It is well established that thymocyte destiny depends on the binding affinity/avidity of the TCR to its self-peptide–MHC ligands on antigen-presenting cells (APC). Thus, although thymocytes expressing a TCR with low affinity for self-peptide–MHC complexes are positively selected and differentiate into conventional T cells, thymocytes that bind self-peptide–MHC complexes with high affinity undergo negative selection and experience cell death in order to remove potentially self-reactive T cells. A third possibility is the existence of thymocytes expressing TCRs with intermediate avidity for self-peptide–MHC, which may differentiate into cells with a regulatory function (tTregs) (33).

Together with the interaction of the TCR with self-peptide–MHC, additional signals are required for thymic induction of Tregs, such as costimulation through CD28 or the signaling provided by the IL-2R-γ_c cytokine family. Thus, deficiency of tTregs was observed in CD80/CD86 and CD28 knockout animals (34). Accordingly, the absence of the Treg population from the thymus and periphery was observed in IL-2R-γ_c knockout animals (35) as well as the emergence of autoimmunity in animals lacking IL-2Rβ, which could be repaired by the transfer of Tregs from control mice (36, 37). Furthermore, in IL-2^−/− or CD25^−/− mice, the expression of FOXP3 in thymocytes was drastically reduced and these animals developed fatal autoimmunity disorders (35).

Regulatory T cells may also be induced to differentiate at a double-positive (CD4⁺CD8⁺) developmental stage (38), since a momentary reduction of FOXP3⁺ thymocytes in newborn mice was observed upon elimination of TGF-βRI in the murine thymocytes at a double-positive (CD4⁺CD8⁺; DP) development stage (39). These results support the hypothesis that thymic Treg selection is enhanced by TGF-β signaling. In this regard, TGF-β signaling has been demonstrated to control apoptosis of self-reactive thymocytes through a Bim-dependent mechanism, raising the Treg precursor cells (40). In addition to the above data in mice, studies on the development of human thymic Tregs have shown that an important percentage of FOXP3⁺ single-positive (SP) cells derive from FOXP3⁺ DP thymocytes (41).

In summary, the majority of the Tregs originate in the thymus at the CD4 SP stage and intermediate signal strength TCR, IL-2, CD28, and TGF-β are needed for FOXP3 upregulation during the process of Treg cell generation. Additionally, Tregs may be differentiated at a previous double-positive stage in the presence of TGF-β.

Experiments based on the transfer of purified FOXP3⁻ CD4⁺ T cells into lymphopenic mice followed by the subsequent analysis of the TCR repertoire of the FOXP3⁺ T cells generated, compared with that of cells which remained FOXP3⁻ demonstrated that “TCR repertoires were only partially overlapping by FOXP3⁺Tregs and FOXP3⁻ non-Treg CD4⁺T cells present in control mice” (42, 43). Furthermore, it was found that the TCRs from Tregs located in the colon were different from those expressed by Tregs from other tissues, and a subset of these colonic TCRs was specific for antigens associated with the gut microbiota (44). The above results suggest that while the differentiation of thymic Tregs is induced by intermediate affinity interactions of TCRs with self-peptide–MHC molecules, the development of peripheral Tregs is probably promoted by encounters with foreign antigens, e.g., foods, gut microbiota, and allergens.

Nevertheless, it has been suggested that the TCRs expressed by pTreg cells probably exhibit high affinity, as shown by the observation that “rare high-affinity antigenic peptides allow for most efficient FOXP3 induction upon stimulation of a cognate transgenic TCR displayed by peripheral CD4⁺T cells in comparison with a less efficient pTreg generation by low-affinity peptide variants” (45). Additionally, it seems that insufficient activation of T cells promotes the induction of FOXP3, since CTLA-4 expression is essential for the induction of FOXP3 in vitro by a mechanism dependent on TGF-β presence (46), while CD28 has the contrary effect (47, 48). Thus, in vitro and in vivo studies suggest that FOXP3 induction and pTreg cell generation require high-affinity TCR signaling together with suboptimal costimulation (high CTLA-4 and low CD28 signaling) (40), and the process is helped by the presence of high amounts of TGF-β (47). Signaling through TGF-βR seems decisive for the expression of FOXP3 in most peripheral CD4⁺ T cells (49).

