SIT and TRIM Determine T Cell Fate in the Thymus¹ Free

Selection of the TCR repertoire within the thymus is a tightly regulated process that ensures development of T cell clones possessing a vigorous reactivity against foreign Ags and at the same time avoids the generation of self-reactive cells. Decisions made in the thymus are mainly dictated by signals transduced via the TCR upon binding MHC-self peptide complexes. An established model postulates that differences in the avidity/affinity are translated into distinct signaling strengths that in turn determine the developmental outcome. According to this model, strong interactions lead to apoptosis of autoreactive clones (negative selection), and intermediate or weak associations promote thymocyte survival and maturation (positive selection), whereas minimal or absent binding induces cell death (death by neglect, nonselection) (1).

The data also suggest that low-intensity signals transduced via the TCR activate the Ca²⁺/calcineurin/NFAT and Ras/MEK/ERK pathways and lead to positive selection, whereas high-intensity signals leading to negative selection may also activate Grb2/Sos, JNK, p38, and PKC (2, 3).

Transmembrane adaptor proteins (TRAPs)³ are integral membrane molecules participating in proximal signaling events downstream of Ag receptors (4). From their position at the plasma membrane, where signals transduced through the Ag receptor are sorted and further propagated into the cell, TRAPs may play an important role in fine-tuning TCR-mediated signal strength required to regulate positive and/or negative selection. However, with the exception of LAT (linker for activation of T cells) (5), little is known about the role of TRAPs in thymocyte development.

Previously, we showed that the Src homology domain containing tyrosine phosphatase 2-interacting transmembrane adaptor protein (SIT) regulates thymic development by setting the signaling threshold for positive selection (6), whereas mice lacking the TCR-interacting molecule (TRIM) showed normal T cell development and function (7). SIT and TRIM are structurally strongly related proteins. Indeed, they both are non-raft-associated homodimeric transmembrane adaptors and both possess several tyrosine-based signaling motifs (TBSMs) within their cytoplasmic domains, two of which are highly conserved between the two molecules (YGNL and YASV in SIT, and YGNL and YASL in TRIM) (8, 9). Interestingly, we have previously reported that the ability of SIT to modulate TCR-mediated signaling in the Jurkat T cell line is determined by these two TBSMs (10). Additionally, SIT and TRIM are both strongly expressed in thymocytes. Based on these similarities and the mild phenotype of the single knockout animals, we proposed that SIT and TRIM adaptors can largely compensate for each other when one of the molecules is lost. To assess this idea, we generated SIT/TRIM double-deficient (DKO) mice. Our data show that SIT and TRIM are indeed functionally redundant during thymic selection. We demonstrate that concomitant loss of SIT and TRIM results in an enhanced MAPK activation and in the conversion of selection processes from nonselection to positive selection or from positive selection to negative selection. Collectively, our studies identified SIT and TRIM as critical regulators of central tolerance.

SIT^−/− (6) and TRIM^−/− (7) mice were intercrossed to generate homozygous DKO mice. DKO mice were further interbred with TCR transgenic mice. OT-I TCR transgenic mice were kindly provided by Dr. Percy Knolle, P14 by Dr. Thomas Kammerthoens, HY TCR transgenic mice by Dr. Gary Koreztky, and F5/β₂-microglobulin (β₂m)⁰/RAG1⁰ by Dr. Dimitris Kioussis. OT-I (11), P14 (12, 13), HY (14), and F5/β₂m⁰/RAG1⁰ (15, 16, 17) were previously described. All experiments involving mice were performed according to the guidelines of the State of Sachsen-Anhalt.

Single-cell suspensions were prepared, stained, and analyzed on a FACSCalibur using the CellQuest software (BD Biosciences) as previously reported (6). The Abs for phospho-ERK1/2, phospho-JNK, and phospho-p38 were purchased from Cell Signaling Technology. Intracellular staining was performed as previously described (6).

Lymph node cells (1 × 10⁵ cells/well) were cultured in RPMI 1640 medium (supplemented with 10% FCS, antibiotics, 2-ME) in U-bottom 96-well plates (Costar) in the presence of plate-bound anti-mouse CD3ε (10 μg/ml, 145-2C11; BD Biosciences), anti-mouse TCRβ (10 μg/ml, H57-597; BD Biosciences), with Con A (2 μg/ml; Calbiochem) or in the presence of 2 × 10⁵ irradiated T cell-depleted splenocytes from wild-type mice loaded with the agonist peptide NP68. Cells were pulsed with 1 μCi [³H]thymidine per well during the last 8 h and harvested as previously described (6).

