The nature of peptides that positively select T cells in the thymus remains poorly defined. Here we report an in vivo model to study the mechanisms of positive selection of CD4+ T cells. We have restored positive selection of TCR transgenic CD4+ thymocytes, arrested at the CD4+CD8+ stage, due to the lack of the endogenously selecting peptide(s), in mice deficient for H2-M and invariant chain. A single injection of soluble agonist peptide(s) initiated positive selection of CD4+ transgenic T cells that lasted for up to 14 days. Positively selected CD4+ T cells repopulated peripheral lymphoid organs and could respond to the antigenic peptide. Furthermore, coinjection of the antagonist peptide significantly inhibited agonist-driven positive selection. Hence, contrary to the prevailing view, positive selection of CD4+ thymocytes can be induced in vivo by agonist peptides and may be a result of accumulation of signals from TCR engaged by different peptides bound to major histocompatibility complex class II molecules. We have also identified a candidate natural agonist peptide that induces positive selection of CD4+ TCR transgenic thymocytes.

T cell ontogeny is a multistep process resulting in the generation of mature peripheral CD4+ helper and CD8+ cytotoxic lymphocytes which recognize antigenic peptides presented by MHC class II or I proteins, respectively. During development of the T cell lineage thymocytes are subject to negative and positive selection processes that shape a repertoire of mature T cells that is tolerant to self-peptide–MHC complexes and at the same time diversified enough to mount an effective immune response 1,2. Selection is based on the interactions of immature CD4+CD8+ thymocytes with peptide–MHC complexes on the surface of thymic stromal cells and results in the three possible outcomes. Thymocytes will die by apoptosis from neglect if their TCRs do not interact with peptide–MHC strongly enough to transduce a signal for positive selection 3. Thymocytes, bearing receptors that interact strongly with self-peptide–MHC, die from activation-induced apoptosis 4,5. Finally, thymocytes that interact weakly with peptide–MHC undergo positive selection and survive to differentiate into single positive CD4+ or CD8+ cells 6,7. While the ability of antigenic (i.e., agonistic) peptides to induce apoptosis of CD4+CD8+ thymocytes in vivo is well established, the relationship between antigenicity of peptides and their ability to induce positive selection is controversial 8,9,10. Conflicting results concerning this issue have been obtained in a variety of in vitro systems 11,12,13,14, but until now no suitable and relatively simple system to study the identity and properties of peptides that mediate positive selection in vivo has been available. Here we describe for the first time an in vivo experimental model in which administration of soluble peptide(s) results in positive selection of mature, peripheral CD4+ T cells. The peptide ligands that initiate positive selection of CD4+ T cells in this model have agonist activity, whereas antagonist peptides only inhibited this process. Using protein sequence search algorithms we have identified an endogenous agonist peptide which also induced positive selection of transgenic T cells. This implies that natural positive selection of CD4+ thymocytes may be driven by high not only low potency ligands.

Mice.

The TCR genes were cloned from T cell hybridoma specific for analogues of the pigeon cytochrome C (PCC[43–58]) peptide and expressed in VA-hCD2 cassette 15,16,17. All TCR transgenic mice (TCRTg) were made by comicroinjection of the respective TCR-α and -β constructs into fertilized eggs of F1 (C57BL/6×CBA/Ca) mice. TCRTg mice were crossed to C57BL6/TCR α chain knockout mice (TCR-α; The Jackson Laboratory) and to mice deficient in H2-M (H2-M; provided by E. Bikoff [Harvard University, Cambridge, MA] and R. Germain [National Institutes of Health, Bethesda, MD]) and Ii (Ii; provided by L. van Kaer, Vanderbilt University, Nashville, TN) to obtain TCRTgTCR-αH2-MIi.

Flow Cytometry Analysis.

Monoclonal antibodies specific for CD4(RM4-5), CD8(53-6.7), Vβ8(F23.1), CD69(H1.2F3), CD44 (IM7), and CD62L(MEL-14) were purchased from BD PharMingen and used according to the manufacturer's recommendations. Cells were analyzed using a FACSCalibur™ instrument (Becton Dickinson) and CELLQuest™ software. For intracellular staining, cells were first stained for CD4 and CD8, fixed in 2% paraformaldehyde, permeabilized in 0.1% Tween-20, and stained with anti–bcl-2 antibody (BD PharMingen) according to the manufacturer's instruction.

Antigen Response and Antagonism Assays of TCR Transgenic Cells.