The pTreg cell generation requires the combined action of soluble factors, such as TGF-β and IL-2, in the microenvironment and the presentation of the antigens by appropriate APCs. Furthermore, the presence of all-transretinoic acid (ATRA) in the Tconv environment synergizes with TGF-β, and this effect is great enough to promote pTreg generation even when a high costimulation is being produced. This is particularly evident in lung tissues where resident macrophages (CD45⁺CD11c⁺MHCclass II^lowF4/80⁺) constitutively expressing TGF-β and retinoic acid are the main subset of cells driving pTreg cell induction from naive CD4⁺ Tconv cells (50).

The data discussed so far indicate that pTreg cells generation is influenced by a specific type of TCR signaling, and costimulation, and through cooperation with other signals, such as TGF-β, IL-2, and ATRA. These conditions suggest that pTreg cell differentiation could be restricted to precise locations such us mucosal surfaces where they may regulate immune responses to harmless antigens such as commensal microbiota and prevent allergic inflammation. Supporting these ideas, “studies using Rag-1-deficient T-B monoclonal mice sufficient or deficient in FOXP3 demonstrated that TCR transgenic pTreg cells were sufficient to establish mucosal tolerance and control allergic inflammation induced by the model antigen recognized by the TCR” (5).

Additionally, several studies suggest a role of microRNAs (miRNAs) in the generation and function of Tregs (51, 52). Specifically, miR155 was recently described as a player in Treg cell differentiation and was shown to induce upregulation of several Treg cell-associated genes when transferred to conventional T cells (51, 53). It has also been reported that an acquired decrease of the endoribonuclease Dicer, involved in the generation of miRNAs (54), induces spontaneous autoimmunity in a disease model (52). The progression to autoimmune disease was associated with a marked decrease in Dicer and an increased miR-155 expression which can upregulate CD62L in Tregs. Furthermore, Singh et al. (55) found three miRNAs (miR-15b/16, miR-24, and miR-29a) that regulated the induction of Tregs from naive CD4⁺ T cells, with miR-15b/16 having the highest effect. Important genes regulated by miR-15b/16 were Rictor and mTOR, which encode components of the mTOR signaling pathway involved in controlling the generation of Tregs.

In summary, the requirements to promote induction of pTregs include high-affinity interaction of TCRs with peptide: MHC complexes; suboptimal activation of dendritic cells (DCs); mucosal administration of peptide; signaling by cytokines, such as TGF-β and IL-2; and regulation by the appropriate miRNAs (49, 51–53, 55, 56).

Innate lymphoid cells are a population of mucosal innate cells characterized by a lack of antigen specificity (absence of T- and B-cell receptors) and by shared developmental origin and phenotypic traits with T cells. ILC2 produce large amounts of Th2 cell cytokines and are linked to allergic disorders, such as asthma, chronic rhinosinusitis, and atopic dermatitis (28).

It has been demonstrated that peripherally induced Tregs effectively suppress the production of the ILC2-driven, pro-inflammatory cytokines IL-5 and IL-13, both in vitro and in vivo by blocking of ICOS:ICOS-L interaction on ILC2 cells. Inducible T cell Costimulator (ICOS) is a receptor expressed by ILC2 cells, Tregs, and others. It is well known that ICOS:ICOS-L interactions on ILC2 cells have a role in cell function and survival and are involved in controlling Th2 cytokine release, airway hyperreactivity (AHR), B-cell differentiation, and IgE class-switching in mice (136, 137). A recent study by Maazi et al. (138) demonstrated that the expression of ICOS and ICOS-L in ILC2s from human peripheral blood was increased by in vitro culture in the presence of IL-2 and IL-7 but not of IL-33. It has been demonstrated that cell contact is required for suppression of ILC2 mediated by Treg, TGF-β, and IL-10. Thus, in vitro stimulation of human ILC2s with IL-2, IL-7, and IL-33 and subsequent blockade of ICOS:ICOS-L interactions decreased the release of IL-5 and IL-13 cytokines (139). Additionally, it has been shown that human pTregs suppress syngeneic human ILC2s via ICOS-L to control airway inflammation in a humanized ILC2 mouse model (140) (Figure 2).

