Therapeutic potential of FOXP3⁺ regulatory T cells and their interactions with dendritic cells

Abstract

FOXP3⁺ regulatory T cells, a unique subset of T cells, are critical for orchestrating an immune response and preventing self-reactivity. With the increasing prevalence and unsatisfactory treatment of autoimmunity, allergic diseases, cancer and chronic infections, much attention has been focused on understanding their mechanisms of action in order to manipulate their function. One goal is to develop drugs or biologics that can enhance or abrogate their functions. Another approach is to utilize Tregs in adoptive cell-based therapy to treat autoimmune diseases or transplant-related complications. This review will focus on their therapeutic potential and mechanisms of action, particularly their interaction with dendritic cells.

1. Introduction

Regulatory T cells (CD4⁺FOXP3⁺, Tregs) are central to the maintenance of self-tolerance and the control of immune homeostasis [1,2]. Mutations in the FOXP3 transcription factor result in a rare human disorder called immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX), and an absence of Tregs [3]. Numerous mechanisms have been proposed to explain the suppressive functions of Tregs, mostly in murine studies, but none appears to be dominant [4]. Proposed mechanisms have included secretion of suppressor cytokines (IL-10, TGF-β, IL-35), CTLA-4/CD80-CD86 interactions, transfer of cAMP from suppressors to responders via gap junctions, generation of adenosine, IL-2 consumption and cell contact–mediated suppression by a yet-uncharacterized membrane molecule. Another important aspect of great controversy is what cell is the primary target of Treg-mediated suppression. A few studies have indicated a Treg–T-responder cell interaction [5], whereas others have strongly supported a Treg–dendritic cell (DC) interaction [6]. In this review, we will focus our discussion on the therapeutic potential of Tregs, problems involved in the purification and expansion of Tregs for use in adoptive cellular immunotherapy, and the role of the DC as the primary cellular target of Treg suppression.

2. Therapeutic potential of FOXP3⁺ Tregs

2.1. Autoimmunity

Of all the autoimmune conditions including autoimmune lymphoproliferative syndrome (ALPS) and autoimmune polyendrocrinopathy candidiasis ectodermal dystrophy (APECED), the IPEX syndrome has the most severe autoimmune phenotype affecting the gut, skin, endocrine and hematologic system. As both the IPEX syndrome in human and the scurfy mouse are caused by mutations in FOXP3 resulting in a deficiency in Tregs, it is now widely accepted that Tregs are critical for the prevention of autoimmunity and maintenance of self-tolerance early in life. A recent publication [7] has shed light into the continued importance and dominant role of Tregs in preventing fatal autoimmunity throughout the lifespan of mice. A diminished frequency or dysfunction of Tregs has been reported in many human diseases, including systemic lupus erythematosus (SLE) [8], rheumatoid arthritis [9], type 1 diabetes [10], multiple sclerosis (MS) [11], aplastic anemia [12], idiopathic thrombocytopenic purpura [13], and graft-versus-host disease (GVHD) [14], as well as transplant rejection [15]. However, many of these studies are flawed by the lack of convincing data on the purity of the Tregs based on FOXP3 analysis compared with findings in healthy controls. Therefore, it is still unclear whether Tregs play a major role in the pathogenesis of these diseases.

In contrast, the evidence is strong regarding the utility of Tregs for the treatment of autoimmunity in murine models. In murine studies, adoptive immunotherapy with Tregs has been shown to be effective in the prevention of experimental autoimmune encephalomyelitis (EAE) [16], type 1 diabetes [17], SLE [18], autoimmune gastritis [19], inflammatory bowel disease (IBD) [20], asthma [21], aplastic anemia [22], graft rejection [23], and GVHD [24]. However, there are still many unanswered questions. The majority of murine studies have used adoptive Treg immunotherapy for the prevention of autoimmune disease. It is unclear whether adoptive Treg immunotherapy can treat a disease once it has begun. Another major issue is whether antigen-specific Tregs are necessary for treating a disease or whether a polyclonal population is sufficient [25]. Antigen/organ-specific Treg preparations would most likely have a higher safety and efficacy profile. Aside from the positive effects of Tregs, the potential detrimental effects must also be considered. There is accumulating evidence suggesting that Tregs might play a negative role by inhibiting anti-tumor immunity [26] and maintaining chronic infections [27]. Until we have a better understanding of these issues, caution must be applied when using Tregs as cell-based therapy. The incorporation of a suicide gene as a safety measure would be prudent [28].

