To establish an immunocompetent TCR repertoire that is useful yet harmless to the body, a de novo thymocyte repertoire generated through the rearrangement of genes that encode TCR is shaped in the thymus through positive and negative selection. The affinity between TCRs and self-peptides associated with MHC molecules determines the fate of developing thymocytes. Low-affinity TCR engagement with self-peptide–MHC complexes mediates positive selection, a process that primarily occurs in the thymic cortex. Massive efforts exerted by many laboratories have led to the characterization of peptides that can induce positive selection. Moreover, it is now evident that protein degradation machineries unique to cortical thymic epithelial cells play a crucial role in the production of MHC-associated self-peptides for inducing positive selection. This review summarizes current knowledge on positive selection–inducing self-peptides and Ag processing machineries in cortical thymic epithelial cells. Recent studies on the role of positive selection in the functional tuning of T cells are also discussed.
The thymus is an organ that produces functionally competent T cells capable of responding to diverse foreign Ags. The diversity of the Ag recognition specificities in T cells is primarily generated in immature thymocytes through the V(D)J rearrangement of TCR genes. T lymphoid progenitors are induced in the thymic cortex to become cortical CD4+CD8+ (double-positive [DP]) thymocytes that express TCR-αβ complexes with individual specificities. The initially generated repertoire of TCR specificities in immature thymocytes is selected upon the interaction between TCRs and self-peptides associated with MHC molecules provided in the thymic cortical microenvironment (1, 2). Low-affinity TCR interactions with self-peptide–MHC complexes transduce signals for the survival of DP thymocytes and their further differentiation into CD4+CD8− or CD4−CD8+ single-positive (SP) thymocytes. This process, termed positive selection, enriches a potentially useful repertoire of TCR specificities. Cortical DP thymocytes are unable to undergo further development in the absence of TCR signals or upon high-affinity TCR interaction with self-peptide–MHC complexes. Self-peptide–MHC complexes presented by cortical thymic epithelial cells (cTECs) are essential for inducing the positive selection of T cells (3–5).
Positively selected thymocytes express the chemokine receptor CCR7 and migrate to the thymic medulla in a CCR7-dependent manner. In the thymic medulla, medullary thymic epithelial cells (mTECs) and dendritic cells (DCs) provide an additional set of self-peptide–MHC complexes (6–8). The high-affinity TCR interaction of developing thymocytes with self-peptide–MHC complexes presented in the thymic medulla further trims the repertoire of TCR specificities to install self-tolerance through the apoptotic elimination of self-reactive thymocytes and the generation of regulatory T (Treg) cells. Consequently, multistage selection processes via cortical positive selection and medullary tolerance induction are key to establishing a functionally competent and self-tolerant repertoire of T cells. In this review, we focus on the unique protein degradation machineries of cTECs, which are important for producing self-peptide–MHC complexes that induce positive selection of thymocytes.
Antigenic peptide variants and positive selection
Early experiments that examined the nature of thymic selection–inducing ligands showed that limited concentrations of antigenic TCR ligands can induce positive selection, and this led to the proposal of the avidity model, in which the avidity of ligand-occupied TCRs determined by the affinity and the availability of TCR–ligand interactions critically affects the outcome of thymic positive and negative selection (9–11). However, positive selection induced by high-affinity antigenic peptides occurs only when the affinity falls within a certain range (12, 13). Moreover, T cells induced by the high-affinity peptides are sometimes functionally incompetent (14) or are driven to unconventional T cell lineages, including Treg cells, NKT cells, and CD8αα+ T cells (15–19). In contrast, it was shown that the positive selection of conventional T cells was efficiently triggered by low-affinity TCR ligands in a dose-independent manner (13, 20). Thus, the affinity between TCRs and peptide–MHC complexes critically determines whether immature thymocytes are positively selected or not.