Mast cells are essential to initiate the immediate phase of allergic reactions (Figure 1, Step 6), and their degranulation initiates the triggering of allergic symptoms (141). It has recently been demonstrated that the symptomatic phase of allergic disorders can be controlled by constitutive FOXP3⁺ Tregs in mice (142), although mast cells increased the IL-6 secretion. This inhibition by Tregs is produced via direct cell-to-cell contact between OX40 expressed on Tregs and OX40 ligand on mast cells, which leads to increased intracellular levels of cyclic AMP (cAMP) and results in blockage of extracellular Ca²⁺ (Figure 2). However, the suppression of IL-6 secretion by mast cells seems to be controlled via TGF-β (143). In vitro studies have shown that “IL-4 and TGF-β1 had balancing effects on mast cell survival, migration and FcεRI expression, with each cytokine cancelling the effects of the other. Dysregulation of this balance may impact allergic disease and be an objective of targeted therapy” (65, 144).

Dendritic cells are key initiators and master regulators of the allergen-specific immune response by processing and presenting antigens to Tconv and secreting cytokines that control T cell differentiation to a certain effector type. Tregs could directly act on DCs by downmodulating their surface expression of CD80/CD86 and subsequently blocking generation of an allergen-specific Th2 cell immune response. Suppression of DCs appears to be mediated through LAG-3, CTLA-4, leukocyte function-associated antigen 1 (LFA-1), and other molecules (Figure 2).

Thus, Tregs may express LAG-3, a homologous of the CD4 molecule that acts as a coreceptor of the MHCII complex, although with higher affinity. The binding of MHCII and LAG-3 reduces the maturation and costimulatory capacity of DCs and, as a consequence, diminishes their capacity for antigen presentation to Tconv (131, 145, 146). Activated human T cells express MHCII molecules, and interaction of Tregs via LAG-3 on Teff might also induce immunosuppression (147).

Additionally, murine and human Tregs exhibit a constitutive expression of the coinhibitory molecule CTLA-4, which is detected on the surface of Teff upon activation (148). The deficiency of the CTLA-4 gene produces serious autoimmune disorders similar to those induced by defective FOXP3, demonstrating that CTLA-4 is essential for Treg function (149). CTLA-4 binds to the same ligands CD80/86 as CD28 (T cell costimulatory antigen) but with a higher binding affinity. The interaction between CD28 and its ligands CD80/86 on DCs is required for T-cell activation. However, CTLA-4 binding to the same ligands blocks such activation and induces the generation of anergic T cells. CTLA-4 may also suppress or decrease the surface expression of CD80/86 molecules, decreasing the activation of Tconv. Furthermore, the interaction of CTLA-4 with CD80/86 in Tregs can promote the generation of indoleamine 2,3-dioxygenase (IDO) that catalyzes degradation of the essential amino acid tryptophan to kynurenine, provoking the starvation of Teff and cell cycle arrest. Additionally, IDO induces pTreg generation (150). Also, CTLA-4 may be involved in the reduced glutathione synthesis observed in murine DCs, which promotes a Redox milieu unfavorable for the proliferation of conventional T cells (151). In relation to that, several studies have shown that certain polymorphisms in the CTLA-4 gene are significantly associated with susceptibility to autoimmunity (152).

Regulatory T cells can control Tconv activation by reducing their interaction with DCs (153). In fact, it has been described that the aggregation of CD25^hi Tregs around DCs, via CTLA-4, downregulated CD80/86 molecules expression (149, 154). Thus, there is a competition between Tregs and Tconv for the interaction with DCs reducing their capability to activate Teff (105). This downregulation of CD80/86 expression by Tregs is also partly dependent on the adhesion molecule LFA-1, thereby indirectly impeding the activation of Tconv cells by APCs in vitro (155, 156) and in vivo (149, 157) (Figure 2).