2.2. Transplantation

Currently, enhancing graft acceptance and preventing GVHD are probably the most fruitful areas for the application of Treg immunotherapy in human. A polyclonal population might be more beneficial to use as Treg immunotherapy during allogeneic hematopoietic stem cell transplantation (HSCT). An ongoing clinical trial is testing the effect of Treg immunotherapy during HSCT to promote graft acceptance and to prevent acute GVHD [29]. A related condition that might greatly benefit from Treg immunotherapy is chronic GVHD (cGVHD). Approximately 30–50% of patients receiving allogeneic HSCT will develop cGVHD, which can be life threatening or can severely impair the patient’s quality of life. With improvement in postgrafting immunosuppressive regimens, the number of individuals at risk for cGVHD is increasing. Treatment remains unsatisfactory and corticosteroids are still the mainstay of therapy. The pathophysiology and molecular mechanisms of this disease are still unclear. However, the loss of self-tolerance resulting in autoimmune manifestations may be a major component.

Given the dominant role of Tregs in maintaining immune homeostasis and self-tolerance, a deficiency or dysfunction of Tregs might play an important role in the pathogenesis of cGVHD. Several studies have attempted to evaluate the frequency and function of Tregs in cGVHD, but the data remain conflicting [30]. Some studies have suggested that the frequency or function of Tregs is reduced [24,31], whereas others have not found any difference [32,33]. However, one aspect of Treg physiology that has not been considered is whether Tregs develop a normal T-cell receptor (TCR) repertoire after HSCT. Tregs develop in the thymus and their TCR repertoire is shaped by interaction with thymic epithelial or dendritic cells [34]. Tregs have been shown to express a diversified TCR repertoire similar to, but also distinct from, that of CD4⁺FOXP3⁻ conventional T cells [35,36]. As the thymus microenvironment in patients who have received myeloablative treatment may be compromised, one major factor that might contribute to the development of cGVHD is that the development and selection of the TCR repertoire in Tregs, compared with conventional T cells, is abnormal or skewed [37,38]. A loss of diversity of TCR expression in the donor’s Treg population would results in an inability of the Tregs to inhibit host-reactive effector cells, leading to the development of cGVHD. As current therapy for cGVHD is inadequate, a clinical trial of cellular biotherapy with donor-derived Tregs should be undertaken if there is evidence of a restricted Treg population.

3. Purification and expansion of human Tregs

3.1. Expansion of Tregs

A major pitfall in the study of human Tregs and in obtaining sufficient numbers for cellular immunotherapy has been the inability to consistently achieve a highly purified FOXP3⁺ Treg product after ex vivo expansion secondary to the outgrowth of contaminating non-Tregs [39]. A CD4⁺FOXP3⁺ population of more than 90% purity can be isolated by fluorescence-activated cell sorting (FACS) of the top 2–4% of CD4⁺ T cells with low levels of CD127 and high levels of CD25 expression (CD127^lowCD25^hi) from peripheral blood, but frequently the percentage of FOXP3⁺ T cells decreases to 75% after 1 week and to 50% after 2 weeks of in vitro expansion. A highly purified FOXP3⁺ Treg product would reduce any potential adverse reactions from contaminating effector T cells and would enhance the efficacy and interpretability of clinical trials.

