1. Basel Institute for Immunology, Grenzacherstrasse 487, CH-4005 Basel, Switzerland

Correspondence to: Ed Palmer Correspondence and requests for materials should be addressed to G.W. (e-mail: Email: werlen@bii.ch) or E.P. (e-mail: Email: palmer@bii.ch ).

To investigate the role of the -CPM in mediating positive selection, we studied Rag-2^-/-mice expressing transgenic wild-type or -CPM mutant OT-1 TCRs. These receptors are referred to as wild-type and mutant, respectively. The OT-1 TCR specifically recognizes SIINFEKL, an octameric peptide from ovalbumin (OVA_257–264) presented by the class I MHC molecule, H-2K^b (ref. 4). The wild-type OT-1 thymus contains significant numbers of CD4^-CD8 ⁺ single-positive (SP) thymocytes, as expected for a class I MHC-restricted TCR (Fig. 1a and ref. 4). In contrast, thymi from mice expressing the mutant receptor contain CD4 ^-CD8^- double-negative (DN) and CD4⁺CD8 ⁺ double-positive (DP) thymocytes and very few SP cells. As observed with an -CPM mutant class II MHC-restricted TCR³, the mutant OT-1 receptor is deficient in positive selection. A defect in TCR expression cannot account for the impaired positive selection, as DP thymocytes from wild-type and mutant mice bear comparable amounts of surface receptors (Fig. 1b). In the periphery, wild-type mice have a markedly higher percentage and number of CD8⁺ T cells than do mutant mice (Fig. 1c). As with mice expressing the -CPM mutant class II MHC-restricted TCR³, the block in positive selection of DP thymocytes expressing the mutant -CPM OT-1 receptor is leaky.

Figure 1: Analysis of thymocytes and peripheral cells from B6.Rag-2^-/- wild-type and mutant -CPM mice.

a, Thymocytes were stained with monoclonal antibodies against CD4 and CD8, and analysed by flow cytometry as described³. The numbers in the dot plots indicate the percentage of CD8⁺CD4 ^- SP thymocytes. WT, wild type. b, Surface expression of wild-type (white) and mutant (black) TCRs was assessed on DP thymocytes using flow cytometry after staining cells with an anti-V2 monoclonal antibody. c, Lymph node cells were stained with monoclonal antibodies against CD8 and V2. The numbers in the dot plots indicate the percentage of CD8⁺V2 ⁺ T cells among lymph node cells.

High resolution image and legend (135K)

To study the ability of wild-type or mutant TCRs to transduce signals leading to positive or negative selection, we bred our transgenic mice to 2-microglobulin deficient mice (2m^-/-), which cannot generate CD4 ^-CD8⁺ SP thymocytes owing to the absence of class I MHC molecules⁵. Analysis of thymi from 2m^-/- animals expressing the mutant or wild-type TCR showed developmental arrest at the DP stage (Fig. 2a). Most (> 90%) thymocytes from these mice express comparable levels of surface TCR (Fig. 2b ). In fetal thymic organ cultures (FTOCs) derived from OT-1 transgenic mice, SIINFEKL induced negative selection whereas the related ligand, EIINFEKL, triggered positive selection^{4, 6}. Using 2m^-/-, Rag-2^-/-mice, we tested how the mutant thymocytes would respond to SIINFEKL or EIINFEKL by measuring the downregulation of CD4 and CD8 co-receptors in response to TCR stimulation⁷. Mutant and wild-type thymocytes responded similarly to SIINFEKL although mutant thymocytes are slightly less responsive (14-fold) at low peptide concentrations (10 aM–1 pM) (Fig. 2c). Thus, cells expressing the mutant receptor can respond to an agonist, indicating that mutant -CPM TCRs may be functional. However, mutant thymocytes were markedly less effective in responding to EIINFEKL, which required a 4,000-fold higher concentration to trigger an equivalent response to that observed in wild-type cells (Fig. 2d).

Figure 2: Analysis of CD4, CD8 and TCR expression in thymocytes from wild-type and mutant B6.2m^-/-, Rag-2^-/- mice.

a, Thymocytes were stained with monoclonal antibodies and analysed as described in Fig. 1a. b, Surface expression of wild-type and mutant TCRs was assessed as described in Fig. 1b. c, d, Analysis of CD4/CD8 downregulation. Thymocytes bearing wild-type (open circles) or mutant TCRs (filled squares) were co-cultured for 16 h with T2-K^b antigen-presenting cells (APCs) in the presence of the indicated amounts of SIINFEKL (c) or EIINFEKL (d), respectively. Cells were collected, stained and analysed for CD4 and CD8 expression as described³.

