Cytokine-mediated immunomodulation of osteoclastogenesis

Abstract

Cytokines are an important set of proteins regulating bone homeostasis. In inflammation induced bone resorption, cytokines, such as RANKL, TNF-α, M-CSF, are indispensable for the differentiation and activation of resorption-driving osteoclasts, the process we know as osteoclastogenesis. On the other hand, immune system produces a number of regulatory cytokines, including IL-4, IL-10 and IFNs, and limits excessive activation of osteoclastogenesis and bone loss during inflammation. These unique properties make cytokines powerful targets as rheostat to maintain bone homeostasis and for potential immunotherapies of inflammatory bone diseases. In this review, we summarize recent advances in cytokine-mediated regulation of osteoclastogenesis and provide insights of potential translational impact of bench-side research into clinical treatment of bone disease.

1. Introduction

Bone homeostasis is critical for health and disease. As an organ harboring a variety type of cells, bone is continuously remodeling through the dynamic equilibrium between osteoclasts, responsible for bone resorption, and osteoblasts/osteocytes, responsible for bone formation. This process is tightly regulated by local and systemic factors, including the body's innate and adaptive immune cells, through direct interaction and via intermediate cellular response. Interrupted bone and skeletal homeostasis often lead to diseases, such as osteoporosis, osteoarthritis (OA), rheumatoid arthritis (RA) and Paget disease, which immensely affect the quality of life and increase the risk of mortality in patients. Notably, osteoporosis is one of the most common degenerative diseases in middle-aged and elderly people, causing a vast health burden worldwide. Characterized by significant bone loss and deterioration of skeletal microstructure, osteoporosis results in weak bone loading function and a significantly increased risk of bone fracture. Nearly one third of women and 1/5 men aged over 50 are suffering from or have a higher chance of experiencing fragility fracture. These account for almost 9 million fragility fractures that occur each year, involving nearly 1.6 million hip fractures, which can cause life-threatening complications [1].

The crosstalk between skeletal and immune systems is essential for the onset and development of osteoporosis [2]. Indeed, osteoimmunology, a concept coined in 2000, has revolutionized our understanding of the osteoporosis pathophysiology and demonstrated a key role of inflammatory cytokines in influencing the fine-tuned balance between bone formation and resorption [3,4]. Simply put, excessive bone resorption caused by hyperactivated osteoclasts (OC) impairs the normal bone remodeling which laid on the foundation of the balance between osteoclasts and osteoblasts, resulting in the significant bone loss in osteoporosis [5]. These osteoclasts are originated from the myeloid cell progenitors, particularly monocytes and macrophages, which are part of innate immune cells (Fig. 1). During osteoclast differentiation, or termed as osteoclastogenesis, myeloid cells receive signals guided by key cytokines such as receptor activator of nuclear factor kappa-B ligand (RANKL), and undergo transcriptional alterations with the ultimate fate as the multinucleated mature osteoclast cells. Adaptive immune cells, mainly T cells, either promote this process or inhibit it from osteoclastic onset of differentiation. Activated T cells are one of the major sources of receptor activator of nuclear factor kappa-B ligand (RANKL) that is essential for osteoclastogenesis during tissue inflammation. Conversely, regulatory T (T_REG) cells inhibit osteoclastogenesis and subsequently reduce bone loss through either contact-directed or intermediate cytokine-modulated regulation, or a combination of both [6,7]. These indispensable connections between skeletal and innate and adaptive immune systems provide a unique perspective to understand the biology of bone remolding, which ultimately allows us to modulate osteoclastogenesis by immunomodulation to treat bone diseases.

Fig. 1. Immune cells produce cytokines to modulate osteoclastogenesis.

Both innate and adaptive immune cells regulate osteoclastogenesis by production of cytokines. T_H17 cells produce RANKL and IL-17 during bone inflammation in diseases such as osteoporosis and rheumatoid arthritis. Macrophages are also the major source of proinflammatory cytokines such as IL-1β and TNF-α. Cytokines produced by these immune cells markedly enhanced the activation of osteoclastogenesis. On the other hand, regulatory T (T_REG) cells, T_H2 cells and innate lymphoid cells are counterpart of this process and negatively regulate osteoclastogenesis by producing cytokines such as IL-10, IL-4, IL-5 and IL-13. These immune cells in together, maintain the dynamics of bone remodeling and are key immunomodulators to treat inflammatory bone diseases.

