Osteoimmunology: cytokines and the skeletal system

Abstract

It has become clear that complex interactions underlie the relationship between the skeletal and immune systems. This is particularly true for the development of immune cells in the bone marrow as well as the functions of bone cells in skeletal homeostasis and pathologies. Because these two disciplines developed independently, investigators with an interest in either often do not fully appreciate the influence of the other system on the functions of the tissue that they are studying. With these issues in mind, this review will focus on several key areas that are mediated by crosstalk between the bone and immune systems. A more complete appreciation of the interactions between immune and bone cells should lead to better therapeutic strategies for diseases that affect either or both systems.

INTRODUCTION

Recent reviews have highlighted the interactions between bone and immune cells as well as their overlapping regulatory mechanisms (1, 2). For example, cells related to osteoblasts-the bone-forming cells of the body-regulate hematopoietic stem cell niches from which all blood and immune cells are derived. On the other hand, osteoclasts, which function to resorb bone, are derived from the same myeloid precursor cells that give rise to macrophages and myeloid dendritic cells. Furthermore, many of the soluble mediators of immune cells, including cytokines, chemokines, and growth factors, regulate the activities of osteoblasts and osteoclasts. This increased recognition of the complex interactions between the immune system and bone led to the development of the interdisciplinary osteoimmunology field, which seeks to translate an understanding of the mechanisms governing the interface between the skeletal and immune systems into therapeutic strategies for the treatment of disorders characterized by bone loss (1, 3).

The regulation of bone by hematopoietic and immune cells produces a variety of physiologic and pathologic effects. It is likely that developing hematopoietic cells regulate bone turnover and maintain the marrow cavity by interacting with osteoblasts and osteoclasts. Conversely, during inflammatory states induced by cytokines, activated immune cells mediate increased bone turnover and the bone pathologies associated with diseases such as rheumatoid arthritis and inflammatory bowel disease. Indeed, if one views bone marrow spaces as “loosely compartmentalized lymphoid organ”, it won’t be difficult to conceptualize how intense immune cells and bone cells interact and influence each other (Fig. 1) (1). We are only beginning to understand the breadth of these interactions and this review is by no means complete. Examining the interface between these systems, however, should contribute to a scientific foundation for novel therapeutic strategies to treat disease states mediated by both systems.

Fig. 1.

A schematic diagram of the bone microenvironment. Hematopoietic stem cells (HSC) are maintained in the stem cell niche that is suggested to be provided by osteoblast lineage cells. Mesenchymal stem cells give rise to adipocytes and osteoblast-lineage cells. Cells derived from HSC differentiate, mature and migrate to the periphery though the vascular system. Memory T and B cells or activated T cells come back to the bone microenvironment, and provide factors influencing the bone cells such as osteoblasts and osteoclasts. Osteoclasts are derived from monocyte lineage cells, and resorb bone. Although not depicted in the diagram, there are numerous interactions among the cells in the bone microenvironment; hence it can be viewed as a loosely compartmentalized secondary lymphoid organ.

Here, we will provide a brief description of the current understanding of the interactions underlying the development of components of both the skeletal and immune systems. We will then focus on the cytokines that regulate bone cells and affect bone metabolism.

Osteoclasts and osteoblasts

Osteoclasts, which form as multinucleated cells following fusion between mononuclear precursor cells, are unique in their capacity to efficiently resorb bone (4, 5). In the 10 years since the osteoclast differentiation factor receptor activator of NF-κB ligand (RANKL) was discovered, a number of reviews have described the molecular pathways underlying the maturation of osteoclasts from bone marrow precursors (1, 2, 4, 5). Therefore, we will concentrate here on recent progress in the identification of the bone marrow cell population that serves as osteoclast precursors.