Proliferation of lymph node cells isolated from TCRTg TCR-α and TCRTgTCR-αH2-MIi mice injected with the selecting peptide was measured in response to antigen. Response of the TCRTgTCR-α lymph node cells to different agonist peptides was measured by proliferation assay in a 96-well plate. 105 responder cells were stimulated with peptides presented by 5 × 105 irradiated C57BL6 or H2-MIi splenocytes. Agonist and control peptides were used at concentrations of 0.01, 0.1, 0.5, 1.0, 2.5, 5.0, 10.0, and 20.0 μM. After 3 d cells were pulsed with 1 μCi of [3H]thymidine added to each well and after 14–16 h thymidine incorporation was measured. The sequences of peptides used were: PCC50V (AEGFSYTVANKNKGIT), PCC50L (AEGFSYTLANKNKGIT), PCC50V54A (AEGFSYTVANKAKGIT), PCC46A49A50V54A (AEGASYAVANKAKGIT), PCC50F54A (AEGFSYTFANKAKGIT), neutral ceramidase (AGFFQYTLYILASEG), and IgGVH (NADFKTPATLTVDKA). All peptides were synthesized by fluorenyl methoxycarbonyl chemistry and purified by reversed-phase high performance liquid chromatography. To compare the antigenic response of CD4+ TCRTg cells lymph node cells were collected from TCRTgTCR-α and TCRTgTCR-αH2-MIi mice 2 wk after peptide injection. Single cell suspensions were prepared and CD4+ cells were sorted using magnetic beads coated with anti-CD4 antibodies (MACS®). Purity of sorted cells exceeded 85%. 2 × 104 of purified CD4+ cells per well was used in proliferation assay. Cells were stimulated with increasing concentrations of agonist peptides presented by APCs from wild-type or H2-MIi mice. IgGVH-derived peptide was used as a negative control. Proliferation was measured after 3 d by thymidine incorporation by pulsing cells with 1 μCi of [3H]thymidine for 16 h. Antagonism assay was done by prepulsing 5 × 105 relevant APCs per well for 2 h with 5 μM of the antigenic peptide PCC50V54A, washing, and then exposing them to varying concentrations of peptide analogues with 2 × 104 purified CD4+ TCRTg lymphocytes as responder cells 18. For antagonism assays the percentage of proliferation inhibition was calculated as 100–100* (thymidine incorporation in the presence of antagonist/thymidine incorporation in the absence of antagonist; reference 18).

Peptide Injections and Surgical Procedures.

TCRTgTCR-α and TCRTgTCR-αH2-MIi were injected intraperitoneally with agonist and control peptides dissolved in PBS. Highly hydrophobic neutral ceramidase peptide was initially dissolved in DMSO followed by dilution in PBS. After indicated time mice were killed, thymus and lymph nodes were collected and single cell suspensions were prepared. Pups were injected with 2 μg of peptide subcutaneously 12–24 h after birth and thymi were transplanted into recipient mice 24 h after injection. 5–8-wk-old TCRTgTCR-αH2-MIi mice were used for thymectomy or as thymic transplant recipients. Surgical procedures were performed as described previously 19. Mice were handled accordingly with the institutional guidance.

Protein Sequence Analysis.

Protein sequence database at the National Center for Biotechnology Information was searched using BLAST, Psi-BLAST, and FASTA search algorithms with the protein motif (AF)XX(AT)(VLFI)AXX(AN) as a query.

Characterization of Peptide Agonists for the PCC-specific Transgenic αβ TCR.

To examine the role of peptides during positive selection in vivo, we generated transgenic mice expressing class II–restricted TCR specific for analogues of a PCC(43–58) peptide presented by the Ab molecule 16. This receptor recognizes analogues of PCC(43–58) in which aspartic acid in position 50 is replaced by amino acids with neutral/hydrophobic (PCC50V, PCC50V54A, PCC46A49A50V54A, PCC50L, PCC50F) side chains (Fig. 1). The TCR transgenic mice were backcrossed to C57BL6 TCR-α knockout mice so that almost all T cells expressing transgenic TCR become CD4+ T cells (Fig. 2A and Fig. B). Initially we examined the capacity of agonist peptides to induce negative selection of transgenic T cells. We found that the efficiency of negative selection correlated with the potency of individual agonist peptides. However, an injection of 20 μg of any tested agonist peptide, in particular a moderate and a strong agonist (PCC50V and PCC50V54A, respectively) did not induce negative selection of transgenic thymocytes as assessed by thymus cellularity, annexin V staining, and TUNNEL assay (Fig. 2C and Fig. D, and data not shown). Strong agonists PCC50L and PCC50V54A (Fig. 2 F) induced profound negative selection when injected at 200 μg per mouse, while moderate agonist PCC50V induced only marginal deletion of CD4+CD8+ cells (Fig. 2 E).

Peptide Agonists Induce Positive Selection of Transgenic CD4+ T Cells.

To examine the potential effect of agonist peptides on in vivo positive selection of transgenic thymocytes, we followed the ontogeny of these cells in a nonselecting thymic environment where Ab molecules are devoid of selecting peptides. For that purpose, we crossed TCRTgTCR-α mice to Ii and H2-M mice to obtain TCRTg on a triple knockout background (TCRTgTCR-αH2-MIi). The development of the majority of CD4+ thymocytes is severely impaired in mice lacking H2-M and Ii molecules, two molecular chaperones that participate in peptide loading to MHC class II molecules 20,21,22. As expected, the thymic development of the transgenic T cells was arrested at the stage of CD4+CD8+ thymocytes and only very few transgenic CD4+ T cells were detected in the periphery (Fig. 3 A). The lack of the natural positively selecting Ab–peptide complex(es) resulted in a block in thymocyte development and increased thymic cellularity in TCRTgTCR-αH2-MIi mice. After these observations, we attempted to restore positive selection in TCRTgTCR-α2H2-MIimice by providing exogenous peptides. A number of irrelevant Ab-binding peptides (IgGVH[59–74], OVA[323–339], Eα[52–68]), and analogues of PCC (50A, 50N, 50E, 52Q) without agonist properties had no effect on thymic selection (data not shown). As shown in Fig. 3B–D intraperitoneal injection of a nondeleting dose of PCC50V54A agonist peptide restored selection of CD4+ single positive thymocytes. Simultaneously, a number of CD4+CD8+ thymocytes upregulated CD69 and bcl-2 expression (Fig. 3E and Fig. F). Positive selection of CD4+ thymocytes was sustained for 14 d after a single injection of the selecting peptide ligand (Fig. 3 D). Contrary to the recent report that agonist peptides induce selection of regulatory CD4+CD25+, cells in our model newly selected CD4+ thymocytes were CD25 (data not shown, and reference 23). The cellularity of the thymus and the number of apoptotic cells detected by TUNNEL assay and annexin V staining were the same in controls and in mice injected with 20 μg of agonist peptide PCC50V54A (data not shown). Four other analogues of the PCC peptide, PCC50V, PCC50L, PCC46A49A50V54A, and PCC50F54A, injected at the same dose (20 μg), also restored positive selection of transgenic thymocytes (data not shown). These results prove that positive selection of CD4+ T cells can be induced in vivo by different agonist ligands. An injection of soluble agonist into TCRTgTCR-α→H2-MIi radiation chimeras also resulted in positive selection of transgenic CD4+ T cells, despite the expression of wild-type Ab–peptide complexes on bone marrow–derived thymic stromal cells (data not shown).

