Regulatory T cells directed to the site of the action

Autoimmune disease occurs when the mechanisms that control immunity break down, allowing T and B cells to respond to self antigens in a destructive manner (1). Treatments to combat the associated inflammation and tissue destruction have been limited to broad and nonspecific immune suppression. Moreover, little progress has been made in efforts to eliminate or inactivate self-reactive cells because of the diversity and breadth of the immune response. Recently, the field has undergone a paradigm shift wherein therapies seek to exploit the control mechanisms inherent within the immune system to create a durable state of immune tolerance (2). This momentum has been driven, in large part, by the discovery that regulatory T cells (Tregs) play an essential role in controlling immunity. The report by Wright et al. (3) in a recent issue of PNAS presents strategies for generating antigen-specific Tregs by using retroviruses to deliver TCR α- and β-chains. This approach, pioneered as a means to augment responses in cancer and HIV immunotherapy, results in the expression of tissue-specific receptors on Tregs, leading to regulation at the site of inflammation. The approach is likely to be applicable to many autoimmune diseases and transplant settings.

Tregs constitute 5–7% of peripheral CD4⁺ T cells, where they function to maintain homeostatic balance in the immune system (4). The loss of Tregs, as a result of failed development or experimental depletion, leads to multiorgan autoimmune disease (5). Conversely, augmenting Tregs has been shown to prevent, and in certain autoimmune settings, reverse ongoing disease (6). In these instances, the transferred antigen-specific Tregs are often more potent than polyclonal Tregs at abrogating disease (7). Expression of the key lineage transcription factor forkhead box P3 (FoxP3) is characteristic of Tregs and is required for their generation and maintenance (8). Tregs suppress immune responses through the expression of an array of contact-dependent negative regulators [i.e., via cytotoxic T-lymphocyte antigen 4 (CTLA-4)], as well as through the production of the immunoregulatory cytokines IL-10 and TGF-β (9). More recently, attention has been directed toward the capacity of Tregs to limit tissue inflammation directly by influencing biochemical pathways (10–12). The requirement for antigen recognition for T cell entry into tissues was recently demonstrated by Lennon et al. (13), who showed that autoreactive T cells entered the islets in a model of type 1 diabetes (T1D), whereas T cells expressing irrelevant TCRs did not. Thus, the effective application of Tregs relies on the capacity to direct these cells to the site of action [i.e., the joints in rheumatoid arthritis (RA) or the pancreas in T1D], based on the specificity of the TCR.

Efforts to obtain antigen-specific Tregs have been limited by two major hurdles. First, the antigenic targets in tissue-specific autoimmune diseases may be poorly defined (this is particularly the case in RA). Second, when antigen targets are known, only a limited number of antigen-specific cells can be isolated from circulation, given the low percentage of Tregs coupled with the low frequency of autoreactive T cells. Furthermore, discriminating Tregs from potentially destructive effector T cells is challenging, given the potential for overlapping surface markers. This raises the question of how to isolate and expand a sufficient number of these cells from the peripheral blood of patients. Wright et al. (3) addressed these limitations through two alternative strategies. First, the authors delivered a model TCR (OT-II) recognizing a peptide fragment of ovalbumin (Ova) to bona-fide natural Tregs (nTregs) by using a retrovirus encoding both α- and β-chains of the TCR (Fig. 1). The alternative approach involved starting with a large pool of conventional CD4⁺ T cells and targeting a TCR in combination with FoxP3, thus generating de novo Tregs. In an elegant series of studies, the investigators showed that both populations displayed characteristic markers of nTregs (CD25, CTLA-4, and GITR) and demonstrated suppressive activity. To address the antigen specificity of the cells, the investigators took advantage of the observation that Tregs require TCR triggering for acquisition of suppressor function, but once activated are capable of suppressing in a nonspecific fashion (14). Both TCR-redirected nTregs or FoxP3 and TCR-redirected CD4⁺ T cells were indeed capable of suppressing third-party responders. Although the experimental system used was highly artificial, it supports the notion that activated antigen-specific Tregs can suppress a broad array of responses.

