Although the etiology of rheumatoid arthritis (RA) is currently unknown, proinflammatory cytokines, such as interleukin-6 (IL-6), tumor necrosis factor α (TNFα), and interleukin-1 (IL-1), have been demonstrated to play key roles in the pathogenesis of RA (1). Furthermore, anti-TNFα therapy and therapies that inhibit IL-1 or IL-6 signaling have been shown to be effective in the treatment of RA. However, it is unknown whether these anticytokine therapies have modes of action distinct from each other. Importantly, a substantial proportion of patients show insufficient responses to a given anticytokine therapy. Thus, a clearer understanding of the protective mechanism of these therapies is important both in understanding the pathogenesis of autoimmune arthritis and in predicting an effective treatment for an individual patient.
Collagen-induced arthritis (CIA) is the most widely studied animal model of RA and has been useful not only in aiding understanding of the pathogenesis of this disease, but also in the development of new therapies (2). In mice with CIA, CD4 T cells are important in disease induction, and Th1 cells have been considered to be the major mediator of the disease. However, the notion that CIA is a Th1-mediated disorder has been challenged by studies using Th1-defective mice. Mice lacking interferon-γ (IFNγ), IFNγ receptor, or IL-12p35 develop accelerated arthritis after induction of CIA (1). Furthermore, recent studies have suggested that highly proinflammatory IL-17–producing Th17 cells, rather than Th1 cells, are central to the pathology of autoimmune arthritis (1, 3, 4). Consistent with the findings of these studies, neutralization (5) or genetic deletion (6) of IL-17 in mice inhibits the development of CIA. In patients with RA, IL-17 has been detected in synovial fluid, while IFNγ is rarely detectable (1).
To date, in vitro experiments using cultured murine CD4 T cells have revealed that the inflammatory cytokine milieu in the presence of transforming growth factor β (TGFβ) strongly induces Th17 differentiation (7). While some studies have demonstrated a role for IL-1 and TNFα in the development of Th17 cells in vitro (7, 8), the presence or absence of IL-6 is likely to be of particular importance, because stimulation of CD4 T cells in vitro with IL-6 plus TGFβ potently induces Th17 differentiation, whereas stimulation with TGFβ alone triggers the development of FoxP3+ immunosuppressive Treg cells (7, 9, 10). In humans, although the mechanism of Th17 differentiation is still poorly understood compared with that in mice, IL-6 has been shown to promote Th17 differentiation in vitro (11, 12). However, an IL-6–independent pathway of Th17 differentiation has also been reported (13, 14). Thus, the in vivo role of IL-6 in the development of Th17 cells remains to be established.
In this study, we investigated the protective effect of IL-6 blockade in vivo with respect to T helper cell development in CIA. We also examined the effect of TNFα blockade on Th17 differentiation in this model and compared it with that of IL-6 blockade.
- Top of page
- MATERIALS AND METHODS
- AUTHOR CONTRIBUTIONS
In this study, we demonstrated that Th17 cells represent the dominant effector T helper cell subset in CIA. Our findings are consistent with the crucial role of IL-17 in the pathology of CIA demonstrated by studies of IL-17–deficient mice or mice treated with anti–IL-17 antibody (5, 6). We also showed that partial depletion of Treg cells in CIA following administration of a neutralizing anti–IL-2 antibody results in delayed but enhanced arthritis, indicating that Treg cells play a significant protective role in reducing the severity of CIA. Thus, since the development of these 2 T helper subsets in vitro are modulated by proinflammatory cytokines, including IL-6 and TNFα, we compared the protective effect in vivo of IL-6 blockade with that of TNFα blockade in CIA, with respect to T helper cell development.
MR16-1 treatment on day 0 inhibited the induction of Th17 cells, indicating that IL-6 is a critical cytokine regulating Th17 development not only in vitro, but also in vivo, during the development of CIA. Moreover, the inhibitory effect of MR16-1 treatment on day 0 on Th17 development was also observed after the second CII immunization, and these mice developed less severe arthritis than did control mice with CIA. In contrast, when MR16-1 was administered on day 14, neither Th17 cell induction nor arthritis development was inhibited. These results indicate that the protective efficacy of MR16-1 is correlated with the inhibition of initial Th17 differentiation. These results also may explain why MR16-1 treatment at later points is much less effective than that at the early stage of CIA (15).
We found that MR16-1 treatment in vivo did not increase, but instead reduced, the frequency of Treg cells in CIA. Since the induction of CIA results in an increase in the frequency of Treg cells, possibly as a counteracting mechanism to limit inflammation, the reduced severity of arthritis we observed in MR16-1–treated mice may in turn lead to a reduction in Treg cell frequency in these mice. Although some evidence indicates that IL-6 inhibits TGFβ-induced generation of adoptive Treg cells in vitro, it remains unclear whether IL-6 interferes with the development of thymus-derived natural FoxP3+ Treg cells, a major population of Treg cells in vivo. In this context, we recently found that the frequency of FoxP3+ Treg cells is not reduced in IL-6–transgenic mice (Fujimoto M, et al: unpublished observations). Thus, it is likely that IL-6 plays a minor or limited role in the regulation of natural Treg cell development in vivo.