Positively Selected CD4+ Thymocytes Repopulate Peripheral Lymphoid Organs and Respond to Antigens.

To determine whether positively selected CD4+ T cells leave the thymus as functional, mature T cells, we analyzed lymph node cells 2 wk after peptide administration. As shown in Fig. 3 D, TCRTgTCR-αH2-MIimice injected with peptide had ∼20% CD4+T cells in the lymph nodes when compared with 1–2% CD4+ T cells found in control animals (Fig. 3 C). Newly selected CD4+ T cells had normal levels of TCR and CD4, and the phenotype of naive T cells (CD44CD62L+; Fig. 3 G). Also the levels of adhesion molecules CD2 (LFA-2) and CD11c (LFA-1) were the same on CD4+ lymphocytes from transgenic and peptide-injected TCRTgTCR-αH2-MIi mice (Fig. 3 H). Peptide-selected cells responded to the agonist peptide, though with much lower potency than CD4+ transgenic cells isolated from mice expressing wild-type Ab (Fig. 3 I).

Since the selecting peptide is an agonist, one could argue that the observed phenomenon results from the expansion of a small number of peripheral transgenic CD4+ T cells rather than from induced positive selection 24. To test this possibility, two types of experiments were performed. In the first experiment, the TCRTgTCR-αH2-MIi mice were thymectomized and then injected with the PCC50V54A peptide. 2 wk after injection, the mice were killed and the number of peripheral CD4+ T cells was counted. As shown in Fig. 4 A, the number of CD4 transgenic T cells was very low and did not increase in comparison with control mice that were thymectomized, but did not receive the PCC50V54A peptide. In the second experiment, TCRTgTCR-αH2-MIi neonates were injected with the PCC50V54A peptide and after 24 h thymi from injected and control neonates were transplanted under the kidney capsules of H2-MIiTCR-α mice, which are devoid of T cells 25. After 10 d, recipient mice were killed and the presence of transgenic CD4+ T cells in the transplanted thymus and host lymph nodes was determined by FACS® analysis. As shown in Fig. 4 B, transgenic CD4+ T cells were found only in mice grafted with thymi from neonates injected with the selecting peptide. Therefore, we conclude that CD4+ T cells appear in the peripheral lymph nodes of TCRTgTCR-αH2-MIi injected with the agonist peptide as a result of positive selection by this peptide.

Antagonist Peptides Do Not Induce Positive or Negative Selection but Inhibit Positive Selection Induced by Agonist Selecting Ligands.

It has been shown that the peptide component of the TCR/MHC/peptide complex influences the strength of interaction between a thymocyte and thymic stromal cell and hence determines the fate of the thymocyte 26. Minor alterations in critical amino acid residues that are exposed towards the TCR may have a profound effect on the outcome of recognition of MHC–peptide complex by the T cell and change the activation properties of a peptide ligand from agonist to an antagonist 27. Peptide ligands positively selecting CD8+ thymocytes were initially described as having antagonist properties. However, later experiments showed that CD8+ thymocytes expressing transgenic TCR were selected by peptides with different activation potencies 28,29. Similar experiments investigating thymocytes expressing MHC class II–restricted TCRs in fetal thymic organ cultures (FTOCs), showed that antagonist peptides either inhibit thymic positive selection or induce negative selection 30,31. We have investigated what is the capacity of antagonist peptides to influence thymic selection using our in vivo model. As shown in Table, we have initially found that analogues of PCC(43–58) with p5 position occupied by E or N have an antagonistic effect on our transgenic CD4+ T cells. Biological properties of agonist and antagonist peptides were preserved in both wild-type and H2-MIi mice (Fig. 5a,Fig. b,Fig. c). The responses of transgenic CD4+ T cells to agonist peptide PCC50V54A were reduced by >50% in the presence of each of these analogues, which is the criteria used to identify a given peptide as antagonist 27. These antagonist variants of PCC peptide inhibited the responses of transgenic CD4+ T cells on APCs derived from wild-type or H2-MIi mice (Fig. 5b and Fig. c). Subsequently, various doses of PCC50E and PCC50N were tested for their ability to affect thymic selection in TCRTgTCR-αH2-MIi mice. Regardless of a wide range of administered doses (up to 50 μg), injection of the antagonist peptides did not result in increased number of CD4+ transgenic T cells in the thymus nor induced deletion of CD4+CD8+ thymocytes (Fig. 6 A). In conclusion, contrary to the “weak affinity” peptide ligands that mediate positive selection of CD8+ T cells, we have found that only agonist not antagonist peptides mediate in vivo positive selection of CD4+ T cells.