Partial substitutions of amino acid residues in antigenic peptides recognized by a given TCR can alter the affinity between the TCR and the peptide–MHC complexes. The combination of OT-I–TCR-transgenic thymocytes and chicken OVA-derived antigenic peptide SIINFEKL in the context of H-2Kb is the most widely examined TCR–peptide–MHC complex (21, 22). It was initially noted that altered peptides that antagonize the Ag-induced activation of mature T cells were often (21–23) but not always (24–26) associated with the capability to induce positive selection. It was later established that among the altered peptides, those with low affinity induced positive selection, whereas high-affinity peptides were efficient in inducing negative selection (27–29). The affinity ranges that induce positive and negative selection are broad, but the transition from positive selection to negative selection occurs across a very narrow affinity range (13, 30). Coreceptor scanning dependence and TCR–ligand dwell-time thresholds are critical for distinguishing the positive and negative selection of thymocytes (13, 31).
How can such a small difference in TCR affinity initiate different TCR signaling cascades that result in distinct fates of developing thymocytes? The difference in subcellular compartmentalization of Ras/MAPK signaling molecules is important for discriminating signaling cascades for positive and negative selection of thymocytes (13). Themis, a component of the TCR signaling pathway, sets the signal threshold for positive and negative selection by specifically attenuating the TCR signals induced by low-affinity TCR ligation through the recruitment of Src homology phosphatase-1 (32–34).
Naturally processed peptides and positive selection
In an attempt to understand the nature of naturally processed MHC-associated self-peptides that induce positive selection in the thymus, peptides bound to MHC molecules have been isolated from various cell types, sequenced by mass spectrometry, and tested for the capability to induce positive selection in fetal thymus organ cultures of TCR-transgenic thymocytes (35–38). Regarding positively selecting peptides for MHC class I–restricted CD8+ T cells, H-2Kb–associated peptides derived from F-actin capping protein (Cappa1) and β-catenin (Catnb), isolated from EL4 T lymphoma cells and LB27.4 B lymphoma cells, were found to induce positive selection of OT-I–TCR-transgenic thymocytes (36, 38). These peptides had lower affinity to OT-I–TCR than did the previously studied SIINFEKL variants that induced positive selection (39). It remains unclear whether these positive selection–inducing naturally processed peptides share biochemical characteristics with antigenic epitope peptides (35, 37, 38).
Naturally occurring peptides that induced positive selection of MHC class II–restricted AND-TCR–transgenic and 5C.C7-TCR–transgenic CD4+ T cells were identified from I-Ek–associated peptides isolated from CH12 B lymphoma cells (40, 41). These TCRs recognize an identical peptide derived from moth cytochrome c bound to I-Ek (41, 42). However, the endocytic receptor protein-derived peptide gp250, which positively selects AND-TCR–transgenic thymocytes, and six peptides identified to induce positive selection of 5C.C7-TCR–transgenic thymocytes are not overlapped, suggesting that the positive selection–inducing peptides for these two individual TCR-expressing thymocytes are distinct. The affinity between AND-TCR and the gp250–I-Ek complex is very low (43). Some of the 5C.C7 positively selecting peptides, but not gp250, are partially homologous to antigenic moth cytochrome c peptide at TCR-contacting residues (40, 41).
Thus, the structural similarity between antigenic epitope peptides and naturally processed positively selecting peptides seems limited rather than universal. More importantly, none of the naturally processed peptides identified so far was isolated from freshly isolated cTECs, even though one study isolated MHC-associated peptides from the transformed thymic stromal cell line 427.1, which morphologically resembled cTECs (35). As we discuss below, cTECs carry unique Ag-processing machineries for the production of positive selection–inducing MHC-associated peptides. Naturally processed peptides that induce positive selection should be examined in cTECs.
cTEC-specific production of self-peptides for positive selection of CD4+ T cells
Peptide presentation by MHC class II molecules involves the endosomal–lysosomal system. Newly generated MHC class II molecules are assembled with invariant chain (Ii) in the endoplasmic reticulum and are transported through the Golgi apparatus to the endosomal system. In the late endosomes, Ii undergoes proteolytic degradation, leaving a short fragment called CLIP in the peptide-binding groove of MHC class II molecules. H-2M in mice, or HLA-DM in humans, catalyzes the replacement of CLIP with peptides available in the endosomes and generates peptide-loaded MHC class II molecules that are eventually transported to the cell surface for Ag presentation. cTECs carry a number of unique protein degradation machineries that seem to affect the repertoire of MHC class II–associated self-peptides that are important for inducing positive selection of thymocytes (Fig. 1).