Statistical analyses were performed using GraphPad Prism software. Mean values ± SEM are shown in each graph. Asterisks represent p-values of an unpaired two-tailed Student’s t test.

Previously, we showed that loss of SIT resulted in a partial alteration of thymic selection (6), whereas the loss of TRIM had no effect upon T cell development (7). SIT^−/−CD4⁺CD8⁺ (double-positive, DP) thymocytes expressed higher levels of CD5 and CD69 and displayed a partial conversion from positive to negative selection in TCR transgenic mice. To explore the possibility whether SIT and TRIM, which appear to be structurally related, are functionally redundant during thymocyte selection, we bred SIT^−/− to TRIM^−/− mice to generate DKO mice. We initially analyzed the activation markers CD5 and CD69, whose expression was up-regulated in DP thymocytes from SIT-deficient mice (6). As shown in Fig. 1,A, CD5 and CD69 expression was notably augmented on DP thymocytes lacking both SIT and TRIM when compared with SIT-deficient mice. Moreover, DP cells from DKO showed a marked up-regulation of TCRβ levels that was not observed in either SIT^−/− or TRIM^−/− mice (Fig. 1 A). Taken together, these data suggest that SIT and TRIM may compensate for one another in lymphocyte development.

It is well established that the levels of TCRβ, CD5, and CD69 are strictly regulated during selection processes and that their expression reflects the intensity of the signal transduced via the TCR. Indeed, mice lacking negative regulators of TCR-mediated signaling such as c-Cbl (18) and SLAP (Src-like adaptor protein) (19) showed up-regulation of TCRβ, CD5, and CD69, whereas knockout mice for positive regulators of intracellular signaling pathways such as Tec-family kinases (20) or calcineurin B1 (21) show reduced levels of TCR, CD5, and CD69. Thus, on the basis of our observations, it appears as if SIT and TRIM both function as negative regulators of TCR signaling strength.

Upon TCR/MHC + self peptide interaction, immature DP thymocytes differentiate into mature CD4⁺ or CD8⁺ SP (single-positive) cells through a complex and ordered process that includes 1) up-regulation of TCRβ, CD5, and CD69, 2) down-regulation of one of the coreceptors, and 3) final commitment to either the CD4 or CD8 lineage (22, 23, 24).

To further investigate the effects of SIT and TRIM deficiency upon T cell development, we stained thymocyte suspensions with CD4 and CD8 mAbs. As presented in Fig. 1,B, we found that DKO mice showed a striking difference in the distribution of thymocyte subsets in comparison to wild-type or single knockout mice. In fact, we observed an increase in the proportion of CD4⁺ cells expressing low level of CD8 in DKO mice (Fig. 1,B). These cells likely represent an intermediate subset of DP thymocytes that upon receiving a signal through the TCR is committed to the SP stage. In agreement with this assumption, we also found that DKO mice display a 2-fold increase in the absolute number of SP thymocytes when compared with the other genotypes (Fig. 1 C). Collectively, our data suggest that in the absence of both SIT and TRIM TCR-mediated signaling is considerably strengthen in DP thymocytes, resulting in a strong up-regulation of markers for positive selection and in an augmented commitment of DP thymocytes to the SP stage.

The number of early T cell precursors, defined by the lack of both CD4 and CD8 expression (double-negative (DN) cells) (Fig. 1 C) and the differential expression of CD25 and CD44 (data not shown), are not affected in DKO mice. These data corroborate our previous observation that both SIT and TRIM are dispensable during early stages of T cell development (6, 7).

