Histopathologic studies of the bone–pannus junction and subchondral bone marrow of patients with rheumatoid arthritis (RA) (1) indicate that osteoclasts play a pivotal role in the focal marginal and subchondral bone loss of inflammatory arthritis. Osteoclasts are multinucleated cells formed by fusion of mononuclear precursors of the monocyte/macrophage family under the influence of cell interactions and cytokines (2).
Physiologic osteoclastogenesis is driven by the tumor necrosis factor α (TNFα) family member RANKL (3), principally as a membrane-bound protein on the surface of marrow stromal cells or osteoblasts (2). Osteoprotegerin (OPG), a soluble decoy receptor, competes with RANK for binding to RANKL, preventing its osteoclastogenic effect (3).
In the context of inflammation, activated T cells and RA synovial fibroblasts express RANKL (4–10) and have the capacity to induce osteoclast differentiation (6, 8, 10, 11). In addition, several proinflammatory cytokines induce multinucleation of osteoclast precursors and/or commitment to the osteoclast phenotype (TNFα, interleukin-1β [IL-1β], IL-15, and IL-17 [12–16]) and may act synergistically with RANKL. In fact, the macrophage can serve both as osteoclast progenitor and as a source of osteoclastogenic cytokines (17, 18).
The cytokine IL-15 is expressed in both monocyte/macrophages and T cells (19–22) and has recently been described to enhance osteoclast differentiation (13). IL-15 secretion to the extracellular space has rarely been demonstrated (23–25); in contrast, IL-15 can be expressed on the cell surface, where it is able to exert biologic functions through cell contact–dependent mechanisms (26–28). IL-15 has been detected in the synovial fluid (SF) (29) and synovial membrane (22, 29) of RA patients. In addition, in vitro and in vivo studies (30, 31) suggest that IL-15 may be a major player in the pathogenesis of RA. Administration of soluble IL-15 receptor α (IL-15Rα) (30) or an antagonist mutant IL-15/Fc protein (31) prevents collagen-induced arthritis in mice and effectively reduces inflammation, synovial hyperplasia, and bone erosions. In addition, a phase I–II clinical trial in humans using a fully human anti–IL-15 monoclonal antibody (mAb) indicates that neutralization of IL-15 in patients with RA is effective and safe (32).
Phenotypic and functional differences have been described between peripheral blood (PB) T cells and monocytes of RA patients and cells of healthy controls, whereas RA SF T cells and monocytes demonstrate a clearly activated phenotype (33, 34). Our objective was to investigate the role of T cells from the PB and SF of RA patients in osteoclast differentiation of autologous monocytes and to study the cytokines implicated in this process. Our early arthritis clinic allowed the study of cells from patients with early RA who had not received disease-modifying antirheumatic drugs (DMARDs) or steroids, thereby minimizing interference of drugs with in vitro T cell and monocyte responses.
- Top of page
- PATIENTS AND METHODS
The role of T cells in RA has been controversial (33), and the presence of T cell–derived cytokines in the PB and SF of RA patients has been difficult to demonstrate (33). Here we have shown that freshly isolated PB T cells from patients with early RA and SF T cells from patients with established RA, but not PB T cells from healthy controls, demonstrate a significant surface expression of the recently described T cell–derived cytokines RANKL and IL-15. In addition, we observed that PB T cells from patients with early RA and SF T cells from patients with established RA, but not PB T cells from healthy controls, can induce osteoclastogenesis in cocultured autologous monocytes.
RANKL-mediated osteoclastogenesis has been described in cultures of unstimulated PB mononuclear cells from patients with psoriatic arthritis (41), multiple myeloma (42), and postmenopausal osteoporosis (43), but not in patients with RA. RANKL expression in synovial tissue (6, 7) and SF (44) T cells from RA patients has previously been reported, together with raised levels of the soluble form of RANKL (sRANKL) and decreased levels of OPG in RA SF (6). In addition, Kotake et al demonstrated that mitogen-activated human T cells induce RANKL-dependent osteoclast formation from human monocytes (6). Here we have extended this observation by showing for the first time that osteoclastogenesis in monocytes is driven by autologous freshly isolated T cells from the PB of patients with early RA and from the SF of patients with established RA, without additional exogenous stimulation.
Monocyte/macrophages may serve both as osteoclast progenitors and as a source of osteoclastogenic cytokines (17, 18). In fact, we observed that freshly isolated PB monocytes from patients with early RA and SF monocytes from patients with established RA, but not PB monocytes from healthy controls, demonstrated significant surface expression of IL-15, indicating an activated state that may facilitate stimulation of cocultured T cells and osteoclastogenesis. In addition, the higher levels of TNFα and IL-1β observed in cocultures of T cells/monocytes from RA patients compared with those from healthy controls reflect an increased monocyte cytokine production favored by the preactivated state of both RA T cells and RA monocytes.
In our system, experiments with OPG-Fc and neutralizing antibodies demonstrate that RANKL plays a pivotal role in ex vivo T cell–mediated osteoclastogenesis, and the cytokines IL-17, IL-15, TNFα, and IL-1β are important contributors that potentiate this effect. This is consistent with previous observations indicating a cooperation of cytokines in inflammation-mediated osteoclastogenesis (12, 14, 16). A low-level constitutive RANKL seems to be mandatory for osteoclastogenesis to take place, as indicated by experiments with RANKL-knockout mice (45), and the effect of even trace amounts of RANKL is potentiated by proinflammatory cytokines (46). In contrast, it has been reported that TNFα is able to induce osteoclastogenesis independently of RANKL (47). TNFα is an important player in inflammatory osteolysis (41,48) and synergistically cooperates with RANKL (46). In fact, TNFα induces RANKL synthesis by marrow stromal cells (46), and RANKL prompts TNFα expression by osteoclast precursors (17). In addition, the action of TNFα on RANKL expression has been described to be mediated by IL-1β (14). The hypothesis of cooperation among osteoclastogenic cytokines is supported by in vivo observations. Blockade of either IL-1β or TNFα does not completely arrest the periarticular erosions of inflammatory arthritis, whereas combined inhibition of both cytokines is significantly more effective (49).
