Regulatory T Cell Differentiation: Turning Harmful into Useful

In this issue of Immunity, Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar demonstrate a surprisingly broad “window of opportunity” for regulatory T (Treg) cell induction and provide evidence for an avidity model of Treg cell differentiation versus deletion. In this issue of Immunity, Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar demonstrate a surprisingly broad “window of opportunity” for regulatory T (Treg) cell induction and provide evidence for an avidity model of Treg cell differentiation versus deletion. A series of landmark studies in the late 1980s indicated that through positive and negative selection “the thymus selects the useful….and destroys the harmful” (von Boehmer et al., 1989von Boehmer H. Teh H.S. Kisielow P. Immunol. Today. 1989; 10: 57-61Abstract Full Text PDF PubMed Scopus (317) Google Scholar). A few years later, the avidity model of thymic selection provided a conceptual framework to resolve the apparent paradox that two diametrically different thymocyte fates (survival versus death) are both orchestrated through T cell receptor (TCR) signals emanating from recognition of “self” (Jameson et al., 1995Jameson S.C. Hogquist K.A. Bevan M.J. Annu. Rev. Immunol. 1995; 13: 93-126Crossref PubMed Scopus (549) Google Scholar). Nevertheless, after two decades, we are still far from a precise understanding of how an analog input (incrementally varying TCR signals) is translated into a digital cell-fate decision (positive or negative selection). In fact, the situation becomes even more perplexing with the discovery that for developing CD4+ T cells there is yet another TCR-driven developmental option, namely the differentiation into the Foxp3+ regulatory T (Treg) cell lineage. The production of Treg cells is now generally recognized as “the third function of the thymus” (Seddon and Mason, 2000Seddon B. Mason D. Immunol. Today. 2000; 21: 95-99Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). How do Treg cell differentiation and clonal deletion relate to each other? There is little evidence that negative selection and Treg cell differentiation segregate in a temporal or spatial manner or differ in their requirement for a dedicated antigen-presenting cell (APC)-type. Instead, a modified version of the avidity model has become a popular hypothesis for explaining the choice between Treg cell differentiation and clonal deletion (Simons et al., 2010Simons D.M. Picca C.C. Oh S. Perng O.A. Aitken M. Erikson J. Caton A.J. J. Leukoc. Biol. 2010; 88: 1099-1107Crossref PubMed Scopus (29) Google Scholar). Consistent with the idea that Treg cell differentiation occurs within an avidity window between positive and negative selection, a positive correlation between the amount of antigen expressed and the number of emerging Treg cells is observed in a series of TCR × neo-antigen double-transgenic systems, whereby “very high” antigen doses eventually favor deletion (Simons et al., 2010Simons D.M. Picca C.C. Oh S. Perng O.A. Aitken M. Erikson J. Caton A.J. J. Leukoc. Biol. 2010; 88: 1099-1107Crossref PubMed Scopus (29) Google Scholar). Similarly, attenuation of cognate antigen presentation by medullary thymic epithelial cells (mTECs) through tissue-specific silencing of major histocompatibility complex (MHC) class II diminishes negative selection and increases the emergence of specific Treg cells (Hinterberger et al., 2010Hinterberger M. Aichinger M. Prazeres da Costa O. Voehringer D. Hoffmann R. Klein L. Nat. Immunol. 2010; 11: 512-519Crossref PubMed Scopus (185) Google Scholar). Because these studies focused on avidity, several open questions pertaining to the role of TCR affinity remain: How do varying TCR affinities influence thymocyte cell-fate decisions at a given intensity of antigen expression? How broad is the range of TCR affinities that are permissive for Treg cell differentiation? How do these parameters relate to the phenomenon of intraclonal competition between Treg cell precursors of identical specificity? In this issue of Immunity, Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar report on experiments that address these questions through TCR “retrogenic” technology, which allows for the relatively rapid retroviral introduction of TCRs into thymocytes or bone marrow cells. In contrast to previous studies that mostly used a fixed transgenic TCR specificity and varied