Many studies have investigated the hypothesis that activated T cells directly or indirectly modulate the formation and function of osteoclasts in the bone resorption associated with rheumatoid arthritis (RA), ever since RANKL, as a factor inducing osteoclastogenesis, was cloned in 1997–1998. We and other investigators have shown that activated T cells expressing RANKL induce osteoclastogenesis (1–3). In addition, in 1999, we demonstrated that interleukin-17 (IL-17), which was cloned in 1995, potently induces osteoclastogenesis via the expression of RANKL on osteoblasts (4). In 2000, it was reported that activated T cells also produce interferon-γ (IFNγ), which potently inhibits osteoclastogenesis from monocytes, leading to degradation of tumor necrosis factor (TNF) receptor–associated factor 6 (TRAF6) (5). The role of activated T cells in osteoclastogenesis has thus been controversial.
The notion that IL-17–producing T cells may be a distinct T cell lineage emerged in 2003 from gene-targeted mouse models of experimental autoimmune encephalomyelitis (EAE) or collagen-induced arthritis (CIA) (6). In 2005 in simultaneous studies from 2 laboratories, it was revealed that IL-17–producing CD4+ T cells (also referred to as Th17 or ThIL-17 cells) are a distinct effector lineage (7, 8). In 2006, Sato et al reported that murine Th17 cells function as a murine osteoclastogenic T helper cell subset in the presence of osteoblasts (9). Thus, in studies utilizing murine cells and experimental arthritis models, it has been revealed that Th17 cells induce osteoclastogenesis, whereas Th1 cells producing IFNγ inhibit osteoclastogenesis.
In this issue of Arthritis & Rheumatism, the study by Rifas and Weitzmann (10) gives us very important insights into the role of T cells in human osteoclastogenesis. Those authors previously reported the occurrence of novel cytokine-like activities in culture medium conditioned with activated T cells, which potently induced osteoblastic IL-6 production (11) and directly stimulated osteoclastogenesis in a manner independent of RANKL (12). In the current study, Rifas and Weitzmann identified the factor secreted by T cells that is responsible for these activities, which they named secreted osteoclastogenic factor of activated T cells (SOFAT). Their study included 3 important points as follows: 1) they cloned a novel cytokine, in the absence of a classic cytokine-like motif, that could induce both human and mouse osteoclastogenesis, independent of both RANKL and osteoblasts; 2) the novel factor was found to be produced by T cells; and 3) the factor was cloned with the use of human cells.
SOFAT as a novel cytokine in the induction of osteoclastogenesis
In 1999, Rifas and Avioli reported that a novel T cell cytokine that is resistant to cyclosporin A (CSA) stimulates IL-6 in human osteoblastic cells (11). They used cultures of activated human T cells and human osteoblasts from rib specimens to study the possibility that lymphokines act on osteoblasts to produce IL-6. Purified T cells activated with a combination of anti-CD3 and anti-CD28 antibodies were cocultured with human osteoblasts in direct physical contact or separated by a transwell system. Conditioned media were assayed for IL-6 production. After a 72-hour incubation period, activated T cell–human osteoblast interaction resulted in a 100-fold increase in the production of IL-6 over basal levels. Treatment with CSA inhibited the production of TNFα and IL-6 from T cells but did not inhibit the T cell induction of IL-6 from human osteoblasts. Assays of activated T cell–conditioned medium with human osteoblasts revealed that a soluble factor, not cell–cell contact, was the major inducer of IL-6. Thus, in that study, Rifas and Avioli found that activated T cells produce a novel, potent IL-6–inducing factor (11).
In 2001, Weitzmann et al reported that T cell activation induces human osteoclast formation via RANKL-dependent and -independent mechanisms (12). Those authors investigated the ability of activated T cells to support human osteoclastogenesis. When T cells were activated by phytohemagglutinin, they stimulated human osteoclastogenesis from peripheral blood stem cells, a blood product enriched in CD34+ stem cells. Although both soluble macrophage colony-stimulating factor (M-CSF) and RANKL were detected in activated T cell supernatants, the presence of M-CSF was not essential for macrophage survival or RANKL-dependent osteoclastogenesis, suggesting that other soluble T cell–derived factors were capable of substituting for this cytokine. Saturating concentrations of osteoprotegerin (OPG) failed to neutralize 30% of the observed osteoclastogenesis. In addition, the T cell–conditioned medium was capable of eliciting a further ∼30% increase in osteoclastogenesis over that induced by saturating the concentration of RANKL and M-CSF. Taken together, these findings suggest that additional redundant mechanisms for osteoclastogenesis by activated T cells, which do not involve RANKL or M-CSF, are present (12).