cTECs strongly express the endosomal–lysosomal protease cathepsin L, which plays a critical role in Ag presentation through the MHC class II pathway by degrading Ii and generating presented peptides, whereas other major APCs, such as DCs and mTECs, predominantly express other lysosomal proteases, including cathepsin S (44–47). In cathepsin l–deficient cTECs, Ii is incompletely processed and cell-surface MHC class II expression is not reduced (44, 45). The cellularity of CD4SP thymocytes and peripheral CD4+ T cells is reduced by 60–80% in cathepsin l–deficient mice (44, 45). The repertoire of MHC class II–bound peptides is different in fibroblasts expressing either cathepsin S or cathepsin L (48), suggesting that cathepsin L contributes to the generation of MHC class II–bound self-peptides that are uniquely presented by cTECs and are central to positive selection of CD4+ T cells in the thymus. Autoimmunity seems less severe in cathepsin l–deficient mice (49, 50). A human homolog of mouse cathepsin L, known as cathepsin V or cathepsin L2, is strongly expressed in cTECs in humans (51). The defective generation of CD4+ T cells in cathepsin l–deficient mice is restored by the transgenic expression of human cathepsin V, suggesting that cathepsin V participates in positive selection in humans (52). The altered expression of and polymorphisms in the cathepsin V gene in patients with type 1 diabetes and myasthenia gravis further suggest its association with the pathogenesis of autoimmune disease in humans (51, 53, 54).
Thymus-specific serine protease.
Thymus-specific serine protease (TSSP; Prss16) is an endosomal protease that is strongly expressed in cTECs (55, 56). TSSP-deficient mice show normal numbers but an altered TCR repertoire of CD4+ T cells, which are poor in particular Ag responses (57, 58). TSSP deficiency results in resistance of NOD mice to diabetes (59–61) and spontaneous colorectal cancer (62). The polymorphisms in the TSSP locus in human populations are associated with the susceptibility to type 1 diabetes (63–65). Thus, it is likely that TSSP is involved in producing self-peptides in cTECs for positive selection, in a manner different from cathepsin L.
Autophagy is the process of recycling amino acids by degrading cytoplasmic components, including aggregated proteins and damaged organelles, and contributes to MHC class II Ag processing of cytosolic proteins (66–71). Autophagy is induced under stress conditions, such as starvation, whereas a basal level of autophagy is essential for homeostasis in almost all cell types (67, 71). Autophagosomes formed by engulfing cytosolic contents are fused with lysosomes, and the delivered proteins are processed for MHC class II–associated peptide presentation (66, 68–70). Therefore, autophagy influences the repertoire of peptides presented by MHC class II (68). Autophagy is constitutively active in a fraction of TECs even without stress induction (67). Autophagosomes are detectable in >60% of cTECs and ∼10% of MHC class II–high mTECs (72). A deficiency in Atg5, which is required for the formation of autophagosomes, causes impaired positive selection of a fraction of MHC class II–restricted TCR-transgenic CD4+ T cells (72), although the development of polyclonal CD4SP thymocytes appears unaffected in the thymus that lacks either Atg5 or Atg7, another gene involved in the autophagosome formation (72, 73). Thus, autophagy seems to play a role in positive selection of a certain specificity of MHC class II–restricted CD4+ T cells. It is also reported that autophagy in TECs, including mTECs, contributes to negative selection of self-reactive T cells and the prevention of the onset of autoimmune diseases (74, 75).
CD83 and March8.