It is well established that the activation of the Ras-Raf-MEK-ERK pathway is critical during positive selection (2). Therefore, we next assessed whether SIT and TRIM regulate ERK activation in thymocytes, by performing intracellular flow cytometric analyses using an Ab directed against the phosphorylated (and hence activated) forms of ERK1 and ERK2 (25). We analyzed DP thymocytes expressing low CD5 levels so that ERK activation can be solely attributed to cells that, upon TCR-MHC-self peptide engagement, are initiating thymic selection and not to postselected DP thymocytes that express CD5 at higher levels. To determine whether the magnitude of ERK phosphorylation directly correlates with positive selection, thymocyte suspensions prepared from mice of each genotype were rapidly fixed in formaldehyde to preserve the phosphorylation status of ERK. As shown in Fig. 2 and in agreement with previous observations (26), no significant ERK activity was detected in DP cells from wild-type or even single mutant mice, thus indicating that relatively few thymocytes initiate positive selection in normal mice. Conversely, we found that there was considerably more ERK phosphorylation in DP thymocytes from DKO mice (Fig. 2). These data show that the enhanced positive selection in DKO thymi correlate with an enhanced ex vivo ERK phosphorylation. Importantly, we previously showed that the more efficient positive selection/shift to negative selection observed in SIT^−/− HY TCR transgenic female mice also correlated with stronger ex vivo ERK activity in DP thymocytes (6). Collectively, these data indicate that SIT and TRIM regulate thymic selection by modulating the degree of ERK activation.

Fig. 2 also shows that DKO DP thymocytes display significantly higher levels of phosphorylated p38 and JNK when compared with single mutant and wild-type mice. Studies performed using knockout mice or specific MAPK inhibitors had suggested that JNK and p38 may be important during negative selection (2). On the basis of these observations, higher JNK and p38 activities should result in the conversion from positive to negative selection in DKO thymi. This hypothesis is also supported by our previous observation that female HY TCR transgenic SIT^−/− mice show a partial conversion from positive to negative selection (6).

To analyze whether loss of both SIT and TRIM resulted in an enhanced conversion from positive to negative selection, we interbred DKO mice with mice expressing transgenic TCRs. Fig. 3,A shows that the cellular pattern of positive selection in female HY DKO mice is dramatically altered and noticeably resembles that observed in HY TCR male mice where negative selection occurs (Fig. 3,A, upper right panel). Indeed, thymi of DKO HY TCR transgenic female mice are completely devoid of DP cells, contain few CD8 SP thymocytes expressing lower levels of CD8 than normal HY TCR transgenic female mice, and are comprised of almost exclusively DN thymocytes (Fig. 3,A). One result of negative selection in male HY mice is the overall reduction of thymic cellularity. Comparison of total cell numbers revealed another striking similarity between DKO HY female and HY male mice. As shown in Fig. 3 B, DKO thymi are indeed markedly hypocellular like those of male HY mice.

These results show that the concomitant deletion of SIT and TRIM results in the complete conversion from positive to negative selection of HY TCR⁺ thymocytes in female mice.

For some time, it was thought that elimination of autoreactive T cells by deletion was the main pathway for the maintenance of self-tolerance. More recent data, however, have revised this paradigm and suggested that nondeletional pathways also contribute to tolerance. For example, autoreactive T cells may escape deletion by down-regulating the coreceptor (27, 28, 29, 30). In this way, autoreactive T cells reprogram the activation threshold and become unresponsive to self Ags. Coreceptor down-regulation has been described to take place on both thymocytes (31, 32) and peripheral mature T cells (30, 33, 34, 35) and allows the generation of large numbers of coreceptor low or DN peripheral T cells. The observation that CD4⁻CD8^low and CD4⁻CD8⁻ TCR⁺ peripheral T cells are abundant in mice carrying transgenic TCRs on negatively selecting genetic background (i.e., HY TCR transgenic male mice) reinforces the idea that coreceptor down-regulation is required to maintain tolerance (14, 36, 37). However, the physiological relevance of these cells in TCR transgenic mice has been questioned due to the premature expression of the αβ TCR (38).

The data shown in Fig. 3,A indicate that the level of CD8 is reduced on thymocytes from DKO HY female mice. To investigate whether mechanisms alternative to deletion are set in motion to reprogram the activation threshold and maintain tolerance in DKO mice, we investigated CD8 expression levels in peripheral T cells. Fig. 3,C shows that, while HY SIT^−/− female mice display a partial down-regulation of CD8, HY DKO female mice exhibit a striking down-regulation that is even more pronounced than that seen in male HY mice (Fig. 3,C, compare lower left and upper right dot plots), and DKO females also accumulate large numbers of T3.70⁺ DN cells (Fig. 3 D). These data are in agreement with our recent observation demonstrating that the loss of SIT markedly enhanced the homeostatic expansion of peripheral T3.70⁺ T cells from HY transgenic female mice (39).

