Type 1 regulatory T cells specific for collagen type II as an efficient cell-based therapy in arthritis

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by synovial inflammation and destruction of joint cartilage and bone and mediated by persistent synthesis of proinflammatory cytokines and matrix metalloproteinases. Proinflammatory cytokines such as interleukin 6 (IL-6), tumor necrosis factor α (TNF-α) and IL-1β are critical mediators in the inflammatory process of arthritis [1, 2]. In the past several years, biologic drugs have been developed to antagonize the effector cytokines, and neutralizing TNF-α or IL-6 has been proven to be successful in the treatment of RA. Despite the clinical benefit of such biologics aimed at ensuring broad immunosuppression, a nonnegligible proportion of patients eventually escape. For example, treatment failures can be related to the development of an immune response against the biologic itself, thus leading to loss of efficacy over time [3‐5]. As a consequence of these failures, there is still a need for new therapies with the aim of proactively restoring immune balance and reestablishing tolerance to joint antigens while avoiding systemic immune suppression.

Regulatory T (Treg) cells have been shown to play a crucial role in inhibiting autoimmune diseases and could be a valuable, interesting tool for use in therapeutic interventions, including in RA treatment. Indeed, Treg cells are ideal for this purpose because they suppress inflammation in an antigen-specific manner and can achieve selective and durable inhibition of pathologic inflammation without blocking protective immune responses against infection. The results of many animal model studies [6‐10], as well as clinical studies, have indicated a link between the efficacy of therapies against arthritis and the increase in the number or function of Treg cell populations [11‐14]. In addition, oral tolerization protocols developed several years ago have shown disease reduction in RA murine models and have recently been associated with the development of a population of Treg cells that suppress inflammation via IL-10 production [15, 16]. More importantly, treatment of RA patients with anti-TNF antibodies has been shown to induce differentiation of a potent population of Treg cells with suppressive activity that is dependent upon transforming growth factor β (TGF-β) and IL-10 [12, 13].

Because of the heterogeneity of human Treg cells, there is no consensus to date about which Treg cell population is optimally suitable for clinical use. Investigators in several phase I clinical trials have tested the ability of ex vivo–expanded CD4⁺CD25⁺ Treg cells to prevent graft versus host disease after allogeneic bone marrow transplantation [17‐19]. The use of such Treg cell populations has been hampered by the low number of circulating Tregs in humans and by the need to expand large amounts of suppressive cells for clinical use. Based on the therapeutic potential of type 1 regulatory T (Tr1) cells in inflammatory bowel disease (IBD) models [20] and on the possibility of generating large numbers of antigen-specific Tr1 cells in vitro [21, 22], a phase I/II clinical trial in which the researchers used ovalbumin-specific type 1 regulatory T (ova-Treg) cells has been performed in severe Crohn’s disease patients with production of very promising clinical data and indications of safety and efficacy [23]. These results and the possibility of isolating and expanding functional collagen type II–specific type 1 regulatory T (Col-Treg) cell clones from RA patients [22], suggesting that Tr1 cell–based therapy could be a novel therapeutic strategy for treating patients with RA.

Tr1 cells are characterized by the secretion of high amounts of IL-10. Adoptive transfer of antigen-specific Tr1 cells has been shown to be efficacious in suppressing inflammation in a variety of animal models for chronic inflammatory and autoimmune diseases, such as colitis, type 1 diabetes, multiple sclerosis, allergy and transplantation [20, 24‐28]. In these settings, the putative mechanism of action of antigen-specific Tr1 cells can be summarized in three steps: (1) homing to the site of inflammation, (2) local recognition of the specific antigen presented by local antigen-presenting cells and (3) local suppression of inflammation mediated by soluble factors and immunomodulatory surface molecules. By their mode of action, Tr1 cells are able to deliver local immunosuppression to the inflamed tissue.

Our goal in the present study was to evaluate the therapeutic potential of Col-Treg cells in experimental models of arthritis in order to propose their use in a new cell-based immunotherapy in joint inflammatory diseases, especially for patients with RA refractory to available treatments.