The turnover of peptide–MHC class II complexes is regulated by E3 ubiquitin ligase March family proteins. In cTECs, the March8-mediated ubiquitination and degradation of MHC class II complexes is interfered by CD83, an Ig superfamily cell-surface protein that facilitates the stabilization of surface MHC class II molecules and contributes to the efficient positive selection of MHC class II–restricted CD4+ T cells (76–78). The CD83-mediated maintenance of the stability of self-peptide–presenting MHC class II complexes is crucial in cTECs for inducing positive selection of CD4+ T cells.
cTEC-specific production of self-peptides for positive selection of CD8+ T cells
The MHC class I pathway of Ag presentation involves proteasomes in the cytoplasm. Cytoplasmic peptides generated by the proteasomes are the major source of peptides that are presented in the association with MHC class I molecules. The core barrel-shaped proteasome complex is composed of four heptameric rings: two outer α rings and two inner β rings (79, 80) (Fig. 2A). Three forms of proteasomes with different catalytic activities have been identified on the basis of the composition of the β subunits: constitutive proteasome (β1, β2, and β5), immunoproteasome (β1i, β2i, and β5i), and thymoproteasome (β1i, β2i, and β5t [Psmb11]) (Fig. 2A).
β5t-containing thymoproteasomes are expressed in most cTECs in a highly specific manner, suggesting that cTECs display a unique set of self-peptides associated with MHC class I molecules (81, 82) (Fig. 1). In mice deficient in β5t-containing thymoproteasomes, cTECs are not reduced in cellularity and express other forms of proteasomes, including β5i-containing immunoproteasomes, without disturbing the corticomedullary architecture of the thymus (81, 82). Strikingly, however, the numbers of CD8SP thymocytes and peripheral CD8+ T cells are reduced to ∼20–30% of normal cellularity in β5t-deficient mice (81, 83). Positive selection of CD8SP thymocytes that express different MHC class I–restricted TCRs is differently affected in the thymus of β5t-deficient mice, whereas the development of CD4+ T cells is not affected (83). Thus, β5t-containing thymoproteasomes are essential for positive selection of the major repertoire of MHC class I–restricted CD8+ T cells.
The variety of the thymoproteasome dependence among the different TCR-expressing thymocytes, the MHC class I expression independence of the thymoproteasome expression, and the thymoproteasome independence of the positive selection of CD4+ T cells together suggest that the thymoproteasome contributes to positive selection of CD8+ T cells by producing a unique set of positive selection–inducing self-peptides in cTECs. The spectrum of MHC class I–associated self-peptides, detected by the cross-reactivity with mAb 25-D1.16, in cTECs is different between control and β5t-deficient mice (83–85). Indeed, the immunoprecipitation of H-2Kb and H-2Db MHC class I molecules from fibroblasts that were managed to predominantly express either thymoproteasomes or immunoproteasomes, the elution of MHC-associated peptides, and the sequence determination of the peptides by mass spectrometry revealed that ∼30% of identified peptides were specific for either thymoproteasome- or immunoproteasome-expressing cells (86). Interestingly, the list of MHC class I–associated peptides identified in thymoproteasome-expressing fibroblasts (86) contained many naturally processed MHC class I–associated peptides previously identified for the capability to induce the positive selection of several TCR-transgenic thymocytes (Cappa1 and Catnb for OT-I–TCR, BP peptide for F5-TCR, and RR peptide for P14-TCR) (35, 36, 38). The thymoproteasome-dependent MHC class I–associated peptides identified from fibroblasts may indeed contain positive selection–inducing peptides presented in cTECs.
How do thymoproteasome-dependent peptides induce positive selection of CD8+ T cells?
It has become evident that β5t-containing thymoproteasomes expressed in cTECs play an important role in inducing positive selection of CD8SP thymocytes in the thymus. Furthermore, it seems reasonable to assume that the thymoproteasomes do so by producing a unique set of MHC class I–associated peptides in cTECs. However, how the thymoproteasomes govern the unique function of cTECs in inducing positive selection of CD8+ T cells remains an unanswered question. Several recent studies have attempted to answer this question.
MHC binding stability.