Thus, the loss of both SIT and TRIM lowers the signaling threshold in T cells and results in a massive deletion of DP cells in the thymus. Therefore, T3.70⁺ T cells reprogram their activation threshold by down-regulating CD8 expression in response to a persistent strong signal.

To further corroborate the effects of SIT and TRIM on positive selection, we used two additional TCR transgenic models expressing either the P14 or OT-I TCRs (Fig. 4). Previously, we showed that TRIM deficiency did not affect selection processes in either P14 or OT-I TCR transgenic thymocytes (7), whereas the loss of SIT modestly affected positive selection only in the P14 TCR transgenic mice (6). In agreement with our previous observations (6), the results in Fig. 4, A and E, show that P14 SIT^−/− mice show a modest reduction in the number of total Vα2⁺ and positively selected Vα2^high thymocytes. Remarkably, DKO mice display a stronger decrease in the proportion and absolute numbers of thymocytes expressing the transgenic TCRα-chain (Fig. 4, A and E). In contrast, the number of nontransgenic Vα2⁻ thymocytes are not decreased but rather slightly increased, indicating that these cells may undergo expansion upon deletion of transgenic thymocytes. The decrease in the total number of clonotypic TCR⁺ thymocytes appears to be due to a reduction of both DP and CD8 SP thymocytes (Fig. 4,B and Table I), thus indicating that, similar to the HY TCR transgenic model, positive selection is likely converted to negative selection in the absence of SIT and TRIM also in the P14 system. To further corroborate this hypothesis is the observation that thymocytes from DKO mice are hyperactivated as they up-regulate both CD5 and CD69 (Fig. 4,C). In agreement with an altered positive selection, the number of peripheral mature Vα2⁺ T cells in lymph nodes of DKO mice are also markedly reduced in comparison to control or SIT-deficient mice (Fig. 4, D and F). These data suggest that the loss of SIT and TRIM results in an enhanced TCR-mediated signaling that converts positive to negative selection.

Subsequently, we confirmed these results and analyzed selection processes in the OT-I transgenic mice. The data presented in Fig. 4,G–L demonstrate that loss of SIT and TRIM severely affects positive selection also in this TCR transgenic model. Indeed, DKO mice showed an enhanced TCR-mediated signaling in DP cells, as suggested by the higher levels of CD5 and CD69 expression (Fig. 4,I), a marked decrease in the number of total thymocytes, total transgenic Vα2⁺ cells, and positively selected Vα2^high thymocytes (Fig. 4, G and K). Moreover, DP and CD8 SP thymocytes are also markedly depleted in DKO mice (Fig. 4,H and Table I). Finally, a reduction in mature Vα2^high lymph node CD8⁺ T cells is also observed in DKO mice (Fig. 4, J and L). In agreement with our previous observations (6), the loss of SIT had only minimal affects upon positive selection. In fact, the total number of positively selected Vα2^high thymocytes and peripheral CD8⁺ T cells as well as CD5 and CD69 expression are grossly unchanged in SIT^−/− mice. Interestingly, both P14 and OT-I DKO mice down-regulate the CD8 coreceptor on peripheral TCR⁺ cells (Fig. 4, D and J, right panels). Thus, in the absence of SIT and TRIM, CD8⁺ T cells from TCR transgenic mice receive constantly stronger signals via the TCR that trigger coreceptor down-regulation.

Collectively, the data obtained from three different class I-restricted TCR transgenic lines suggest that the loss of SIT and TRIM enhances TCR-mediated signaling and converts positive into negative selection.

Conversely, when we investigated the MHC class II-restricted OT-II TCR transgenic model we did not observe any effect on positive selection (data not shown).

Finally, we investigated whether SIT and TRIM also affect negative selection. However, we found no differences in the CD4/CD8 profile or total cellularity of DKO thymocytes from male HY TCR transgenic mice (data not shown). Moreover, in vitro CD3-mediated deletion of DKO DP cells was normal (data not shown), thus suggesting that SIT and TRIM do not regulate negative selection or thymocyte survival.

In summary, these data demonstrate that SIT and TRIM cooperatively set the TCR signaling threshold for positive selection.