On the basis of the findings from these 2 prior studies using human cells, Rifas and Weitzmann performed the present study (10) and found that a novel T cell cytokine, SOFAT, induces osteoclast formation in a RANKL-independent manner. SOFAT is derived from an unusual messenger RNA splice variant encoded by the threonine synthase–like 2 (THNSL) gene homolog, a conserved gene remnant coding for threonine synthase, which functions as an enzyme only in microorganisms and plants. Interestingly, SOFAT may represent the first in a potential family of novel cytokines possessing biologic activities in the absence of a classic cytokine-like motif (10).
SOFAT is secreted by activated T cells via a calcineurin-independent pathway, because it was observed that treatment of activated T cells with CSA failed to block the production of this cytokine (11), suggesting that production of SOFAT by T cells is stimulated by an intracellular pathway different from that utilized by RANKL (Figure 1). SOFAT induces IL-6 production by human osteoblasts. In addition, SOFAT induces human osteoclast formation from monocytes in the absence of exogenous RANKL or osteoblasts. The osteoclast formation induced by SOFAT is sensitive to the effects of both CSA and FK-506, suggesting that SOFAT signaling is associated with the NFAT signal transduction pathway, like that of RANKL (10).
SOFAT may be an important osteoclastogenic factor in the bone destruction associated with RA (Figure 1). In the study by Rifas and Weitzmann (10), TNFα potently amplified the osteoclastogenesis induced by SOFAT. It is possible that production of IL-6 by osteoblasts stimulated through SOFAT induces RANKL expression by osteoblasts, thus inducing osteoclastogenesis from monocytes. We have previously demonstrated that IL-6 and soluble IL-6 receptors in synovial fluid from patients with RA are responsible for osteoclastogenesis (13). In addition, IL-6 may modulate the production of T cell–derived osteoclastogenic cytokines such as IL-17 and RANKL (10), leading to a feedback mechanism between T cells and osteoblasts (Figure 1), as was also demonstrated by Wong et al in a study using the antigen-induced arthritis model in mice (14). Indeed, it has been demonstrated that IL-6 is required for the differentiation of Th17 cells (6).
Characterization of human osteoclastogenic T cells
Expression of RANKL on T cells, as well as on osteoblasts, induces osteoclastogenesis. Two groups, Kong et al and Horwood et al, using mouse studies, demonstrated that T cells directly induce osteoclastogenesis by inducing the expression of RANKL (1, 2). In 2001, we also demonstrated, in a study using human T cells and monocytes, that human T cells induce osteoclastogenesis from peripheral human monocytes through the expression of RANKL (3). In addition, we observed that CD4+ T cells in synovial tissue from patients with RA express RANKL, and that the ratio of levels of RANKL to those of OPG in synovial fluid is elevated (3). Thus, in the joints of patients with RA, monocytes are likely to differentiate into osteoclasts.
In strong support of the hypothesis that activated T cells induce osteoclastogenesis, 2 groups demonstrated important findings in 2006. First, Miranda-Carus et al reported that peripheral blood T cells from patients with RA promote osteoclastogenesis from autologous monocytes through the expression of RANKL on T cells, although the T cells simultaneously produced both OPG and IFNγ (15). Thus, they concluded that T cells are important contributors to the pathogenesis of bone erosion in RA through an interaction with monocytes. Second, Geusens et al, in a 5-year study of 92 patients with early active RA (16), observed that the ratio of circulating OPG levels to RANKL levels in early RA is predictive of the development of later joint destruction. Thus, these reports on the association between osteoclastogenesis and bone destruction in patients with RA support the hypothesis that activated T cells induce osteoclastogenesis.
The pathogenesis of periodontitis is similar to that of RA, because both diseases show chronic inflammation. Mogi et al demonstrated that the levels of RANKL are elevated in the gingival crevicular fluid of patients with periodontitis (17). In addition, in 2006, Vernal et al reported that T cells from patients with periodontitis express RANKL (18). Interestingly, in 2000, Teng et al reported that transplantation of human T cells expressing RANKL into NOD/SCID mice induces alveolar bone destruction (19). In 2007, Kawai et al reported that the induction of an adaptive immune response to orally colonizing nonpathogenic Pasteurella pneumotropica by immunization with a phylogenetically closely related bacterium, Actinobacillus actinomycetemcomitans, results in T cell–derived RANKL-dependent periodontal bone loss in mice (20). Thus, T cells expressing RANKL play a critical role in bone destruction associated with periodontitis.