A single thymocyte is likely exposed to numerous different peptides bound to MHC and it has been proposed that the sum of signals produced by these interactions determines thymocyte fate. It was shown that antagonist peptides can affect thymocyte selection, however these results were controversial. In one study antagonist peptide inhibited negative selection of thymocytes while in another study the opposite result was reported 30,31. In addition, an antagonist peptide was also described as being capable of blocking positive selection of CD4+ TCR transgenic thymocytes, although it could not be ruled out that this peptide instead induced late deletion at the stage of single positive cells 32. To examine how the antagonist peptide affects positive selection we coinjected antagonist peptide PCC50E together with the selecting agonist peptide PCC50V54A at the same low dose (3 μg), and evaluated the outcome of selection 4 d later. As shown in Fig. 6b and Fig. c the number of CD4+ thymocytes which were positively selected by PCC50V54A agonist peptide was significantly reduced. This effect was also reproduced for the second antagonist peptide PCC50N. It is unlikely that the reduced number of CD4+ thymocytes in mice injected with agonist/antagonist mixture is a result of late negative selection because a much higher dose of the same antagonists administered alone did not induce negative selection of CD4+CD8+ thymocytes. This result argues that a single thymocyte accumulates signals received by interaction with different MHC–peptide complexes.

Identification of a Candidate Natural Peptide that Mediates Positive Selection of TCRTg Thymocytes.

Since the sequence of the exogenous selecting peptide(s) has been determined, we hypothesized that the natural selecting ligand exists that has agonist properties. By testing different PCC analogues we have determined that amino acids A or F in position 46, A or T in position 49, A in position 51, and A or N in position 54 are important for binding to Ab. Moreover, amino acids V, L, F, I in position 50 were important to stimulate TCRTg lymphocytes. We have used the peptide motif (AF)XX(AT)(VLFI)AXX(AN) to search the nonredundant protein and EST protein databases using different computer algorithms. These searches resulted in identification of a mouse proteins which may encode natural peptides with agonist properties for our TCR. One of these proteins was the neutral ceramidase that contained amino acid motif FXXTLYXXA where only Y does not match the original motif 33. This protein is ubiquitously expressed in many tissues, including epithelial cells. Subsequently, we synthesized the peptide AGFFQYTLYILASEG which contained the homologous motif sequence. This peptide when tested in vitro acted as weak agonist eliciting proliferation of naive TCRTg CD4+ T cells (Fig. 7 A). Injection of 50 μg of neutral ceramidase peptide into TCRTgTCR-αH2-MIi mice resulted in positive selection of transgenic CD4+ thymocytes (Fig. 7 C) and the appearance of CD4+ lymphocytes in the peripheral lymph nodes (Fig. 7 D). These peripheral CD4+ cells weakly proliferated when stimulated with agonist peptide PCC50V (Fig. 7 B). Hence using biocomputing analysis we have been able to identify a candidate natural agonist peptide which after injection into TCRTgTCR-αH2-MIi mice induces positive selection of CD4+ T cells. Our results imply that the natural peptides bound to MHC class II that select thymocytes are recognized by the relevant TCRs with higher affinity than have been previously postulated.

Many studies have analyzed a role of peptides in positive selection of MHC class I– and II–restricted T cells. However, despite great effort these studies have yielded conflicting results on the nature of the positively selecting peptide ligand. Development of CD8+ thymocytes in FTOCs from β2-microglobulin or transporter associated with antigen processing 1 knockout mice showed that positive selection can be peptide specific and that the avidity of the TCR for the MHC–peptide complex determines the fate of immature thymocytes 34,35. The response of CD8+ thymocytes selected in FTOCs by altered peptide ligands was often compromised 10,11,36 and none of the selecting peptides identified in vitro supported positive selection of the respective transgenic receptor in vivo 37,38. The selection of CD4+ T cells in organ cultures have shown that peptide agonists either delete or select class II–restricted TCR transgenic thymocytes towards the CD8+ lineage 39. Peptide antagonists inhibited the generation of CD4+ thymocytes and promoted the selection of CD8+ cells or induced negative selection 30,31,40. Positive selection of CD4+but not CD8+ T cells was achieved using either hybrid monoclonal antibodies or monovalent antibodies against TCR and CD4 41,42. It has been reported that agonist ligands can induce differentiation of CD4+ thymocytes in reaggregation cultures or in vivo after the intrathymic injection of the recombinant adenovirus encoding the respective peptide 9,13,43. In the latter study agonist, antagonist, and even peptides with unrelated amino acid sequence were capable of selecting TCRs with defined antigenic specificity 9. In contrast, our data strongly suggest that peptides capable of selecting CD4+ T cells may have differing primary sequence, but they are required to possess agonist activity. Antagonist peptides were not able to induce positive or negative selection in wide range of tested concentrations. In addition, antagonists blocked the activity of agonists and inhibited positive selection.