Another surface molecule involved in immunosuppression is T cell immunoglobulin and ITIM domain (TIGIT), which is highly expressed on Treg and Teff. A protein named Poliovirus receptor (PVR, NECL5, or CD155), highly expressed on DCs and others, was identified as a high-affinity coreceptor for TIGIT. Upon interaction of Treg cell with DCs through TIGIT, secretion of the suppressive cytokines IL-10 and TGF-β by DCs was observed (158).

Neuropilin-1 is another molecule expressed by Tregs and which elongates their contact with DCs reducing the presentation of antigen to Tconv. These results were confirmed by using an anti-Nrp-1 antibody to abrogate Treg-mediated suppressive activity (159, 160). However, as other CD4⁺ cells express Nrp-1 (161) it is possible that the anti-Nrp-1 antibody was interfering with cell activation rather than Treg function (146).

Finally, ICOS:ICOS-L interactions between mostly plasmacytoid DCs and Tconv could result in their differentiation into IL-10-secreting Tregs (28).

During the activation process, T cells follow several differentiation pathways, acquiring specific properties and functions. Th cells could differentiate into Th1 (IFN-γ-secreting cells), Th2 (IL-4/IL-5-secreting cells), Th17 (IL-17-secreting cells), and other subsets. Th1 cells are specialized in the elimination of intracellular microorganisms; Th2 cells are required for fighting extracellular pathogens; and Th17 cells protect against extracellular fungal and bacterial pathogens and have a role in autoimmune tissue injury. Regarding that, it has been shown that Tregs can avoid allergy by suppression of the effector Th1, Th2, and Th17 cells (162, 163) (Figure 2). Nevertheless, such results are not clear since it has been published that TGF-β1 secreted by FOXP3⁺ Tregs is necessary to block Th1 and support Th17-cell generation (164). On the other hand, it has been suggested that Tregs induced by nitric oxide can suppress Th17 but not Th1 cell development and function (165). Furthermore, it has been published that Tregs can also block the synthesis of the cytokine IFN-γ without inhibiting Th1 cell differentiation (166) and that FOXP3⁺ Tregs induce Th17 cell differentiation in vivo through IL-2 modulation (167). Additionally, it has been found that the cytokines IL-6 and TGF-β promote the generation of pathogenic Th17 cells (65, 168). Further studies should be performed to clarify the importance of Tregs in the suppression of specific Th subsets.

Recent evidence suggests the existence of a balance between Tregs and Th17 cells during the first stages of naive T cell development. Thus, the presence of IL-6 and TGF-β promotes the differentiation of Tconv into Th17 cells. However, without IL-6 in the milieu, T cells differentiate into Tregs. Additionally, IL-21 has a role in the generation of the Th17 subset and inhibits FOXP3. Given the contribution of Th17 cells in asthma, the inhibition of the Th17 population by Tregs is crucial to maintaining the immune homeostasis (146).

Furthermore, it has been observed that Tregs suppress the TCR-mediated proliferation and IL-2 release of Tconv cells (132). The mechanism by which Tregs suppress murine (169) or human (168) Tconv proliferation can be directly mediated by immunosuppressive factors or by a contact-dependent action. Tregs can also block Tconvs indirectly by controlling the activation of APCs as previously described (14).

Another mechanism to control immune activation would be effector cell-death induction by Tregs. In fact, it was observed that human tTregs express Gzm A and kill activated CD4⁺ T cells and other cells by perforin-dependent mechanism (170). Another study reported a partially Gzm B-dependent inhibition of Tconv proliferation by murine Tregs in vitro, although perforin was not involved (130). A further example of suppression of T effector cells by Tregs was demonstrated in vitro and in vivo in a model of transplantation in mice with the involvement of the death receptor TRAIL (171).

At the same time, high-level IL-2R expression on Tregs, indispensable in Treg cell homeostasis (172), could deprive Teff of IL-2 and inhibit their proliferation (173). Nevertheless, there is a controversy about the role, if any, of the decreased availability of IL-2 because of its consumption by Tregs that exhibit high expression levels of the CD25 molecule and may depend on the specific setting and stimulation conditions of the cells (14).