Bone marrow, peripheral blood, and spleen cell populations can form osteoclast-like cells (OCLs) in various in vitro culture systems (4, 5). Cultures with ST2 stromal cells, RANKL, and macrophage colony-stimulating factor (M-CSF) were used to demonstrate that murine bone marrow cell populations expressing c-kit can form OCLs (6). The authors concluded that the identified c-kit⁺c-fms⁺CD11b^low bone marrow cells included multipotential progenitor cells that frequently gave rise to osteoclasts. These RANK⁻ (the receptor for RANKL) progenitor cells expressed RANK in response to M-CSF. Interestingly, the precursors were not restricted to osteoclastogenesis; in methylcellulose cultures, they differentiated into macrophages and mononuclear TRAP⁺ cells. We later found that the osteoclast precursor cells were negative for CD3 and CD45R, and confirmed they did not or only weakly expressed the monocytic marker CD11b/Mac-1 (7). c-kit expression further separated this population into cells that rapidly formed OCLs in vitro when cultured with M-CSF and RANKL (c-kit^high cells) and cells that formed OCLs more slowly in vitro (c-kit^low or c-kit⁻ cells). Although we initially found that the most efficient osteoclast precursors were CD11b⁻ or CD11b^low, culture with M-CSF and RANKL induced the mononuclear precursor cells to transiently express high levels of CD11b.

The relationship between osteoclasts and antigen-presenting dendritic cells has also been detailed. Human and murine cells expressing early markers for the myeloid dendritic cell lineage can differentiate into osteoclasts in vitro (8). Additionally, murine bone marrow cells that were able to present antigen to T lymphocytes following cytokine treatment formed OCLs in culture when they were treated with M-CSF and RANKL (9). Speziani et al., however, found that neither mature myeloid dendritic cells generated in vitro nor plasmacytoid dendritic cells generated in vivo formed OCLs in culture (10).

Although investigators previously suggested that common progenitor cells can differentiate into macrophages, osteoclasts, and myeloid dendritic cells (8), single cell clones from murine bone marrow were only recently isolated and differentiated into macrophages and dendritic cells (11). There is good evidence that these myeloid precursor cells can also differentiate into OCLs in vitro (J. A. L. unpublished data). Interestingly, commitment of the common precursors to the osteoclast lineage occurs relatively quickly (within 24 hours) after the cells are treated with RANKL (12).

Outside of the bone marrow, expression of the myeloid-specific antigen CD11b as well as Gr-1 was used to identify circulating osteoclast precursor cells, the cell number of which is regulated by the inflammatory state of the organism and in particular tumor necrosis factor-α (TNF-α) (13, 14). In human peripheral blood, RANK⁺ osteoclast precursors were identified using CD14 expression and the lack of CD16 expression (15–17). Migration and adhesion of these human CD14⁺ monocytes to sites of inflammation may be mediated through activation of microvascular endothelial cells by proinflammatory cytokines (18).

Interestingly, cells with cell-surface phenotypes similar to those of bone marrow osteoclast precursors were identified in the spleen, even though splenic osteoclastogenesis does not occur under any known condition, possibly because the splenic cells lack factors that are required for osteoclast differentiation. This hypothesis, however, seems unlikely because multiple investigators have established the osteoclastogenic potential of splenocytes in vitro. Another possibility is that the splenic microenvironment lacks critical signaling molecules that define an adherent condition for osteoclastogenesis (19). This hypothesis is supported by the recent findings that osteoclast differentiation and activation require various costimulatory molecules, which act in concert with M-CSF and RANKL (1, 2). These signals are transduced through immunoreceptor tyrosine-based activation motif domain-containing adaptor proteins, such as DAP12 and FcRγ, and at least four receptors have been found to associate with either FcRγ (OSCAR and PIR-A) or DAP12 (TREM-2 and SIRP β1) (20, 21). The ligands for these receptors are currently unknown.

Osteoblasts are derived from multipotential mesenchymal progenitor cells that also differentiate into marrow stromal cells and adipocytes (22). The regulatory signals that drive the progenitor cells to an osteoblast fate have not been fully elucidated. A number of critical paracrine signals and cell autonomous transcription factors, however, have been identified, including the transcription factors Runx2 and osterix as well as bone morphogenic proteins (BMPs), which initiate osteoblast differentiation (23–25). Wnt signaling also contributes to mesenchymal progenitor cells becoming either adipocytes or osteoblasts (26–28).

Cytokines and local immune cell factors as regulators of bone cells

In this section, we describe the current understanding of the roles of various inflammatory cytokines in bone metabolism.

RANKL

Characterizations of the functions of RANKL and its receptors (RANK and osteoprotegerin [OPG]) have elucidated the interplay between active immunity and bone homeostasis. A number of recent reviews about the diverse physiologic functions of the RANKL-RANK-OPG signaling axis in bone and it central roles in osteoimmunology have been published (1, 2, 4, 5). Therefore, we will briefly summarize the RANKL axis in Fig. 2.

Fig. 2.