In 1999, we demonstrated that IL-17 in mice potently induces osteoclastogenesis from murine monocytes through the expression of RANKL on osteoblasts (4). In addition, we recently reported that IL-17 in humans induces osteoclastogenesis from human monocytes even in the absence of osteoblastic cells or soluble RANKL, through both inductively expressed TNFα and constitutively expressed RANKL on human monocytes (21). During osteoclastogenesis, the synergistic effect of TNFα and RANKL plays an important role; the expressed level of each cytokine alone is too low to induce osteoclastogenesis (21). This synergism has also been reported by 2 other groups, Lam et al in 2000 (22) and Zou et al in 2001 (23). More recently, Miranda-Carus et al reported that peripheral blood T cells from patients with early RA promote osteoclastogenesis in autologous monocytes in the absence of exogenous cytokines or osteoblasts, and that osteoclastogenesis is significantly inhibited by neutralizing monoclonal antibodies to IL-17 (15). Thus, IL-17 induces osteoclastogenesis from monocytes both in the presence and in the absence of osteoblasts.
Synovial tissue in patients with RA contains an increased number of T cells. However, until 1999, it was very difficult to detect cytokines derived from T cells, such as IL-2 or IFNγ, in synovial tissue or synovial fluid with the use of a conventional enzyme-linked immunosorbent assay (ELISA), because of the sensitivity of this technique. This absence of recognized T cell–derived cytokines was an enigma to be solved in understanding the pathogenesis of RA. In 1999, using a conventional ELISA, we demonstrated that a sufficient amount of a T cell–derived cytokine, IL-17, is detectable in the synovial fluid of patients with RA (4). In 2004, Lubberts et al showed that treatment of mice with a neutralizing anti-mouse IL-17 antibody after the onset of CIA reduces joint inflammation, cartilage destruction, and bone erosion (24). In addition, in 2003, Nakae et al observed that IL-17 plays an important role in the pathogenesis of arthritis in another murine model (25). In 2007, Sarkar et al demonstrated that dendritic cells genetically modified to express IL-4 exert a therapeutic effect on CIA by targeting IL-17 (26). More importantly, in 2005, Raza et al reported that in patients with early RA of ≤3 months' duration (mean 9 weeks), the disease is characterized by a distinct and transient synovial fluid cytokine profile of T cells, including IL-17, but not IFNγ (27); this latter study underlines the fact that the disease duration is very important in studies on the role of cytokines in the pathogenesis of diseases. Thus, IL-17 plays an important role in both human RA and murine arthritis models.
A novel subset of T helper cells was identified in 2005. T helper cells expressing IL-17, being distinct from Th1 or Th2 cells, were demonstrated and designated Th17, ThIL-17, or inflammatory TH cells. The notion that IL-17–producing T cells may be a distinct T cell lineage emerged from gene-targeted mouse models of EAE or CIA (6). Furthermore, in 2005, reports from 2 laboratories simultaneously described IL-17–producing CD4+ (Th17, ThIL-17) T cells as a distinct effector lineage (7, 8). In 2006, it was reported that naive T cells are primed to become Th17 cells by IL-6 and transforming growth factor β (TGFβ), and that IL-23 stimulates the production of IL-17 from primed T cells (6).
In 2007, Acosta-Rodriguez et al reported that human Th17 cells exhibit a distinct migratory capacity, in that human Th17 cells could express both CCR6 and CCR4 and had antigenic specificities for Candida albicans (28). Results from 3 different studies using mouse models of infection, Ye et al in 2001 (29), Happel et al in 2005 (30), and Rudner et al in 2007 (31), have suggested that Th17 cells develop to mediate protection against extracellular bacteria and Pneumocystis carinii. In addition, Sato et al reported that Th17 functions as an osteoclastogenic T helper cell subset that links T cell activation and bone destruction (9). In 2007, Hirota et al found that autoimmune arthritis in SKG mice is highly dependent on the development of Th17 cells (32). In addition, IL-17 plays an important role in the pathogenesis of periodontitis, as described above. Thus, Th17 cells may be involved in bone destruction in RA and in periodontitis, as well as in the pathogenesis of autoimmune disorders in mouse models such as EAE and autoimmune arthritis, whereas Th17 cells develop to mediate protection against fungi and extracellular bacteria.
IFNγ produced by Th1 cells plays an important role in the host defense against viruses or intracellular microorganisms. It has been reported that IFNγ inhibits osteoclastogenesis (33, 34). In 2000, IFNγ was shown to inhibit mouse osteoclastogenesis through the degradation of TRAF6 in vitro and to repress bone resorption in calvariae from prepubertal mice in vivo (5).