the antigen dose, Lee et al. pursued an approach that combines fixed conditions of intrathymic autoantigen expression with varying TCR affinities (wherein the term “affinity” is used to describe the relative responsiveness of a given TCR rather than true biophysical properties). The authors use the well-characterized rat insulin promoter-membrane ovalbumin (RIP-mOVA) model, in which OVA expression is specifically and exclusively confined to mTECs and thereby recapitulates features of “promiscuous” expression of peripheral-tissue antigens. Previous work has shown that these modalities of OVA expression efficiently induce Treg cell differentiation of DO11.10 TCR transgenic thymocytes, allowing Lee at al. to employ the DO11.10 TCR as a point of reference throughout the study. Using a sophisticated approach involving TCR sequencing of Foxp3+ or Foxp3− CD4+ T cells from mice that transgenically express the TCRβ chain of the DO11.10 TCR, the authors selected a total of eight closely related OVA-specific TCRs that all contained the β chain of the DO11.10 TCR and harbored identical TCRVα and Jα elements, but differed in one or two amino acid residues within the TCRα complementarity determining region 3 (CDR3). When tested for their dose-response characteristics in vitro, these TCRs showed a broad distribution of responsiveness to stimulation with OVA peptide, with the most sensitive TCR resembling the DO11.10 TCR, whereas the least sensitive TCR required more than 1,000-fold more OVA peptide for half-maximal stimulation. Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar then asked whether any of these TCRs would mediate OVA-independent Treg cell differentiation in vivo. This was done using retroviral TCR transfer into CD4 CD8 double-negative thymocytes from Rag2−/− mice followed by intrathymic injection into wild-type (WT) mice. Somewhat surprisingly, three of the TCRs, including the one with the highest responsiveness to OVA, gave rise to Treg cells even in the absence of OVA. However, since all three TCRs had initially been cloned out of Foxp3+ cells from WT mice, a plausible explanation is that they were cross-reactive to endogenous self-antigens. The finding that “OVA-specific” Treg cells can arise in the absence of OVA is not a trivial observation (at least for those who do not think about TCR cross-reactivity on a daily basis), because it indicates that high-affinity “foreign-specific” Treg cells can be educated on self, which has implications for a variety of settings, such as infection or vaccination. Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar then go on to show that of the remaining five TCRs that did not facilitate Treg cell differentiation in WT recipients, four did so when TCR-transduced precursors were injected into RIP-mOVA thymi. The least sensitive of these “Treg cell-promoting” TCRs was at least 100-fold less responsive to OVA in vitro when compared to the DO11.10 TCR. Hence, under fixed conditions of antigen expression, Treg cell differentiation can occur over a remarkably broad range of TCR affinities. However, the efficacy of Treg cell differentiation was directly proportional to the TCR affinity, so that a relatively large fraction of cells bearing high-affinity TCRs became Treg cells, whereas only a small percentage of cells with low-affinity TCRs did so. Recent studies in mice transgenically expressing Treg cell-derived TCRs of unknown specificity indicated that the fraction of Foxp3+ cells among clonotype positive cells inversely correlated with their precursor frequency (Bautista et al., 2009Bautista J.L. Lio C.W. Lathrop S.K. Forbush K. Liang Y. Luo J. Rudensky A.Y. Hsieh C.S. Nat. Immunol. 2009; 10: 610-617Crossref PubMed Scopus (198) Google Scholar; Leung et al., 2009Leung M.W. Shen S. Lafaille J.J. J. Exp. Med. 2009; 206: 2121-2130Crossref PubMed Scopus (112) Google Scholar). Conceivably, parameters such as limited access to antigen or competition for other permissive and/or instructive factors (e.g., cytokines) in particular thymic microenvironments may be involved in this intraclonal competition. To address a potential link between TCR affinity and Treg cell “niche” size, Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar retrovirally introduced two OVA-specific TCRs and the DO11.10 TCR into bone marrow cells and generated mixed chimeras with graded doses of TCR transgenic precursors. This confirmed previous work