Eosinophils are secondary effector cells involved in the pathogenesis of allergy (Figure 1B). It has been reported that Tregs can inhibit their function (174), and a negative correlation has been found between the percentage of FOXP3⁺ cells in bronchoalveolar lavage fluid (BALF) from tolerant mice and the number of eosinophils detected in that fluid (65, 175). The mechanism by which Tregs inhibit eosinophils activation is mediated by IL-10 released by Tr1 cells (176, 177).

Additionally, Tregs might directly control allergic responses by induction of neutrophil apoptosis and/or by promoting an immunosuppressive phenotype on these cells (Figure 2) that generate IL-10, TGF-β1, IDO, heme oxygenase-1 (HO-1), and the suppressor of cytokine signaling 3 molecule (SOCS3) (178). These anti-inflammatory neutrophils may inhibit Th17 cell induction, which depends on the presence of both TGF-β1 and IL-6 (179). Additionally, phagocytosis of apoptotic neutrophils by macrophages and DCs results in a reduction in IL-23 release by these cells, which in turn leads to lower IL-17 secretion by CD4⁺ T cells (180). Hence, cooperation of activated Tregs with neutrophils might result in inhibition of Th17 response, providing an important control of inflammatory responses. Supporting this idea, one study (181) has shown that the reduced capacity of weanling, as compared with neonatal, mice to develop inducible bronchus-associated lymphoid tissue (iBALT) in response to LPS can be reversed by the elimination of Tregs. This was associated with a high expression of IL-17A and CXCL9 and with increased neutrophilic inflammation in the lungs.

Activated Tregs may directly suppress effector B-cells through the release of Gzm and perforin (182). As a consequence of this, Tregs can control IgE production and the posterior mast cell-mediated inflammation. In fact, the large amounts of IL-10 and TGF-β that are secreted by Tregs drastically inhibit IgE release, although a simultaneous increase in the production of IgG4 and IgA by B-cells has been observed. This isotype unbalance has also been reported in individuals naturally exposed to large allergen doses. Thus, beekeepers with multiple stings and patients with chronic helminthic infections have tolerance mediated by IL-10 and increased levels of antigen-specific IgG4 (183). Recently, the identification of IL-10-producing B regulatory cells with immunosuppressive function has been reported; these cells may also control the inflammatory reactions mediated by T cells (184) and may participate in the generation of peripheral CD4⁺CD25⁺ cells by inducing the development or elongating the survival of such cells (146).

Several researchers have suggested that as with conventional CD4⁺ helper T cells, Tregs display various phenotypically and functionally diverse subsets and that their location in different tissues is critical for their ability to interact with and regulate the different subsets of Teff (82). Tregs may modulate a Th cell subset specifically by expressing the characteristic transcription factor as well as adhesion and chemoattractant receptors of said specific subset that would target them to the same tissues and inflammatory sites (185). These comprise chemokine receptors such as CXCR3, CCR8, and CCR6, involved in the migration of cells to locations of Th1-, Th2-, or Th17-mediated inflammatory responses, as well as other general receptors such as CCR2 and CCR5 (185, 186). In this regard, it was shown that upregulation of T-bet in a Treg subset upon IFN-γ secretion was required for the control of inflammatory responses mediated by Th1 cells (187). Similarly, the expression of IRF-4 (associated with Th2 and Th17 cells) in Tregs was essential for the inhibition of immune responses mediated by Th2 cells (188). Moreover, the presence of the transcription factor STAT3, characteristic of Th17 cells, in Tregs was critical for the suppression of the intestinal inflammation associated with Th17 cells (189). In accordance with this, the absence of GATA-3, transcription factor associated with Th2 cells in Tregs, led to autoimmunity, defective FOXP3 expression, and elevated cytokine levels specific of Th1, Th2, and Th17 cells (190). In fact, Tregs lacking GATA-3 expression usually transform into a Th17 subset. In relation with that, Bcl6, which acts by suppressing GATA-3 transcriptional activity independently of IL-4 and STAT6, is crucial for the control of Th2 inflammation by Tregs (191–193). The mechanism used by Tregs to suppress each subset of Teff might include deprivation of limiting factors since Tregs accumulate at the location of the immune response. Accordingly, IRF-4 or STAT3-deficient Tregs lack suppressive function in vitro. In contrast, the presence of certain transcription factors in Tregs might inhibit FOXP3 expression with the subsequent blocking of Tregs suppressive function. Thus, STAT3 is a transcription factor induced by the release of various cytokines that downregulate Tregs and FOXP3, such as IL-6, IL-23, or IL-27 (192, 194). In fact, the lack of STAT3 in T cells in a model of induced colitis promoted the development of Treg and diminished the symptoms of the disease (195). Probably, the different CD4⁺ T cell subsets have the ability to experiment high levels of plasticity between them, although there is controversy about the stability of the Tregs in vivo (22, 196). Nevertheless, several studies indicate an evident plasticity between Th17 cells and pTregs (197), and supporting this idea, a transitory stage with simultaneous expression of RORγt and FOXP3 has been detected (198). In the context of a specific cytokine environment, the release of IL-17 in murine and human Tregs, which might maintain immunosuppressive activity, was described (199, 200).