IL-17 is a novel cytokine produced by activated T cells (15,50), and elevated levels of IL-17 have been described in SF from RA patients but not in SF from patients with osteoarthritis (50). In addition, CD4+,CD45RO+ T cells in synovial tissue of RA patients are immunoreactive with anti–IL-17 antibodies (15). IL-17 contributes to osteoclastogenesis by altering the RANKL/OPG balance (16), has the capacity to induce joint destruction in an IL-1–independent manner, and can bypass TNFα-dependent arthritis (51). Anti–IL-17 is of interest as a new therapeutic option for RA (16), particularly in patients in whom elevated IL-17 might attenuate the response to other biologics such as anti-TNFα and anti–IL-1β agents.
Experiments with Transwell inserts indicate that direct cell contact is mandatory to initiate the T cell/monocyte crosstalk resulting in osteoclastogenesis. Once intercellular crosstalk is initiated, sRANKL, IL-17, TNFα, and IL-1β are liberated that contribute to augmenting the osteoclastogenic effect, while OPG acts to neutralize RANKL. A decreased OPG:RANKL ratio in our RA coculture supernatants is an additional contributor (3). Soluble RANKL is liberated by a TNFα-converting enzyme-like protease that cleaves surface T cell RANKL (52), although membrane-bound RANKL has been demonstrated to be significantly more effective than its soluble form (53). The observed coculture-induced up-regulation of monocyte RANK, which mediates RANKL action on osteoclast precursors, further contributed to augmenting osteoclastogenesis. In our system, antiosteoclastogenic cytokines IL-4 (54, 55) and IFNγ (56) present in coculture supernatants of RA patients were not able to counterbalance the effect of proosteoclastogenic factors.
Consistent with previous reports (23–25), no secretion of IL-15 could be detected in our coculture supernatants. IL-15 acts through a heterotrimeric receptor consisting of a specific high-affinity binding α-chain (IL-15Rα) plus the IL-2R β- and common γ-chain that mediate signaling (23). The high affinity of IL-15Rα conditions an extremely rapid uptake of secreted IL-15, preventing detection of IL-15 in culture supernatants (24). Most of the IL-15 detected on cell surfaces is bound to IL-15Rα (24) and can stimulate in trans both βγ- and IL-15Rαβγ–bearing cells (24). The presence of surface IL-15Rα–bound IL-15 is synonymous with active IL-15 secretion, and the level of expression of surface IL-15 in a given cell population may reflect the rate of internalization of the IL-15–IL-15Rα complex. Thus, in contrast to IL-2, IL-15 can be expressed on the cell surface, where it is able to exert biologic functions through cell contact–dependent mechanisms (26–28).
Although not detected initially (57), T cells were later shown by more sensitive techniques to express IL-15 mRNA (19, 20) and protein (20, 21), and Thurkow et al described IL-15 protein expression in synovial tissue T cells from RA patients (22). IL-15 has been shown to enhance osteoclast differentiation, whereas IL-2, which shares receptor components with IL-15, has no effect on osteoclastogenesis (13). In our system, surface IL-15 on RA T cells and RA monocytes appears to be an important contributor to the observed ex vivo osteoclastogenesis.
MTX was effective at decreasing ex vivo osteoclast formation, consistent with results presented by Lee et al (58). MTX is the most commonly used drug in RA, and its mechanism of action is still being investigated (59). Low-dose MTX, as used for RA treatment, induces the release of adenosine to the extracellular space, and this autacoid seems to mediate the pharmacologic effect of MTX (37, 59). In our system, experiments with adenosine receptor antagonists suggest that the effect of MTX on ex vivo osteoclastogenesis is mediated through adenosine, acting on A2A receptors.
An in vivo antiresorptive effect of MTX has been demonstrated in rat adjuvant-induced arthritis (60). Clinical studies indicate that changes in radiographic progression in RA patients are directly related to fluctuations in disease activity (61). In fact, early and aggressive antirheumatic drug treatment affects the association of HLA class II alleles with progression of joint damage in RA (62). This interaction was independent of other prognostic factors, such as RF and baseline disease activity, suggesting that early and aggressive DMARD treatment can modify the dysregulated immune process (62). Accordingly, we observed that the number of new erosions at 1 year was significantly lower in patients in whom disease remission was achieved.
Our experiments with PB T cells and monocytes from patients with early RA who subsequently received oral MTX with or without low-dose prednisone showed that control of disease activity is associated with a down-regulation of surface RANKL and IL-15 together with decreased ex vivo osteoclastogenesis. At the same time, the baseline expression of surface T cell RANKL, T cell IL-15, and monocyte IL-15, together with baseline ex vivo osteoclastogenesis, were not associated with the number of new bone erosions at 1-year followup. This reflects the fact that treatment with DMARDs is able to modify the natural course of the disease and prevent joint destruction (62), and indicates that surface RANKL and IL-15 expression together with ex vivo osteoclastogenesis are markers of disease activity rather than independent predictors of radiographic progression.
The results presented here extend our understanding of the pathogenesis of bone erosion in RA, and provide an experimental basis for the use of biologic anticytokine agents or their combinations in patients who do not respond to conventional therapy.