Several lines of evidence indicate a link between Tr1 cells and clinical improvement in RA patients. First, clinical response to anti-TNF therapy has been associated with the differentiation of a potent population of iTreg cells with suppressive activity that is dependent on TGF-β and IL-10 [12]. Also, disease remission in RA patients is associated with restored numbers of peripheral blood plasmacytoid dendritic cells that are implicated in the maintenance of tolerance through the induction of IL-10–secreting Tr1 cells [39]. In the present study, we evaluated, for the first time to the best of our knowledge, the clinical benefit of adoptively transferred in vitro differentiated Tr1 cells specific for a locally expressed antigen in two preclinical models of arthritis.

A key question about the use of Treg cells in therapy is to determine whether antigen-specific or polyclonal Treg cells are more suppressive in various clinical settings. In order to induce suppression, Treg cells must first be activated via TCRs by recognition of specific antigen major histocompatibility complexes. Authors of several reports have demonstrated that high TCR diversity appears to be crucial for optimal Treg-cell capacity [40]. The improved suppressive potential is probably due to increased numbers of activated Treg cells, which is underscored by the facts that (1) wild-type CD4⁺CD25⁺ Treg cells have higher suppressive capacity in lymphopenic mice [7] and (2) relevant tissue-specific antigen must be present in the periphery in order to generate fully protective Treg cells [41]. Concerning antigen-specific Tr1 cells, as clearly demonstrated in an IBD experimental model, the antigen specificity of the Tr1 cells is required for their local activation to trigger their suppressive function, even if the antigen (ova) is not implicated in the inflammation of the colon. In our study, we chose a self-antigen specifically expressed in the joint—Col II—to trigger local activation of Tr1 cells.

First, we demonstrated the feasibility of producing mouse Col II–specific Tr1 cells using an experimental protocol previously used for the production of ova-specific Tr1 cells. The Tr1 cell identity of Col-Tregs was confirmed by their cytokine secretion profile (IL-10^highIL-4^negIFN-γ^int) and by their in vitro immunosuppressive activity, which was dependent on IL-10 and TGF-β production [20, 26]. In contrast to thymus-derived CD4⁺CD25⁺ nTregs, Col-Treg cells upregulate FoxP3 and CD25 upon TCR activation and do not express CD62L, underlining their iTreg phenotype. However, as described for nTregs [42], activated Col-Treg cells do not express high levels of CD127. Overall, these mouse Col-Treg cells share similar characteristics with the human Col-Treg cells that we previously obtained from RA patients [22] and with ova-Treg cells derived from Crohn’s disease patients lymphocytes [21].

The data presented herein indicate that a single Col-Treg cell infusion leads to a significant decrease in clinical symptoms in both acute and chronic models of arthritis as measured by reductions in paw swelling, arthritic scores, clinical incidence and body weight loss, as well as by reduction of Col II–specific antibody titers. Interestingly, the results of a recent comparative study of antigen-specific nTregs and iTregs implicated greater efficacy for the latter, with a dose range similar to that which we used in our present study [43]. In both models used, the Col-Treg cells are detected in inflamed joints and axillary LNs following injection. These data provide evidence that Col-Treg cells are delivered to their site of action, (that is, in the arthritic joints), thus leading to a decrease in local inflammation. Our observations are similar to those of previous researchers who conducted distribution studies of infusion of CD4⁺CD25⁺ Treg cells in CIA, for which inhibition of local inflammation is associated with a decrease in IgG antibodies [6, 7].

IL-10 has been described as the main mediator of Tr1 anti-inflammatory effects. Systemic and locally administered IL-10 have previously been reported to have clear therapeutic potency in CIA [44‐50]. Interestingly, in those studies, the therapeutic effect of IL-10 on anti–Col II antibodies was associated with contradictory results that could be explained by the dose and timing of the IL-10 treatments. In our study, we found that the therapeutic effect of Col-Treg cells injected at the onset of clinical signs was associated with a global impact on anti–Col II IgG antibodies, whereas a preferential decrease in the anti–Col II IgG2a isotypes was monitored following earlier injection. IL-10 is also a potent suppressor of macrophage activation, and intraarticular IL-10 secretion has been reported to be associated with decreased proinflammatory cytokines [51]. The local IL-10–mediated decrease in TNF-α is a likely mechanism underlying the Tr1-protective effect observed in the CAIA experimental model, in which neutrophils and macrophages are important inflammatory cells and TNF-α and IL-1β secretion is pathogenic.