The data above indicate that SIT and TRIM are critical regulators of T cell development. In all models tested, we found that positive selection is either enhanced or converted to negative selection. The enhanced MAPK activation, together with the up-regulation of CD5 and CD69, is consistent with the idea that TCR-mediated signaling is stronger in DKO mice. Therefore, we next investigated whether the loss of SIT and TRIM enables positive selection of T cells under experimental conditions in which TCR avidity is abrogated and, thus, positive selection should not occur. As a model system, we used β₂m-deficient/RAG1-deficient mice expressing the class I-restricted F5 transgenic TCR (F5/β₂m⁰/RAG1⁰) (17). In these mice, T cell development is blocked at the DP stage and thymic selection does not occur, as class I MHC heavy chains are expressed at extremely low levels.

To evaluate whether the loss of SIT and TRIM may alter cell fate by diverting thymocytes from nonselection to positive selection, DKO mice were bred onto F5/β₂m⁰/RAG1⁰ mice. As shown in Fig. 5,A, neither the loss of SIT or TRIM alone induced the generation of mature SP thymocytes. Indeed, the CD4/CD8 profiles of single knockout mice resemble those of control F5/β₂m⁰/RAG1⁰ mice. Strikingly, however, significant numbers of CD8⁺ SP thymocytes were found in F5/β₂m⁰/RAG1⁰ mice lacking both SIT and TRIM (Fig. 5, A and B). Analysis of the CD5 and CD69 expression levels (Fig. 5,A) indicated that positive selection in DKO mice is likely initiated by TCR-mediated signaling, as the expression of both molecules is induced on DP thymocytes from DKO mice. Even more surprisingly, we also detected mature CD8⁺ T cells in peripheral secondary lymphoid organs of DKO/F5/β₂m⁰/RAG1⁰ mice (Fig. 5 C). However, the efficiency of positive selection remained very low, with <10% of the number of CD8⁺ T cells recovered from the lymph nodes from DKO/F5/β₂m⁰/RAG1⁰ mice (0.244 × 10⁶ ± 0.021 × 10⁶; n = 3 mice) as compared with F5/β₂m^+/−/RAG1^+/− mice (3.610 × 10⁶ ± 0.215 × 10⁶; n = 3 mice).

Therefore, it appears that the loss of SIT and TRIM likely lowers the TCR-mediated signaling threshold for positive selection, thereby enabling the maturation of T cells by TCR/low-avidity ligand-driven signals.

To test whether the loss of SIT and TRIM might have resulted in the spontaneous development of T cells in a TCR-independent manner, we investigated DKO RAG-deficient mice. RAG1⁰ and DKO/RAG1⁰ mice both displayed a developmental block at the DN3 stage (data not shown), thus reinforcing the idea that the loss of SIT and TRIM alone does not result in the autonomous maturation of T cells.

Unexpectedly, DKO/F5/β₂m⁰/RAG1⁰ mice displayed CD4⁺ SP thymocytes (Fig. 5,A) and peripheral CD4⁺ T cells (Fig. 5,C). Interestingly, also DKO mice expressing the MHC class I-restricted P14 TCR showed an increase in the number of Vα2⁺CD4⁺ SP thymocytes (Fig. 4,B and Table I). Therefore, it is likely that the loss of SIT and TRIM also alters T cell lineage commitment, as the quantity of TCR-mediated signals not only influences selection processes, but also lineage commitment (40).

Having shown that DKO/F5/β₂m⁰/RAG1⁰ mice develop peripheral T cells, we next determined whether these T cells were also functional. Initially, total lymph node cells from wild-type C57BL6/J and DKO/F5/β₂m⁰/RAG1⁰ mice were stimulated in the presence of plate-bound CD3 and TCRβ mAbs or Con A. As shown in Fig. 5,D, DKO/F5/β²m⁰/RAG1⁰ cells robustly responded to all mitogenic stimuli. Finally, to determine whether DKO/F5/β₂m⁰/RAG1⁰ T cells specifically responded to the F5 TCR cognate peptide, lymph node cells from mice of each genotype were stimulated with irradiated APCs loaded with increasing amount of the influenza virus peptide NP68. Fig. 5 E shows that only lymph node cells from DKO/F5/β₂m⁰/RAG1⁰ and from F5/β₂m^+/−/RAG1^+/− control mice responded to the NP68 agonist peptide in a dose-dependent manner. Collectively, our data demonstrate that DKO F5 TCR transgenic T cells positively selected on a β₂m-deficient genetic background are functional. In summary, these results demonstrate a unique joint role for SIT and TRIM as inhibitory TRAPs that lower TCR signals, thereby directly influencing the fate of developing thymocytes.