The data reported in this paper suggest that positive selection is peptide specific and, in a natural environment, mediated by agonist peptide ligands. In contrast, analysis of thymocyte development in mice expressing high density of Ab–single peptide complexes implied that selection of multiple TCRs does not require specific peptide 15,44. Results obtained in both experimental models are consistent with the avidity model of positive selection which postulates that a T cell might be positively selected by multiple low affinity or few high affinity MHC–peptide complexes 6. The CD4+ T cells found in mice expressing single AbEp complex had TCRs with undetectable affinity to AbEp (to avoid negative selection), in contrast to transgenic CD4+ T cells described in this paper that are positively selected by few high affinity (agonist) Ab–peptide complexes. An alternative explanation of T cell selection in mice expressing Ab–single peptide complex postulated that most CD4+ lymphocytes are selected on contaminating, low abundance peptides 45 which is consistent with the reported here positive selection by agonist peptides. The ongoing studies on the complexity of peptides present in mice expressing various covalent MHC class II–peptide complexes in vivo should soon reveal how CD4+ T cells develop in these mice.

One of the intriguing phenomenons we have observed is the decreased proliferation of peptide selected CD4+ T cells after TCR stimulation. This effect is not peptide specific since the same cells respond poorly to anti-CD3 antibody stimulation and does not reflect lower expression of the TCR or costimulatory molecules (unpublished data). The discussed feature is also not a characteristic of our experimental system. Cells expressing the same transgenic TCR are very efficiently selected on endogenous peptides present in Ii mice and have similarly decreased response to TCR stimulation (data not shown). We are currently investigating if this phenomenon is caused by lower MHC class II expression in H2-MIi or Ii mice, the lack of the selecting peptide ligand in the periphery, decreased life span, or inherently elevated activation threshold due to the nature of the selecting peptide ligand.

Because αβ TCRs expressed on CD4+CD8+ thymocytes likely interact with multiple MHC–peptide complexes, the fate of a thymocyte may depend on the net result of all these interactions 32,46. Experiments described in this report, where coinjected antagonist peptide inhibited positive selection support this hypothesis and raise interesting questions about how the balance of agonist and antagonist ligands impacts T cell development. The doses of selecting agonist and antagonist peptides were chosen to be low to exclude the possibility that the observed phenomenon is due to competition for Ab binding. Instead, the unique signal induced by both agonist and antagonist could result in the inhibition of positive selection. Since the transition from CD4+CD8+ to CD4+ thymocytes requires 24–48 h experiments where injection of the antagonist peptide is delayed or proceeds the injection of agonist peptide may determine the course of events that lead to inhibition of positive selection by antagonist MHC–peptide ligand.

Recent experiments in reaggregation cultures raised the possibility that bone marrow dendritic cells deliver a strong agonist signal necessary to induce CD4+ T cell differentiation 47. We have investigated the development of transgenic CD4+ T cells in chimeras made by reconstituting H2-MIi mice with the bone marrow from TCRTg TCR-α mice (TCRTgTCR-α→H2-MIi chimeras). The development of thymocytes in such chimeras is arrested at the CD4+CD8+ stage despite the presence of selecting wild-type Ab–self-peptide complexes on bone marrow–derived dendritic cells. Mature CD4+ T cells appeared only after injection of the selecting peptide PCC50V54A (data not shown). This suggest that in our in vivo model, in contrast to the reaggregation cultures 47, radioresistant thymic epithelial cells are required to provide an agonist signal that induces thymocyte differentiation towards CD4+ lineage in vivo.

We have induced positive selection of transgenic thymocytes using a peptide dose that did not delete these cells in wild-type mice. The presentation of injected agonist peptides by wild-type Ab molecules present on bone marrow cells in TCRTgTCR-α→H2-MIi radiation chimeras also did not induce negative selection (data not shown). One could then postulate that agonist peptides differ in their relative capacity to induce positive versus negative selection. Some of these peptides induce negative selection even at a very low dose, but others induce positive selection at a low dose and negative selection only when used at a high dose. Under physiological conditions, it is likely that positively selecting peptides are presented at very low concentration so it seems plausible that these peptides would have agonist activity. Peptide-specific positive selection would favor interaction with a narrow range of more potent agonists while positive selection of more promiscuous TCRs may involve collective interaction with a broader range of peptides sharing lower agonist potency 16,48. Alternatively, selection of CD4+ T cells was possible in our system because the positively selecting agonist ligand sets the threshold for negative selection, and negative selection requires an interaction with the more potent agonist ligand 49,50. Signaling studies also support the notion that agonist peptides may be natural ligands for selection of the class II–restricted thymocytes 51,52, based on the observation that commitment to the CD4 lineage requires a stronger signal via TCR than commitment towards the CD8 lineage. Previously published reports describing natural peptides capable of selecting αβ TCR transgenic T cells used mass spectroscopy analysis to identify peptides eluted from MHC molecules 28,29. The advantage of this method is that it identifies peptides physically bound to MHC, but because of limited sensitivity of detection this analysis may be biased in favor of the most abundant peptides. Here, we have attempted to use an alternative method to find peptides that may be the natural selecting ligands. We have searched protein database for the presence of a sequence motif that we identified present in agonist selecting peptides. A peptide fragment derived from the identified protein had both agonist properties and positively selected transgenic thymocytes. Although, this method does not guarantee that this particular peptide is presented in the thymus, the expression pattern of neutral ceramidase suggests that this is a good candidate protein to yield selecting peptides 33,53. Our approach may become even more useful in the future as more sequences are available in the databases.