High resolution image and legend (39K)

We also analysed thymocyte subpopulations in FTOCs derived from wild-type or mutant mice. The addition of EIINFEKL to wild-type thymic lobes resulted in a marked reduction in DP thymocytes and a substantial increase in CD4 ^-CD8⁺ SP cells (Fig. 3), supporting previous studies⁴. In contrast, EIINFEKL did not significantly increase the number of SP thymocytes in mutant FTOCs, documenting the block of positive selection in vitro (Fig. 3). In the presence of SIINFEKL, wild-type and mutant FTOCs generated similar proportions and numbers of DN and CD4^-CD8⁺ SP thymocytes (Fig. 3). These SP cells are immature CD8⁺ cells that have not undergone positive selection⁶. Binding of the EIINFEKL/K^b and SIINFEKL/K^b tetramers to the mutant TCR was comparable to that seen with the wild-type receptor (data not shown), indicating that a deficiency in TCR–ligand interaction cannot account for the impaired positive selection. It seems more likely that the block in positive selection is due to impaired signalling generated on binding of EIINFEKL/K^b to the mutant TCR.

Figure 3: Analysis of thymocyte development induced by SIINFEKL or EIINFEKL in transgenic Rag-2^-/-, 2m^-/- fetal thymic organ cultures expressing wild-type or mutant TCRs.

Thymic lobes were excised from fetal mice at gestational day 15 and cultured as described⁴ in the presence of 5 g ml ^-1 human 2m and 20 M of EIINFEKL, SIINFEKL or no peptide. After 7 days of culture, thymocytes were collected, stained with monoclonal antibodies against V2, CD4 and CD8, and analysed for CD4 and CD8 expression by flow cytometry. Numbers indicate the percentage of cells in the quadrant.

High resolution image and legend (65K)

In thymocytes, the Ras/Raf/MEK/ERK cascade is thought to be involved in positive selection^{8, 9, 10, 11, 12}, whereas the MKK6/p38 pathway regulates negative selection¹³. The activation of the c-Jun-amino-terminal kinase (JNK) subfamily is relevant in peripheral T cells^{14, 15, 16, 17, 18, 19}. To investigate whether a particular MAPK cascade is preferentially activated to promote positive or negative selection, we challenged wild-type or mutant thymocytes from Rag-2^-/-, 2m^-/- mice with SIINFEKL or EIINFEKL peptides, presented by glutaraldehyde-treated T2-K ^bcells. Glutaraldehyde treatment eliminates MAPK activities from these antigen-presenting cells (APCs) (Figs 4c and 5c). Although 2 M SIINFEKL activated p38 in wild-type thymocytes (Fig. 4a), as expected for a peptide that promotes negative selection, SIINFEKL could also activate ERK and JNK. These results indicate that TCR engagement by an agonist does not discriminate between the MAPK signalling cascades. Interestingly, whereas the kinetics of JNK and p38 activation reached a plateau at 5–30 min, ERK activation by SIINFEKL was faster, peaking around 2 min and decreasing thereafter. SIINFEKL activated the MAPK cascades in mutant thymocytes with kinetics equivalent to those of wild-type cells (Fig. 4a). Comparing wild-type and mutant thymocytes, each MAPK activity showed a similar dose response to SIINFEKL stimulation (Fig. 4b). Together, these results indicate that the mutation of the -CPM domain may not affect the activation of various MAPK cascades in response to SIINFEKL.

Figure 4: Analysis of MAPK activation induced by a negative selection peptide.

a, Thymocytes were incubated for the indicated times with 2 M SIINFEKL presented by glutaraldehyde-treated T2-K^b cells. Induction of p38 and ERK phosphorylation was analysed by immunoblotting. JNK activity was measured using an immunocomplex kinase assay¹⁷. b, Thymocytes were stimulated with 0, 0.2 pM, 200 pM, 20 nM or 2 M SIINFEKL. Induction of ERK activity was analysed after 2 min; p38 and JNK activities were measured after 15 min of stimulation. c, Glutaraldehyde treatment eliminates MAPK activities from the APCs. Control and glutaraldehyde-treated T2-K^bcells were stimulated for 30 min with 2 M anisomycin, and p38 phosphorylation was analysed by immunoblotting.

High resolution image and legend (63K)

Figure 5: Analysis of MAPK activation induced by a positive selection peptide.

a, Thymocytes were stimulated with 2 M EIINFEKL for the indicated times. Activation of ERK, p38 or JNK was measured as described in Fig. 4a. Stimulation of thymocytes with 100 nM PMA was used as a positive control for ERK activation (asterisk). b, Thymocytes were stimulated with 0, 200 pM, 20 nM, 200 nM or 2 M EIINFEKL. Activation of ERK, p38 or JNK was analysed as described in Fig. 4b. c, Glutaraldehyde treatment eliminates MAPK activities from the APCs. Control or glutaraldehyde-treated T2-K^b cells were stimulated for 15 min with 100 nM PMA, and ERK phosphorylation was assessed by immunoblotting.