Cytokines are a broad range of secreted proteins necessary for cell signaling. Through the interaction with cell surface receptors, cytokines initiate complex downstream signaling cascades that can direct cell proliferation, survival, differentiation and metabolism. Immune cells are one of the major sources of cytokines. As important bone-affecting factors, cytokines maintain bone homeostasis by direct modulation of bone forming-osteoblasts and bone resorbing-osteoclasts. The excess or insufficiency of key cytokines, for instance RANKL, tumor necrosis factor alpha (TNF-α), and IL-6, are often considered as major immune-related factors causing inflammatory bone diseases. For example, in osteoporosis, a number of proinflammatory cytokines, such as RANKL, IL-1β, TNF-α, IL-6 and IL-17, are increased, which rapidly promotes osteoclast formation. On the other hand, the presence of regulatory cytokine IL-10, which negatively regulates OC development, is often declined in osteoporosis (Table 1). These abnormalities, together, tip the balance towards the inflammatory bone loss and the severe illness of osteoporosis. To be noted, most of these pro-inflammatory cytokines were also increased in patients infected with SARS-CoV-2 [8,9]. Therefore, further understanding of cytokines' role in regulating bone remodeling also enables us to adapt to the fast-ever-changing global health situation in response to COVID-19 and have necessary knowledge available to prevent and treat potential infection induced bone loss. Cytokines are also implemented in clinical diagnosis or evaluation of treatment efficacy in patients with acute or chronic bone disorders. More importantly, beyond diagnosis, emerging progress has been made in developing novel immuno-therapeutics either straightforwardly using cytokines as an immunologic agent or targeting them to restore the balance between bone formation and bone resorption [10]. These potential cytokine-based immunotherapeutics can provide broader options in addition to the current treatments in patients with osteoporosis or broad bone disorders, which are mainly chemical or physical based approaches.

Table 1.

Cytokines in osteoclastogenesis.

Regulation	Factor	Main source	Action	Reference
Positive regulation	RANKL	Osteoblasts, osteocytes, TH17, ILC3	Induces osteoclastogenesis, survival, proliferation and fusion.	[19-38]
	M-CSF	Osteoblasts, MSCs	Supports osteoclast precursor survival, proliferation and fusion.	[2,44-49]
	IL-1β	Monocytes, macrophages	Induces osteoclastogenesis by RANK/RANKL signaling via TRAF6 and increases RANKL expression.	[50-56]
	IL-7	Stromal cells, DC	Induces osteoclastogenesis by inducing STAT5 activation; enhance the production of TNF-α and RANKL by T cells.	[82,83]
	IL-8	Macrophages	Induces osteoclastogenesis by inducing RANKL-induced NFATc1 activation.	[84-89]
	IL-15	Monocytes, macrophages	Induces osteoclastogenesis by activation of RANK/RANKL signaling; induces RANKL expression.	[92-97]
	IL-17	TH17, ILC3	Induces RANKL, IL-1β, IL-6, IL-23 and TNF-α expression.	[5,39,57-63]
	TNF-α	Macrophages, T/B/NK cells	Induces M-CSF, RANKL and RANK expression.	[48,69-80]
Negative regulation	IL-4	TH2, ILC2	Inhibits osteoclastogenesis by affecting NF-kB activation in a STAT6-dependent manner or modulating Ca2+ signaling and decreases IL-1, IL-6, TNF-α expression.	[27,28,102,106-109]
	IL-5	TH2, ILC2	Inhibits osteoclastogenesis.	[110]
	IL-9	TH2, ILC2	Inhibits osteoclastogenesis.	[100]
	IL-10	Macrophages, DC, regulatory T/B cells	Inhibits osteoclastogenesis by reducing NFATc1 expression and level of RANKL and M-CSF, and upregulating OPG expression; suppresses IL-1β and TNF-α expression.	[6,7,116-120]
	IL-12	Macrophages, monocytes, DC, B cells	Inhibits osteoclastogenesis by inducing Fas/FasL or NO mediated apoptosis.	[131-134]
	IL-13	TH2, ILC2	Inhibits osteoclastogenesis.	[28,101,108]
	IL-27	Macrophages, monocytes, DC, B cells	Inhibits osteoclastogenesis by suppressing NFATc1-AP-1-MAPK signaling pathway; inhibits STAT1 dependent c-Fos activation; downregulation of RANKL and TREM-1 expression.	[135-137]
	IFN-β	Macrophages, DC	Inhibits osteoclastogenesis by c-Fos dependent RANK signaling via TRAF6 and suppresses RANKL expression; increases CXCL11 expression.	[26,126-129]
	IFN-γ	T cells, NK	Inhibits osteoclastogenesis by RANK/RANKL signaling via TRAF6.	[3,138-142]
Ambiguous regulation	IL-6	Macrophages, monocytes	Regulates RANK/RANKL signaling via JAK/STAT3 and increases RANKL expression; Suppresses the osteoclast precursor differentiation via NF-kB signaling.	[145,146] [147]
	IL-23	Macrophages, monocytes, DC, B cells	Promotes osteoclast formation by upregulation of RANK expression; activation of DAP12; induces RANKL and IL-17 production.	[150-153]
			Inhibits osteoclastogenesis by induction of GM-CSF.	[154,155]
	TGF-β1	Macrophages, monocytes, T/B/NK/DC cells	Regulates osteoclast differentiation via SMAD2/3, which mediated by TRAP6-TAB1-TAK1 complex at low-concentration or early stage of differentiation	[161-163,165,166]
			Inhibits osteoclastogenesis through regulating RANKL/OPG ratio secreted by osteoblasts.	[161,163,164,166]

T_H2: Type 2 helper T cells, ILC2: Group 2 innate lymphoid cells; T_H17: Type 17 helper T cells; ILC3: Group 3 innate lymphoid cells; NK: natural killer cells; DC: dendritic cells.