Mice transgenic for MHC class II–restricted αβ TCR and lacking H2-M and Ii molecules likely constitute a nonselecting environment for most of MHC class II–restricted transgenic receptors studied so far. We show, for the first time, that in vivo administration of soluble peptide may restore positive selection of CD4+ thymocytes in such mice. This strategy may be used to identify and determine the properties of peptides capable of selecting TCR transgenic CD4+ T cells with different antigen specificities in vivo. Since in this experimental model, the onset of positive selection of a large number of thymocytes is known, it may be particularly useful for analysis of gene expression patterns associated with positive selection and lineage commitment of CD4+ thymocytes within hours after the delivery of the positively selecting signal. Furthermore, in this study we have identified a number of selecting and nonselecting peptides so it will be possible, using soluble recombinant proteins, to determine the relationship between the affinity of the TCR for different peptide–MHC class II complexes and the capacity to induce positive selection in vivo. Our finding challenges the existing dogma that low-affinity peptides are predisposed to induce positive selection while high-affinity ligands only induce negative selection. Instead, for CD4+ T cells the latter not the former peptides seem to be predisposed to induce positive selection.

We thank Dr. Markowitz for editing the manuscript. The authors also thank G. Pacholczyk for help in producing transgenic mice, and Drs. E. Bikoff, R. Germain, D. Kioussis, and L. van Kaer for their generous gifts of knockout mice and DNA constructs. All work involving animals was conducted under protocols approved by the Animal Care and Use Committee at Medical College of Georgia.

This work was supported by National Institutes of Health basic research grants AI41145-01A1 and HD36302-02 to L. Ignatowicz. P. Kisielow is supported by Howard Hughes Medical Institute.