Different strategies have been developed to optimize the purity of Tregs. One method is the addition of rapamycin to the expansion cultures. Several reports [40,41] have demonstrated that FOXP3⁻ T cells are more susceptible to rapamycin-induced apoptosis, whereas Tregs are more resistant because of their constitutive expression of pim 2, a serine/threonine kinase with anti-apoptotic effects. Based on our experience, the addition of 25 nmol/l rapamycin only on day 0 does enhance the FOXP3 purity with an average of 81% (73–88%) on day 7, 71% (55–87%) on day 14, and 64% (42–84%) on day 21, but the large variations in purity remain an issue. Even when rapamycin was given more frequently and at a higher concentration, populations of expanded FOXP3⁺ with purity greater than 90% could not be achieved after 14 days of expansion [42,43]. The use of CD4⁺CD45RA⁺CD25^hi Tregs as the starting population also has been advocated to improve the purity of Tregs during the expansion [39]. However, it is difficult to isolate CD45RA⁺ Tregs from adult peripheral blood, as the vast majority (>80%) of FOXP3⁺ T cells are CD45RA⁻ memory cells [44]. Another concern is that any contaminating CD4⁺CD45RA⁺FOXP3⁻ cells would be highly susceptible to conversion to FOXP3⁺ non-Tregs by TGF-β in the culture medium used for expansion (see section 3.2) [45]. A recent publication has claimed that the vast majority of Tregs are CD49d⁻ and therefore depletion of CD49d⁺ cells allows for purification of Tregs free of contaminating effector cells [46]. In our healthy adult population, the majority of Tregs are not CD49d⁻ and there is high variability, as CD49d is upregulated with activation, particularly in patients with disease or recent infection. Equally important to the starting population of Tregs is the type of stimulation use for expansion. Although anti-CD3/CD28 conjugated beads are convenient, DCs [47] or artificial antigen presenting cells (APCs) [48] might be more efficient and physiologic. The ideal method for expansion of Tregs would be optimal yield and purity while maintaining function.

A highly purified Treg product after ex vivo expansion would not only reduce any potential adverse reactions, but would also enhance the efficacy and interpretability of multi-center clinical trials on Treg immunotherapy. Even if an expanded population is uniformly FOXP3⁺, it is unclear whether it is homogeneous or contains a mixture of bona fide Tregs and FOXP3⁺ non-Tregs (see section 3.2). We do observe a significant percentage of FOXP3⁺ cells in our expansion cultures that are cytokine producers (Fig. 1A). Therefore, we have developed a novel method to repurify the bona fide Tregs from the FOXP3⁺ non-Tregs and the FOXP3⁻ contaminating effector T cells based on selective expression of surface markers unique to the Tregs (manuscript submitted). Although several markers that have been implicated to be specific for Tregs, latency-associated peptide (LAP) [49] and GARP (LRRC32) [50], but not neuropilin-1 [51], are preferentially expressed on activated Tregs and not at their resting state (Fig. 1B). LAP is a homodimeric propeptide synthesized from the cleavage of TGF-β proprotein by a furin-like convertase [52]. LAP noncovalently associates with the dimeric mature TGF-β to produce the small latent TGF-β complex that is inactive. Therefore, the detection of surface LAP on activated Tregs indicates the presence of latent TGF-β. On the other hand, GARP is an 80-kDa transmembrane protein with an extracellular region composed primarily of 20 leucine-rich repeats [53]. Its function is unknown but has been shown to be expressed in megakaryocytes and platelets [54] as well as activated Tregs [50]. Using magnetic bead purification technology, it is feasible to purify and expand human Tregs, and then to use these markers to consistently re-isolate in large numbers a highly purified Treg product (>90% FOXP3⁺) that should be ideal for cell-based therapy (Fig. 1C). Most importantly, the purified Tregs are completely free of contaminating FOXP3⁺ and FOXP3⁻ non-Tregs. Regardless of what method is used to expand Tregs, our novel technique allows for the re-isolation of large numbers of highly purified Tregs that should facilitate the rapid advancement of the therapeutic application of Tregs in human disease.

Fig. 1.

