The immune system uses several mechanisms of central and peripheral tolerance in order to prevent the activation of T lymphocytes toward self-antigens. Although the importance of immune self-tolerance has been established for a long time, some essential cellular and molecular mechanisms of T-cell tolerance have only been recently revealed. Once thought to be a recycling system, protein ubiquitylation by E3 ligases has now emerged as a regulated and crucial modulator of immune responses, and more importantly as a key signaling pathway involved in T-cell tolerance. In this review, we highlight our current understanding of the transcriptional and molecular signaling mechanisms involved in ubiquitylation-mediated T-cell tolerance.
Upon activation, the immune system orchestrates robust and potentially destructive responses intended to eliminate the immunogenic antigen. Thus, immune system activation is under stringent control to prevent a detrimental immune response against self-antigens. The processes that ensure unresponsiveness toward self-antigens are known as immunological tolerance and, in their absence or failure, autoimmunity occurs 1. Central tolerance mechanisms operate in the thymus to eliminate potentially self-reactive thymocytes and generate (Treg) suppressor cells 2. In addition, several peripheral tolerance mechanisms act to prevent self-reactivity once mature T cells have reached the periphery. These peripheral mechanisms of tolerance include the following: (i) immunological ignorance to self-antigens expressed at low levels or in immunoprivileged sites; (ii) active immunosuppression by Foxp3+ Treg cells; (iii) activation-induced cell death of autoreactive T cells and (iv) induction of a state of unresponsiveness known as T-cell anergy 3.
During the past few years, extensive research has revealed new insights into the cellular and molecular mechanisms necessary to induce and maintain T-cell tolerance. In particular, ubiquitylation of proteins by E3 ligases has emerged as a novel and indispensable signaling pathway that regulates T-cell tolerance toward self-antigens 4–6.
The ubiquitylation process
Ubiquitylation is an important mechanism of post-translational modification of proteins by which the 76-aa polypeptide ubiquitin, Ub, is covalently attached to a substrate protein. This process occurs in a stepwise manner, requiring the participation of three enzymes: a ubiquitin-activating enzyme E1, responsible for the first binding and activation of the Ub; a second ubiquitin-conjugating enzyme E2, which receives the activated Ub from the E1; and a third ubiquitin-ligase E3 that catalyzes the final covalent transfer of the Ub from the E2 enzyme to the protein substrate 7. Once thought to only mediate proteosomal degradation, the ubiquitylation of a protein can have multiple effects, including altered subcellular localization, regulation of protein–protein interactions, and functional modulation. The fate of the tagged substrate protein will depend on the number of Ub residues added as well as on the lysine residue involved in the formation of polyubiquitylation chains. In particular, polyubiquitin chains formed by the linkage of ubiquitin molecules through lysines in position 48 (K48) leads to the degradation of the substrate protein in the proteosome. In contrast, lysine-63 linked ubiquitin chains (K63) target the substrate for non-proteolytical modifications, such as endosomal trafficking, DNA repair, enzymatic activation and/or protein translation 8.
By binding and selecting the target substrate, E3 ligases determine the specificity of the ubiquitylation process. E3 ligases are generally classified into two groups depending on the enzymatic domain that participates in ubiquitylation: HECT (homology to E6-associated protein carboxyl terminus) and RING (really interesting new gene) type E3 ligases 6. Diverse biochemical and genetic evidence has implicated E3 ligases of both HECT and RING types as crucial molecular regulators of both innate and adaptive immunity 9, 10.
Transcriptional regulation of E3 ligases in T-cell tolerance
NFAT in anergy and T-cell tolerance induction
The concept of anergy is broadly used to define a cell-intrinsic state of functional unresponsiveness that occurs when a T cell has been presented with the antigen in suboptimal activation conditions, either strong TCR stimulation without co-stimulation or with a low-affinity ligand in the presence of co-stimulation 11–13. Once the T cell becomes anergic, it remains viable but unresponsive to further stimulation. Early studies revealed that anergy does not constitute a simple loss of signaling molecules, but an active process where “anergic factors” are being synthesized to establish and maintain the unresponsive state 11, 14. Subsequently, genome wide screening analysis further confirmed the differential transcriptional profile between anergic and activated T cells, and identified the calcium/NFAT signaling pathway as a key pathway for the induction of at least one genetic program associated with T-cell anergy 15.