The proteolytic specificity of thymoproteasomes is different from that of constitutive proteasomes or immunoproteasomes. In particular, the chymotrypsin-like activity of thymoproteasomes to produce the hydrophobic C termini of the peptides is lower than that of other types of proteasomes (81). The hydrophobicity of the C termini of the peptides contributes to the stable anchoring of the peptides in the peptide-binding groove of MHC class I molecules (87, 88). Those results led to speculations that thymoproteasomes may preferentially generate peptides that are loosely bound to MHC class I molecules and the instability of the binding between peptides and MHC class I molecules may contribute to the efficient positive selection of CD8SP thymocytes by cTECs (89, 90) (Fig. 2B). Xing et al. (91) examined whether OVA peptide variants that contained alterations at the C-terminal amino acids might improve positive selection of OT-I–TCR-transgenic thymocytes. The results showed that some of those variant peptides were indeed unstable in the association with H-2Kb, but none of those peptides was capable of inducing the positive selection of OT-I–TCR-transgenic thymocytes. Therefore, it seems unlikely that the stability at the C termini of the peptides is responsible for the thymoproteasome-dependent optimal positive selection.
Optimal TCR affinity.
The unique proteolytic specificity of thymoproteasomes not only affects the hydrophobicity at the C-terminal cleavage site but also alters the choice of amino acids at the residues proximal to the C termini of the produced peptides. A major fraction of thymoproteasome-dependent H-2Kb–bound peptides possess proline at the sixth residue and acidic amino acids at the seventh residue, the residues that face the TCR CDR3 region (86). Amino acid replacement in the OVA peptide SIINFEKL according to these structural motifs drastically lowers the affinity to OT-I–TCR. These peptide variants almost completely lose agonistic activity toward mature OT-I T cells but instead induce efficient positive selection of OT-I–TCR-expressing CD8SP thymocytes in fetal thymus organ culture (86). These results are in agreement with the possibility that thymoproteasome-expressing cTECs display a unique set of self-peptides that carry optimal TCR affinities for triggering positive selection of CD8+ T cells (92) (Fig. 2B).
The thymoproteasome-dependent production of a unique set of MHC class I–associated peptides specifically in cTECs has also led to the speculation of another possibility that the difference between positively selecting self-peptides in the thymic cortex and other self-peptides expressed everywhere else in the body, including the thymic medulla, may create a window of positively selected TCR repertoire that can escape from stringent negative selection (89). At least three laboratories have attempted to test this possibility, and those attempts have yielded superficially different conclusions so far. The first examination used CCR7-deficient thymocytes (83). CCR7 is a chemokine receptor, of which positive selection–inducing TCR signals induce the expression in DP thymocytes. CCR7 guides positively selected thymocytes into the medullary region where mTECs produce CCR7 ligand chemokines (93–95). It was shown that CD8SP development in thymoproteasome-deficient mice remained defective in the absence of CCR7-mediated cortex-to-medulla migration of thymocytes (83). It was also shown that the deficiency in thymoproteasomes reduced the cellularity of MHC class I–engaged postselection DP thymocytes, which are localized in the thymic cortex (91). These results indicate that the thymoproteasome regulates positive selection of CD8+ T cells within the thymic cortex, without the contribution of the microenvironment different from the thymic cortex and without subsequent negative selection in the thymic medulla.
The contribution of the switch in proteasome types was examined by producing mice carrying β5i-encoding gene knocked in at the β5t-encoding locus (91). These mice were further crossed with β5i-knockout mice to produce mice in that cTECs expressed β5i-containing immunoproteasomes instead of β5t-containing thymoproteasomes, and all other cells in the body, including mTECs, expressed β5-containing constitutive proteasomes. These mice therefore provided an in vivo situation where proteasomes were switched between cTECs and other cells, despite the loss of thymoproteasomes. It was found that the loss of thymoproteasomes resulted in defective CD8SP development, even though the difference in proteasome types was created between immunoproteasome-expressing cTECs and constitutive proteasome–expressing other cells. These results indicate that the role of the thymoproteasome in thymic selection is not merely to generate peptides distinct from other thymic APCs but to generate unique peptides that are specialized to promote positive selection (91).