A simplified view of osteoclast differentiation. RANKL produced by osteoblasts induces the differentiation of osteoclasts, and further regulates the activation and survival of mature osteoclasts. Some of the key transcription factors (PU.1, NF-κB, c-Fos, NFATc1) and signaling molecules (MAPK, c-Src, PI3K, AKT) are also shown in the diagram. HSC, hematopoietic stem cells.

TNF-α

TNF-α was shown to stimulate osteoclast formation and bone resorption in vivo, and enhance the formation of OCLs in bone marrow cultures (1, 2, 4, 5). The ability of TNF-α to stimulate osteoclast formation in mixed stromal cell/osteoclast precursor cell cultures was shown to be IL-1-dependent (29), whereas TNF-α-induced osteolysis was found to be dependent on M-CSF (30). Cultures of cells from RANK-deficient mice suggested that TNF-α directly stimulated osteoclast formation independent of RANK (31), although TNF-α administration produced only a few osteoclasts in RANK-deficient mice (32). TNF-α also inhibits osteoblast differentiation and collagen synthesis (33–35). Additionally, TNF-α is potently pro-apoptotic for osteoblasts (36), possibly through Fas-Fas ligand (FasL) signaling (37). Interestingly, however, mice deficient for both TNF receptor 1 and TNF receptor 2 do not have abnormal bone phenotypes, suggesting that TNF-α affects bone during inflammatory states, rather than during normal development.

Other members of the TNF superfamily

Precursor and mature osteoclasts express Fas and FasL (38). Treatment of mature osteoclasts with Fas induced apoptosis, whereas FasL treatment of cultured M-CSF/RANKL-treated osteoclast precursor cells increased osteoclast formation (39). In contrast, RANKL and FasL have counterregulatory roles in the apoptosis of mature osteoclasts; at high concentrations, RANKL inhibited the ability of FasL to induce this response (40). Additionally, the effects of FasL deficiency on bone mass are controversial; different studies found this index to increase or decrease in FasL-deficient mice (39, 41). Studying bone mass in Fas- or FasL-deficient mice, however, is difficult, because these models have generalized lymphroliferative disorders, which activate a wide variety of immune responses that affect bone. Interestingly, estrogen receptor-α was found to regulate FasL production in murine osteoclasts, which, in turn, mediated the bone loss induced by estrogen withdrawal (42).

Treatment of osteoclasts with TNF-related apoptosis inducing ligand (TRAIL) induces apoptosis (43). Accordingly, injection of young murine bone with TRAIL increased bone mass, which was associated with increased levels of the cyclin-dependent kinase inhibitor p27^Kip1 (44). TRAIL may also contribute to the effects of myelomas on osteoblasts (45).

CD40 ligand (CD40L) is involved in the differentiation of T_H1 effector cells. In humans, CD40L deficiency results in X-linked hyper IgM syndrome, which causes spontaneous bone fractures and osteopenia, possibly via activated T lymphocytes that are defective for interferon (INF)-γ production (46). Additionally, CD40L expression in the synovial cells of rheumatoid arthritis patients induces RANKL expression in these cells and enhances their osteoclastogenic activity (47).

M-CSF

Mutant op/op mice, which lack osteoclasts and show defective macrophage/monocyte formation, were found to be deficient for M-CSF. M-CSF injections or osteoblast-specific M-CSF expression in op/op mice rescued the osteoclast formation and bone resorption defects (1, 2, 4, 5). Consistently, bone resorption stimulators increase M-CSF production in bone (48, 49). Interestingly, membrane-bound M-CSF facilitated osteoclast differentiation (48, 50), whereas soluble M-CSF inhibited 1,25-dihydroxyvitamin D₃-stimulated OCL formation in bone marrow cultures (51). In osteoclast precursor cells, M-CSF is a potent stimulator of RANK expression and proliferation (6, 7).

M-CSF also regulates osteoclast apoptosis as the addition of M-CSF to mature osteoclast cultures prolongs their survival (52). This may contribute to the osteopetrotic phenotype of op/op mice, because transgenic expression of apoptosis-inhibiting Bcl-2 in myeloid cells partially rescued the osteoclast and macrophage developmental defects in these animals (53).