However, our analyses left open the possibility that MR16-1 may contribute to the induction of adoptive Treg cells in CIA. In CII-responsive CD4 T cells identified by CFSE dilution assay, in which CII-responsive adoptive Treg cells were presumably enriched, we found that MR16-1 treatment on day 0 slightly, although not significantly, increased the frequency of FoxP3+ Treg cells. However, even in this CII-responsive population, we observed a stronger effect of MR16-1 on Th17 cells than that on Treg cells. Thus, while MR16-1 may have an antiinflammatory effect by shifting the balance of Th17 and Treg cells, we believe at present that the major inhibitory action of MR16-1 on CIA is the suppression of Th17 development.
The present study did not reveal any protective mechanisms of IL-6 blockade other than the regulation of initial T helper responses. IL-6 is a pleiotropic cytokine involved in a variety of inflammatory responses. For example, IL-6 induces mononuclear cell recruitment at the inflammatory site. IL-6 can function as a growth factor for T cells and stimulate B cells to induce antibody production (21, 22). IL-6 can promote bone destruction by inducing synovial cell proliferation and osteoclast differentiation (23). These functions of IL-6 are believed to play a pathologic role in the effector phase of autoimmune arthritis and, presumably, are attenuated by IL-6 blockade. A clear limitation of MR16-1 (rat IgG1) compared with TNFR-Fc in CIA is that repeated injection of this antibody in mice is not possible because it induces an anaphylactoid reaction (Mihara M, et al: unpublished observations). However, in a previous study of glucose-6-phosphate isomerase–induced arthritis, another model of human RA, a single administration of MR16-1 either before or after disease onset showed protective efficacy (24), suggesting an effect of MR16-1 that is distinct from that on T helper cell differentiation. Nevertheless, the results of the present study suggest that the modulation of T helper responses, and particularly the inhibition of Th17 cell development, by the early administration of MR16-1 is the key protective mechanism for the observed inhibition of CIA.
As a comparison with MR16-1 treatment, mice with CIA were treated with TNFR-Fc, an established therapy for experimental arthritis models (25) and human RA. In accordance with the results of our in vitro experiments, TNFR-Fc treatment from day 0 to day 14 failed to inhibit Th17 development in CIA. Moreover, TNFR-Fc treatment at this phase did not suppress arthritis. However, unlike MR16-1 treatment, TNFR-Fc treatment after the second immunization successfully inhibited CIA without affecting the development of Th17 cells. These results clearly indicate that TNFα blockade by TNFR-Fc in vivo in CIA has no inhibitory effect on the induction of Th17 cells and, thus, suggest a critical role of TNFα in the effector phase of CIA. The effect of TNFα blockade cannot be attributed to the induction of Treg cells, since their frequency in mice treated with TNF-Fc beginning on day 21 was not increased, but rather, was slightly reduced, presumably reflecting the amelioration of arthritis in these mice. Taken together, these results suggest that there are considerable differences in the protective mechanisms of action between IL-6 blockade and TNFα blockade in vivo. Thus, the blockade of IL-6 may represent an alternative treatment for RA patients whose disease has previously failed to respond to TNFα blockade.
Clinical trials of anti–IL-6R antibody (tocilizumab) demonstrated that IL-6 blockade is highly efficacious in the treatment of established RA (26–28), and tocilizumab has been approved as a therapy for RA in Japan. It is likely, however, that antirheumatic drugs generally work better in early disease than in established RA (29). In addition, there is the hypothesis of “window of opportunity,” a concept that early diagnosis coupled with early therapy might induce clinical remission (29). In this context, anti–IL-6R antibody, as a modulator of initial T helper responses, appears to be an attractive therapy for early RA to induce clinical remission. A more aggressive therapy such as the combination of anti–IL-6R antibody and anti-TNF agents might be more advantageous, since these agents appear to have nonoverlapping mechanisms.
Future studies are needed to establish the effect of anti–IL-6R therapy on T helper responses in human RA and to assess its clinical impact on the outcome and prognosis of the disease.
- Top of page
- MATERIALS AND METHODS
- AUTHOR CONTRIBUTIONS
Dr. Naka 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.
Study design. Fujimoto, Kishimoto, Naka.
Acquisition of data. Fujimoto, Serada, Mihara, Uchiyama, Yoshida, Koike, Ohsugi, Nishikawa, Kimura, Naka.
Analysis and interpretation of data. Fujimoto, Naka.
Manuscript preparation. Fujimoto, Ripley, Naka.
Statistical analysis. Fujimoto.