A recent report showed the results of analysis of mice deficient in all four components of cell type–specialized proteasomes, that is, β1i, β2i, β5i, and β5t (4KO mice) (96). All cells in the 4KO mice, including cTECs, mTECs, thymic DCs, and other APCs, express conventional proteasomes only without immunoproteasomes or thymoproteasomes. It was shown that the 4KO mice had a profound defect in the generation of CD8+ T cells and that the decrease in CD8 lineage thymocytes was more profound at the later stages of SP thymocyte development than the earlier stages of recently positively selected DP thymocytes. These results suggest that the development of CD8+ T cells in the absence of all specialized proteasomes may be still blocked at the later stages after positive selection and coincident with the negative selection in the thymic medulla (96). Through experiments using mice deficient in Bim, which is required for the apoptosis-mediated negative selection of self-reactive thymocytes (97), it was found that the Bim-mediated rescue of thymocytes was equivalent in the presence or absence of thymoproteasomes in one study (91) and was more pronounced in the 4KO mice than the control mice in another study (96). Nonetheless, the possible contribution of the proteasome switch, and thereby the self-peptide switch, in CD8+ T cell production was formally supported (96).
Note that even the study that supported the peptide switch mechanism noted that positive selection was defective in the absence of the thymoproteasome (96). Thus, the thymoproteasome governs positive selection at least by generating unique self-peptides that are specialized to promote positive selection by cTECs, for example, by optimizing the affinity to TCRs. To what extent the peptide switch additionally contributes to the generation of the CD8+ T cell repertoire is an interesting question to be addressed.
Positive selection affects Ag responsiveness in T cells
Recent studies have revealed a novel aspect of positive selection, in which the TCR affinity for peptide–MHC complexes during positive selection in the thymus influences the functional capability of mature CD4+ and CD8+ T cells during immune response (98–101). Positive selection affects T cell function by skewing the TCR repertoire (98–100) and by fine-tuning T cell responsiveness (101).
A recent report has described that the CD5 expression levels of monoclonal TCR-transgenic CD4+ T cells are variable and correlated with basal TCRζ phosphorylation levels (98). CD5 is a membrane protein that is elevated upon TCR signaling and can negatively regulate TCR signaling (102–104). The affinity between TCR and self-peptide–MHC complexes affects CD5 expression in positively selected thymocytes and mature T cells (104). CD5 expression levels in CD4+ T cells are positively correlated with TCR binding strength and T cell response to foreign Ags, suggesting that positive selection contributes to effective immunity by skewing the mature TCR repertoire toward the efficient recognition of pathogens (98).
Allen and colleagues (99, 105) examined two TCR-transgenic mouse lines, listeriolysin O (LLO)56 and LLO118, whose CD4+ T cells are identical in the recognition of the same epitope of Listeria monocytogenes virulence factor LLO190–205 associated with I-Ab but are different in immune response. LLO56 CD4+ T cells are more susceptible to activation-induced cell death than LLO118 CD4+ T cells during primary Ag response, and they exhibit stronger recall response to secondary infection (105). Upon TCR stimulation, LLO56 CD4+ T cells induce greater ERK phosphorylation and produce larger amounts of IL-2 than do LLO118 CD4+ T cells (99). Interestingly, LLO56 CD4+ T cells express higher levels of CD5 and greater basal phosphorylation of TCRζ than do LLO118 CD4+ T cells, whereas the expression levels of many other cell-surface molecules are similar to each other (99). These results suggest that self-reactivity during thymic selection, which can be presumed from the CD5 expression levels, affects the functional difference between LLO56 and LLO118 T cells (99).
CD8+ T cells also contain CD5high and CD5low cells, which proliferate in the response to IL-2 and IL-7 robustly and poorly, respectively (106–108). A recent analysis of CD5high and CD5low subpopulations within CD44low naive CD8+ T cells further showed that CD5highCD8+ T cells expanded more effectively than did CD5low CD8+ T cells after infection (100). Additionally, TCR affinity for self-peptides was further shown to affect the function of Treg cells. The difference in TCR affinity for self-antigens during thymic selection drives the differentiation of Treg cells into distinct subsets with nonoverlapping regulatory activities (109).