Additional colony stimulating factors

GM-CSF and IL-3 also affect osteoclast differentiation (51, 54). Both factors inhibit RANKL-mediated osteoclastogenesis (55, 56), whereas they promote osteoclast precursor development (57); these proteins drive myeloid precursor cells to lineages other than osteoclasts (55), and inhibit TNF receptor expression on myeloid precursor cells (58). IL-3 also inhibits osteoblast differentiation, which may contribute to the effects of multiple myelomas on bone (59).

Systemic G-CSF injections decrease rodent bone mass, likely because of increased osteoclast formation and decreased osteoblast activity (60, 61). G-CSF also mobilizes hematopoietic precursor cells from bone marrow (62) and increases the number of circulating osteoclast precursors (63), which are likely related to increased osteoclast activity. Accordingly, G-CSF overexpression in mice inhibited the ability of osteoblasts to respond to BMP and increased bone resorption (64, 65).

IL-1

IL-1-a potent stimulator of bone resorption-is produced in bone and acts on osteoclasts directly and indirectly via enhanced RANKL production and activity (1, 2, 4, 5). Additionally, both RANKL- and 1,25-dihydroxyvitamin D₃-stimulated osteoclast formation in vitro is mediated in part by IL-1 (66, 67). IL-1 also increases prostaglandin synthesis in bone (68, 69), which may account for some of its resorptive activity (70). Stimulation of osteoclastogenesis by IL-1 in mixed murine stromal cell/hematopoietic cell cultures was dependent on RANKL but not TNF-α (71). IL-1-mediated RANKL production in osteoblasts and osteoclasts survival depends on myeloid differentiation factor 88 (MyD88), PI3-kinase/AKT, and ERK, but not Toll/interleukin-1 receptor domain-containing adaptor inducing inter-feron-β (TRIF) (72, 73). A recent report found decreased bone mass in mice deficient for bioactive type I IL-1 receptor (IL-1R) (74), although our experiments did not support this result (75).

IL-6 family cytokines

IL-6-a multipotent cytokine with a wide variety of activities-is produced by osteoblastic cells and BMSCs (76, 77). Although IL-6-mediated bone resorption varies depending on the in vitro assay system (78, 79), it has been shown to regulate the development of mature osteoclasts (80), and directly stimulate the production RANKL and OPG mRNA as well as prostaglandins in bone (81, 82). IL-6 appears to mediate the increased bone resorption and pathologies that characterize various clinical syndromes, including Paget’s disease (83), hypercalcemia associated with malignancy (84), fibrous dysplasia (85), giant cell tumors of bone (86), and Gorham-Stout disease (87). Finally, there is conflicting data about the role of IL-6 in the PTH-mediated responses of bone (88, 89).

In response to resorptive stimuli, bone cells produce IL-11, which stimulates osteoclast formation and bone resorption in vitro (90–93). Interestingly, it has no effect on isolated mature osteoclasts. Mice deficient for the IL-11 receptor showed increased trabecular bone mass, which may reflect the decreased bone turnover, osteoclast formation, and resorption activity observed in vitro (94).

LIF is produced by bone cells in response resorption stimuli (95), resulting in variable effects on bone resorption. In some in vitro systems, LIF stimulated prostaglandin-dependent resorption (96), whereas, in others, it produced inhibitory effects (97, 98). In neonatal murine calvaria cultures, LIF stimulated both RANKL and OPG expression (82). Furthermore, local injections of LIF augmented parameters of bone resorption and formation, as well as the thickness of the treated bones (99). Mice lacking specific LIF receptors showed reduced bone volume and increased osteoclast numbers (100).

Although oncostatin M stimulates multinuclear cell formation in bone marrow cultures, these cells appeared to be macrophage polykaryons and not osteoclasts (101). On the other hand, oncostatin M inhibited 1,25-dihydroxyvitamin D₃-stimulated OCL formation in human bone marrow cultures (101), and decreased bone resorption rates in fetal mouse long bone cultures (102). Moreover, oncostatin M overexpression in transgenic mice induced a osteopetrotic phenotype (103). Hence, oncostatin M likely inhibits osteoclast formation and bone resorption.

The roles of the IL-6 cytokine family in osteoclast formation should be examined based on data demonstrating that mice lacking the gp130 activator protein have increased osteoclast numbers (104). Because gp130 transduces signalling for all of the IL-6 family members, this result argues that at least some of these factors inhibit osteoclast formation and bone resorption. The available data implicate oncostatin M (102) and possibly LIF (97, 98) for this function.