that showed that Treg cell differentiation inversely correlates with clonal precursor frequency, but more importantly, it revealed that at a given precursor frequency, a higher TCR affinity results in a larger clonal niche. The authors hypothesized that the niche size may be determined by the number of APCs that present sufficient antigen to trigger Treg cell differentiation at the respective TCR affinity. In this context, it is worth mentioning that RIP-driven OVA, like “true” peripheral-tissue antigens, may only be expressed by a small fraction of mTECs, so that the chances of antigen encounter may indeed be a limiting determinant of Treg cell differentiation. How do these findings relate to negative selection? Even with the highest-affinity TCRs, such as the DO11.10 TCR, Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar observed only weak signs of clonal deletion in the RIP-mOVA thymus. Therefore, the present work does not allow for an accurate quantification of the “affinity margin” that separates the minimal requirements for Treg cell differentiation and negative selection, respectively. Nonetheless, it appears fair to conclude that clonal deletion requires a substantially stronger interaction than Treg cell differentiation. The work by Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar may pave the way for a more precise quantitative integration of the parameters that specify the decision-making by MHC class II-restricted thymocytes. Clearly, the affinity and avidity rules of thymocyte selection, as they have mostly been established for MHC class I-restricted thymocytes, cannot be “linearly” transferred to MHC class II-restricted T cells. Thus, whereas for developing CD8+ T cells there appears to be a very sharp affinity threshold separating positive and negative selection (Daniels et al., 2006Daniels M.A. Teixeiro E. Gill J. Hausmann B. Roubaty D. Holmberg K. Werlen G. Holländer G.A. Gascoigne N.R. Palmer E. Nature. 2006; 444: 724-729Crossref PubMed Scopus (456) Google Scholar), it seems that for CD4+ T cells an additional, surprisingly broad, and in part overlapping window of opportunity for Treg cell differentiation separates these two options (Figure 1) . The present work also raises a number of other questions: What is a low-affinity Treg cell good for, considering that Treg cells need TCR stimulation to exert suppressive function? Do precursors carrying low- and high-affinity TCRs compete for the same Treg cell niche, and if so, will high-affinity TCRs out-compete low-affinity TCRs? An utterly unexpected finding of Lee et al., 2012Lee H.M. Bautista J.L. Scott-Browne J. Mohan J.F. Hsieh C.S. Immunity. 2012; 37 (this issue): 475-486Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar is that at the lowest precursor frequencies tested (0.1% of CD4+SP cells), the Treg cell niche for high-affinity TCRs “collapses,” so that no Treg cell differentiation is detectable any longer. Future work is needed to explain this observation, in particular in light of the fact that this experimental frequency of 1 in 1,000 still far exceeds recent estimates of the physiological frequency of antigen-specific thymocytes in the polyclonal repertoire (Moon et al., 2011Moon J.J. Dash P. Oguin 3rd, T.H. McClaren J.L. Chu H.H. Thomas P.G. Jenkins M.K. Proc. Natl. Acad. Sci. USA. 2011; 108: 14602-14607Crossref PubMed Scopus (89) Google Scholar). In sum, the study by Lee et al. provides strong evidence in favor of an avidity model of Treg cell differentiation versus negative selection and sheds novel light on how the thymus not only selects the useful and destroys the harmful, but also fulfills its third essential function: turning harmful cells into useful ones. A Broad Range of Self-Reactivity Drives Thymic Regulatory T Cell Selection to Limit Responses to SelfLee et al.ImmunityAugust 23, 2012In BriefThe degree of T cell self-reactivity considered dangerous by the immune system, thereby requiring thymic selection processes to prevent autoimmunity, is unknown. Here, we analyzed a panel of T cell receptors (TCRs) with a broad range of reactivity to ovalbumin (OVA323-339) in the rat insulin promoter (RIP)-mOVA self-antigen model for their ability to trigger thymic self-tolerance mechanisms. Thymic regulatory T (Treg) cell generation in vivo was directly correlated with in vitro TCR reactivity to OVA-peptide in a broad ∼1,000-fold range. Full-Text PDF Open Archive