These results together suggest that TCR affinity for positively selecting peptides affects immune responsiveness of both CD4+ T cells and CD8+ T cells by skewing the TCR repertoire. Note, however, that CD5 expression levels in CD4+ and CD8+ T cells are regulated not only by TCR reactivity during thymic selection but also by TCR interactions with self-peptide–MHC complexes in the periphery (110). Therefore, CD5 expression levels in T cells do not always reflect TCR affinity during positive selection in the thymus.
Using fetal thymus organ cultures of P14-TCR–transgenic thymocytes, naturally processed peptides were examined for their capability to induce positive selection (111). It was shown that two peptides, DBM and RPP, which were searched from mouse protein database on the basis of homology to antigenic lymphocytic choriomeningitis virus–derived peptides, were capable of inducing positive selection of P14-TCR–transgenic thymocytes. Interestingly, P14-TCR–transgenic CD8 SP thymocytes positively selected by the DBM peptide did not show the ability to respond to the DBM peptide, whereas RPP-selected P14-TCR–transgenic CD8 SP thymocytes retained their ability to respond to a high concentration of RPP peptides (111). These results indicate that the difference in the peptides that induce positive selection in the thymus affects the function of generated T cells.
We have recently demonstrated that TCR affinity for positively selecting peptides is regulated by the thymoproteasome and contributes to the determination of the function of CD8+ T cells (101). In mice deficient in thymoproteasomes, cTECs present an altered set of self-peptide–MHC complexes. Monoclonal TCR-transgenic CD8+ T cells, including OT-I, P14, and F5 T cells, positively selected by thymoproteasome-deficient cTECs are altered in TCR responsiveness. OT-I–TCR-transgenic CD8+ T cells generated in the thymoproteasome-deficient thymus preferentially gave rise to KLRG1highCD44high short-lived effectors upon antigenic OVA-expressing L. monocytogenes infection. Importantly, the recently TCR-signaled CD69+CCR7− stage of OT-I–TCR-transgenic DP thymocytes expressed smaller amounts of CD5 and CD69 in the thymoproteasome-deficient thymus than did the equivalent stage of OT-I–TCR-transgenic DP thymocytes generated in normal control thymus, indicating that thymoproteasome-independent positively selecting self-peptides are altered in TCR affinity. It was further shown that OT-I–TCRhigh CD8SP thymocytes positively selected by the peptides with different TCR affinity exhibited different TCR responsiveness in a positive correlation with TCR affinity for positively selecting peptides (101). These results indicate that TCR affinity for positively selecting self-peptides produced by thymic epithelium determines the subsequent Ag responsiveness of mature CD8+ T cells in the periphery, and that thymoproteasomes critically optimize TCR affinity for positively selecting self-peptides (Fig. 3).
This review summarizes current knowledge of positive selection–inducing self-peptides and highlights the Ag processing machineries in cTECs. The role of positive selection in the functional tuning of T cells is also discussed. It is now established that TCR affinity for self-peptide–MHC complexes is crucial for positive selection of T cells and that the Ag processing machineries in cTECs, including thymoproteasomes and cathepsin L, play an important yet somewhat mysterious role in optimizing the TCR affinity for self-peptide–MHC complexes for inducing positive selection. The origin of self-peptides presented by cTECs may not be limited to the genome within cTECs but may possibly extended to exogenous Ags, because previous results showed the accessibility of circulating proteins into the thymic cortex to affect the selection of DP thymocytes (112, 113). Further discoveries in the biology of cTECs are expected to advance our understanding of T cell development, a fundamental issue in immunology.
This work was supported by Ministry of Education, Culture, Sports, Science and Technology–Japan Society for the Promotion of Science Grants 23249025, 24111004, and 16H02630 (to Y.T.) and 26460576 (to K.T.).
Abbreviations used in this article:
cortical thymic epithelial cell
medullary thymic epithelial cell
thymus-specific serine protease.
The authors have no financial conflicts of interest.