Calcium signaling in T-cell biology is well characterized 16, 17. TCR engagement couples to PLCγ1-dependent synthesis of inositol 3-phosphate. Inositol 3-phosphate promotes the release of calcium from the endoplasmic reticulum into the cytosol, which in turn, opens the calcium release activated channels at the plasma membrane 16, 17. The sustained influx of extracellular Ca2+ through calcium release activated channels (CARC) activates the phosphatase calcineurin, which subsequently binds and dephosphorylates the cytosolic NFAT family of transcription factors. Upon dephosphorylation, NFAT proteins translocate to the nucleus to induce different gene expression programs 18, 19. The Ca2+/NFAT signaling pathway is required for the proper activation of T cells. However, in T cells activation, CD28 co-stimulation strongly activates additional pathways, in particular the Ras/MAPK/AP–1 signaling pathway, which are otherwise poorly stimulated by TCR engagement. Following co-stimulation, NFAT cooperates with the AP-1 family of transcription factors, and possibly others, to induce the expression of activation-associated genes 20, 21. In contrast to normal activation, anergic conditions appear to lead to an unbalanced activation of NFAT and translocation of NFAT proteins to the nucleus in the absence of AP-1, thereby inducing the expression of anergy-associated genes 15.
In addition to anergy intrinsic to the T cell, NFAT proteins have been implicated in the regulation of gene expression of two other major mechanism of peripheral tolerance – Treg suppression and activation-induced cell death 22. For instance, it has been reported that NFAT1 interacts with the transcription factor Foxp3 in the epigenetic regulation of genes involved in Treg differentiation. Disrupting this NFAT1:Foxp3 complex prevents the immunosuppressive function of Foxp3+ Treg cells 23. Furthermore, CD4+ T cells from NFAT1/NFAT4 double-deficient mice are resistant to Treg-mediated immunosuppression 24. Additionally, NFAT1/NFAT4 double knockout mice show a defect in TCR-induced T-cell apoptosis, most probably due to an absence of FasL upregulation 25, 26. Thus, NFAT proteins are emerging as key T-cell transcription factors that can regulate the different cell fates of a T cell.
Transcriptional regulation of E3 ligases
The NFAT-induced anergy-associated genes encode, among other proteins, tyrosine phosphatases, proteinases, cell cycle inhibitors, transcriptional regulators, diacylglycerol kinases and E3 ligases 15, 19. These “anergic factors” target different molecules involved in T-cell activation and proliferation to maintain the long-term unresponsive state 27. In particular, the upregulated E3 ligases Cbl-b, Grail and Itch bind and ubiquitylate different key TCR signaling molecules to fine tune peripheral T-cell responses (Fig. 1) 4, 5.
Intensive research is now focusing on the molecular mechanisms involved in NFAT-dependent transactivation of E3 ligase genes. Cbl-b gene expression is reported to be dependent on the early growth transcription factors Erg2 and Erg3, which are themselves induced under anergic conditions in an NFAT-dependent manner 15, 28, 29. Erg3−/− T cells are resistant to anergy induction and Erg3-deficient mice have increased susceptibility to autoimmunity, phenocopying Cbl-b-deficient T cells in vitro and the phenotype of Cbl-b mutant mice in vivo, respectively 28, 29. Overexpression of Erg2 or Erg3 in T cells leads to higher Cbl-b expression levels, whereas Erg3−/− T cells treated with ionomycin (a calcium ionophore used to induce anergy) are unable to upregulate Cbl-b 29. However, it is still not clear if Erg2 and Erg3 bind directly to the Cbl-b promoter or if other transcription factors are involved, including NFAT itself 28, 29. The transcription factor Foxp3 has also been reported to transcriptionally regulate the expression of Erg3 and Cbl-b, but the mechanism is not yet known 30. In the case of Grail, a recent publication shows that homodimers of NFAT1 can directly bind to the Grail promoter and activate gene expression, without any other transcriptional partners 31. It is tempting to speculate, that NFAT hetero- or homodimers may regulate the gene expression of additional E3 ligases involved in T-cell biology; tough further research is needed to examine this possibility.
