Development of regulatory T cells—a look back

More than 50 years ago, it was shown that thymectomy in mice results in a wasting disease and autoimmune phenomena [1, 2] suggesting the presence of immune regulatory cells. In 1995, Sakaguchi described a subset of CD4+ T lymphocytes that express the IL-2Rα (CD25) and suppress autoimmune disease in thymectomized mice and other models of autoimmunity [3]. Isolated CD25+ cells were anergic but suppressed the proliferation of naïve T cells in vitro. Since the 1980s, numerous laboratories confirmed the existence of CD4+ T cell that expresses the IL-2r and is capable of suppressing autoimmunity upon transfer of CD4+CD25+ cells. Mice lacking the IL2Rα or IL2Rβ chain suffer from systemic autoimmunity [4, 5]. The key lineage-defining transcription factor (TF) Forkhead Box P3 (Foxp3), together with other transcriptional regulators, was found to control the expression of gene programs, which define and maintain Treg cell identity and function [6, 7]. The importance of Foxp3 is also illustrated by Foxp3 gene mutations. Immune dysregulation, polyendocrinopathy, enteropathy, and X-linked (IPEX) syndrome in humans and scurfy mutant mice, both bearing Foxp3 mutations, develop severe autoimmunity [8, 9] such as inflammatory bowel disease, and allergy accompanying hyperproduction of IgE [9,10,11]. Depletion of Treg cells in adults also leads to similar autoimmune pathology, demonstrating that Treg cells are needed for lifelong maintenance, of immune self-tolerance and homeostasis [12]. Ectopic expression of Foxp3 confers suppressive capacity on conventional T cells [13]. In addition to Foxp3 expression itself, an increasing amount of publications highlight the role of the epigenome for the development and identification of the Treg cell lineage, such as DNA methylation, nucleosome positioning, or histone modifications. Treg cell development requires the establishment of Treg-specific DNA hypomethylation pattern [14]. DNA hypomethylation is linked to transcriptionally permissive states, which enable transcription factors to bind to their target gene loci [15]. Foxp3 has also been shown to function as a transcriptional repressor [16] and contributes to Treg function by exploiting a preexisting enhancer network which is established during Treg cell development [17].

Regulation of Foxp3

Three main post-translational modifications have been described for the Foxp3 protein, namely acetylation, phosphorylation, and ubiquitination, which affects the DNA-binding capacity of Foxp3 and thus Treg function. The binding of Foxp3 expression to its targets and stable Foxp3 expression is promoted by acetylation of lysine residues by lysine acetyltransferases, such as TIP60 and p300 [18]. Phosphorylation is performed at serine and threonine residues by several kinases, including PIM-1, PIM-2, and CDK2 [19,20,21].

Transcriptional regulation of Treg cell development

Foxp3 mainly acts as a transcriptional repressor and thereby contributes to the characteristics of Treg cells. Foxp3 target genes are predominantly activated by TCR stimulation such as Zap70, Ptpn22, and Itk. Foxp3 also represses IL-2 production and also Satb1 to inhibit the expression of pro-inflammatory cytokines [22,23,24,25]. Foxp3 is also involved in the upregulation of key genes, such as IL2Rα, CTLA-4, and TNFRSF18 [25, 26]. Similar to other TFs Foxp3 interacts with a number of other cofactors, which are required for Treg phenotype and function in physiological and pathological conditions [7, 27,28,29,30]. The requirement to interact with cofactors indicates that they also need to be expressed in Treg cells for Foxp3-dependent transcription but are also direct targets of Foxp3 itself which indicates that Foxp3 and some cofactors also positively regulate each other [31,32,33]. Five TF have been shown to cooperate with Foxp3 to generate most of the Treg-type gene expression, namely Eos, Gata1, IRF4, Satb1, and Lef1 [6]. A more global analysis suggests that Foxp3 binding partner build a network of multi-protein complexes that bind to pre-existing DNA enhancers and regulate gene transcription [34, 35]. Foxp3 also interacts with genes that mediate epigenetic modifications, such as TIP60, sirtuins, and HDAC7 and thereby modulate TF binding and gene transcription [36].