Kisielow
P.
,
von Boehmer
H.
Development and selection of T cellsfacts and puzzles
Adv. Immunol.
58
1995
87
209
[PubMed]
Sebzda
E.
,
Mariathasan
S.
,
Ohteki
T.
,
Jones
R.
,
Bachmann
M.F.
,
Ohashi
P.S.
Selection of the T cell repertoire
Annu. Rev. Immunol.
17
1999
829
874
[PubMed]
Surh
C.D.
,
Sprent
J.
T-cell apoptosis detected in situ during positive and negative selection in the thymus
Nature.
372
1994
100
103
[PubMed]
Kisielow
P.
,
Bluthmann
H.
,
Staerz
U.D.
,
Steinmetz
M.
,
von Boehmer
H.
Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes
Nature.
333
1988
742
746
[PubMed]
Sha
W.C.
,
Nelson
C.A.
,
Newberry
R.D.
,
Kranz
D.M.
,
Russell
J.H.
,
Loh
D.Y.
Positive and negative selection of an antigen receptor on T cells in transgenic mice
Nature.
336
1988
73
76
[PubMed]
Ashton Rickardt
P.G.
,
Bandeira
A.
,
Delaney
J.R.
,
Van
K.L.
,
Pircher
H.P.
,
Zinkernagel
R.M.
,
Tonegawa
S.
Evidence for a differential avidity model of T cell selection in the thymus
Cell.
76
1994
651
663
[PubMed]
Hogquist
K.A.
,
Jameson
S.C.
,
Bevan
M.J.
The ligand for positive selection of T lymphocytes in the thymus
Curr. Opin. Immunol.
6
1994
273
278
[PubMed]
Liblau
R.S.
,
Tisch
R.
,
Shokat
K.
,
Yang
X.
,
Dumont
N.
,
Goodnow
C.C.
,
McDevitt
H.O.
Intravenous injection of soluble antigen induces thymic and peripheral T-cell apoptosis
Proc. Natl. Acad. Sci. USA.
93
1996
3031
3036
[PubMed]
Nakano
N.
,
Rooke
R.
,
Benoist
C.
,
Mathis
D.
Positive selection of T cells induced by viral delivery of neopeptides to the thymus
Science.
275
1997
678
683
[PubMed]
Jameson
S.C.
,
Hogquist
K.A.
,
Bevan
M.J.
Specificity and flexibility in thymic selection
Nature.
369
1994
750
752
[PubMed]
Hogquist
K.A.
,
Jameson
S.C.
,
Bevan
M.J.
Strong agonist ligands for the T cell receptor do not mediate positive selection of functional CD8+ T cells
Immunity.
3
1995
79
86
[PubMed]
Berg
R.E.
,
Irion
S.
,
Kattman
S.
,
Princiotta
M.F.
,
Staerz
U.D.
A physiological ligand of positive selection is recognized as a weak agonist
J. Immunol.
165
2000
4209
4216
[PubMed]
Yasutomo
K.
,
Doyle
C.
,
Miele
L.
,
Germain
R.N.
The duration of antigen receptor signalling determines CD4+ versus CD8+ T-cell lineage fate
Nature.
404
2000
506
510
[PubMed]
Wang
F.
,
Huang
C.Y.
,
Kanagawa
O.
Rapid deletion of rearranged T cell antigen receptor (TCR) Vα-Jα segment by secondary rearrangement in the thymusrole of continuous rearrangement of TCR α chain gene and positive selection in the T cell repertoire formation
Proc. Natl. Acad. Sci. USA.
95
1998
11834
11839
[PubMed]
Ignatowicz
L.
,
Rees
W.
,
Pacholczyk
R.
,
Ignatowicz
H.
,
Kushnir
E.
,
Kappler
J.
,
Marrack
P.
T cells can be activated by peptides that are unrelated in sequence to their selecting peptide
Immunity.
7
1997
179
186
[PubMed]
Kraj
P.
,
Pacholczyk
R.
,
Ignatowicz
L.
α/β TCRs differ in the degree of their specificity for the positively selecting MHC/peptide ligand
J. Immunol.
166
2001
2251
2259
[PubMed]
Zhumabekov
T.
,
Corbella
P.
,
Tolaini
M.
,
Kioussis
D.
Improved version of a human CD2 minigene based vector for T cell-specific expression in transgenic mice
J. Immunol. Meth.
185
1995
133
140
Lyons
D.S.
,
Lieberman
S.A.
,
Hampl
J.
,
Boniface
J.J.
,
Chien
Y.
,
Berg
L.J.
,
Davis
M.M.
A TCR binds to antagonist ligands with lower affinities and faster dissociation rates than to agonists
Immunity.
5
1996
53
61
[PubMed]
Coligan, J.E., A.M. Kruisbeek, D.H. Margulies, E.M. Shevach, and W. Strober. 1997. Current Protocols in Immunology. J.P. Reeves and P.A. Reeves, editors. Wiley, New York. 1–11.
Tourne
S.
,
Miyazaki
T.
,
Wolf
P.
,
Ploegh
H.
,
Benoist
C.
,
Mathis
D.
Functionality of major histocompatibility complex class II molecules in mice doubly deficient for invariant chain and H-2M complexes
Proc. Natl. Acad. Sci. USA.
94
1997
9255
9260
[PubMed]
Kenty
G.
,
Martin
W.D.
,
Van Kaer
L.
,
Bikoff
E.K.
MHC class II expression in double mutant mice lacking invariant chain and DM functions
J. Immunol.
160
1998
606
614
[PubMed]
Kovats
S.
,
Grubin
C.E.
,
Eastman
S.
,
deRoos
P.
,
Dongre
A.
,
Van Kaer
L.
,
Rudensky
A.Y.
Invariant chain-independent function of H-2M in the formation of endogenous peptide-major histocompatibility complex class II complexes in vivo
J. Exp. Med.
187
1998
245
251
[PubMed]
Jordan
M.S.
,
Boesteanu
A.
,
Reed
A.J.
,
Petrone
A.L.
,
Holenbeck
A.E.
,
Lerman
M.A.
,
Naji
A.
,
Caton
A.J.
Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide
Nat. Immunol.
2
2001
301
306
[PubMed]
Martin
S.
,
Bevan
M.J.
Antigen-specific and nonspecific deletion of immature cortical thymocytes caused by antigen injection
Eur. J. Immunol.
27
1997
2726
2736
[PubMed]
Takeda
S.
,
Rodewald
H.R.
,
Arakawa
H.
,
Bluethmann
H.
,
Shimizu
T.
MHC class II molecules are not required for survival of newly generated CD4+ T cells, but affect their long-term life span
Immunity.
5
1996
217
228
[PubMed]
Williams
O.
,
Tanaka
Y.
,
Tarazona
R.
,
Kioussis
D.
The agonist-antagonist balance in positive selection
Immunol. Today.
18
1997
121
126
[PubMed]
De Magistris
M.T.
,
Alexander
J.
,
Coggeshall
M.
,
Altman
A.
,
Gaeta
F.C.
,
Grey
H.M.
,
Sette
A.
Antigen analog-major histocompatibility complexes act as antagonists of the T cell receptor
Cell.
68
1992
625
634
[PubMed]
Hogquist
K.A.
,
Tomlinson
A.J.
,
Kieper
W.C.
,
McGargill
M.A.