Physiological and molecular roles of E3 ligases in T-cell tolerance
The RING-type E3 ligase Cbl-b (Casitas B-cell lymphoma-b) was the first identified E3 ligase involved in T-cell activation and tolerance in vivo32–34. Ablation of Cbl-b renders peripheral T cells hyperproliferative and able to be fully activated in the absence of CD28 co-stimulation, suggesting that Cbl-b is a critical modulator of T cells, uncoupling T-cell activation from the requirement of co-stimulation 32, 33, 35. At the molecular level, Cbl-b-mediated ubiquitylation of p85, the catalytic subunit of PI3K, alters its intracellular localization, thereby preventing its interaction with CD28 and TCRζ at the plasma membrane. 36. This results in the inhibition of Akt and PKCθ-mediated NFκB activation 37, and additionally, prevents Vav1-mediated cytoskeleton rearrangements required for receptor clustering and synapse formation 38. Furthermore, Cbl-b has been reported to destabilize the immune synapse by interfering with the CrkL/C3G signaling pathway 39. Cbl-b-dependent ubiquitylation of CrK-L prevents its association with C3G, therefore, inhibiting CrkL/C3G-driven clustering of cell adhesion molecules, in particular LFA-1 39. In accordance with early studies showing Cbl-b upregulation in anergized T cells, Cbl-b−/− T cells are resistant to ionomycin-induced anergy 34, 40. Under anergic conditions, Cbl-b upregulation correlates with PLCγ1 and PKCθ ubiquitylation and an ensuing calcium mobilization defect, suggesting these two molecules may be crucial targets for Cbl-b anergic functions 34, 40.
Additionally, Cbl-b controls susceptibility to Treg-mediated immunosuppression and possibly T-cell activation-induced apoptosis. Naturally occurring Treg cells develop normally and can function properly in the absence of Cbl-b 41, 42. However, Cbl-b−/− T cells are partially resistant to Treg and TGF-β-mediated immunosuppression 41, 42. Moreover, Cbl-b deficiency in Th1 subtype cells provides resistance to activation-induced apoptosis when Th1 cells are induced by CD3 in the absence of co-stimulation 43. Interestingly, in a recent publication, microarray analysis revealed that anergy-associated genes, particularly Erg 2, Erg3 and Cbl-b are upregulated during CD8+ clonal T-cell deletion, indicating that the molecular programs of these two mechanisms of tolerance have common features, in particular, the requirements for E3 ligases 44.
The strongest evidence for the crucial role of Cbl-b in peripheral T-cell tolerance arises from genetic studies. Cbl-b knockout mice develop spontaneous autoimmunity, characterized by generalized organ infiltration and high serological levels of autoantibodies 33. Importantly, these mice are highly susceptible to developing peptide-induced experimental autoimmunity, such as experimental encephalomyelitis 32, arthritis 34 and autoimmune diabetes 45. In accordance with in vitro data, Cbl-b-deficient mice cannot be tolerized in vivo34. More strikingly, Cbl-b-mediated T-cell tolerance determines the survival of the mice upon rechallenge with the same antigen. Repetitive challenge of P14 TCR transgenic mice with high doses of the soluble cognate peptide p33 induces a protective immunotolerance, where P14+ T cells cannot further proliferate upon re-stimulation with p33. In contrast, rechallenge of Cbl-b−/− P14+ mice with p33 is lethal, due to massive activation of T cells and the ensuing cytokine storm 34. Thus, Cbl-b has emerged as a master regulator of immunity, controlling T-cell tolerance and autoimmunity.
Mice deficient in the HECT-E3 ligase Itch develop a spontaneous severe dermatitis-type inflammatory disorder with constant scratching of the skin 46. This itching phenotype is characterized by a Th2 bias in T-cell development and a concomitant increase in the IL-2-dependent cytokines IL-4 and IL-5, as well as in serological levels of immunoglobulins IgG1 and IgE 47. This phenotype inversely correlates with that of mice lacking JunB, the AP1 transcriptional factor responsible for gene regulation in Th2 cell differentiation 48. Further studies suggest that Itch controls Th2 differentiation by binding and mediating JunB ubiquitin-dependent degradation 47. However, as Itch is ubiquitously expressed, the contribution of other cellular and molecular Itch-dependent mechanisms to this complex phenotype cannot be excluded. Indeed, Itch deficiency can also affect the differentiation of epidermal keratinocytes and γδ T-cell-dependent IgE production 49, 50.
Beside its role in T-cell differentiation, Itch has also been implicated in the regulation of T-cell tolerance. As in the case of Cbl-b, Itch is upregulated in vitro under anergizing stimuli 40. In accordance, Itch−/− Th2 cells are resistant to ionomycin-induced anergy as well as high doses of tolerizing antigen in a mouse model for asthmatic airway inflammation 51. Moreover, Itch−/− CD4+ T cells are resistant to Treg-dependent immunosuppression, and Itch−/− Treg cells differentiated in the presence of TGF-β have low expression of Foxp3 and are unable to suppress airway inflammation 52. In this study, it was proposed that Itch participates in the TGF-β-mediated differentiation of Treg cells by ubiquitiylating and activating the transcription factor TIEG1, which in turn appears to bind to and transactivate the Foxp3 promoter.