Epigenetic regulation of Treg cells

Stable expression of Foxp3 and its cofactors is linked to certain epigenetic modifications, which include, for example, DNA methylation and histone modifications, which are important for stable Foxp3 expression: While methylation of CpG residues interferes with TF binding, demethylation increases the accessibility. Treg specific DNA demethylated regions (TSDRs) are present in Foxp3 gene but also in other Treg signature genes, such as Ikzf2, Ikzf4, Tnfrsf18, and CTLA4 and correlate with Treg cell stability [37, 38]. The FOXP3 gene hast three conserved noncoding enhancer regions, CNS1–3 [39]. CNS2 has binding sites for Runx1-CBF-β TF complexes that are important for Foxp3 stability and prevention of autoimmunity. In line, CNS2 deficient mice develop signs of autoimmunity. CNS2 contains a conserved CpG island (TSDR), which is hypomethylated in Treg cells. The methylation pattern allows distinguishing Treg cells from conventional T cells [40,41,42,43].

Histone modification is another important regulator of gene transcription and cell identity [44]. TF binding depends on the chromatin formation. Determination of histone modifications, such as monomethylation, dimethylation, and trimethylation of histone H3 at Lys4 (H3K4me1, H3K4me2, and H3K4me3, respectively), acetylation, and trimethylation of histone H3 at Lys27 (H3K27ac, H3K27me3), and acetylation of histone H3 at Lys9 (H3K9ac), allows to define permissive and repressive chromatin states. Studying these histone modifications further allows the definition of the status of gene transcription and enhancer activity [44, 45]. Treg cells have a unique pattern of histone modifications. While permissive histone modifications are found in upregulated genes in Treg cells, repressive modifications are associated with downregulated genes, partially controlled in a Foxp3 dependent manner. Foxp3 target genes are characterized by repressive marks [23]. In contrast, promoters of Treg signature genes are marked with permissive histone modifications and correlate with gene expression, such as DNA demethylation at TSDRs, suggesting that Treg cell-specific DNA demethylation and histone modifications have similar roles in the maintenance of Treg cells [44]. This epigenetic landscape needs to be established before or with the expression of lineage-specifying TFs [46]. In addition, cell type-specific super-enhancers have been identified, that define cell identity and lineage specification [47,48,49]. Recent data show that super-enhancers in Treg cells are gradually established and activated in early stages of Treg cell development before the expression of Treg specific DNA demethylation. Treg specific super-enhancers were associated with Treg signature genes, such as Foxp3, CTLA4, and IL2Rα [50].

Types of Treg cells

Different types of Treg cells have been described, which can be classified according to their developmental origin: thymus-derived or formally called naturally occurring Treg cells develop in the thymus as a separate lineage at the stage of CD4+ single-positive thymocytes and are thought to show enrichment for T cell antigen receptors (TCRs) with high affinity for self-peptides [51]. Peripherally derived Treg (pTreg) cells are generated in the periphery upon encounter to antigens in the presence of additional factors such as IL-2 and TGF-beta [36] as compared to in vitro generated Treg (iTreg) cells [52]. One of the most exciting advances in the field of Treg cells was the understanding of tissue-resident Treg cells, which will contribute to the resolution of local inflammation. The most recent findings are reviewed elsewhere [36, 53].

Function of Treg cells

A wide range of Treg cell-mediated suppressive mechanisms have been discussed as mediators for Treg cell function, including CTLA-4, IL-10, TGF-β, ITGb8, IL-35, granzyme, perforin, CD39, CD73, and TIGIT [54], and the overall view in the field is that there is no one universal mechanism. Treg cell can mediate their suppressive function depending on the particular situation [55]. CTLA-4 has been described as a key molecule for Treg function. Loss of CTLA-4 in Treg cells leads to the development of fatal autoimmunity in mice [56,57,58]. Similarly, haploinsufficiency of CTLA-4 leads to severe autoimmune syndrome, similar to the IPEX syndrome [59, 60]. A variety of possible mechanisms have been shown, but the exact role of CTLA-4 in Treg function remains not fully understood. Treg cells may, for example, downregulate CD80 and CD86 on dendritic cells in a CTLA-4 dependent manner to inhibit effector cell activation [61]. CTLA-4 ligation may also lead to expression of indoleamine 2,3-dioxygenase [62]. TGF-β, IL-10, and IL-35 are involved in the direct suppression of effector signaling and are the main regulatory cytokines released by Treg cells [55]. Treg cells may also exert their suppressive function through granzymes and are thereby able to effector functions through apoptosis [63].

Another key feature of Treg cells is their inability to produce IL-2, which is essential for proliferation and differentiation of effector T cells. Binding to the IL-2 receptor of Treg cells leads to IL-2 deprivation from other T cells and therefore represents one mechanism of immune-mediated suppression. Overexpression of CTLA-4 and repression of IL-2 in effector T cells resembles Treg-mediated suppressive features [64].