High resolution image and legend (57K)

Does the -CPM mutation affect the signal transduction induced by a ligand triggering positive selection? Although 2 M EIINFEKL induced sustained activation of ERK in wild-type thymocytes, no ERK activation was detected in mutant cells (Fig. 5a). Interestingly, EIINFEKL activated the p38 and JNK cascades similarly in mutant and wild-type thymocytes (Fig. 5a, b), indicating that mutation of the -CPM domain specifically affects signalling to the ERK pathway. In wild-type thymocytes both the positive- and negative-selecting peptides activate the ERK pathway, but with different kinetics ( Figs 4a and 5a). Thus, sustained ERK activation may be a prerequisite for positive selection. As p38 and JNK are equivalently activated by both peptides in wild-type and mutant thymocytes, they do not appear to be important for positive selection (Figs 4, 5), in agreement with previous work^{13, 14, 15, 16, 19}.

What components link the TCR to sustained activation of the ERK pathway? We found that all of the CD3 subunits were expressed in mutant thymocytes (data not shown). However, as we previously observed³, the mutant TCR failed to co-immunoprecipitate CD3 ( Fig. 6a), a CD3 subunit implicated in the generation of mature T lymphocytes²⁰. Unlike the wild-type TCR, the mutant receptor was unable to co-immunoprecipitate P-p21 (Fig. 6b), even though this active form of CD3 is found in mutant thymocytes (Fig. 6c, soluble fraction). Propagation of TCR-derived signals requires the formation of a DIG-associated complex containing signalling and adaptor proteins^{21, 22, 23}. Analysis of tyrosine-phosphorylated proteins revealed that in wild-type thymocytes, a phosphorylated form of Lck (P-Lck), P-p21 and a broad band with relative molecular mass 75,000 (probably comprising ZAP-70 and SLP-76) were transiently recruited to the DIGs after 15 min of EIINFEKL stimulation (Fig. 6c). Although LAT, an adaptor protein required for complete activation of ERK²⁴, was constituitively associated with the DIGs (Fig. 6d), phospho-LAT appeared in lipid rafts only after 30 min of peptide stimulation (Fig. 6c). This indicates that LAT phosphorylation by ZAP-70 and Lck may ensue after these tyrosine kinases appear in the DIG-localized signalosome. In contrast, in mutant thymocytes, neither activated Lck, P-p21 nor ZAP-70 are recruited to DIGs, which may explain why LAT is not phosphorylated despite its constitutive association with DIGs (Fig. 6c, d). The signalosome formed subsequent to EIINFEKL stimulation of mutant thymocytes is incomplete and probably accounts for the defective ERK activation. Together, our observations support a model in which a low-avidity ligand uses the -CPM and CD3 to promote recruitment of activated Lck, P-p21 and ZAP-70 to a DIG-associated signalosome that includes LAT. Phosphorylation of LAT may allow the recruitment to the lipid raft of components specifically activating ERK for a sustained period, which leads to positive selection. The activation of the JNK and p38 pathways that potentially mediate negative selection are not affected by the -CPM mutation.

Figure 6: Analysis of proximal signalling events.

Thymocytes (5  10⁷) were stimulated for the indicated times with 2 M EIINFEKL. The composition ( a) and tyrosine-phosphorylation (b) of the immunoprecipitated TCR complex were analysed by immunoblotting as described³. The position of the Ig light chain (LC) from the antibody used for immunoprecipitation is indicated. c, Thymocytes were stimulated with 2 M EIINFEKL. The Brij58-containing cell lysates were fractionated as described²³; proteins were concentrated, resolved by SDS–PAGE and analysed by immunoblotting with an anti-phosphotyrosine antibody. d, Membranes were stripped and separately reprobed with anti-ZAP-70 monoclonal antibody, anti-Lck, anti-CD3³ and anti-LAT²³ antibodies. n.d., not determined.

High resolution image and legend (38K)

There is increasing evidence that cell survival or death in a number of cellular systems depends on the strength, duration and balance of various MAPK cascades^{25, 26, 27}. In PC12 pheochromocytoma cells, transient activation of ERK induced by epidermal growth factor stimulates proliferation whereas sustained ERK activation triggered by nerve growth factor induces differentiation^{25, 26}. The MAPKs target different substrates, which might explain the selectivity of the ERK pathway in positive selection. Our results provide evidence that the TCR, by distinguishing the affinity of the ligand, modulates the composition of a DIG-associated signalling complex that controls the triggering of the ERK pathway, whereby sustained activation leads to positive selection and thymocyte survival.

References

Acknowledgements

We thank F. Carbone, M. J. Bevan and D. Kioussis for transgenic mice expressing wild-type OT-1 receptors; A. Peter for generating mutant -CPM TCR constructs; U. Müller for generation of mutant transgenic mice; T. Potter for T2-K ^b cells; M. Daniels, K. Hogquist and S. Jameson for tetramers; L. Samelson for LAT antisera; S. Stotz and T. Harder for discussions; and E. Jacinto, T. Baldari, P. Kisielow and J. Kirberg for reading the manuscript. The Basel Institute for Immunology was founded and is supported by F. Hoffmann-LaRoche Ltd.