterized by chronic joint inflammation along with bone damage, such as localized bone erosion and periarticular bone loss, and synovial hyperplasia that eventually lead to the destruction of cartilage and bone [78]. RANKL is highly expressed in the synovium of RA patients. RANKL-mediated generation of osteoclasts is responsible for RA-related bone destruction [79, 80]. The expression of RANKL is detected in synovial fibroblasts and inflammatory cells such as monocytes and T cells isolated from the synovial fluid [81–84]. Although T cells extensively infiltrate the inflamed synovium and, therefore, T cell-derived RANKL was thought to contribute to the pathogenesis of RA by facilitating osteoclastic activity, T cell-derived RANKL can be counteracted by T cell-derived anti-osteoclastogenic factors such as IFN-γ, IL-10, IL-4, and OPG [85–87]. Instead, it appears that synovial fibroblasts are the primary source of RANKL in joints with RA [88]. It has further been determined that Th17 cells, which promote autoimmune diseases by producing IL-17, are a subset of RANKL-expressing T cells and are responsible for inducing RANKL expression in synovial fibroblasts via the production of IL-17 as well as IL-1, TNF-α, and IL-6 [79, 89]. Additionally, Th17 has also been reported to stimulate mature non-resorptive osteoclasts to begin resorbing bone via cell–cell contact, which can be attenuated by RANKL-neutralizing antibodies. This suggests that RANKL is not merely a factor in the terminal differentiation of osteoclasts but also plays a role in regulating the bone-resorptive function of fully matured osteoclasts [90].

Recent studies have also identified a mechanism of B cell-mediated pathogenesis in autoimmune diseases including RA. The implication of B cell-mediated pathogenesis in RA was obtained by studies of B cell depletion therapy, which has been used effectively to treat many autoimmune diseases [91]. Several studies have shown that B cell depletion with monoclonal antibodies against CD20 such as rituximab significantly reduces clinical symptoms and inflammation in RA alongside inhibiting disease progression by increasing bone formation and decreasing bone resorption [92–94]. In RA, activated T cells that recognize autoantigens help B cells differentiate into plasma cells that produce autoantibodies such as anti-cyclic citrullinated peptide antibodies (ACPAs) and rheumatoid factors, which are known to promote osteoclastogenesis [95–97]. However, the therapeutic effect of B cell depletion may not be solely due to the depletion of B cells/plasma cells producing autoantibodies. In fact, a report showed no correlation between changes in levels of autoantibodies and the clinical responses [92]. In addition, several studies have reported that plasma cells lack expression of CD20 [96, 98, 99]. Rather, the therapeutic effect of B cell depletion may be due to the blunting production of immunoregulatory cytokines. B cells in the RA synovium are capable of producing pro-inflammatory and bone-destructive cytokines, including TNF-α, IL-6, and RANKL [100–102]. RANKL expression is detected in Fc receptor like 4 (FcRL4)⁺ B cells enriched in the synovium that also express CD11c, CD20, CD95, and costimulatory molecules CD80 and CD86 [101, 103, 104]. RANKL expression in B cells is triggered by activation via B-cell receptor and CD40, and it is further augmented by IFN-γ, which may be produced by T cells enriched in RA or B cells [104–106]. In a murine model of arthritis induced by proteoglycan, B cell-specific IFN-γ-deficient mice are resistant to the induction of arthritis, suggesting a critical role of B cell-derived cytokines in the pathogenesis of RA [105]. Recently, a report showed that RANKL-expressing plasma cells are increased in the bone marrow of arthritic mice and exhibit the capacity to induce osteoclastogenesis. RANKL in B-lineage cells has been shown to be required for periarticular bone loss in the murine model of autoimmune arthritis, suggesting the importance of RANKL in plasma-cells in periarticular bone loss [107].