IL-7

IL-7, which plays nonredundant roles in B- and T-cell lymphopoiesis, also regulates bone homeostasis (105); the mechanisms by which IL-7 affects bone cells, however, are controversial. Systemic IL-7 administration enhances osteoclast formation in peripheral blood by increasing osteoclastogenic cytokine production in T cells (106). Moreover, IL-7 did not induce bone loss in T-cell-deficient nude mice (107). Additionally, ovariectomy enhances T-cell development through IL-7, which may underlie ovariectomy-induced bone loss (108). Thus, interpreting the effects of IL-7 treatment is complicated by secondary effects in vivo, which result from T cells producing bone-resorbing cytokines (106, 107).

In contrast, we observed that IL-7 inhibited osteoclast formation in murine bone marrow cells cultured with M-CSF and RANKL (109). Moreover, the osteoclast number markedly increased and trabecular bone mass decreased in IL-7-deficient mice (110). Additionally, trabecular bone loss after ovariectomy was similar in wild-type and IL-7-deficient mice (110). Curiously, IL-7 mRNA levels in bone increased following ovariectomy, which may result from altered osteoblast function after estrogen withdrawal (105, 111). IL-7 also inhibited bone formation in newborn murine calvaria in vivo and in vitro (105). When IL-7 was locally overexpressed by osteoblasts, however, trabecular bone mass increased (112). Furthermore, targeted IL-7 overexpression in IL-7-knockout mice rescued their osteoporotic bone phenotype (113). These studies indicated that the effects of IL-7 on bone cells depend on whether IL-7 is delivered systemically or locally.

IL-8 and chemokines

Chemokines, which can be divided into the CXC, CC, C, and CX₃C subtypes based on the sequence motif containing the first cysteine residue, act through G-protein-coupled receptors to initiate cytoskeletal rearrangement, adhesion, and directional migration (114). IL-8, a CXC chemokine produced by osteoclasts, stimulates osteoclastogenesis and bone resorption independent of the RANKL pathway (115, 116). Additionally, IL-8 produced by certain cancers stimulates lytic bone lesions in metastatic disease (115, 116). The effects of IL-8 on bone may be partly mediated by upregulated osteoclast nitric oxide synthase expression (117).

CCL3 (macrophage inflammatory protein-1α), which is expressed in bone and bone marrow cells, directly stimulates osteoclastogenesis through the receptors CCR1 and CCR5 (118–120). CCL3 also mediates the osteolytic activity of multiple myelomas (121, 122). Interestingly, CCL3 and IL-8 stimulate the motility but suppress the resorption activity of mature osteoclasts (123).

Expression of CCL9 (macrophage inflammatory peptide-1γ) and its receptor CCR1 in osteoclasts is induced by RANKL (124). Moreover, CCL9 and other chemokines that bind CCR1 (CCL3, CCL5, and CCL7) are produced by osteoclasts, osteoblasts, and their precursors in bone; expression of these chemokines in differentiating osteoblasts is induced by proinflammatory cytokines (IL-1 and TNF-α) (125). Additional chemokine receptors that are produced by osteoclasts include CCR3, CCR5, and CX₃CR1 (118, 126).

Injections of M-CSF into osteopetrotic tl/tl rats, which lack M-CSF, induce osteoclastogenesis, bone resorption, and a rapid upregulation of CCR1 and CCL9 expression in bone (127). Furthermore, anti-CCL9 antibodies ameliorated M-CSF-stimulated osteoclastogenesis in these rats. Inhibiting CCR1 expression suppressed the migration of RAW 264.7 osteoclast precursor cells and murine bone marrow cells (128). Furthermore, blocking ligand binding by CCR1 in murine bone marrow cultures prevented OCL formation (128), and neutralizing antibodies against CCL9 inhibited RANKL-induced osteoclastogenesis (124).

CXCL12 (stromal cell derived factor-1) and its receptor CXCR4 contribute to hematopoietic cell homeostasis and immune responses. Osteoclast precursor cells express CXCR4 (129), which is downregulated as these cells differentiate to the osteoclast lineage (130). Treatment of RAW 264.7 cells with CXCL12 induces matrix metalloproteinase 9 (MMP9) expression, which may contribute to the migration of precursor cells to bone (129). In human osteoclast precursor cells, CXCL12 enhanced the migration and osteoclastogenesis induced by RANKL and M-CSF (129, 130). CXCL12 expression is upregulated in osteoclasts differentiating on a calcium phosphate matrix (130). CXCL12 may also be involved in precursor cell recruitment to giant cell tumors in bone (131) and the increased osteolysis associated with multiple myelomas (132).