Selective expression of LAP and GARP on activated human Tregs in expansion cultures allows their repurification. (A) CD25⁺ cells were purified with magnetic beads and expanded in vitro by stimulation with anti-CD3/CD28 Dynabeads and 100 U/ml IL-2 for 14 days. The cells were then restimulated for 5 hours with phorbol myristate acetate/ionomycin in the presence of brefeldin A and analyzed for expression of FOXP3 (clone 236A/E7) and intracellular cytokines. (B) Human FACS-sorted CD4⁺CD127^lowCD25^hi (CD25^hi) and CD25⁺ bead purified cells were expanded in vitro for 12 days, restimulated for 48 hours with anti-CD3/CD28, and then stained for surface LAP (clone 27232, R&D Systems), GARP (clone Plato-1, Alexis Biochemicals) or Neuropilin-1 (clone 446921, R&D Systems) expression and intracellular FOXP3. (C) CD25⁺ cells were expanded in vitro for 12 days and then restimulated for 48 hours with anti-CD3/CD28. Purified LAP⁺ and LAP⁻ or GARP⁺ and GARP⁻ fractions were isolated from the expanded cells using magnetic beads against LAP or GARP. Surface GARP, LAP, or isotypes and intracellular FOXP3 staining was performed.

3.2. De novo generation of Tregs

An alternative strategy for developing Treg immunotherapy would be the generation of Tregs from FOXP3⁻ T cells. This method would be more convenient than expanding natural Tregs and more feasible for generating antigen/organ-specific Tregs. In murine studies, induced Tregs (iTregs) can be generated from CD4⁺FOXP3⁻ T cells by TCR stimulation in the presence of IL-2 and TGF-β [55]. These iTregs have all of the phenotypic and functional properties of Tregs and exert potent suppressor function both in vitro and in vivo. We have failed to generate bona fide Tregs from naïve CD4⁺ T cells isolated from adult peripheral blood or cord blood [45]. Although it was possible to induce more than 70% of naïve CD4⁺FOXP3⁻ T cells to express FOXP3, they still lack regulatory functions. These FOXP3⁺ cells likely represent TGF-β–induced non-Tregs [45] generated during the expansion (Fig. 1A).

Thus far, all of our attempts to generate functional human Tregs in vitro from FOXP3⁻ precursors have failed. We tried multiple rounds of induction and have also tested various cytokines and agents such as retinoid acids, trichostatin A, and 5-azacytidine without much success. Although there are strong evidences in murine studies for the induction of Tregs by various subsets of DCs [56-59], we have not been successful in our human studies using various APCs from peripheral blood. It might be that tissue-derived DCs are critical for generating functional and stable induced Tregs. An obvious difference between the mouse and human TGFβ-induced FOXP3⁺ T cells is that optimal induction of FOXP3 in the mouse is seen earlier in the induction cultures at days 1–2, whereas significant percentages of human FOXP3⁺ T cells are observed only after 4–5 days of stimulation. Thus, the human cells may have committed to other pathways of differentiation before expression of FOXP3. Moreover, the level of expression of FOXP3 in iTregs in mice is similar to thymic-derived Tregs, whereas in human beings it is significantly lower. A recent study has demonstrated that the amount of FOXP3 per cell is critical determinant for the acquisition of Treg phenotype [60]. Therefore, it is essential to understand why the induction in the human beings results in a more delayed and lower expression of FOXP3. We are currently investigating whether differences in the regulation of TGF-β signaling exist between mice and human beings. The ability to generate antigen/organ-specific Tregs in vitro would create enormous therapeutic potentials.

4. Mechanism and cellular target of Treg-mediated suppression

4.1. Cellular target of Treg-mediated suppression

From a therapeutic standpoint, it is critical to not only analyze the frequency and function of Tregs but also to determine and characterize the target cells, such as the DCs and responder T cells. If the frequency or function of Tregs is abnormal, Treg immunotherapy would be a therapeutic option. However, if the target cells are resistant to suppression, or if a cytokine (IL-6, TNFα) or surface molecule from the target cells is subverting Treg function, a different therapeutic strategy must be taken. In this case, developing agents that can block or neutralize the subverting molecule or the use of tolerogenic DCs might be of greater therapeutic efficacy than Treg cellular therapy.