Clinical evidence suggests that aberrant RANK signaling in B cells contributes to the induction of B cell autoimmunity and malignancy [108–112]. Somatically acquired mutations of RANK in which the intracellular signaling domain Lys at amino acid 240 changes to Glu (RANK^K240E) have been detected in human diffuse large B cell lymphoma specimens. This mutation has been suggested to lead to gain-of function mutations, and the forced expression of RANK^K240E variant in murine B cells disrupted B-cell tolerance and induced a fully penetrant systemic lupus erythematosus (SLE)-like disease in addition to the development of chronic lymphocytic leukemia (CLL) in mice [113]. SLE is one of B-cell mediated autoimmune diseases that result from disrupted B cell tolerance. These observations suggest that B cell-intrinsic aberrant RANK signaling is sufficient to disrupt B-cell tolerance in vivo, driving SLE-like autoimmune disease with the progression of B cell malignancy.

The central nervous system (CNS) is an immune privileged site due to the blood-brain barrier (BBB) protection, which is mainly composed of tight junction. It seals the endothelium of meningeal vessels and prevents the diffusion of cells and large molecules from the blood into the cerebrospinal fluid (CSF) [114, 115]. The choroid plexus in the brain separates the peripheral blood and CSF. The choroid plexus has a unique structure; it is composed of a layer of fenestrated epithelium. T cells enter into the CSF for normal immune patrolling exclusively through the choroid plexus fenestrated epithelium under physiological conditions. However, in case of CNS inflammation, the vascular endothelium loses its barrier properties and recruits leukocytes into the CNS unselectively [116]. T cells that have entered CNS infiltrate into the CNS parenchyma when they are reactivated by the resident antigen-presenting cells. RANKL–RANK signaling has been revealed to be critical in controlling T-cell migration in the CNS under inflammation. RANKL on T cells mediates the interaction between T cells and RANK-expressing astrocytes, induces the secretion of the C-C type chemokine ligand 20 (CCL20) by astrocytes, promotes the infiltration of the C-C motif chemokine receptor 6 (CCR6)⁺ T cells into the CNS parenchyma, and induces EAE pathogenesis [117].

Periodontitis is one of common infectious diseases triggered by species residing within the bacterial biofilm on the tooth surface [118]. Periodontal infection is characterized by inflammation of periodontal tissues, followed by osteoclast-mediated alveolar bone destruction, leading to tooth loss. The osteoclastic bone damage has been shown to depend on RANKL [118]. Periodontal tissue inflammation recruits a number of inflammatory infiltrates with inflammatory cytokines such as IL-1, TNF, IL-6, and IL-17 [119, 120], generating Foxp3⁺ T cell-derived Th17 (exFoxP3Th17) cells that have been shown to crucially contribute to the pathogenesis of asthma and rheumatoid arthritis [121, 122]. exFoxP3Th17 cells and inflammatory cytokines induce RANKL expression by osteoblasts, osteocytes, and periodontal ligament cells, resulting in osteoclast generation and subsequent bone loss. Memory B cells have also been shown to be involved in periodontal bone loss through the expression of RANKL [123]. OPG, which is expressed in osteocytes in periodontal tissues, can be proteolytically cleaved by oral-bacterial and osteoclastic proteases, resulting in the promotion of osteoclastogenesis [124–126]. Both the degradation of OPG and the increasing expression of RANKL may contribute to osteoclastic bone damage and periodontal bone loss.

RANKL in tumorigenesis, metastasis, and tumor microenvironment

Cumulative evidence reveals importance of RANKL and RANK in tumor development from the initial oncogenesis process to the establishment of the distant metastasis. The RANKL–RANK pathway is often overexpressed in tumors, including breast, lung, endometrial, renal cell, and gastric cancer, and it is correlated with poor prognosis [127]. Disturbance in coordinated RANKL–RANK axis has been shown to contribute to tumorigenesis. Overexpressing RANK protein under the control of mouse mammary tumor virus (MMTV) promoter in mammary glands resulted in the formation of pre-neoplasia foci and increased mammary tumor formation. By contrast, RANKL inhibition by RANK-Fc attenuated mammary tumor development [128]. Genetically engineered mice carrying breast cancer 1 (Brca1) and p53 mutation have been shown to generate malignant tumors, which can be prevented by the inactivation of the RANKL–RANK pathway [129]. The overexpression of RANK has also been shown to induce stemness and epithelial-mesenchymal transition (EMT), a cellular program that is critical in the progression of cancers, in mammary epithelial cells [130–133]. These data revealed that RANKL–RANK signaling promotes the initial stage in breast cancer development.

