TCR Comodulation of Nonengaged TCR Takes Place by a Protein Kinase C and CD3γ Di-Leucine-Based Motif-Dependent Mechanism ¹

The TCR comprises the clonotypic Tiα and β chains and the invariant CD3γ, ε, δ, and ζ chains (1, 2, 3, 4, 5, 6). The α- and β-chains are responsible for the specific recognition of peptide-MHC complexes and the signaling events following Ag recognition are mediated by the CD3 and ζ-chains (7, 8, 9).

Regulation of TCR expression levels is most probably a very important mechanism that allows T cells to calibrate their responses to different levels of stimuli. TCR internalization takes place both in resting T cells as part of constitutive TCR cycling and following TCR triggering and protein kinase C (PKC) ³ activation. During constitutive internalization, TCR cycling between the cell surface and the endosomes results in a steady-state distribution of the TCR with 70–85% of the cycling TCR pool at the cell surface and 15–30% of the pool inside the cell (10, 11, 12, 13). Recently, it has been shown that constitutive TCR cycling is dependent on the di-leucine-based (LL-based) motif in CD3γ (14). This motif consists of the DxxxLL sequence and the mechanism whereby it mediates TCR internalization has been described in detail previously (15, 16). In completely assembled TCR complexes, the CD3γ LL-based motif is not fully exposed. However, subsequent to phosphorylation of CD3γ serine126 the LL-based motif becomes exposed and the clathrin-associated AP-2 binds to the motif. Binding of AP-2 leads to internalization of the TCR by the clathrin-dependent internalization machinery (16). In addition to constitutive TCR cycling, both PKC- and ligand-induced TCR down-regulation are dependent on the CD3γ LL-based motif (14, 17, 18). During constitutive TCR cycling and following PKC-mediated TCR internalization, the TCR are recycled back to the cell surface in a functional state (19). In contrast, when the TCR are internalized by the ligand-induced mechanism, which is dependent on protein tyrosine kinase (PTK) activity, a large fraction of the TCR is targeted for lysosomal degradation (17, 20, 21, 22, 23).

After contact between a T cell and an APC loaded with the relevant peptide, TCR are actively recruited to the contact zone between the T cell and the APC (24). A well-organized structure, the immunological synapse (IS), is formed within minutes of contact, with adhesion molecules in the periphery and TCR located at the center of the IS (25, 26). Only a small fraction of the TCR in the IS appears to be bound to specific peptide-MHC complexes (24). This is interesting taking into consideration that up to 90% of the TCR on the surface of the T cell can be rapidly internalized after ligand stimulation. Based on such observations, Lanzavacchia and coworkers (27) proposed the TCR serial triggering model in which one peptide-MHC complex is able to stimulate and down-regulate a large number of TCR (up to 200) within a relatively short period. However, some studies have challenged the TCR serial triggering model by showing that TCR down-regulation might include both directly engaged as well as nonengaged TCR (28, 29, 30, 31, 32). A few studies suggested that a physical association of multiple TCR might be the mechanism behind comodulation (28, 29), whereas other studies have indicated that trans-acting phosphorylation of nonengaged TCR plays a crucial role (30, 31, 32). Still other studies have not been able to detect TCR comodulation (3, 27, 33). Thus, whether comodulation of nonengaged TCR actually takes place and if so the mechanisms behind comodulation are still not fully known.

The aim of this study was to investigate whether comodulation of nonengaged TCR takes place and if so to determine the role of ligand affinity and the mechanisms involved in TCR comodulation.

The Vβ8⁺ chicken OVA/I-A^d-specific mouse T cell hybridoma DO11.10 (34), the I-A^d+ mouse B cell lymphoma M12.B5 (35), the Vβ8⁺ human Jurkat T cells with SV40 virus large T Ag (JTag), and the human T cell Jurkat variant JBN, which lacks the Tiβ chain, were cultured in complete medium (RPMI 1640 medium supplemented with 0.5 IU/L penicillin, 500 mg/L streptomycin, and 10% FBS) at 37°C in 5% CO₂. The anti-mouse Vβ8.2 mAb F23.1 and the F23.1 variants 20 and 32 (36) were kindly provided by Dr. K. Karjalainen (Institute for Research in Biomedicine, Bellinzona, Switzerland). The anti-human Vβ8 mAb MX6 was kindly provided by Dr. A. Boylston (University of Leeds, Leeds, U.K.). The PE-, FITC-, and nonconjugated anti-mouse Vβ8 mAb (F23.1), anti-mouse Vβ2 mAb (B20.6), PE-conjugated anti-human Vβ8 mAb (BV8), and anti-human Vβ3 mAb (BV3S1) were obtained from BD PharMingen (San Diego, CA). Protein A was purchased from Amersham Pharmacia Biotech (Uppsala, Sweden). The chicken OVA peptide OVA_323–339 was obtained from Schafer-N (Copenhagen, Denmark). 2′,7′-bis-(2-carboxyethyl)-5-(and 6)-carboxyfluorescein acetoxymethyl ester (BCECF/AM) was purchased from Molecular Probes (Eugene, OR). The broad PKC inhibitor Ro 31-8220 (37) was kindly provided by Dr. D. Bradshaw (Roche Research Center, Welwyn Garden City, U.K.). The scr family-specific inhibitor PP1 and the broad PKC inhibitor bisindolylmaleimide I were obtained from Calbiochem (San Diego, CA).