To date, with the exception of LAT, deletion of any other known transmembrane adaptors has had negligible effects on thymocyte development. By showing that the two related transmembrane adaptor proteins SIT and TRIM together regulate thymic selection, we revealed important insights into the role of TRAPs in T cell development and in the establishment of central tolerance. We propose that SIT and TRIM function as suppressors of signal amplification and regulate thymic selection by modulating the degree of MAPK activation. Furthermore, our data also show that SIT and TRIM function in a cooperative manner, thus indicating that TRAPs may represent a group of molecules with overlapping functions.

As summarized in Fig. 6, our data are consistent with the signaling threshold model for thymocyte selection (41). This model postulates that the TCR affinity/avidity for the selecting ligand is in a direct relationship with the strength of the TCR-mediated signaling and, in turn, with the outcome of selection. Therefore, the outcome of selection of thymocytes expressing a specific TCR can be influenced by increasing or decreasing the TCR signal potential. Indeed, as predicted by the signaling threshold model, our data show that in a TCR transgenic system where the avidity for the selecting ligand is thought to be extremely low (such as in the F5/β₂m⁰/RAG1⁰ system), the loss of SIT and TRIM converts nonselection into positive selection (Fig. 6). Despite the fact that loss of SIT alone is not sufficient to induce the shift into positive selection, DP thymocytes from SIT^−/−/F5/β₂m⁰/RAG1⁰ mice displayed enhanced CD5 levels. This observation suggests that also in SIT^−/−/F5/β₂m⁰/RAG1⁰ mice the TCR-mediated signaling strength is potentiated, as CD5 expression is directly regulated by TCR signal intensity (24). When we tested transgenic mice expressing a TCR with a weak ligand avidity (i.e., HY), where positive selection occurs inefficiently, the loss of SIT alone is sufficient to induce a more efficient positive selection and a partial shift to negative selection (6), whereas the loss of both SIT and TRIM results in a complete conversion to negative selection (Fig. 6). Positive selection of thymocytes expressing a TCR with a higher affinity for its selecting ligand than HY (i.e., P14) is partially converted to negative selection by the loss of SIT alone, and the additional deletion of TRIM further enhances the shift to negative selection (Fig. 6). Regardless of the outcome of selection (i.e., efficient positive selection such as in OT-I TCR transgenic mice or negative selection such as in male HY TCR transgenic mice), over a certain level of avidity and TCR signaling strength, SIT seems to be dispensable during selection processes and the further deletion of TRIM converts efficient positive selection into negative selection only in OT-I TCR transgenic mice (Fig. 6).

Models based on signal strength postulate that negative, positive, and nonselection lie on a continuum of signals emanating from the TCR and that amplification of the signals at a very proximal level determines the right signal intensity for the specific developmental outcome (40). Our data strongly support this idea and further suggest that SIT and TRIM both function as suppressors of signal amplification. Therefore, it is likely that they are located at the very apical part of the TCR signaling cascade. Interestingly, the effects of SIT/TRIM double deficiency on thymocyte development are opposite to those observed in mice carrying mutations in the TCR-associated ζ-chain, a well-known amplifier of TCR-mediated signaling. Indeed, reduction of TCR signal potential by substituting the number of ITAMs within TCR-ζ or by knocking out TCR-ζ resulted in a shift from negative to positive selection or even converted positive into nonselection (Fig. 6) (41, 42, 43, 44).

In light of these observations, it appears that thymocytes are provided with signaling components such as the TCR ζ-chain or TRAPs such as SIT and TRIM that possess opposing regulatory functions. Therefore, signal amplification by multiple ITAMs within the TCR/CD3 complex appears to be counterregulated by an inhibitory system based on several TBSMs within transmembrane adaptors. In fact, in analogy with the effects of an incremental reduction in the number of TCR-ζ ITAMs, our data suggest that also the sequential deletion of TRAPs gradually affects selection processes. It appears that the coordinated integration of the regulatory activity of multiple positive and negative signaling motifs within membrane-associated proteins is required for shaping the T cell repertoire. As this regulatory system appears to be crucial for cell fate specifications in the thymus, it is not surprising that the molecules controlling the process are physically located at the plasma membrane as a part of, or in close proximity to, the receptor. In fact, we have shown that TRIM is an integral component of the TCR/CD3 complex (8, 45), whereas SIT was identified as a protein co-immunoprecipitating with CD3 (9).