The dissemination of cancer cells from sites of primary tumor growth is responsible for metastasis, which occurs in the bone most frequently. The disseminated tumor cells form a favorable microenvironment for the development of the metastatic spread [134]. In the bone tissue tumor microenvironment, cancer cells produce the parathyroid hormone-related peptide (PTHrP), which alters RANKL/OPG expression [135, 136]. They also directly express RANKL and pro-osteoclastogenic factors such as IL-1α, IL-6, MCSF, or prostaglandin E2 (PGE2) [137, 138]. All of them promote osteoclast differentiation and survival, resulting in local osteolysis, which supports metastatic growth. Afterwards, growth factors such as insulin-like growth factor (IGF) and transforming growth factor-β (TGF-β) and calcium are released from bone, stimulating the further growth of cancer cells. This “vicious cycle” accelerates osteolytic metastasis and cancer cell proliferation [139]. Bone metastasis increases the risk of skeletal-related events (SREs) such as pathologic fracture, spinal cord compression, and necessity for radiation or surgery for the bone. RANKL regulates bone metastasis through the stimulation of the migration of cancer cells to the bone. RANKL acts directly on RANK-expressing tumor cells and promotes cell migration toward the bone [140–143]. Blocking RANKL through the administration of OPG or AS2676293, a small-molecule inhibitor of RANKL, attenuates bone metastasis of breast cancer cells and malignant melanoma [144, 145]. Indeed, the level of RANK expression in primary breast cancer and renal cell carcinoma has been reported to be positively correlated with the frequency of bone metastasis in human clinical studies [142, 146]. RANKL is a membrane-bound cytokine, but it is also proteolytically cleaved by metalloproteinases and its soluble form is produced. A recent study demonstrated that the soluble form of RANKL is responsible for bone metastasis by promoting the migration of RANK-expressing tumor cells without affecting bone resorption [147] (Figure 2).

Figure 2.

RANKL regulates osteolytic bone metastasis through the stimulation of cancer cell migration toward the bone. RANKL acts directly on RANK-expressing tumor cells and promotes cell migration. Tumor cells induce RANKL expression on bone marrow stromal cells, leading to osteoclastic bone resorption. Increased bone resorption resulted in the release of factors that promote tumor growth (vicious cycle). Denosumab prevents osteoclastic bone destruction. Denosumab also has the potential to treat tumors in combination with immune checkpoint inhibitors.