During the immune response, extracellular ATP acts as a danger signal and may exert its effects on DCs. Cell damage induces the release of the intracellular ATP since the nucleotide is present within the cells in high concentration. “Extracellular ATP can be sensed by purinergic P2 receptors such as CD39. This molecule is the main ectoenzyme in the immune system, hydrolyzes ATP or ADP to AMP and is expressed by B cells, DCs, all mouse Treg cells, and about 50% of human Treg cells” (201). Thus, another anti-inflammatory mechanism that may be used by Tregs could be catalytic inactivation of extracellular ATP by CD39 (201). Supporting this idea, CD39 knockout Tregs showed decreased suppressive capacities in vitro and in vivo (133). In fact, CD39 expression was suggested to identify a highly suppressive human Treg subset (202), and inhibition of Tconv proliferation by this subset could be partially abolished by suppression of ectonucleotidase activity (203). CD73 is another ectoenzyme, also expressed by Tregs, that degrades AMP to adenosine (204). Adenosine binds to the A2A receptor and may suppress DCs and/or Teff, e.g., by increasing cAMP (205). In vivo, signaling through A2A receptor might lead to anergy and induce pTreg development (206). In conclusion, adenosine seems to contribute to the regulatory function of certain Treg subsets. Thereby, cAMP is transferred through gap junctions from the Tregs into Teff where it activates protein kinase A that suppresses proliferation and IL-2 release by triggering the activation of inducible cAMP early repressor (ICER) (207).

Despite the critical role of TGF-β and IL-10 in various in vivo models, the specific contribution of immunosuppressive cytokines in Treg-mediated regulation is still poorly understood (Figure 2).

microRNAs have recently been described as critical regulators of Treg development and function. These are small double-stranded RNAs that negatively regulate gene expression at a post-transcriptional stage (225). Furthermore, an exosomal pathway has been described that can capture miRNAs from certain cells and transfer them to other cells (226), providing a mechanism for cell communication. In this sense, Okoye et al. (53) have observed that Tregs are capable of releasing miRNA-containing exosomes. Specifically, Tregs released and transferred Let7d to Th1 cells, regulating Th1 cell proliferation and IFN-γ release. Furthermore, generation of miRNA and the release of exosomes by Tregs were both required for suppression of Th1 cell activation in vivo and for the prevention of systemic inflammatory disorders. Supporting this new mechanism of control by Tregs, it has been demonstrated that exosomes isolated from Tregs can inhibit Teff. However, this suppressive effect was not as potent as that of Tregs, suggesting that exosome transfer and other mechanisms are necessary for optimal regulation (53). Additionally, miR-21 expression has been associated with Tregs deficient in Bcl6 which exhibit a defective ability to control Th2 inflammation by limiting the transcriptional activity of GATA-3 (193).

In summary, the mechanisms of Treg cell-mediated suppressive function open up the possibility that Tregs capture and deliver different miRNAs and other molecules to different cells at different times depending on the specific situation.