A key drawback of current medicines on osteoporosis is the severe adverse effect. The risk of potential side effects is particularly higher in those who require long-term clinical therapy [1]. To date, anti-resorptive drugs bisphosphonates and RANKL monoclonal antibody (Denosumab) are most used medications in clinical practice, taking advantages of their relative effectiveness, long-term tolerance and affordable price [11]. However, studies suggest that long-term use of these medicines might induce serious side effects, although rare, including atypical leg fractures, osteonecrosis of the jaw, and esophageal cancer, cardiovascular disorder, osteoarthritis, hypercalcemia and hypocalcemia [12]. Moreover, anabolic drugs parathyroid hormone (NATPARA) and sclerostin monoclonal antibody (Romosozumab) were approved by FDA as additional treatments for osteoporosis in postmenopausal women with high risk of fracture [13-17]. Notwithstanding, potential issues such as osteosarcoma, osteonecrosis and relatively high cost still exist, making these medications “second-line” treatments in patients with severe osteoporosis [13,15]. Moreover, these medications were developed to broadly turn down the osteoclastogenesis and activated immune response, which might induce immunosuppression with increased risk of cancer and opportunistic infection [18]. Therefore, additional and alternative novel clinical treatment options are greatly needed.

In this review, we focus on osteoimmunological cytokines and summarize the advances and current knowledge of cytokine-mediated immunomodulation of cellular and molecular process in osteoclastogenesis. In particular, we comprehensively review how cytokines regulate osteoclastogenesis, a key process for bone resorption. In addition, we will provide insights on how the current studies and their findings have the potential to translate into clinical design of novel immunotherapeutics, and how they might benefit the patients with bone-related disorders.

2. Positive regulation of osteoclastogenesis by cytokines

2.1. RANKL

The receptor activator of nuclear factor kappa-B ligand (RANKL) is a member of the tumor necrosis factor (TNF) superfamily of ligands and instrumental for the ultimate differentiation of osteoclast precursors into osteoclasts. RANKL is secreted or expressed by both immune cells such as type 17 helper T (T_H17) cells and skeletal cells such as osteoblast and osteocytes. RANKL interacts with its receptor RANK on osteoclast surface, where trimerized RANK-RANKL complex recruits downstream adaptor molecules such as tumor necrosis factor (TNF) receptor-associated factor (TRAF) 6 to its cytoplasmic domain, resulting in the activation of multiple signaling pathways including IKK complexes (IKKα, IKKβ, IKKγ, and NIK-IKKα) and mitogen-activated protein kinases (MAPKs: extracellular signal-regulated kinase (ERK), p38, and c-Jun terminal kinase (JNK)) in premature osteoclasts [19,20]. Together with activated CaMKIV-CREB signaling pathway, RANKL-transduced signaling cascades promote AP-1 activation, followed by the robust induction of the key transcriptional factor of osteoclastogenesis, nuclear factor of activated T cells cytoplasmic 1 (NFATc1) (Fig. 2) [20-23]. Cytosolic calcium, which can be regulated by signaling receptors including FcγRIII, PIR-A, OSCAR, TREM2 and SIRPβ1, is also essential to activate NFATc1 [5] [24,25]. NFATc1 promotes the expression of Blimp1, which represses anti-osteoclastogenic genes, such as Mafb, BCL-6 and IRF8 [23]. In addition, STAT family members are also critical during the RANKL-mediated osteoclastogenesis. For instance, interferon-β negatively regulates osteoclast differentiation through the STAT1 pathway [19,26]. IL-4 abrogates osteoclastogenesis through STAT6-dependent suppression of NF-kappaB [27,28]. Moreover, STAT5 negatively regulates osteoclastogenesis, by suppressing Dusp1 and Dusp2, two key phosphatases in MAPK pathway [22]. Additionally, sialic acid-binding immunoglobulin-like lectins family molecule Siglec-15 was shown to regulate RANKL induced-osteoclastogenesis [29-31]. Deciphering how cytokines might affect these molecules during osteoclastogenesis could help us develop further treatments to intervene the OC formation.

Fig. 2. Critical signaling pathways activated by cytokines in osteoclasto-genesis.

RANKL binds to RANK which recruits TRAF6 to activate signaling pathways including MAPK, IKK and Ca⁺/CAMKIV. These signaling cascades ultimately induces the translocation and activation of key transcriptional factors such as NFATc1 that targets and initiates the expression of osteoclastogenic genes. Cytokines such as TNF-α, MCSF and IL-1β facilitate osteoclastogenesis by activating other critical signaling pathways such as JAK/STAT, PI3K, NF-κB to promote the maturation, differentiation and survival of osteoclasts.

RANKL not only stimulates the differentiation of osteoclast but also accelerates bone resorption by enhancing function of mature osteoclasts and prolonging their survival [32]. Indeed, mice with injected soluble RANKL showed increased measurable bone resorption within 60 min [33]. Notably, RANKL signaling can be potently suppressed through the negative feedback action coordinated by osteoprotegerin (OPG), a secreted receptor also belonging to the TNF receptor superfamily. OPG, largely produced by bone marrow mesenchymal stem cells (MSCs), osteoblast lineage cells, B cells and dendritic cells, functions as a decoy receptor for RANKL. OPG binds to RANKL to prevent its interaction with RANK receptor [34], thereby limiting osteoclastogenesis. RANKL/OPG ratio is therefore recognized as the one of the hallmarks for patients with osteoporosis and other various bone related diseases, including rheumatoid arthritis, Paget's disease, metastatic bone disease as well as glucocorticoid-induced bone loss [35,36]. In addition to the secretion of OPG, bone marrow MSCs also regulates the OC formation through the signaling involves Cyclin-dependent kinase 8 (CDK8), by which the alteration of CDK8 level resulted in changed STAT1 signaling downstream of RANKL induced osteoclastogenesis [37,38].