The tumor microenvironment protects cancer cells by modulating anti-cancer immune responses, thereby allowing the tumor cells to escape immune surveillance. Tumor cells develop several strategies to evade immune surveillance: reducing infiltration by cytotoxic T cells or natural killer (NK) cells and increasing the recruitment of immune suppressive cells such as CD4⁺Foxp3⁺ regulatory T cells (Tregs), tumor associated macrophages (TAMs), and tumor-associated neutrophils (TANs). In the tumor microenvironment, immune cells that express RANKL and RANK are commonly found, and in many situations, they are important drivers to create an immunosuppressive microenvironment, thereby promoting tumor progression. In the tumor microenvironment, RANKL expression is detected by infiltrating T lymphocytes. In the mouse model of breast carcinoma, the main intratumoral source of RANKL has been revealed to be Tregs [148]. On another front, RANK expression has been detected in TAMs, myeloid-derived suppressor cells (MDSCs), and dendritic cells (DCs) in various mouse tumor models and in human tumors [149–152]. In human extramammary Paget’s disease, a rare intraepithelial adenocarcinoma, M2-macrophage-type TAMs accumulate in the microenvironment, and the RANKL ligation on RANK expressed on M2 macrophages activates the production of a series of chemokines that recruit immunosuppressive cells such as Tregs and Th2 cells, which thereby induces the maintenance of the tumor microenvironment [153]. MDSCs in the bone marrow of mice bearing mammary cancer and bone metastasis have been shown to be converted to immunosuppressive osteoclast-type cells through RANK signaling [154, 155]. Although the MDSC-derived osteoclast-type cells have not been fully characterized, given the rich expression of RANKL in bone marrow niche, MDSCs could potentially promote tumor development by contributing to a “vicious cycle” of continued bone resorption and the resultant release of pro-tumorigenic factors such as TGF-β, which suppress anti-tumor immunity. The role of RANK-expressing DCs in cancer is not well defined. DCs had been initially suggested to promote their survival and conduct T-cell stimulatory activities through ligation with RANKL. However, a tolerogenic role of RANK-expressing DCs has been demonstrated in diverse non-malignant pathological and physiological settings [2, 68, 156–158], implying a potential tolerogenic effect of RANK signaling in DCs in the tumor microenvironment. Tumor expressing-RANK has recently been demonstrated to play a role in the crosstalk between tumor cells and immune cells in the tumor microenvironment [159]. RANK deficiency in the mouse oncogene-driven mammary tumor model (mouse mammary tumor virus-polyoma middle tumor-antigen: MMTV-PyMT) revealed that tumor transplants from the RANK^−/− background had a higher immune infiltration, enriched in an anti-tumoral population such as CD8⁺ T cells and with lower macrophage and TANs infiltration than that from the control (RANK^+/+) background [159]. RANK^+/+ tumor cells promoted immune suppression through TANs, while RANK^−/− tumor cells activated CD8⁺ T cell-mediated anti-tumor response [159]. These observations suggest that RANK signaling in tumor cells promotes the establishment of an immunosuppressive environment.

Targeting the RANKL–RANK axis: The application of the anti-RANKL antibody

The discovery of an obligate role of RANKL–RANK signaling in osteoclastogenesis led to the development of RANKL inhibitors, which culminated in the development of the RANKL-targeted antibody denosumab [160]. Denosumab is a fully human IgG2 monoclonal antibody that binds human RANKL with a high affinity and blocks RANKL from binding to and oligomerizing its receptor RANK, thereby suppressing osteoclastogenesis and inhibiting bone resorption [160]. Denosumab binds both soluble and membrane-bound primate RANKL but fails to recognize mouse or rat RANKL [160].

Post-menopausal osteoporosis is the most common bone disease, characterized by bone microarchitecture deterioration, low bone mass, and increased bone fragility due to an imbalance in osteoclast activity versus new bone formation [161, 162]. Osteoporosis occurs due to an alteration in hormone expression, nutrition, mobility, and/or senescence, and it has been linked to increased RANKL levels on bone marrow cells of women exhibiting osteoporosis [24]. Denosumab has been shown to be a potent and safe therapy for osteoporosis in many clinical trials (Figure 3). In the phase III FREEDOM clinical trial, denosumab significantly reduced osteoporotic fractures in post-menopausal women with osteoporosis [80]. Denosumab has superior pharmacological properties compared with other RANKL inhibitor OPG-Fc or osteoclast inhibitor bisphosphonates. Long-term denosumab administration induces a continuous increase in BMD as well as a reduction in cortical porosity, which are not observed with classical antiresorptive drugs [161]. Denosumab has been approved by the FDA at the subcutaneous dose of 60 mg every 6 months for the treatment of osteoporosis.

Figure 3.

RANKL–RANK binding increases osteoclast formation, activity, and survival. Human anti-RANKL monoclonal antibody denosumab has been clinically used for treatment of osteoporosis and prevention of skeletal-related events (SREs). Several studies have shown the potential therapeutic benefits of denosumab for osteoclastic bone loss such as rheumatoid arthritis.