Use of Tregs as a therapeutic agent requires a critical understanding of their mechanisms of suppression and cellular target(s). Although it is well accepted that the primary suppressive mechanism of Tregs requires cell–cell contact, the critical molecules involved in this process are unknown. Moreover, there is controversy regarding the cellular target for Treg-mediated suppression. Some studies have strongly supported a Treg–T-responder cell interaction [5,61], whereas others favor a Treg–APC interaction [6,19]. Several murine studies have demonstrated both in vitro and in vivo the critical interaction between Tregs and DCs for activation and regulation of immune responses [62-64]. Therefore, identifying the primary target cell and the adhesion molecules involved in Treg cell contact–mediated suppression would provide a valuable opportunity to design therapeutic protocols for manipulating Treg function. However, it has been difficult to address these questions, as one of the major problems to defining the molecules involved in Treg-target cell interaction is that adhesion molecules, such as LFA-1/ICAM-1, are also necessary for the interaction of responder T cells with APCs. Blocking integrins such as LFA-1 in an in vitro suppression assay would affect the activation of the responder cells and the ability to measure suppression.

We have used the methods described above to develop a novel in vitro suppression assay using both resting and expanded, activated human Tregs as suppressors and murine CD4⁺CD25⁻ T cells and DCs as responders (Fig. 2A) [65]. We demonstrate that human Treg-mediated suppression can cross species (Fig. 2B) and requires LFA-1 on the Tregs, as suppression can be reversed with blocking anti-human CD11a or -CD18 mAb or using Tregs from LFA-1 deficient patients (Fig. 2C). We were able to bypass the requirement of LFA-1/ICAM-1 interaction for optimal activation of human Tregs by activating them with plate-bound anti-CD3/CD28 in the cocultures or using preactivated human Tregs. We could then define a critical role for LFA-1(CD11a-CD18)/ICAM-1(CD54) interactions in the cell contact process of Treg suppression. This human Treg–murine T-responder cell hybrid assay also provided a unique opportunity to decipher whether the Tregs predominantly target the DCs or the responder T cells. Because ICAM-1 is present on both T cells and DCs, we used responder CD4⁺ T cells from mice deficient in ICAM-1 expression and demonstrated significant suppression of murine CD4⁺ responder T cells under conditions in which human LFA-1 can interact only with ICAM-1 on the murine DCs and not on the responder T cells (Fig. 2D).

Fig. 2.

Human Tregs mediate their suppression by targeting DCs via an LFA-1/ICAM-1 dependent interaction to inhibit the upregulation of CD80/CD86 on the DCs and their ability to activate responder T cells. (A) Schematic of our novel human-murine suppression assay. Human FACS-sorted CD4⁺CD127^lowCD25^hi Tregs (hCD25^hi) are activated in the co-cultures by plate-bound anti-hCD3 (5 μg/ml) and anti-CD28 (2.5 μg/ml) mAbs and murine CD4⁺CD25⁻ responders (mCD4⁺) are stimulated with murine splenic dendritic cells (mDC) in the presence of soluble anti-mCD3 (0.25 μg/ml) mAbs. (B) mCD4⁺ T cells (50,000) are stimulated with 10,000 mDCs and soluble anti-mCD3 alone (black square) or with increasing number of mouse (mCD25⁺) or human (hCD25^hi) Tregs. (C) mCD4⁺ T cells are stimulated with mDCs and anti-mCD3 alone (red bar) or at 1:1 ratio with 48 hours preactivated hTregs (+ act.hCD25^hi, blue bar) in the presence of isotype control or blocking anti-hCD11a (efalizumab) or −hCD18 (TS1/18) mAbs. (D) mCD4⁺ T cells from wild-type (WT) or ICAM-1^−/− mice are stimulated with wild-type mDCs and anti-mCD3 alone (red bar) or at 1:1 ratio with 48-hour preactivated hTregs (+ act.hCD25^hi, blue bar). Proliferation was measured by [³H]-TdR incorporation on day 3 (B–D). Preactivated hTregs were stimulated for 48 hours with plate-bound anti-hCD3/CD28 and IL-2 then used in the suppression assay without restimulation during the co-cultures (C, D).