T_H17 cells promote osteoclastogenesis in inflammation-induced osteoporosis and rheumatoid arthritis, amongst many RANKL-expressing activated CD4⁺ T cells [39]. This cellular regulation is elegantly coordinated by both secreted RANKL as well as surface expression of RANKL on T_H17 cells that can lead to a short but potent synapse-like cell-cell interaction [40]. Importantly, IL-17, the signature cytokine produced by T_H17 cells synergizes RANKL produced by T_H17 cells and subsequently promotes osteoclastogenesis and bone resorption. Interestingly, a subset of CCR6⁺ innate lymphoid cells also expresses RANKL on cell surface [41]. Although lacking antigen specific receptors, these group 3 innate lymphoid (ILC3) cells act similarly to their T cell counterparts, T_H17 cells, which produce IL-17 and IL-22 in response to IL-1b and IL-23 stimulation during tissue inflammation in gut, lung and skin. However, whether and how these RANKL expressing cells regulate osteoclastogenesis in bone microenvironment remains unaddressed.

2.2. M-CSF

M-CSF plays an important role during the early proliferation, survival and differentiation of osteoclast precursor cells [42,43]. Produced by various cell populations derived from mesenchymal cells, M-CSF binds to its receptor CD115 (c-Fms) and initiates the downstream signaling pathways including phosphoinositide 3 kinase (PI3K), phospholipase C gamma (PLCγ), as well as extracellular signal-regulated kinases 1 and 2 (ERK1/2) signaling (Fig. 2) [2]. These signaling pathways are indispensable for the development of osteoclast precursors during early differentiation.

Indeed, animal studies found that mice with depleted M-CSF receptor c-fms showed reduction in the number of osteoclasts which associated with increased risk of osteopetrosis [44]. Moreover, high level of M-CSF was often observed in pathological bone diseases including osteoporosis, inflammatory osteolysis and rheumatoid arthritis, along with an increased number of osteoclasts [45-47]. In an inflammatory osteolysis model, TNF-α induced-M-CSF production in TNF-responsive stromal cells contributed to the increased osteoclastogenesis [48]. On the other hand, blockade of M-CSF using monoclonal antibodies was able to markedly ameliorate inflammatory and arthritic pain in arthritic diseases in mice [49]. Targeting M-CSF appears to be an alternative approach to limiting excessive activation of osteoclastogenesis. However, due to its equally important function in polarization of macrophages, such as M1 macrophages against infection, immunomodulation of osteoclastogenesis using M-CSF neutralizing antibodies or inhibitors should be particularly cautious with the consideration of the context and regimen to avoid severe immunosuppression.

2.3. IL-1β

IL-1β is a pro-inflammatory cytokine that induces bone destruction in inflammatory bone disease through the activation of osteoclast. IL-1β promotes osteoclast differentiation through its synergetic function with RANKL/RANK signaling to enhance TRAF6 and its downstream signal transduction [50]. IL-1β also facilitates osteoclastogenesis through its receptor IL-1RI and strongly recruits MITF to the promoter region of TRAP and OSCAR, which are osteoclast marker genes (Fig. 2) [51]. In vivo mice studies suggested that IL-1 stimulates osteoclastogenesis since IL-1β deficiency increased bone mass, resulting from decreased osteoclast numbers. Interestingly, the numbers of osteoblasts and areas of osteoid surface per bone surface (OS/BS) were significantly reduced in IL-1β knock-out mice [52]. RANKL expression on stromal cells can be induced by IL-1β, which directly links to the activation of osteoclast precursors [53]. These notions were also observed in patients with postmenopausal osteoporosis and RA patients, where the serum IL-1β and RANKL expression were largely increased in the active phase of diseases, while the reduction of these cytokines were observed after the remission of osteoporosis or RA symptoms in patients [54,55]. Interestingly, different from its common roles in resolving inflammation, T_REG cells were found to be osteoclastogenic and became RANKL-expressing T_REG cells in the presence of abundant IL-1β to accelerate osteoclast formation and bone erosion [56]. These results further demonstrate the strong immunomodulatory effects of cytokines on osteoclastogenesis.