Surprisingly, such an elaborated system for fine-tuning TCR-mediated signals does not seem to play a role during peripheral T cell activation, as mutations in the number of TCR-ζ ITAMs (41, 42) or loss of SIT and TRIM (data not shown) only minimally affect activation of mature T cells. These observations indicate that other regulatory mechanisms, such as those mediated by costimulatory molecules, are required to adjust peripheral T cell responses.

The mechanism by which SIT and TRIM regulate signaling threshold remains elusive. A major obstacle encountered in elucidating this question is the development of compensatory mechanisms in mutant mice. Indeed, we showed that thymocytes from both SIT^−/− (6, 39) and DKO mice overexpress CD5, a negative regulator that dampens TCR-mediated signaling (24, 46, 47, 48). Changes in the level of CD5 expression is the best known example of TCR signaling tuning and hence may indicate sensory adaptation (27, 49). It is thought that CD5 expression is regulated to modulate the activation threshold in response to an altered signaling strength. In fact, cells that express a high level of CD5 display a parallel lowered responsiveness to TCR triggering (39, 47, 49). In addition to upregulating CD5, modulation of T cell responsiveness can be also achieved by down-regulating the CD8 coreceptor (29). Interestingly, T cells from SIT-deficient (39) and DKO mice also down-regulate CD8. Thus, the loss of SIT and TRIM resulted in the development of sensory adaptation by up-regulating CD5 and down-regulating CD8.

In accordance with these observations, it is therefore not surprising that we did not observe an enhanced TCR-mediated signaling when DP thymocytes from DKO mice were stimulated with CD3 Ab in vitro (data not shown). It is likely that, in these cells, the enhanced TCR-mediated signaling caused by the loss of SIT and TRIM is counterbalanced by CD5 overexpression or perhaps even by other negative regulatory molecules. The generation of additional mouse model such as SIT^−/−/TRIM^−/−/CD5^−/− triple knockouts is required to investigate how SIT and TRIM modulate TCR-mediated signaling when the counterbalancing effects of CD5 are no longer present.

Recent observations from our laboratory suggest that SIT and TRIM may act in concert to regulate PI3K activation. Indeed, we found that, under particular conditions of TCR stimulation, SIT^−/− T cells showed enhanced Akt phosphorylation, a measure of PI3K activation (B. Schraven and L. Simeoni unpublished results). We have previously shown that also T cells from TRIM^−/− mice displayed enhanced Akt phosphorylation upon TCR triggering (7). It appears that by binding the p85 regulatory subunit of PI3K, TRIM could inhibit PI3K activation, thus sequestering p85 outside of the lipid rafts. However, how SIT regulates PI3K remains as yet unknown. In fact, conversely to TRIM, SIT does not bind p85 directly. We suggest that SIT could regulate Gab association with p85 by controlling Gab phosphorylation via the tyrosine phosphatase SHP-2 (50). Interestingly, we have recently shown that SIT also shares redundant function with the third non-raft transmembrane adaptor protein LAX (linker for activation of X cells) in B1 cells and T lymphocytes (our unpublished results). Similarly to TRIM, LAX binds the p85 regulatory subunit of PI3K and the loss of LAX also resulted in an enhanced Akt phosphorylation in B and T lymphocytes (51). Based upon these data, we speculate that the non-raft adaptors SIT, TRIM, and LAX may functionally converge to regulate PI3K activation (52). Whereas SIT and TRIM appear to play a major role in thymocyte development, SIT and LAX seem to regulate B1 cell development, T cell activation, and autoimmunity. In conclusion, our studies have begun to shed light onto the redundant function of the non-raft adaptors in lymphocyte development and activation. Additional efforts are now required to reveal how these molecules regulate Ag receptor-mediated signaling in concert.

We are grateful to Vilmos Posevitz and Ines Meinert for excellent technical assistance, to Dr. Jonathan Lindquist for critically reading the manuscript and helpful discussion, and the employees of the animal facility for maintenance of the animals. We thank Dr. Dimitris Kioussis for providing F5/β₂m⁰/RAG1⁰ mice.

The authors have no financial conflicts of interest.