The interaction of the human Tregs with murine DCs is accompanied by a marked inhibition of the expression of CD80/CD86 on the DCs and their capacity to optimally activate responder T cells. Numerous studies in human beings [66-68] and in mice [69-72] have shown the unique ability of Tregs to inhibit the upregulation of activation markers and proinflammatory cytokines on DCs. Of particular interest is the ability of human Tregs not only to suppress costimulatory molecules and T-cell stimulatory activity on monocyte-derived DCs, but also to increase their expression of CCR7 and HLA-DR and to enhance their ability to migrate in response to CCL19 [68]. A recent study showing the importance of LFA-1 on murine Tregs and their interaction with DCs supports our findings [73]. In contrast to a recent report [74] of a murine study asserting that CTLA-4 on Tregs played a major role in mediating the inhibition of CD80/CD86 expression on murine DCs, we were unable to abrogate the human Treg–mediated suppression of murine DCs with blocking antibodies to human CTLA-4 or with anti-IL-10 or anti-TGFβ. Although the LFA-1/ICAM-1 interaction is required as the first step in Treg mediated downregulation of DC function, the characterization of other cell surface molecules on Tregs and their cellular receptors on DCs that are involved in this process remains to be determined. A complete understanding of this process should offer a potential therapeutic approach to augment or reverse Treg suppressor function. Our findings also caution against the use of biologic agents such as efalizumab, a humanized monoclonal anti-CD11a approved for the treatment of psoriasis and in clinical trials for other autoimmune diseases, which might have a detrimental effect if there is an unbalanced inhibition of Treg function over effector T-cell activation.

5. Conclusions

It has been more than 10 years since the rediscovery of Tregs and 5 years since the identification of FOXP3 as the transcription factor that controls many aspects of their differentiation and function. Although much progress has been made in characterizing their phenotype and demonstrating their broad range of functions, their mechanism(s) of suppression is still unclear. The prevailing view [75] is that FOXP3⁺ Tregs can use multiple mechanisms of suppression, and their choice of mechanism may depend on the context of the response that they are attempting to regulate. Thus, a sterile inflammatory response may require a cell contact–mediated mechanism, whereas severe inflammatory response such as those driven by intestinal bacteria may require that the Tregs produce anti-inflammatory cytokines. Tregs may use TGF-β bound to their cell surface to amplify their function through a mechanism of infectious tolerance by converting effector T cells into iTregs [76]. The apparent diversity of Treg suppressor mechanisms has hindered the development of therapeutic agents to either enhance or abrogate their function.

A number of studies suggest that Tregs, at least in vitro, can directly target virtually every bone marrow-derived cell population including mast cells [77], natural killer cells [78], invariant NKT cells [70], and B cells [80]. On the other hand, our studies using the human–murine hybrid assay strongly suggest that DCs are their major target. The capacity of Tregs to modulate the ability of DCs to present antigens or upregulate costimulatory molecules is consistent with the ability of this relatively minor population of lymphocytes to exert potent suppressor function in vivo. Modulation of DC function by Tregs also explains the unique ability of Tregs to mediate “bystander suppression” in which activation of Tregs by one antigen presented by a DC results in inhibition of the presentation of other antigens associated with the same DC. In contrast, in an environment in which potent activation of effector cells must pre-dominant, DCs can counter-regulate the Tregs by upregulating inhibitory molecules to block their functions or by producing large quantities of cytokines that can hyperactivate the effector cells and render them resistant to suppression [72,81]. Neutralizing of these molecules or adoptive immunotherapy with tolerogenic DCs might be beneficial to recondition the Tregs. Therefore, it is critical to comprehensively understand the interaction between the DCs and Tregs in order to design effective treatments. If the DCs are highly activated and resistant to suppression, infusion of Tregs into patients might not be effective. Given the current lack of effective treatments for autoimmunity and transplant-related complications, both of which inflict a huge medical and socioeconomic burden, there is a motivating drive to develop novel therapies to modulate or cure a disease. To achieve this goal, we will have to gain more insights into a disease to develop and tailor our therapies appropriately. Although the road is not straight and smooth, there are great promises and excitement over the horizon.