CCL2 (monocyte chemoattractant protein-1)-the ligand for CCR2-is highly expressed in osteoblasts associated with induced inflammatory lesions (133), an effect that is mediated by proinflammatory cytokines (134). CCL2 may also contribute to the eruption of teeth via its expression in dental follicle cells (135), although CCL2-deficient mice demonstrated that CCL2 is not required for tooth eruption (136). In mononuclear precursor cells, CCL2 expression is induced by RANKL and enhances OCL formation (137). It was also recently shown that treating osteoblasts with PTH increased CCL2 expression and enhanced the fusion of preosteoclasts (138).

IL-10

IL-10 produced by activated T and B lymphocytes directly inhibits osteoclastogenesis and osteoblastogenesis via increased tyrosine phosphorylation of a variety of proteins in osteoclast precursor cells, suppressed production of osteoblastic proteins, and inhibition of the onset of mineralization (139–141). The direct effects of IL-10 on RANKL-stimulated osteoclastogenesis are associated with decreased NFATc1 expression and nuclear translocation (142) as well as suppressed c-Fos and c-Jun expression (143). Thus, IL-10 may be therapeutically useful to control wear-induced osteolysis (144). IL-10 also appears to indirectly affect osteoclastogenesis, because in vitro IL-10 treatment inhibited and enhanced RANKL and OPG production, respectively, in dental follicle cells (145). Recently reported results suggest that some effects of IL-10 on osteoclasts are mediated by interactions between the inducible T-cell costimulatory molecule 4-1BB its ligand (146).

IL-12

Although IL-12 is thought to inhibit osteoclastogenesis, the underlying mechanism is controversial. One study demonstrated that IL-12 directly inhibited RANKL-stimulated osteoclastogenesis in purified primary osteoclast precursors and RAW 264.7 cells, which was associated with suppressed NFATc1 expression (147). Interestingly, IL-12 did not inhibit osteoclastogenesis in cells pretreated with RANKL. In contrast, others have found that the inhibitory effects of IL-12 on osteoclastogenesis are indirect; one group showed that IL-12 inhibition was mediated by T lymphocytes and did not involve INF-γ (148), whereas a second group found that IL-12 inhibited osteoclastogenesis in T lymphocyte-deficient cultures or nude mice (149). The second study also reported that anti-INF-γ antibodies partially blocked IL-12-mediated inhibition of RANKL-stimulated osteoclastogenesis.

IL-15

IL-15, a member of the IL-2 superfamily, has been shown to increase osteoclast progenitor cell numbers in culture (150). Moreover, IL-15 production by T lymphocytes is involved in the increased osteoclastogenesis and bone destruction seen patients with rheumatoid arthritis (151).

IL-17 and IL-23

The IL-17 cytokines constitute a six-member family (IL-17A-F) that are central for adaptive immune responses (152). They are products of the T_H17 subset of CD4⁺ T lymphocytes, which have a high IL-17-dependent osteoclastogenic activity (153). IL-17A was initially found to stimulate osteoclastogenesis in mixed hematopoietic cell/osteoblasts cultures via prostaglandin synthesis and RANKL expression (154). Moreover, IL-17A mediates the activation of osteoclasts and bone destruction in joints affected by rheumatoid arthritis (154), effects that are enhanced by TNF-α (155). IL-17A inhibition in an antigen-induced arthritis model inhibited joint and bone destruction and decreased the levels of RANKL, IL-1-β, and TNF-α in the pathologic lesions (156).

IL-23 is an IL-12-related cytokine that, along with TGF-β and IL-6, is critical for T_H17 differentiation and proliferation (157). LPS-induced inflammatory bone destruction was markedly attenuated in mice deficient for either IL-17 or IL-23 (153). The authors also observed IL-23 mRNA expression in the synovial tissues of patients with rheumatoid arthritis, which suggests that similar mechanisms underlie rheumatoid arthritis-induced bone loss.