In bone metastasis, RANKL contributes to bone pathogenesis by driving the vicious cycle as an essential cytokine for osteoclastogenesis. Denosumab has been demonstrated to have a remarkable effect on the blockade of SREs in patients with bone metastasis and multiple myeloma. In addition, denosumab acted against SREs in patients with breast cancer and prostate cancer more efficiently than bisphosphonates [163, 164]. Denosumab is FDA approved at a dose of 120 mg subcutaneously every 4 weeks for the prevention SREs including fracture, spinal cord compression, orthopedic surgery interventions, and palliative radiation to the bone in patients with multiple myeloma (MM), bone metastasis from solid tumors, and giant cell tumor of bone (GCTB) (Figure 2).

Given the essential role of RANKL in RA-associated bone lesions, anti-RANKL therapy for the prevention of RA-related bone destruction has been attempted. The effects of denosumab in patients with RA have been analyzed in clinical studies (phase II clinical trial of the effects of denosumab, phase III DESIRABLE study), revealing its protective effects on RA-associated bone lesion as well as its impact on increasing BMD [80]. Denosumab is considered a putative therapeutic option for RA, and it is approved for “inhibition of the progression of bone erosion associated with RA” in Japan [80] (Figure 3).

Recent data has also proposed the possibility of further repurposing denosumab from an anti-resorptive agent in bone diseases to cancer immunotherapy in combination with checkpoint inhibition. Immune checkpoint inhibitors (ICI) such as anti-CTLA4 antibody, anti-PD-1 antibody, and anti-PD-L1 antibody can restore and augment anti-tumor immune responses by blocking the immune-suppressive molecules, and these are used clinically for the treatment of many cancers. A few case reports of patients with metastatic melanoma treated with denosumab, in combination with ipilimumab (anti-CTLA4 antibody) and pembrolizumab (anti-PD1 antibody), showed significant therapeutic efficacy, suggesting that RANKL inhibition enhances anti-tumor effects [165, 166]. Additionally, retrospective case series revealed the possible synergism of denosumab with ICI in the treatment of melanoma [167, 168]. Enhancement of ICI in combination with the anti-RANKL antibody has also been reported using mouse models of melanoma, prostate cancer, and colon adenocarcinoma cell line [150, 169]. These findings support that RANKL inhibition may enhance the activity of ICI, leading to improvement in the capacity to control tumor in patients. Additionally, the D-BEYOND clinical trial, where the effects of denosumab on tumor immune infiltration were evaluated, showed that denosumab treatment increased tumor infiltrating lymphocytes (TILs) in particular CD8⁺ T cells [159]. These observations resemble the recent preclinical model, which revealed that tumor cells exploit the RANK pathway as a mechanism to evade immune surveillance. The inhibition of the RANKL–RANK pathway by denosumab would lead to an increased infiltration of the anti-tumoral immune population. The immunomodulatory effects of denosumab have raised concerns regarding the risk of malignancy, although no specific toxicity data has been published for denosumab when combined with ICI. A recent systematic review and meta-analysis reported no association of denosumab with the risk of malignancy [170]. RANKL inhibition is unlikely to prohibitively exacerbate immune-related adverse effects of immunotherapies. However, this issue requires to be evaluated in early phase clinical trials.