2.4. IL-17

IL-17 is produced by T_H17 cells and group 3 innate lymphoid cells in inflammation and infection. It is also known that IL-17 promotes osteoclastogenesis in arthritis and osteoporosis with enhanced RANKL/OPG ratio in mice and patients [5,57]. As early as in 1999, IL-17 was found to induce COX-2-mediated PGE₂ synthesis in osteoblasts/stromal cells in synovial tissues, which subsequently directs RANKL expression and secretion in osteoblasts. The increased RANKL further stimulates osteoclastogenesis [57]. IL-17 was found to coordinate local inflammatory cytokine production, which results in an overall amplified proinflammatory tissue microenvironment. Cytokines, such as TNF-α, IL-1β, IL-6 and IL-23, are rapidly secreted into the inflammatory sites after initial elevation of IL-17, which altogether intensifies RANKL expression and RANKL induced osteoclast differentiation. Indeed, in vivo studies suggested that IL-23/IL-17 axis is critical for inflammation-induced bone destruction, where a pronounced reduction of bone destruction was found in IL-17 and IL-23α knock-out mice [39]. On the other hand, IL-17A can also enhance the production of granulocyte-macrophage colony-stimulating factor (GM-CSF) by osteoblasts, which maintains monocytes in an undifferentiated state by downregulating c-Fos, Fra-1, and nuclear factor of activated T cells 1 (NFATc1) [58,59]. Of note, mice with IL-17A deficiency have normal bone homeostasis and development of functional osteoclasts [60]. This is likely due to the dual role of IL-17A in driving RANKL production while maintaining the possibility to induce osteoblast differentiation and osteogenesis by activating JAK2/STAT3 signaling in the mesenchymal stem cells and other bone-derived cells [61-63]. In some postmenopausal women, serum IL-17 levels were elevated and associated with reduced bone mass and estrogen levels [64-66]. As a signature cytokine of T_H17 cells, increased secretion of IL-17 is usually accompanied by other proinflammatory cytokines, such as IL-1β, IL-6, IL-21, IL-22 and IL-23, which all show close relation with osteoclastogenesis and bone disease. This unique feature identifies the amount of T_H17 cells as a good diagnostic marker for diseases like osteoporosis and rheumatoid arthritis as well as key immunomodulatory targets conferring prevention and protection for bone disorders.

2.5. TNF-α

Many immune cells produce TNF-α in inflammation and infection. This includes activated macrophages, T cells, B cells, and NK cells [67]. Meanwhile, TNF-α is a critical factor for bone destruction especially during inflammatory bone diseases [68]. Mechanistically, TNF-α directly binds to its receptor on cell surface and synergize with RANKL signaling to promote osteoclastogenesis (Fig. 2) [69]. Moreover, TNF-a stimulated M-CSF gene expression, in turn, induces RANK expression in osteoclast precursors [48]. Importantly, TNF-α often acts in synergy with RANKL and/or together with other inflammatory cytokines, such as IL-6 to promote osteoclastogenesis and bone resorption under inflammatory conditions [70,71]. Consistently, elevated TNF-α was found in postmenopausal women, which contributes to the increased incidents of osteoporosis [72]. Similar bone erosion phenomenon has been observed in many other inflammatory bone diseases, such as RA and periodontitis, all with increased serum level of TNF-α [73-75]. TNF-α also induces the production of Dickkopf-1 (DKK1), a critical inhibitor of WNT pathway. TNF-upregulated DKK1 suppresses bone formation and thus contributes to inflammatory bone loss, such as in RA [76,77]. Of note, TNF-α can suppress osteoblast differentiation by inhibiting the expression of IGF-1, BMP2, osterix (OSX), and osteocalcin (OCN), which result in degradation of Runx2 and inhibition of SMADs in osteoblasts [78-80]. In addition, Xing and colleagues revealed that Notch activation suppressed TNF-α-mediated inhibition of MSC differentiation into osteoblasts via noncanonical NF-κB signaling in inflammatory arthritis [81]. It is thus clear that TNF-α induces bone erosion by both increasing osteoclastic bone resorption and decreasing osteoblastic bone formation in many inflammatory diseases associated with bone loss. It is thus promising to improve bone health through the use of blocking antibodies (infliximab) and inhibitors of TNF-α or its downstream targets as a complementary approach in patients with inflammatory bone diseases.

2.6. IL-7, IL-8 and IL-15

IL-7 is known for the maintenance of immune cell memory. While during the inflammation, stromal cells and osteoblasts increase the secretion of IL-7 in response to IL-1b and TNFa, and this IL-7 might directly function on osteoclast through the activation of STAT5, independent of RANKL [82]. It was shown that IL-7 also potentiated OC formation by promoting the RANKL and TNF-α secretion from activated T cells [83]. Cytokine IL-8 (CXCL8) is important in recruiting neutrophil, monocytes and other immune cells in response to infection [84]. Experiment has found that IL-8 promoted osteoclastogenesis and bone resorption in osteolysis in conjunction with metastatic breast cancer and multiple myeloma [85-87]. IL-8 also can act as an autocrine regulator of osteoclastogenesis, which mediated by RANKL-induced NFATc1 activation. Human osteoclast precursors secreted IL-8 prior to the formation of mature OC, where IL-8 could strengthen osteoclastogenic effect of RANKL [86]. The inhibition of IL-8 with neutralizing antibody or IL-8 receptors, CXCR1 or CXCR2, attenuated RANKL-induced osteoclastogenesis [86]. Interestingly, two studies reported that the serum level of IL-8 was enhanced in arthritis-specific and anticitrullinated protein/peptide antibodies (ACPAs)-positive RA patients. During the differentiation and activation of osteoclast precursors, circulating ACPAs binds to the mature osteoclasts in bone marrow, leading to an increased osteoclast activity and bone resorption. Interestingly, blocking of IL-8 by neutralizing antibodies markedly suppressed the ACPA-induced osteoclast formation [88,89]. IL-15 is a IL-2 family cytokine, and it shares many structural and biological properties with IL-2 [90]. Many cells secrete IL-15, including monocytes, macrophages, dendritic cells, fibroblasts, epithelial cells, and bone marrow stromal cells [91]. IL-15 induces RANKL expression by osteoblasts and stromal cells that subsequently affect osteoclastogenesis [92]. Cell line studies also found that IL-15 promoted RANKL-induced osteoclastogenesis in RAW264.7 (RAW) cells by activating ERK signaling [93]. The blockade of IL-15 reduces the number of osteoclasts in the joint and local bone destruction during arthritis in mice [94,95]. Notwithstanding, inconsistent results also showed that IL-15 could active NK cells that can induce the apoptosis of osteoclast resulting in reduced bone resorption [96,97]. Studies that elucidate the clear function of these cytokines are urgently awaited.