DO11.10 cells were transfected with the expression vector pCA134 (38) coding for the mouse Vβ2 chain (kindly provided by Dr. B. Malissen, Marseille, France) and JBN cells were transfected with the expression vector pcDNA6-HA1.7-TCRβ-5 (39) coding for the human Vβ3 chain (kindly provided by Dr. L. Wedderburn, London, U.K.) as previously described (40). Cells were plated at 1 × 10⁴ and 5 × 10⁴ cells/ml in 96-well tissue culture plates in complete medium containing either 1 mg/ml G418 sulfate (DO11.10) or 5 μg/ml blasticidin (JBN). After 3–4 wk of selection G418- and blasticidin-resistant clones were expanded and maintained in complete medium.

For lipotransfection of JTag cells, 0.5 ml of RPMI 1640 containing DMRIE-C reagent (16 μg/ml; Invitrogen, Carlsbad, CA) and 0.5 ml of RPMI 1640 containing the expression vectors pcDNA6-HA1.7-TCRβ-5 (4 μg/ml) and either pEGFP-HCD3γ-WT-1 or pEGFP-HCD3γ-LLAA-3 (4 μg/ml) were mixed in each well of a six-well plate. The pEGFP-HCD3γ-WT-1 coded for chimeric wild-type (WT) CD3γ-green fluorescent protein (GFP; CD3γWT-GFP) and pEGFP-HCD3γ-LLAA-3 coded for CD3γLLAA-GFP mutated in the LL-based motif as described previously (14). To allow formation of lipid-DNA complexes, the plates were incubated at room temperature for 30 min. Subsequently, 2 × 10⁶ JTag cells were added to each well. Following 4 h of incubation at 37°C in 5% CO₂, 2 ml of RPMI 1640 containing 15% FBS was added to each well. After 24 h, 1 ml of complete medium was added to each well and the cells were used for experiments 48 h after lipotransfection.

For experiments with the Vβ2⁺Vβ8⁺ DO11.10 transfectants, Maxisorb plates (Nunc, Roskilde, Denmark) were coated with protein A or rabbit anti-rat Ig (10 μg/ml) overnight at 4°C. The plates were washed in PBS and subsequently blocked for 1 h with 2% BSA in PBS at room temperature. Following blocking, the plates were washed in PBS and incubated with various concentrations of anti-mouse Vβ8 or anti-mouse Vβ2 mAb diluted in PBS with 0.2% BSA for 2 h at room temperature. The cells were adjusted to 4 × 10⁵ cells/ml and transferred to the coated plates, centrifuged for 1 min at 500 rpm, and incubated at 37°C. At the indicated time, cells were transferred to ice. The cells were stained directly with PE-conjugated anti-mouse Vβ2 or anti-mouse Vβ8 mAb and analyzed by flow cytometry. For stimulation with peptide/MHC, the APC cells M12.B5 were pulsed for 2 h with the indicated concentrations of OVA_323–339 peptide, stained with BCECF/AM for 10 min, washed four times, and cocultured with DO11.10 cells at a ratio of 1:1 for the time indicated. The cells were subsequently transferred to 4°C and analyzed for TCR expression by incubation with PE-conjugated anti-mouse Vβ2 or anti-mouse Vβ8 mAb and gating out cells with green fluorescence (BCECF/AM). For the experiments using kinase inhibitors, cells were preincubated with PP1 (16 μM, 30 min), Ro 31-8220 (40 μM, 15 min) or bisindolylmaleimide (10 nM,; 15 min) at 37°C. The TCR down-regulation experiments were performed as described above. However, cells treated with the PKC inhibitors were stained with FITC-conjugated Abs.

For experiments with JTag transfectants, plates were coated with anti-human Vβ8 mAb (250 ng/ml) overnight at 4°C and subsequently washed in PBS. The cells were adjusted to 5 × 10⁵ cells/ml and transferred to the coated plates, centrifuged for 1 min at 500 rpm, and placed at 37°C. At the indicated time, cells were transferred to ice. The cells were stained directly with PE-conjugated anti-human Vβ3 or anti-human Vβ8 mAb and analyzed by flow cytometry. TCR down-regulation was determined as ((mean fluorescence intensity (MFI) of stimulated cells)/(MFI of nonstimulated cells)) × 100%.