Plasticity vs stability of Treg cells

Loss of Foxp3 expression in tTreg cells has been observed under various conditions in vitro as well as in different settings in vivo and has been shown to contribute to autoimmunity and inflammation [65,66,67,68,69,70,71,72,73]. Treg cells that lost Foxp3 expression during the development of experimental autoimmune encephalitis (EAE) and diabetes gained effector functions and could induce EAE and could transfer diabetes similar to effector T cells [68, 74]. Under arthritic conditions, exFoxp3 cells transdifferentiate into Th17 cells with an osteoclastogenic potential [75]. However, in order to regulate immune responses, Treg cells also need to adapt to their local environment which requires a certain degree of plasticity. Treg cells also need to upregulate certain TF to control inflammatory responses, while Foxp3 expression is maintained. Expression of the TF T-bet is necessary to control Th1-mediated inflammation [76]. Development of Tbet+Treg cells is dependent on the TF STAT1 and occurred in response to IFN-γ [77]. Loss of Tbet+Treg cells results in severe Th1 autoimmunity [78]. In addition to T-bet expression, Th1-like Treg cells upregulate the chemokine receptor CCR5 and CXCR3. Increased frequencies of IFN-γ producing Treg cells have also been observed in autoimmune diseases such as type 1 diabetes [79], multiple sclerosis [80], and autoimmune hepatitis [81]. STAT3 expression has been shown to be important for the regulation for Th17-mediated diseases [82]. RORC expressing Th17-like Treg cells have also been described in healthy subjects, which secrete IL-17, express CCR4, and CCR6 and maintain suppressive function [83, 84]. The role in human autoimmune diseases remains controversial, since beneficial as well as pathogenic aspects have been described depending on the disease setting [75, 85,86,87,88,89]. Expression of the TF IRF4 or GATA3 has been shown to be a key element in Treg-mediated suppression of Th2-mediated inflammation [90, 91]. Th2 like Treg cells, upregulate IRF-4, and GATA3 and secrete Th2-associated cytokines such as IL-4 and IL-13 and have been observed in mice susceptible to allergy and in patients with food allergy and systemic sclerosis [92, 93]. T helper cell-like Treg cells have a demethylated TSDR in the Foxp3 locus, which suggests a reversible phenotype [79].

Identifications of Treg cells in humans

Human Treg cells were identified in the thymus and the peripheral blood as CD4+ T cells with the highest expression of CD25. Since CD25 and Foxp3 are also expressed by activated CD4+ T cells, much research has been focused on the identification of further markers to precisely distinguish Treg cells from recently activated T cells. Foxp3 as an intracellular protein cannot be used for Treg isolation and functional characterization. However, CD127, the IL-7 receptor a-chain, in combination with CD25 can be used to identify and isolate Treg cells [94]. Foxp3 expression and suppressive capacity are enriched in T cells with low expression levels of CD127. However, also CD127 has its limitation as a marker for Treg cells. It is also downregulated in activated T cells and a high percentage of CD127+ cells express Foxp3 and reciprocally cells with low expression of CD127 did not express Foxp3 [95]. A variety of additional markers were described for the identification of Treg cells, including GITR, CTLA-4, and CD49d [56, 96, 97]. Depletion of CD49d removes effector T cells from CD25+ Treg cells and allows purification of Foxp3+ cells in combination with CD127. Miyara et al. divided the Treg cell compartment into three subpopulations according to the expression of CD45RA, namely suppressive CD45RA+Foxp3low (resting Treg cells), CD45RAFoxp3high (activated Treg cells), and non-suppressive cytokine secreting CD45RAFoxp3low non-Treg cells [98].

Treg cells in human autoimmune diseases

Studies from monogenic conditions reveal the importance of Treg cells for human immune homeostasis. Patients with CD25 deficiency suffer from autoimmune phenomena and immunodeficiency, similar to the already mentioned IPEX syndrome. Although CD25 deficiency does not affect Treg numbers, impaired suppressive activity has been observed due to decreased IL-10 production [99]. Besides CD25, mutations in other crucial genes have been reported, which are associated with autoimmune phenotypes, such as STAT5B, CTLA4, or LRBA [60, 100,101,102]. Human studies on Treg cells clearly suffer from the limitation of a definition and the clear identification of Treg cells. In the last decade, an increasing amount of papers have been published, which address numbers and functions of Treg cells in a variety of autoimmune diseases. Due to the evolving numbers of different Treg markers and the lack of a clear definition of a human Treg cells, partly conflicting results in numbers and or function of Treg cells, isolated from the peripheral blood (ref. Miyara) have been reported in a variety of autoimmune diseases, including type 1 diabetes, multiple sclerosis, systemic lupus erythematosus (SLE), myasthenia gravis, rheumatoid arthritis (RA), and others. We will only summarize the more recent findings of selected systemic and organ-specific autoimmune diseases. An extensive summary of the publications of systemic autoimmune diseases can be found in Table 1.