The gene related to anergy in lymphocytes (Grail) is a type I transmembrane E3 ligase with an endosomal subcellular localization, containing a luminal protease-associated PA domain and a cytosolic enzymatic RING domain 53. As its name suggests, Grail was identified as a transcript highly upregulated in anergized T cells 40, 53. Subsequent studies confirmed its upregulation in different models of in vivo adaptive tolerance, and further showed that constitutive expression of Grail was sufficient to render the transduced T cells anergic 54. However, at that time the mechanisms underlying Grail-mediated T-cell anergy were unclear, as no substrates had been identified. The Rho guanine dissociation inhibitor (RhoGDI) was the first putative Grail substrate identified in a prokaryotic-based E3 ligase screen 55. In this study, Grail-dependent ubiquitylation increased RhoGDI stability while concomitantly inhibiting RhoA GTPase activity. The authors hypothesized that increased RhoGDI levels may sequester Rho molecules in the cytosol, thus preventing Rho signaling pathways involved in T-cell activation, such as cytoskeleton rearrangements and IL-2 production 55. Furthermore, recent studies have proposed that Grail ubiquitylates different transmembrane proteins, including CD40L and the tetraspanin proteins CD151 and CD81, in a novel “across membrane” mechanism, in which the Grail PA domain binds the target protein in the extracellular/luminal compartment while the RING finger domain catalyzes the ubiquitylation process in the cytosol 56, 57.
As with Cbl-b and Itch, Grail also appears to regulate Treg cells. Grail is highly expressed in natural occurring Treg cells, and its expression levels correlate with the immunosuppressive activity of Treg cells induced in vivo through repetitive exposure to superantigens 58, 59. Moreover, retroviral expression of Grail allows the transduced T cells to acquire, together with many of the characteristic cellular markers of Treg cells, the capacity to immunosuppress T-cell responses 59. Thus, Grail appears to be essential for T-cell immunotolerance. Future genetic studies will help establish the in vivo immune functions of Grail.
A novel putative RING-domain E3 ligase, Roquin, has been identified as a key in vivo autoimmune regulator. Mice carrying homozygous mutant alleles of Roquin are susceptible to autoimmune diabetes and develop an autoantibody-driven autoimmune phenotype resembling systemic lupus erythematosus 60. At the cellular level, these mice carry T cells with excessive expression levels of the co-receptor ICOS, together with a marked increase in the number of follicular B helper T cells TFH and ensuing excessive numbers of germinal centers 60, 61. Roquin is believed to participate in the differentiation and function of TFH by negatively regulating the expression of ICOS at the mRNA level, affecting ICOS mRNA stability and/or translation 62. As ubiquitylation is intimately involved in the regulation of mRNA by microRNA, it is highly probable that Roquin regulation of ICOS is dependent on its E3 ligase activity. If this hypothesis is verified, these studies will have undercovered a novel mechanism for E3 ubiquitylation-mediated regulation of autoimmunity.
Molecular regulation of E3 ligases in tolerance
Since E3 ligases are critical mediators of T-cell immunotolerance, their activity must be stringently controlled. Mice deficient in Ndfip-1 (Nedd4 family interacting protein-1) exhibit a phenotype similar to Itch−/− mice: spontaneous inflammation of the skin, a TH2 bias and decreased JunB turnover 63. These observations led to the discovery that Ndfip-1 is upregulated upon T-cell activation and interacts with Itch to promote Itch-dependent ubiquitylation and degradation of JunB, through a yet unknown mechanism 63. Additionally, Itch activity toward JunB can be regulated by phosphorylation; JNK-mediated serine/threonine phosphorylation activates Itch, whereas phosphorylation of tyrosine residues by Fyn results in Itch inactivation 64, 65. In a similar way, CD28 engagement targets Cbl-b for ubiquitylation and subsequent degradation, whereas CTLA-4 binding to B7 upregulates Cbl-b expression 66, 67. However, the molecular intermediates in these regulations need to be clarified.
Moreover, E3 ligases can interact and regulate each other. The HECT-E3 ligases Itch and Nedd4 can bind and ubiquitylate Cbl-b in vitro, targeting it for proteosomal degradation 68. Furthermore, Nedd4 has been recently confirmed to bind and mediate Cbl-b polyubiquitylation in vivo upon CD3/CD28 engagement 69. Hence, Nedd4 appears as a plausible key intermediate between CD28 co-stimulation and Cbl-b degradation, required for promoting T cells activation. There are also data suggesting that the E3 ligase TRAF6 is involved in Cbl-b regulation. TRAF6-deficient T cells fail to upregulate Cbl-b expression under anergizing conditions, and mice carrying this conditional T-cell deletion exhibit similar phenotypic characteristics to Cbl-b−/− mice 70, 71. Whether this regulation is directly mediated by TRAF6 or depends on its E3 ligase activity needs to be explored.