An increased risk of fractures, called rebound vertebral fractures (RVFs), following denosumab discontinuation has been identified [171–174]. The RVFs were spontaneous and occurred shortly after treatment discontinuation in patients with an estimated low risk of facture. The RVFs have been shown to be associated with a vast increase in osteoclast number and activity, which leads to a subsequent profound increase in bone turnover [175]. This new side effect had not been previously associated with other antiresorptive therapies. The mechanisms underlying RVFs have not been identified yet. One possible explanation is that immature bone-unresorbing osteoclasts accumulate in the bone tissue during the robust inhibition of RANKL by denosumab, which leads to a mass increase in osteoclastogenesis and RANKL release after the administration of denosumab is stopped [176]. A recent report described that blocking RANKL-signaling with OPG-Fc cumulates the “osteomorph” population that is generated by osteoclast fission. Osteomorphos are fusion competent and motile and capable of forming bone-resorbing osteoclasts [177]. Cessation of OPG-Fc treatment and RANKL release resulted in the reduction of osteomorphs along with a rebound increase in osteoclasts [177], implying that the accumulation of osteomorphs during the blocking of RANKL signaling contributes to recycling osteomorphs back in osteoclasts, accelerated bone loss, and pathological vertebral fractures observed in RVFs. The molecular and cellular mechanisms underlying the effects of prolonged denosumab treatment and its withdrawal remain to be addressed.

Conclusions

The discovery of the RANKL-RANK-OPG pathway has not only tremendously advanced the understanding of mechanisms regulating osteoclast differentiation and bone homeostasis but also uncovering essential roles of the triad in the immune system, significantly contributing to emergence of the field of osteoimmunology and the development of a fully human anti-RANKL antibody that is now clinically available for treatment of osteoporosis and cancer-induced bone loss. However, it will be important to answer additional questions raised by emerging findings: What is the impact of the RANKL–RANK reverse signaling in regulating the development of the immune system and cancer? What is the relevance of blocking RANKL–RANK signaling by denosumab in central tolerance and anti-tumor immunity? Further research will refine the benefit of targeting the RANKL–RANK pathway as a therapeutic strategy for additional disease conditions.

Fig. 4.

RANKL–RANK binding increases osteoclast formation, activity, and survival. Human anti-RANKL monoclonal antibody denosumab has been clinically used for treatment of osteoporosis and prevention of skeletal-related events (SREs). Several studies have shown the potential therapeutic benefits of denosumab for osteoclastic bone loss such as rheumatoid arthritis.

Table 1.

Summary of dysregulation of RANKL-RANK axis and diseases

The system	Diseases	Dysregulation of RANKL-RANK axis	References
Bone	Osteoporosis	Increased RANKL levels on bone marrow cells	26
	Osteopetrosis	Genetic mutation of RANKL	28
	Bone metastasis	Altered RANKL/OPG expression by cancer cell-derived PTHrP, which contributes to osteoclastic osteolysis	139, 140
		Cancer cell-derived RANKL which promote osteoclastic osteolysis	141, 142
		RANK expression on tumor cells which contributes to tumor cell migration toward the bone	144, 145, 146, 147
	Rheumatoid arthritis (RA)	Increased RANKL levels in synovial fibroblasts	85, 87, 88, 92
The immune system	Rheumatoid arthritis (RA)	Synovium infiltrating inflammatory cell (monocytes, Th17 cells)-derived RANKL	85, 86
		Activated B cell-derived RANKL	104, 105, 106, 108, 111
	Systemic lupus erythematosus (SLE)	Gain-of function mutation of RANK in B cells in mice	117
	Chronic lymphocytic leukemia	Gain-of function mutation of RANK in B cells in mice	117
	Tumor progression and microenvironment	Expression of RANKL and/or RANK in immune suppressive cells such as Tregs, TAMs and TANs	152, 153, 154, 155, 156
		Crosstalk between CD8+ T cells and RANK-expressing Tumor cells, which promotes establishment of an immunosuppressive environment	163
Oral cavity	Periodontitis	Inflammation-induced RANKL expression by osteoblasts, osteocytes and periodontal ligament cells	123, 124
		RANKL production from memory B cells in experimental periodontitis in mice	127
		Proteolytic degradation of OPG by oral-bacterial and osteoclastic proteases	128, 129, 130
Mammary gland	Cancer	Promoted RANKL-RANK signaling in BRCA1/2 mutation-driven breast cancer	132, 141