To obtain cells expressing two distinct TCR, we transfected the Vβ8⁺ DO11.10 T cell hybridomas with plasmids coding for the mouse Vβ2 chain. Double-positive transfectants that stained equally bright for Vβ2 and Vβ8 were selected for subsequent experiments (Fig. 1, A and B). When using Tiβ-specific mAb in comodulation analysis, it is mandatory to know whether the mAb used for stimulation cross-reacts with the other type of Tiβ chain expressed by the double-positive transfectants. Thus, to analyze for possible cross-reaction of the stimulating mAb, the transfectants were preincubated with either nonconjugated anti-mouse Vβ2 or anti-mouse Vβ8 mAb for 2 h at 12°C. The cells were next incubated with either PE-conjugated anti-mouse Vβ2 or anti-mouse Vβ8 mAb for 30 min on ice and subsequently analyzed by flow cytometry. The nonconjugated anti-mouse Vβ8 mAb was able to blocked subsequent binding of PE-conjugated anti-Vβ8 mAb but did not affect the binding of PE-conjugated anti-mouse Vβ2 mAbs (Fig. 1, C and D). Likewise, the nonconjugated anti-mouse Vβ2 mAb blocked subsequent binding of PE-conjugated anti-mouse Vβ2 mAb but did not affect the binding of PE-conjugated anti-mouse Vβ8 mAb (Fig. 1, C and D).

Some studies have shown that TCR triggering can induce comodulation of nonengaged TCR (30, 31, 32). In contrast, other studies did not observe comodulation of nonengaged TCR following stimulation of T cells expressing two different TCR (3, 27, 33). To analyze whether engaged TCR induced comodulation of nonengaged TCR, the Vβ2⁺Vβ8⁺ DO11.10 transfectants were stimulated with plate-bound Vβ8- or Vβ2-specific mAb for various periods of time. Stimulation of the cells with anti-mouse Vβ8 mAb induced down-regulation of both the specifically engaged TCR (Vβ8) and the nonengaged TCR (Vβ2) (Fig. 2,A). Likewise, stimulation of the cells with anti-mouse Vβ2 mAb induced down-regulation of both the specifically engaged TCR (Vβ2) and the nonengaged TCR (Vβ8) (Fig. 2,B). The kinetics whereby engaged TCR and nonengaged TCR were down-regulated differed. The engaged TCR were rapidly down-regulated and reached a plateau of ∼30% after 1 h. The nonengaged TCR were down-regulated to a lesser degree and with a slightly slower kinetics than engaged receptors. To analyze whether comodulation also took place during stimulation with natural peptide/MHC ligands, the Vβ2⁺Vβ8⁺ DO11.10 transfectants were incubated for various times with I-A^d+ M12.B5 cells pulsed with OVA_323–339. Although down-regulation of the directly triggered TCR (Vβ8) was not as efficient as seen after stimulation with the anti-Vβ mAb, comodulation of Vβ2 clearly took place (Fig. 2 C). Taken together, these experiments clearly demonstrated that comodulation of nonengaged TCR took place both following triggering with high-affinity ligands in the form of mAb and following triggering with low-affinity natural ligands in the form of peptide-MHC complexes.

It has been shown that the affinity between the ligand and the TCR affects the activation levels of the T cell (36, 41, 42), and we speculated whether the observed differences in TCR down-regulation observed after stimulation with mAb vs peptide-MHC complexes could be due to the ∼300-fold higher affinity of the mAb compared with the peptide-MHC complexes. To investigate the role of affinity on TCR comodulation, we used a panel of variants of the anti-mouse Vβ8 mAb F23.1 with known affinities (36). Vβ2⁺Vβ8⁺ DO11.10 transfectants were stimulated with the plate-bound Abs for various times and were subsequently tested for the expression of Vβ2 and Vβ8. All mAb applied induced TCR down-regulation of both Vβ2 and Vβ8 (Fig. 3). However, the affinity of the stimulating mAb had a great impact on the degree of TCR down-regulation of both engaged and nonengaged TCR. For the mAb with the highest affinity (K_D = 2.3 nM), an optimum was reached after 30 min for nonengaged TCR and after 60 min for engaged TCR. Interestingly, after reaching a comodulation optimum following ∼30 min of stimulation, the level of comodulation started to decline. After ∼2 h of Vβ8 stimulation, the expression level of nonengaged TCR were similar to the steady-state level of unstimulated cells, whereas the engaged TCR were permanently down-regulated to a level of ∼30% of unstimulated cells (Fig. 3 A). This experiments demonstrated that comodulation of nonengaged TCR takes places in a time-dependent and reversible manner following stimulation with high-affinity ligands. This might explain why some groups did observe comodulation whereas other groups did not observe this phenomenon.