Table 1 Treg numbers and function in patients with selected autoimmune diseases

Systemic lupus erythematosus

SLE represents a systemic autoimmune disease, which can affect multiple organs. In patients with SLE quantitative as well as functional deficiencies of Treg cells have been described. A more recent publication reports low levels of STAT5 phosphorylation upon IL-2 which suggests an inherent Treg cell defect [195]. Major reasons for the discrepancy of these findings are not only the lack of a reliable Treg specific marker but also the heterogeneity of the disease, the small size of the studied cohorts and different methods used for the isolation of Treg cells. In addition, a subset of Foxp3+ cells that lack the expression of CD25, have been described in SLE patients, which was later on also described in patients with RA [196] and MS [197]. While our own group observed phenotypic and functional characteristics of Treg cells within this newly described cell population [103], others report that CD25Foxp3+ Treg cells in SLE patients are activated T cells, rather than a distinct Treg cell population [198]. In addition, we could show that CD4+CD25Foxp3+ T cells are elevated in SLE patients with renal involvement. CD4+CD25Foxp3+ T cells were also detected in urine sediment samples of patients with active glomerulonephritis and correlated with the extent of proteinuria [104]. A recent study confirmed the regulatory features of this cell population, including demethylation of the Foxp3 TSDR and constitutive expression of the TF HELIOS in the majority of the cells and an inability to produce IL-2 [105]. Interestingly treatment of SLE patients with low dose IL-2 led to a twofold to threefold increase in the expression levels of CD25 in Treg cells and a dramatic expansion of CD25high Treg cells among Foxp3+CD127low Treg cells [199].

Rheumatoid arthritis

RA is a systemic autoimmune disease which leads to chronic inflammation and tissue destruction in the joint. Conflicting results have been reported for numbers of Treg cells in the peripheral blood and the synovial fluid [161, 165,166,167,168,169,170,171, 176,177,178, 192,193,194, 200, 201]. In addition to quantitative defects, functional deficiencies have been reported in patients with RA [192,193,194]. Treatment with anti-TNF and anti-IL6 restored the balance of Treg and Th17 cells in RA patients and can affect Treg function [170, 194, 201].

Ankylosis spondylitis

AS represents an inflammatory autoimmune disease, associated with HLA-B27. Similar to other autoimmune diseases, data on Treg numbers and function in AS patients remain controversial [184,185,186, 202]. Similar to RA, recent data indicate that anti-TNF treatment might affect the Th17/Treg ratio [203, 204].

Systemic sclerosis (SSc)

SSc is a connective tissue disease, which is characterized by immune abnormalities, microvascular injury, and fibrosis of the skin and internal organs [205]. Several reports describe defects in the number of Treg cells, although conflicting results exist [116, 145,146,147,148,149, 151,152,153,154,155,156,157,158,159,160, 206, 207]. Similarly, impaired as well as normal Treg function has been reported in SSc patients [145, 146, 152,153,154, 159]. Mathian et al. report that activated and resting Treg cells are not functionally impaired in SSc patients. Numbers of activated Treg cells are decreased in early and resting Treg cells in late stages of the disease [154].

Treg-based therapies

The contribution of Tregs in various human autoimmune diseases has opened up a new therapeutic avenue, which included Treg-based cellular therapies or therapies which aim to restore the balance of Treg and Teff cells such as IL-2 therapies. The concept to use Treg cells as a cell-based therapeutic approach was first demonstrated in murine models, such as EAE or CIA, in which Treg cells has been shown to be involved in the pathogenesis. Transfer of Treg cells could ameliorate the disease [208, 209]. The strategy of Treg cell transfer cannot simply be translated into humans due to the low number of Treg cells. Thus, manufacturing protocols were developed to expand Treg cells in vitro to produce large numbers of polyclonal Treg cells. Phenotypic and functional characterization showed that expanded Treg cells are highly suppressive with high expression levels of Treg-associated markers such as CTLA4, CD25, and Foxp3 and demethylation of the TSDR [210, 211]. Interestingly, Treg cells from RA patients, which were expanded in the presence of rapamycin, maintained their suppressive capacity and Foxp3 demethylation and were more effective in suppressing conventional T cell proliferation compared with their ex vivo counterparts [212]. Several phase I and phase II clinical trials were designed which use ex vivo expanded Treg cells as an approach for the treatment of autoimmune diseases, such as autoimmune hepatitis, GvHD, type I diabetes, SLE, kidney, and liver transplantation [213,214,215,216].