IL-18

The levels of IL-18, a member of the IL-1 superfamily, increase at sites of inflammation, such as those associated with rheumatoid arthritis (158). It is expressed by osteoblastic cells and inhibits osteoclast formation through a variety of mechanisms, including enhanced GM-CSF expression in T cells (159). IL-18, which is also a mitogen for osteoblastic cells in vitro (160), stimulates INF-γ production in bone (161) and its inhibition of osteoclastogenesis is enhanced by cotreatment with IL-12 (162). Additionally, it has been shown to increase OPG production (163). In IL-18-overexpressing transgenic mice, the number of osteoclasts decreased, although so did bone mass, suggesting that IL-18 may also affect bone growth (161). Surprisingly, IL-18 has also been shown to indirectly stimulate osteoclastogenesis through its effects on T lymphocytes (164).

Interferons

INF-γ is a type II INF that inhibits bone resorption in vitro as a result of its effects on osteoclast progenitors (165). INF-γ also inhibits the abilities of 1,25-dihydroxyvitamin D₃, PTH, and IL-1 to stimulate OCL formation in bone marrow cultures (166). INF-γ inhibits RANK signalling by accelerating the ubiquitin/proteasome-mediated degradation of TRAF6 (167); it, however, does not directly inhibit resorption by mature osteoclasts (168). INF-γ is also reported to stimulate resorption through enhanced RANKL and TNF-α production in T lymphocytes (169). Moreover, it inhibits osteoblast proliferation and has variable effects on osteoblast differentiation (35, 160, 170). The in vivo effects of INF-γ on bone are different from its actions in vitro. In rats, intraperitoneal INF-γ injections induced osteopenia (171), whereas administration of INF-γ stimulated bone resorption and appeared to partially reverse the course of osteopetrosis in patients, which may be due to INF-γ stimulation of osteoclast superoxide synthesis or osteoclast formation (172, 173).

As for the type I INFs (INF-α and INF-β), mice deficient for the INF-α/β. receptor component IFNAR1 show reduced trabecular bone mass and increased osteoclast numbers (174). RANKL induces INF-β expression in osteoclasts, which, in turn, inhibits RANKL-mediated osteoclastogenesis by decreasing c-fos expression (174). For osteoclast development, therefore, osteoclast precursors upregulate the expression of the cytokine signaling regulator SOCS3 to suppress the effects of IFNs (175, 176). INF-α has also been shown to inhibit bone resorption, although the mechanism of action is not well understood (177). INF-α had no effect on bone turnover in vivo (178).

Additional cytokines

IL-4 and IL-13 are locally acting inhibitory cytokines that have related effects on osteoblasts and osteoclasts. Transgenic IL-4 overexpression causes osteoporosis due to inhibited osteoclast formation, osteoclast activity, and bone formation (179–181). IL-13 and IL-4 inhibited IL-1-stimulated bone resorption by decreasing prostaglandin levels and cyclooxygenase-2 activity (182). These cytokines also induce osteoblastic cell migration (183) and influence the ability of osteoblasts to regulate osteoclast formation and activity through increased and decreased OPG and RANKL production, respectively (184). Direct IL-4 inhibition of osteoclast precursor cell maturation, which is mediated by STAT6, NF-κB, peroxisome proliferator-activated receptor γ1, mitogen-activated protein kinase signalling, Ca²⁺ signalling, NFATc1, and c-Fos, is stronger than that of IL-13 (185–189).

Macrophage migration inhibitory factor (MIF) is produced by T lymphocytes, pituitary cells, and activated macrophages (190). MIF overexpression causes high-turnover osteoporosis in mice (191). Interestingly, MIF-deficient mice did not lose bone mass or have increased osteoblast or osteoclast numbers following ovariectomy (192), suggesting MIF mediates the effects of estrogen withdrawal on bone. Estrogen downregulates MIF expression in activated macrophages (193), which may explain bone and bone marrow responses following ovariectomy. MIF is also made by osteoblasts, in which its production is upregulated by various growth factors, including TGF-β, FGF-2, IGF-II, and fetal calf serum (194). Finally, MIF increases MMP9 and MMP13 expression in osteoblasts (195) and inhibits RANKL- stimulated osteoclastogenesis in vitro (196).