Extensive studies of the modes of action of Tr1 cells in the two experimental models fall outside the scope of our present study. However, to further investigate the in vivo suppressive function of the Tr1 cells, and notably their impact on the adaptive immune response, we provide evidence that ova-Treg cells significantly dampen the proliferation of antigen-specific CD4 effector T cells in DLNs. The antigen specificity of Tr1 cells is necessary to control their activation; however, we cannot exclude the possibility that the suppressive mechanism implies bystander suppression and impedes the proliferation of non-antigen-specific T effector cells. Although IL-10 has been considered to be the main mediator of a Tr1 anti-inflammatory effect for a long time, the concomitant expression of other anti-inflammatory cytokines, such IL-13, and immunomodulatory molecules, including GITR, CTLA-4, CD39 and granzyme B, suggest that Col-Treg cells display an arsenal of molecules able to suppress the autoimmune response through different modes of action, as previously described for nTregs and iTregs [33, 52]. Indeed, in addition to suppression of effector T cells via IL-10 and TGF-β secretion, expression of PD-1 (programmed cell death 1 cell surface protein) and CTLA-4 [53], adenosine production by CD39⁺ Tr1 cells [54] and granzyme B–dependent killing of myeloid cells has previously been found to be associated with a suppressive function of human Tr1 cells. The latter cytolytic activity on macrophages and neutrophils, which are key inflammatory cells in the CAIA experimental model, is an additional mechanism underlying the therapeutic effect of Col-Tregs in the CAIA model.

The results of the present pharmacokinetic studies of injection of Col-Treg cells in arthritis confirm previous observations in which ova-Treg cells were used in an IBD model [26]. In severe combined immunodeficiency mice reconstituted with CD4⁺CD45RB^high T cells, the injected ova-Treg cells were detected in inflamed tissue after 4 weeks, but not later. In our present study, we demonstrate that Col-Treg cells were maintained after 1 week but barely detectable after 7 to 10 days. These data indicate that infusion of Col-Treg cells did not lead to uncontrolled expansion in vivo, even in the continuous presence of their specific antigens. This observation is crucial in the context of further cell therapy protocols using Col-Treg cells in RA patients because lymphoproliferation could be seen as a major potential pitfall of such treatment. Furthermore, we did not observe uncontrolled proliferation of injected cells in long-term experiments in which Col-Tregs were injected into naïve or arthritic mice, that is, under permanent antigen exposure in the absence or presence of inflamed tissue.

Because of the heterogeneity of human Treg cells, there is no consensus to date about which Treg population is optimally suited for their various clinical uses. One of the major obstacles to the implementation of clinical protocols using the various Tregs is the low frequency of Tregs in the peripheral blood, which necessitates ex vivo multiplication of the cells prior to their use in vivo. The most commonly used expansion protocol at present is based on polyclonal stimulation using anti-CD3/anti-CD28 beads in the presence of high doses of recombinant IL-2, which is supplemented in some protocols with rapamycin. However, the expansion is usually antigen-nonspecific. More appealing for clinical application is the concept of expanding antigen-specific Treg cells to control their activation. Importantly we previously demonstrated the feasibility of producing high numbers of clinical grade functional Col-Tregs in RA patients [22]. Altogether, our results strongly support a further evaluation of the therapeutic potential of Col-Treg cells in human disease. The use of Tr1 cells in human models of disease is enhanced by their multiple mechanisms of action, and their local immunosuppressive action, in contrast to the systemic and single-target approaches used in the development of large numbers of biologics for RA that are already approved or in development. Patients with severe RA that is refractory to biologics represent a high unmet medical need, and antigen-specific Tr1 cells could be an innovative approach to dampening inflammation and restoring immune tolerance to joint antigens in this patient population.

The data presented herein demonstrate the therapeutic potential of antigen-specific Tr1 cells in two experimental models of arthritis and open great perspectives for the development of novel therapeutic approaches in the treatment of RA.