3. Negative regulation of osteoclastogenesis by cytokines

3.1. IL-4, IL-5, IL-9 and IL-13

IL-4, IL-5, IL-9 and IL-13 are cytokines that belong to type 2 inflammation. These cytokines are key immunomodulators during allergic response as well as antiparasitic infections. It is also shown that type 2 inflammation is protective in skewing proinflammation response that occurs in diseases such as obesity [98], diabetes [99], rheumatic arthritis [100] and osteoarthritis [101,102]. Type 2 helper T (T_H2) cells and group 2 innate lymphoid (ILC2) cells are major cells producing these cytokines amongst other immune cells [103,104]. Indeed, studies have suggested that T_H2 cells and ILC2 cells are protective against bone loss during severe inflammation [100-102]. The immunomodulatory functions of these cells in bone remodeling are mainly mediated by production of cytokines. As a signature cytokine of this kind, IL-4 could directly inhibit bone resorbing activity driven by mature osteoclasts through the suppression of NF-kB activation in a STAT6-dependent manner in osteoclast precursors [27,28]. IL-4 can also indirectly inhibit osteoclast formation through regulation of osteoblast to produce greater amount of osteoprotegerin (OPG). OPG is the soluble decoy receptor of RANKL thus reduces the amount of RANKL in the microenvironment [105]. Interestingly, IL-4 deficient mice are sensitive to RANKL induced hypercalcemia, a condition that can lead to the exacerbated bone resorption by osteoclast [106]. Human clinical study also reported inhibitory effect of IL-4 on RANKL-induced osteoclastogenesis through peroxisome proliferator-activated receptor γ1 (PPARγ1), which laid the foundation, in part, for the beneficial effects of the thiazolidinedione class of PPARγ1 ligand on bone loss in diabetic patients [107]. Meanwhile, others revealed that the serum concentration of IL-4 was markedly lower in patients with postmenopausal osteoporosis, which closely correlated with reduction of bone marrow density [108]. IL-13 shares the same gamma receptor with IL-4, which could suppress RANKL expression while promoting OPG expression in osteoblasts in a STAT signaling-dependent manner together with IL-4 [28]. Interestingly, despite the similar inhibitory function on osteoclast formation, IL-4 exerts stronger inhibition than that of by IL-13 [109]. As mentioned above, newly discovered group 2 innate lymphoid (ILC2) cells also possess the regulatory functions to suppress the differentiation of osteoclasts. These regulatory functions are achieved by ILC2-derived IL-4/13, which considerably suppressed IL-1β secretion by macrophages and bone erosion in the inflamed sites by osteoclasts [102]. Similar immunomodulatory function was found in another cohort of patients with rheumatoid arthritis, where ILC2s produced IL-9 and remarkably resolved the inflammation with altered numbers of osteoclasts and improved bone protection [100]. IL-5 also mediates bone formation through the mobilization and activation of osteogenic progenitors and/or inhibition of recruited osteoclasts [110]. In summary, cytokines that belong to the type 2 immunity/inflammation are remarkable candidates that can elicit protective roles in preventing bone loss. However, caution exists because these cytokines also robustly regulate other immune processes, which might induce unexpected complications, such as allergic response [111], dermatitis [112], and promote IgE isotype class switching and induce other autoimmune diseases, such as lupus and asthma [113]. Therefore, further research on how to precisely modulate osteoclastogenesis and bone remodeling process by type 2 cytokines is greatly needed and currently emerging.