We have demonstrated that activated human T cells induce human osteoclastogenesis from monocytes though the expression of RANKL (3), as described above. Moreover, we have shown that human IFNγ-producing T cells (Th1 cells) induce human osteoclastogenesis from monocytes via the production of RANKL from Th1 cells, and that the number of T cells expressing both IFNγ and RANKL is elevated in the peripheral blood of patients with RA (35). In strong support of our findings, Gao et al (a study in which Weitzmann was a coauthor) demonstrated that IFNγ exhibits indirect pro-osteoclastogenic properties in vivo, since IFNγ induced bone resorption in 3 mouse models of osteoporosis under conditions of estrogen deficiency, infection, and inflammation (36). In addition, Miranda-Carus et al reported that peripheral blood T cells from patients with RA promote osteoclastogenesis from autologous monocytes through the expression of RANKL on T cells, although T cells simultaneously produce IFNγ (15). Moreover, Stashenko et al demonstrated, in a murine model of periodontitis, that a Th1-mediated immune response promotes severe bone resorption caused by Porphyromonas gingivalis (37). Thus, IFNγ or Th1 cells may play a critical role in indirect bone resorption and destruction in a murine model in vivo and in patients with RA, whereas IFNγ directly inhibits murine and human osteoclastogenesis from monocytes in vitro.
On the basis of these findings, we would suggest that the term osteoclastogenic T cells be used more frequently to describe activated T cells that are able to induce osteoclastogenesis. Murine Th17 cells are osteoclastogenic T cells; however, it remains to be elucidated whether human Th17 cells are osteoclastogenic T cells. In addition, the characteristics of T cells producing SOFAT should be examined more closely, although it has been suggested that T cells producing SOFAT are both CD4+ T cells and CD8+ T cells (12). It would be interesting to investigate whether SOFAT induces the differentiation of the CD4+ T cells that produce it, since it is known that IFNγ, IL-4, IL-21, and TGFβ play positive feedback roles in the differentiation of Th1, Th2, Th17, and regulatory T cells, respectively (38).
Of note, in the study by Rifas and Weitzmann in this issue of Arthritis & Rheumatism (10), human cells were used to identify the novel factor. In basic science, usually murine cells are studied, because these cells and experimental tools to facilitate their use can be easily obtained. In addition, it is possible to create transgenic and knockout murine models. Indeed, biologic science has made marked progress with the use of murine cells and disease models. In the study by Rifas and Weitzmann (10), the authors showed that recombinant human SOFAT could induce mouse osteoclastogenesis from RAW264.7 cells, thus demonstrating that a mouse model offers significant advantages in terms of experimental manipulation. Conversely, there are many disadvantages to the use of human cells in such studies. For example, it is difficult to obtain human cells, and it is impossible to perform an in vivo study. However, to investigate the pathogenesis of RA, it is critical to investigate osteoclastogenesis using human cells. In bone cell biology, some cytokines show functions that differ between humans and mice. For example, M-CSF induces colony formation in mouse cells, whereas in human cells, M-CSF usually induces the differentiation of progenitor cells of monocytes, rather than colony formation, although Motoyoshi et al reported initially, in 1982, that M-CSF induced colony formation in human macrophages (39).
Furthermore, some cytokines show levels of expression that differ between humans and mice. For example, Kanamaru et al reported that the expression of membrane-bound RANKL is limited in human T cells compared with mouse T cells (40). Moreover, the mouse CD4+CD25+ regulatory T cell subset can be isolated from all CD25+ T cells, regardless of the level of CD25 expression; however, when similar criteria are followed to isolate these cells from human blood, CD25+ cells (high and low combined) do not exhibit an anergic phenotype or significant suppressive function. In 2001, Baecher-Allan et al demonstrated that in humans, CD4+CD25high cells exhibit all of the properties of regulatory T cells (41). In addition, Amadi-Obi et al reported that a major immunologic difference between humans and mice is the presence of CD4+ T cells producing IL-17 (Th17 cells), which was demonstrated in the peripheral blood of healthy humans but was not observed in mice (42). Recently, several studies have also demonstrated the existence of substantial numbers of human CD4+ T cells that are able to produce both IL-17 and IFNγ, and in fact, the term Th17/Th1 cells has been proposed (43). Thus, in the study of human diseases, it is essential to investigate human osteoclastogenesis using human cells; recognition of the differences between the species used in studies is critical for understanding the function of cytokines. We suggest that the term “human osteoclastology” be used to describe studies of human osteoclastogenesis.
The role of T cells in human osteoclastogenesis has been clarified in vitro. However, the function of osteoclastogenic T cells in bone resorption has not been fully elucidated in humans, including patients with RA, whereas stimulation of bone resorption by osteoclastogenic T cells has been demonstrated in a murine model in vivo. Besides Th17 cells, T cells producing SOFAT should be considered osteoclastogenic T cells, as was demonstrated in the study by Rifas and Weitzmann (10). Thus, further investigation is necessary to clarify the role of T cells producing SOFAT in the bone resorption associated with human diseases such as RA.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kotake had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.