Beside the development of Treg cell-based therapies, a lot of existing therapies target Treg cells as well as Teff cells, such as rapamycin, anti-CD3, CTLA-4Ig, or anti-CD25 [217,218,219]. Rapamycin, an inhibitor of PI3Kakt-mTORC1 signaling, promotes the expansion and survival of Treg cells, while suppressing the proliferation of Th1 and Th17 cells [220,221,222]. Treatment with CD3 antibody increased Treg number and stabilized Treg function in type 1 diabetic mouse model [223]. Anti-CD3 treatment preserved residual beta-cell function in patients with type 1 diabetes [224, 225].

One of the most exiting concepts, which evolved during the last couple of years is based on the expression of the IL-2 receptor in Treg cells, which allows a preferential expansion of Treg cells with low amounts of IL-2. Since IL-2 is also a known growth factor for effector cells, the use of IL-2 in autoimmune diseases was thought to be contraindicated. However, murine studies, in which IL-2 signaling was impaired but not abrogated, showed that IL-2R-signaling promoted only the development of Treg cells but not effector T cells, which lead to the investigation of low dose IL-2 as a treatment strategy for autoimmune diseases [226,227,228,229]. Treatment with IL-2 was shown to block the differentiation of naïve CD4+ T cells into effector T cells [230, 231]. Indeed, promising data were already reported for a variety of diseases such as type I diabetes, GvHD, alopecia areata, hepatitis C virus-induced vasculitis, and SLE [199, 226, 232,233,234,235]. In patients with alopecia areata, low dose IL-2 treatment expanded Treg cells in the blood and the hair follicles and led to an impressive hair regrowth [234]. In SLE patients, a dysbalance between Treg and effector T cells have been reported, and a recent paper described Treg defects as hallmarks of IL-2 deficiencies. Lack of IL-2 production by CD4+ T cells thereby accounts for the loss of CD25 expression which could be reversed by IL-2 [199]. Treatment with low dose IL-2 was anticipated as a potentially effective treatment. Indeed, numbers of circulating Treg cells were significantly increased [199] along with a reduction in disease activity [235]. A more recent study by Klatzmann et al. investigated the potential of low dose IL-2 therapy as a new therapeutic approach in 11 different autoimmune diseases, including RA, AS, SLE, psoriasis, Behcet’s disease, granulomatosis with polyangiitis (GPA), Takayasu’s disease, Crohn’s disease (CD), ulcerative colitis, autoimmune hepatitis, and sclerosing cholangitis. In general, low dose IL-2 was well tolerated and led to a Treg specific expansion and activation. These data certainly highlight the potential use of this treatment strategy for various additional autoimmune and inflammatory diseases [236].

Conclusion

Technological progress in the recent years allowed to define and characterize murine Treg cells and helped to understand the balance between Treg plasticity, which is necessary for proper Treg function but also Treg instability, which can drive autoimmunity (Fig. 1). On the other hand, data from the human system, especially in patients with autoimmune or inflammatory conditions are still conflicting and misleading due to a lack of a reliable Treg cell-specific marker. The progress that has been made in the murine system has to be translated in the human system. It will be important to determine the physiological relevance and the mechanism that drive plasticity and stability of Treg cells in patients with autoimmune diseases and the contribution of the epigenetic signature to disease pathology. This is ultimately necessary for the new and exciting treatment approaches, which target Treg cells as a direct Treg-based therapy, using polyclonal Treg cells or as a Treg cell-targeted therapy. Clear identification of patients with functional or numerical Treg deficiencies will be necessary for future successful treatment and monitoring of patients with Treg targeted therapies. A more detailed understanding of the exact role of Treg cells under various inflammatory conditions will help to develop a personalized treatment approach within the next years.

Fig. 1
figure 1

Schematic presentation of the role of Treg cells in autoimmune diseases