3.2. IL-10

IL-10 is a vital immunomodulatory cytokine with potent anti-inflammatory function. Many immune cells involving M2 macrophage, dendritic cells, regulatory B cells, and regulatory subsets of CD4⁺ and CD8⁺ T cells achieve their tuning functions by producing IL-10 [114,115]. As a regulatory cytokine, IL-10 primarily suppresses differentiation and proliferation of the responding cells expressing IL-10 receptor. These include monocytes, macrophages, dendritic cells, neutrophils, B cells and T cells [115]. Numerous studies have found that IL-10 elicited vigorous inhibition on osteoclastogenesis by reducing the intrinsic expression of NFATc1 and the presence of RANKL and M-CSF secreted by other cells [116-118]. In vivo studies found that IL-10 deficient mice developed osteopenia, characterized by decreased bone mass and impaired bone formation [119]. While secretion of a number of inflammatory factors, such as TNF-α, IFN-γ, RANKL, and NO, was remarkably increased in the same mice, decreased expression of markers including OCN, ALP and phosphate-regulating gene endopeptidases (PHRX), known key factors for osteoblasts and osteocytes, was observed [119,120]. In patients with postmenopausal osteoporosis, serum levels of IL-10 are significantly lower than healthy women [121]. IL-10 gene polymorphisms are associated with bone mineral density and the susceptibility to osteoporosis in postmenopausal women [122,123]. This evidence suggests a unique role for those major IL-10 producing cells, such as T_REG cells and IL-10-producing innate lymphoid cells, in controlling excessive osteoclastogenesis during bone inflammation. Indeed, recent studies have highlighted that T_REG cells exceptionally inhibited differentiation of osteoclast through IL-10 and CD80/CD86 mediated immunomodulatory function [6,7]. As a rapidly growing field, a number of T_REG cell-based immunotherapies are currently under clinical trials for treating inflammatory and autoimmune diseases [18,124,125]. Most of these clinical trials require sufficient IL-10 secretion by T_REG cells. It would be thus very interesting to see how these therapeutics could be translated into the treatment of inflammatory bone diseases in the near future.

3.3. Interferon-β (IFN-β)

Interferons are critical in regulating innate and adaptive immune responses against infections and can be sub-grouped into three major classes. Type I IFNs are critical for the regulation of osteoclastogenesis in mice. Induction of c-Fos and TRAF6 activates the NF-kB and JNK pathways, which promotes the production of Interferon-β (IFN-β) in osteoclasts [26]. However, it is also known that IFN-β can initiate negative feedback where it can reduce the activity of c-Fos, followed by inhibited osteoclastogenesis [26,126]. Indeed, administration of IFN-β improved bone loss in ovariectomized osteoporotic mice [26]. This negative feedback loop is important for bone homeostasis as IFNAR1 or IFN-β deficient mice with impaired feedback tend to have osteopenia [26]. Interestingly, gene expression profiling revealed that in human osteoclast in vitro cell culture, addition of IFN-β upregulated CXCL11, a chemokine which could further repress osteoclastogenesis (Fig. 2) [127]. Our group has identified that miR182 is an upstream inhibitor whereas Def6 and PKR are activators of IFN-β expression. These factors form a regulatory network to coordinate and fine tune osteoclast differentiation. The miR182-PKR-IFN-β axis and Def6-IFN-β pathway play important roles not only in bone homeostasis but also in pathological bone loss [128,129].

3.4. IL-12 and IL-27

IL-12 and IL-27 both belong to the IL-12 family of cytokines. These cytokines are secreted by macrophages, monocytes, dendritic and B cells in response to bacterial and parasitic infections [130]. IL-12 can inhibit RANKL/TNF-α/LPS-induced osteoclastogenesis in vitro, which is attributed to the FAS/FASL-mediated apoptosis [131-133]. It was also found that IL-12 reduced TNF-α induced osteolysis in vivo [134]. However, it was still not very clear how IL-12 might affect the osteoclastogenesis along with other cytokines. IL-27 has a pleiotropic immunoregulatory functions. It was shown that IL-27 suppressed RANKL-induced osteoclastgenesis in vitro by inhibiting MAPK signaling pathway and NFATc1 expression [135]. IL-27 also functions on STAT1 which subsequently reduced the activation of c-Fos in osteoclast precursor cells [135,136]. A recent in vivo study revealed that IL-27 treatment prevents estrogen-deficient induced bone loss through promotion of early growth response gene 2 (Egr2) and Id2 in mice [137]. It is tempting to expect that more exciting results will be found from bench-to-bedside studies on IL-12 and IL-27-mediated regulation of osteoclastogenesis.

4. Cytokines that show dual roles in osteoclastogenesis

There are many other cytokines involved in the regulation of osteoclastogenesis, for instance, interferon γ, IL-2, IL-6 and TGF-β. These cytokines often play a multifaceted role during bone remodeling. IFN-γ produced by T cells strongly accelerates JAK-STAT1 signaling transduction to induce rapid degradation of TRAF6, which results in interfering with RANKL-RANK signaling to suppress osteoclastogenesis [3]. Moreover, IFN-γ also reduces the expression of c-fms and RANK on the surface of osteoclast precursors to synergistically inhibit osteoclastogenesis with other factors such as Toll-like receptors TLRs (Fig. 2) [138]. Interestingly, IFN-γ not only inhibits osteoclast differentiation directly, but also reduces osteoclastogenesis by promoting osteoblast-derived NO, which could induce the apoptosis of osteoclast via FAS ligand/Fas signaling [139,140]. On the other hand, however, IFN-γ can promote the fusion of osteoclast precursor required for osteoclastogenesis through the induction of the expression of dendritic cell-specific transmembrane protein (DC-stamp), which is often required for immature osteoclast fusion [141]. In vivo studies further support the binary function of IFN-γ on OC formation, where bone volume was found significantly reduced in IFN-γ or IFN-γ receptor deficiency mice, while exogenous IFN-γ administration can lead to severe bone loss in osteoporotic mice [3,142]. Controversial results were also found with regards to IL-2, where IL-2 was found to stimulate osteoclastic activities [143], but the deficiency of IL-2 in mice could lead to severe bone loss [90,144].