Fig. 1.

The generation of antigen-specific regulatory T cells (Tregs). Retroviral vectors deliver TCR α- and β-chains to naturally occurring polyclonal Tregs, effectively redirecting their specificity. Alternatively, retroviral vectors delivered the Treg-associated transcription factor FoxP3 in combination with a TCR to conventional T cells to create de novo Tregs. Both strategies create antigen-specific suppressors that interact with antigen-presenting cells (APCs) to create a local immunoregulatory milieu. The figure was produced, in part, by using Servier Medical Art.

The investigators next tested the capacity of redirected Tregs to prevent T cell-mediated tissue damage in vivo by using a BSA-induced arthritis (AIA) model. In this context, the adoptive transfer of TCR-redirected nTregs, or FoxP3-converted OT-II TCR-expressing cells, before rechallenge, were shown to specifically suppress disease. Importantly, the investigators used congenic markers to assess the homing, antigen-specificity, and local inflammatory state resulting from arthritis-inducing T cells and TCR-redirected Tregs. The results indicated that adoptively transferred TCR-engineered Tregs substantially decreased swelling and tissue inflammation when the cognate antigen (Ova) was present, but had no effect in the control knee that lacked the antigen for the co-transferred Tregs. Moreover, activated Tregs were capable of suppressing the production of IL-17, a cytokine associated with autoimmunity (15). Interestingly, TCR-expressing nTregs were more potent at suppressing the development of these inflammatory T cells compared with conventional CD4⁺ T cells expressing ectopic FoxP3 and the TCR. Although they did not test the idea directly, the authors speculate that the differences in potency of nTregs compared with FoxP3-transduced cells might result from the expression of additional genes in nTregs that help to confer the full regulatory program (16). Further characterization of the two subsets is likely to enhance our understanding of the requirements for robust and long-lasting immunosuppression.

How can the findings by Wright et al. (3) be translated into the clinic to benefit patients with arthritis, as well as other autoimmune disorders? Although the authors demonstrated the efficacy of redirected Tregs in an experimental model of arthritis, additional studies are clearly warranted in model systems that more closely approximate the pathogenesis in humans. Specifically, it will be essential to examine the ability of redirected Tregs to suppress spontaneous autoimmunity after disease onset. Are these cells required to persist indefinitely, or will a short period of exposure to Tregs influence T cell development in the microenvironment, resulting in long-term tolerance?

Regulatory T cells (Tregs) play an essential role in controlling immunity.

While Wright et al. (3) demonstrate a powerful approach to generating large numbers of antigen-specific cells, important safety concerns remain for the application of this approach. One concern involves the potential for mispairing of the introduced TCR α- and β-chains with the endogenous receptors expressed in mature T cells, leading to unwanted specificities. Wright et al. addressed this concern by including disulfide linkages between the introduced TCR α- and β-chains. Although this strategy allows preferential pairing, it remains unclear whether the structure, and thus immunogenicity, of the engineered TCR is altered in this setting or whether heterologous pairing with endogenous receptors will be totally eliminated (17). Another concern is the inherent risk associated with the use of any integrating vector, including the potential for mutagenesis. Here, the safety profile of vectors will be of paramount importance, given that these cells would be targeted to patients with autoimmunity where there is a low risk of mortality. Finally, it is particularly important that antigen-specific Tregs remain stable and suppressive. Our lab has previously shown that a fraction of Tregs are capable of losing FoxP3 protein and suppressive function (18), a finding that in the context of potentially autoreactive T cells would be particularly dangerous. Little is known about what factor(s) control FoxP3 and Treg stability at sites of inflammation, although cytokines are speculated to play a role. In fact, functional defects in Treg activity have been reported to occur in RA, potentially contributing to disease development (19). Thus, it is essential that the site of action of TCR-redirected Tregs be determined as well as cell stability in the context of inflammation.

In sum, this study brings to light a novel mechanism by which Tregs can be redirected to protect tissues in autoimmune disease, even when the eliciting antigens are unknown. More studies in animal models and using human T cells will advance this effort toward therapeutic treatments with antigen-directed Tregs.