,
Hart
M.C.
,
Naylor
S.
,
Jameson
S.C.
Identification of a naturally occurring ligand for thymic positive selection
Immunity.
6
1997
389
399
[PubMed]
Hu
Q.
,
Bazemore Walker
C.R.
,
Girao
C.
,
Opferman
J.T.
,
Sun
J.
,
Shabanowitz
J.
,
Hunt
D.F.
,
Ashton-Rickardt
P.G.
Specific recognition of thymic self-peptides induces the positive selection of cytotoxic T lymphocytes
Immunity.
7
1997
221
231
[PubMed]
Page
D.M.
,
Alexander
J.
,
Snoke
K.
,
Appella
E.
,
Sette
A.
,
Hedrick
S.M.
,
Grey
H.M.
Negative selection of CD4+CD8+ thymocytes by T-cell receptor peptide antagonists
Proc. Natl. Acad. Sci. USA.
91
1994
4057
4061
[PubMed]
Spain
L.M.
,
Jorgensen
J.L.
,
Davis
M.M.
,
Berg
L.J.
A peptide antigen antagonist prevents the differentiation of T cell receptor transgenic thymocytes
J. Immunol.
152
1994
1709
1717
[PubMed]
Williams
O.
,
Tarazona
R.
,
Wack
A.
,
Harker
N.
,
Roderick
K.
,
Kioussis
D.
Interactions with multiple peptide ligands determine the fate of developing thymocytes
Proc. Natl. Acad. Sci. USA.
95
1998
5706
5711
[PubMed]
Tani
M.
,
Okino
N.
,
Mori
K.
,
Tanigawa
T.
,
Izu
H.
,
Ito
M.
Molecular cloning of the full-length cDNA encoding mouse neutral ceramidase. A novel but highly conserved gene family of neutral/alkaline ceramidases
J. Biol. Chem.
275
2000
11229
11234
[PubMed]
Bevan
M.J.
In thymic selection, peptide diversity gives and takes away
Immunity.
7
1997
175
178
[PubMed]
Jameson
S.C.
,
Hogquist
K.A.
,
Bevan
M.J.
Positive selection of thymocytes
Annu. Rev. Immunol.
13
1995
93
126
[PubMed]
Sebzda
E.
,
Kundig
T.M.
,
Thomson
C.T.
,
Aoki
K.
,
Mak
S.Y.
,
Mayer
J.P.
,
Zamborelli
T.
,
Nathenson
S.G.
,
Ohashi
P.S.
Mature T cell reactivity altered by peptide agonist that induces positive selection
J. Exp. Med.
183
1996
1093
1104
[PubMed]
Goldrath
A.W.
,
Bevan
M.J.
Low-affinity ligands for the TCR drive proliferation of mature CD8+ T cells in lymphopenic hosts
Immunity.
11
1999
183
190
[PubMed]
Levelt
C.N.
,
Mizoguchi
E.
,
Huang
X.
,
Zacks
R.
,
Bhan
A.K.
,
Tonegawa
S.
Inhibition of intrathymic T cell development by expression of a transgenic antagonist peptide
Proc. Natl. Acad. Sci. USA.
95
1998
14349
14354
[PubMed]
Kersh
G.J.
,
Engle
D.L.
,
Williams
C.B.
,
Allen
P.M.
Ligand-specific selection of MHC class II-restricted thymocytes in fetal thymic organ culture
J. Immunol.
164
2000
5675
5682
[PubMed]
Volkmann
A.
,
Barthlott
T.
,
Weiss
S.
,
Frank
R.
,
Stockinger
B.
Antagonist peptide selects thymocytes expressing a class II major histocompatibility complex-restricted T cell receptor into the CD8 lineage
J. Exp. Med.
188
1998
1083
1089
[PubMed]
Takahama
Y.
,
Suzuki
H.
,
Katz
K.S.
,
Grusby
M.J.
,
Singer
A.
Positive selection of CD4+ T cells by TCR ligation without aggregation even in the absence of MHC
Nature.
371
1994
67
70
[PubMed]
Schmitt
S.
,
Muller
K.P.
,
Kyewski
B.A.
Two separable T cell receptor signals reconstitute positive selection of CD4 lineage T cells in vivo
Eur. J. Immunol.
27
1997
2139
2144
[PubMed]
Wang
R.
,
Nelson
A.
,
Kimachi
K.
,
Grey
H.M.
,
Farr
A.G.
The role of peptides in thymic positive selection of class II major histocompatibility complex-restricted T cells
Proc. Natl. Acad. Sci. USA.
95
1998
3804
3809
[PubMed]
Ignatowicz
L.
,
Kappler
J.
,
Marrack
P.
The repertoire of T cells shaped by a single MHC/peptide ligand
Cell.
84
1996
521
529
[PubMed]
Barton
G.M.
,
Rudensky
A.Y.
Requirement for diverse, low-abundance peptides in positive selection of T cells
Science.
283
1999
67
70
[PubMed]
Kisielow
P.
,
Miazek
A.
Positive selection of T cellsrescue from programmed cell death and differentiation require continual engagement of the T cell receptor
J. Exp. Med.
181
1995
1975
1984
[PubMed]
Yasutomo
K.
,
Lucas
B.
,
Germain
R.N.
TCR signaling for initiation and completion of thymocyte positive selection has distinct requirements for ligand quality and presenting cell type
J. Immunol.
165
2000
3015
3022
[PubMed]
Fukui
Y.
,
Oono
T.
,
Cabaniols
J.P.
,
Nakao
K.
,
Hirokawa
K.
,
Inayoshi
A.
,
Sanui
T.
,
Kanellopoulos
J.
,
Iwata
E.
,
Noda
M.
Diversity of T cell repertoire shaped by a single peptide ligand is critically affected by its amino acid residue at a T cell receptor contact
Proc. Natl. Acad. Sci. USA.
97
2000
13760
13765
[PubMed]
Basu
D.
,
Williams
C.B.
,
Allen
P.M.
In vivo antagonism of a T cell response by an endogenously expressed ligand
Proc. Natl. Acad. Sci. USA.
95
1998
14332
14336
[PubMed]
Grossman
Z.
,
Singer
A.
Tuning of activation thresholds explains flexibility in the selection and development of T cells in the thymus
Proc. Natl. Acad. Sci. USA.
93
1996
14747
14752
[PubMed]
Albert Basson
M.
,
Bommhardt
U.
,
Cole
M.S.
,
Tso
J.Y.
,
Zamoyska
R.
CD3 ligation on immature thymocytes generates antagonist-like signals appropriate for CD8 lineage commitment, independently of T cell receptor specificity
J. Exp. Med.
187
1998
1249
1260
[PubMed]
Hernandez-Hoyos
G.
,
Sohn
S.J.
,
Rothenberg
E.V.
,
Alberola-Ila
J.
Lck activity controls CD4/CD8 T cell lineage commitment
Immunity.
12
2000
313
322
[PubMed]
Klein
L.
,
Kyewski
B.
Self-antigen presentation by thymic stromal cellsa subtle division of labor
Curr. Opin. Immunol.
12
2000
179
186
[PubMed]

Abbreviations used in this paper: FTOC, fetal thymic organ culture; PCC, pigeon cytochrome C; TCRTg, TCR transgenic mice.