Factors stimulating Toll-like receptors (TLRs)

TLRs are critical activators of innate immune responses that are highly expressed on antigen-presenting cells, such as dendritic cells and macrophages (197). Receptor ligation by microbial molecules or endogenous “danger” factors results in upregulated expression of costimulatory molecules and inflammatory cytokines. Because macrophages, dendritic cells, and osteoclasts have common progenitors, it is not surprising that TLRs are also detected in bone cells (198). Direct signaling via TLRs on osteoclast precursors, including TLR4, inhibits RANKL- mediated osteoclastogenesis (198), which appears counterintuitive because bacterial infections can cause inflammatory diseases characterized by reduced bone mineral density, such as periodontitis, osteomyelitis, and bacterial arthritis (199). Additionally, LPS may stimulate bone loss in mice via increased osteoclast numbers, and TLR activation can enhance osteoclast differentiation through RANKL and TNF-α expression in osteoblasts (200, 201). Our recent data suggest that TLRs inhibit osteoclast differentiation partly via the expression of type I IFNs; INF-β receptor-deficient monocytes are resistant to the TLR-mediated suppression of osteoclastogenesis (Y. C., unpublished data).

The basis for how TLR stimulation negatively regulates osteoclastogenesis, while at the same time bacterial infections are associated with excessive bone resorption by osteoclasts remains unclear. As described earlier, Gram-negative bacterial infections lead to alveolar bone destruction in periodontitis as a result of T-cell responses, upregulation of RANKL expression, and enhanced osteoclastogenesis (202). In the same study, bacterial infections in immunodeficient mice did not cause significant levels of alveolar bone loss, suggesting that bone loss associated with bacterial infections may be an indirect outcome of exacerbated T cell responses.

Similar to macrophages and dendritic cells, osteoclast precursors produce proinflammatory cytokines in response to TLR ligands (198). Moreover, although TLR stimulation inhibits osteoclast differentiation, osteoclast precursors treated with TLR ligands retain marked phagocytic activities. Therefore, TLR stimulation of osteoclast precursors likely results in a net enhancement of immune responses, which can be achieved by increased cytokine production and inhibiting their differentiation into nonimmune cells, such as mature osteoclasts. Because these cells can differentiate into mature osteoclasts if TLR ligands are removed (198), however, residual activated T cells present after a microbial infection is cleared can drive the phagocytic precursors to differentiate into bone-resorbing osteoclasts. Additionally, TNF-α produced by osteoclast precursors following TLR stimulation enhances bone resorption.

Conversely, the RANKL axis may regulate the inflammatory effects of TLR signaling. For example, a recent report suggested that LPS-induced production of proinflammatory cytokines was reduced in OPG-deficient mice, whereas it increased in RANKL-knockout mice, which increased the lethality if LPS injections (203). Moreover, mice pretreated with RANKL were partially protected from LPS-induced death. These results suggest that RANKL may suppress the cytokine response to LPS or other TLR ligands.

TLRs are thus likely to regulate the balance between immune responses and bone metabolism during acute attacks by various microbes. in vivo stimulation of TLRs, however, may have different effects on bone metabolism depending on the nature of the immune response. For example, continual stimulation of TLRs by commensal bacteria may affect bone metabolism, which is supported by recent data showing that mice deficient for mediators of TLR/IL-1R signaling pathways exhibit altered bone metabolism, although the precise signaling defects are unclear (72, 204).

Conclusion

Key cellular and molecular mechanisms governing homeostasis in the immune and skeletal systems have been described. Despite extensive cross-regulation between bone metabolism and the immune system, however, the mechanisms by which these systems regulate each other are poorly understood, due in part to the challenges typically associated with crossing disciplinary boundaries. While it is difficult for scientists and physicians to keep abreast of advances in their own field, it is even harder to develop the knowledge base and materials necessary to address issues that touch on multiple disciplines. Therefore, it is critical to create environments conducive to the study of multidisciplinary challenges. Awareness of intersystem crosstalk will no doubt contribute to our understanding of how bone and the immune system are physiologically regulated. Moreover, this endeavor may lead to better treatments for pathologies involving both systems, including inflammatory and metabolic bone diseases as well as tumor-induced bone lysis. Many of these processes are being targeted with therapeutics in the absence of a solid scientific understanding of the underlying molecular and cellular processes.

A report from the US Surgeon General on bone health suggested one in two Americans more than 50 years old will be at risk for fractures related to osteoporosis or low bone mass by 2020. These secondary health concerns are becoming more prominent as people live longer and remain more active as they age. Future preventative treatments for chronic bone-related diseases that are often associated with inflammation and impact patients’ quality of life will require a high degree of specificity, especially in populations already suffering from or vulnerable to other age-related ailments. We believe these issues place osteoimmunology in a position of unique clinical significance.