E3 ligases can also be modulated by extracellular factors. It has been shown that human peripheral blood T cells acquire an anergic phenotype if primed with CD3/CD28 in the presence of oxidized phospholipids (Ox-PL) and that under these stimulatory conditions, OxPL induces the upregulation of Erg3 and Cbl-b in T cells 72. Although certainly interesting, the mechanism, as well as the in vivo relevance, of this Ox-PL-anergy regulation remains unknown. In addition, IL-7 has recently been demonstrated to regulate the expression of E3 ligases 73. IL-7 administration in vivo leads to up- and downregulation of Nedd4 and Cbl-b in CD8+ T cells, respectively, suggesting that IL-7 may regulate Cbl-b degradation via Nedd4 73.
Reversing anergy: Deubiquitylation and IL-2 signals
Even though the unresponsive state in anergic T cells is long lasting, it is not permanent. Clonal T-cell anergy is reversed by the addition of exogenous IL-2, whereas in vivo anergy can be reversed if the cognate Ag is removed 74, 75. A recent publication suggests that IL-2-induced signaling inhibits and/or reverses the anergic state by controlling expression of E3 ligases 76. IL-2 receptor signaling, through JAK3 and mTOR, inhibited the calcium-dependent expression of anergy-associated genes, including Cbl-b and Grail E3 ligases. Additionally, IL-2 signaling upregulates AP-1, suggesting that the repression of E3 ligases could be partially explained by a reconstitution in the balance of NFAT and AP-1, which in turn may inhibit the NFAT-dependent anergic gene expression programme 76.
Similar to the way phosphatases control kinase-mediated signaling pathways, deubiquitylating enzymes (DUB) can also reverse the ubiquitylation process 9. Some DUB involved in immune system regulation have already been identified, including A20, CDYLD, DUBA, Otubain-1 and FAM/USP9X 77. However, their involvement in T-cell tolerance is still primarily speculative 9. Among the best-characterized putative DUB involved in E3 ligases signaling are Otubain-1 and FAM/USP9X, which regulate Grail and Itch auto-ubiquitylation, respectively. While FAM-dependent deubiquitylation protects Itch from auto-ubiquitylation-driven degradation, Otubai-1 deubiquitylation of Grail results in its proteolysis 78, 79. Otubain-1 modulation of Grail has been suggested to contribute to the regulation of T-cell anergy and, as recently proposed, may participate in T-cell activation and T-cell proliferation 78, 80.
Concluding remarks and perspectives
The last few years have seen a remarkable progress in our knowledge regarding the molecules and signaling pathways involved in T-cell tolerance. Extensive work of many laboratories has accumulated evidence that E3 ligases are key tolerogenic molecules and that ubiquitylation of proteins by E3 ligases is a novel crucial pathway controlling the development and susceptibility to autoimmunity. Hence, E3 ligases are exciting and novel suitable targets for the treatment of autoimmune diseases and the regulation of immunological tolerance. However, before the modulation of E3 ligases becomes a real therapeutic possibility, several unanswered questions must be considered. For instance, the underlying mechanism(s) in the regulation of the ubiquitylation system in T-cell tolerance has been comparatively poorly explored and further work is needed to undercover the complexity of the system. Importantly, although some in vitro target molecules have already been identified, the identity of the relevant E3 ligase in vivo substrates remains largely unknown.
Given the complexity and great numbers of molecules involved in T-cell activation, and the capacity of E3 ligases to bind several proteins, one can envision that many in vivo substrates are awaiting discovery. For the identification of such novel target molecules involved in T-cell tolerance, the generation of specific E3 ligase activity-dead mice would certainly be of great value. As many E3 ligases are ubiquitously expressed and act at different cellular levels, participating also in many aspects of the innate immunity 10 – though not discussed here – the generation of conditional knockout mice will be also necessary to further dissect the cellular mechanisms that might mask the phenotypes of conventional total body knockout mice. Gaining insights into the mechanisms of E3 ligase-mediated T-cell activation and immunotolerance has not only expanded the universe of T-cell signaling pathways but also provides a unique opportunity for specific therapeutic interventions in immunological diseases.
The authors thank Shane Cronin for critical reading of the manuscript. This work is supported by grant to J.M.P. from Euro-Thymaide, the Austrian Science Found (Spezialforschungbereich -SFB- on immunotolerance), and Innovative Mouse Models for Functional Genomics in Immunology (IMDEMI).
Conflict of interest: J.M.P. is a co-owner of a company that attempts to develop anti-Cbl-b therapies.