IL-6 is another cytokine of such kind, where in some conditions it is pathogenetic for inflammatory bone loss through its promotion of osteoclast precursors into mature osteoclasts [145,146]. On the flip side, IL-6 was suggested to suppress RANKL-mediated NF-kB and JNK activation in during osteoclastogenesis (Fig. 2) [147]. Tocilizumab, the first humanized anti-IL-6 receptor antagonist, has been approved by FDA to treat patients with rheumatoid arthritis [148]. Clinical study also found that Tocilizumab administration in patients with anticitrullinated protein antibody-positive rheumatoid arthritis might prevent bone loss [149], which supports the beneficial effects on bone by blocking IL-6 signaling. Although more clear results should be provided from bed to bench side, IL-6 might serve as good target for bone diseases in the context of severe inflammation.

IL-23 belongs to the IL-12 family of cytokine. Like IL-6, both positive and negative roles were reported regarding to the regulation of OC formation. On the one hand, IL-23 promotes osteoclast formation in vitro by upregulating RANK expression on osteoclast precursors [150]. IL-23 also activates DAP12, which is required for osteoclastogenesis [151]. Indirectly, IL-23 can enhance the osteoclastogenesis by promotion of T_H17 cell polarization and IL-17A production, this will lead to the increased production of RANKL required for osteoclastogenesis [150,152]. The anti-IL-23 antibody reduces synovial inflammation and bone destruction in rats with collagen-induced arthritis [153]. On the other hand, IL-23 was reported to inhibit osteoclastogenesis in vitro [154,155]. Indeed, the genetic IL-23p19 deficient mice shows 30 % decrease in bone mass [154]. Moreover, clinical study has found that anti-IL-12/23 p40 monoclonal antibody (ustekinuma) treatment significantly improved the joint damage in patients with active psoriatic arthritis (PsA) [156]. Hence, IL-23 might have context-dependent regulation of regulating osteoclastogenesis.

Aberrant transforming growth factor-β 1(TGF-β1) signaling induces immune cell abnormalities, which is one of the leading causes of inflammation, allergy, autoimmunity and cancer [157]. Indeed, TGF-β1-deficient mice die within 2–3 weeks due to the severe lymphocyte infiltration into multiple vital organs [157]. Intriguingly, TGF-β1 is one of most abundant cytokines in the bone matrix during bone homeostasis, which has been shown to regulate both bone resorption and bone formation [158-160]. Despite lots of studies investigating the effect of TGF-β on osteoclastogenesis, the conclusions are still controversial and contradictory [161-167]. Some studies reported that endogenously produced TGF-β is an essential autocrine factor for osteoclastogenesis in combination with RANKL and M-CSF [165]. TGF-β binds to its receptor on osteoclasts and activates SMAD2/3, which is mediated by TRAF6-TAB1-TAK1 complex and triggers osteoclast differentiation [162]. On the other hand, a previous study reports that endogenous TGF-β1 inhibits bone resorption partly by inducing osteoclast apoptosis through up-regulating Bim in human osteoclast cultures and also partly by stimulating osteoblastic bone formation [164]. However, previous works also show that low-concentration treatment of TGF-β1 promotes RANKL or TNF-α stimulated osteoclast differentiation via direct action, while high level of TGF-β1 indirectly represses osteoclast differentiation by enhancing OPG secretion from osteoblasts in the cell cultures [161,166]. Another study found that TGF-β1 showed both inhibitory and stimulatory effects on human osteoclast differentiation, and that these opposing functions are mediated by SMAD1 and SMAD3 signaling, respectively [163]. TGF-β1 inhibits osteoclastogenesis when added at the early stages of differentiation, suggesting that timing affects the function of TGF-β1 on osteoclastogenesis [163]. The significantly higher level of human TGF-β1 was noticed in subchondral bone of knee joints throughout of osteoarthritis in patients comparing to that of in healthy donors [168]. Moreover, inhibition of TGF-β signaling attenuated cartilage damage and improved the inflammatory bone loss in OA [168]. Taken together, TGF-β1 is an important cytokine in regulating bone homeostasis, while its clear role on osteoclastogenesis requires genetic evidence and further investigation.

5. Conclusion

In this review, we summarized recent progress in cytokine-mediated immunomodulation of osteoclastogenesis and discussed potential therapeutic possibilities of bench-side research into novel therapies to treat bone disease (Table 1). In a fast-growing field of immunotherapy, cytokines serve as an excellent intermediate to connect skeletal system and immune system. Due to the dual role or even multifaceted roles of cytokines in regulating bone remodeling, it might not be practical to develop a single cytokine-based treatment to treat broad and complex conditions of bone disorders. Taken this into account, the heterogenous nature of human genetics also requires more precise and personalized treatment options to confer protection or improvement of disease symptoms with limited adverse events. Novel techniques, such as single cell RNA-sequencing, CRISPR editing and screening, can facilitate the development of new treatments targeting more than one cytokine for the improvement of bone health in patients with different needs and backgrounds. By doing so, more exciting results from both fundamental science and clinical research of osteoimmunology would emerge and are expected.