Following stimulation with the mAb with a medium affinity (K_D = 28 nM), TCR down-regulation optimum was again reached after 30 min for nonengaged TCR and after 60 min for engaged TCR. The expression level of engaged TCR remained at this level for the rest of the experiment, whereas the expression level of nonengaged TCR slightly increased during the rest of the experiment (Fig. 3,B). For the mAb with the lowest affinity (K_D = 6200 nM), equivalent to the affinity between natural peptide/MHC ligands and the TCR, the kinetics of TCR down-regulation were similar for engaged and nonengaged TCR, reaching an optimum after 60 min and remaining at this level for the rest of the experiment (Fig. 3,C). Interestingly, the kinetics of TCR down-regulation following stimulation with the low-affinity mAb were similar to the kinetics of TCR down-regulation following stimulation with peptide-MHC complexes (Fig. 2,C vs Fig. 3 C). Taken together, these experiments showed that the affinity of the stimulating mAbs had great impact on the kinetics of TCR down-regulation of both engaged and nonengaged TCR. Furthermore, whereas the affinity clearly affected the maximum level of down-regulation of engaged TCR, it did not seem to affect the maximum level of comodulation of nonengaged TCR significantly.

Previous studies have demonstrated that ligand-mediated TCR internalization is dependent on PTK activity (17, 20, 21, 22). In addition, a recent study demonstrated that efficient ligand-mediated TCR internalization also depended on PKC activity (18). We therefore wished to investigate the role of PTK and PKC in comodulation of the TCR. Accordingly, TCR comodulation experiments were performed in the presence of either the scr family-specific inhibitor PP1 or the broad PKC inhibitors Ro 31-8220 and bisindolylmaleimide. Vβ2⁺Vβ8⁺ DO11.10 transfectants were pretreated with either PP1, Ro 31-8220, or bisindolylmaleimide or were left untreated. The cells were next stimulated with plate-bound anti-mouse Vβ8 mAb as described above in the presence or absence of inhibitors and tested for the expression of Vβ2 and Vβ8. As plate-bound anti-mouse Vβ8 mAb, either the low-affinity (K_D = 6200 nM) or the high-affinity (K_D = 2.3 nM) anti-mouse Vβ8 mAb was used. Both PP1 and Ro 31-8220 completely inhibited comodulation of Vβ2 following stimulation with the low-affinity anti-mouse Vβ8 mAb (Fig. 4, A–C). Likewise, down-regulation of engaged Vβ8 was completely inhibited by PP1, whereas treatment with Ro 31-8220 only partially inhibited Vβ8 down-regulation (Fig. 4, A–C). When using the high-affinity mAb, Vβ2 comodulation was almost completely inhibited in cells treated with PKC inhibitor, whereas Vβ8 down-regulation was only partially inhibited (Fig. 4, D and E). PP1 did not significantly inhibit Vβ8 down-regulation or Vβ2 comodulation following stimulation with the high-affinity mAb (data not shown) as previously observed by others following stimulation with high-affinity ligands (32). Taken together, these data indicated that TCR comodulation is strongly dependent on PKC activity whereas down-regulation of engaged TCR only partially depends on PKC activity. Furthermore, both TCR comodulation and down-regulation of engaged TCR are dependent on PTK activity following stimulation with low-affinity ligands.

A previous study on TCR double-positive cells has demonstrated that the CD3γ chain of both engaged and nonengaged TCR becomes phosphorylated following specific stimulation of one of the TCR types expressed (30). Several studies have demonstrated the important role of PKC-mediated activation of the LL-based motif in CD3γ during PKC-induced TCR down-regulation (16, 17, 43). Taken together with the present results on the dependence of TCR comodulation on PKC activity, this implied that the CD3γ LL-based motif might play a central role in TCR comodulation. To test this, the Vβ8⁺ Jurkat T cell line JTag was cotransfected with Vβ3 and chimeric constructs of WT CD3γWT-GFP or mutated CD3γLLAA-GFP. To exclude cross-reactivity between the anti-human Vβ8 mAb and Vβ3, the Tiβ chain-negative Jurkat variant JBN was transfected with Vβ3. Vβ3⁺ JBN cells and Vβ8⁺ JTag cells were then stimulated with plate-bound anti-human Vβ8 mAb and subsequently analyzed for the expression of Vβ3 and Vβ8, respectively. Stimulation with anti-human Vβ8 mAb induced TCR down-regulation in Vβ8⁺ Jurkat cells, whereas no effect was seen on Vβ3⁺ JBN cells, demonstrating that the anti-human Vβ8 mAb did not cross-react with Vβ3 (Fig. 5,A). Subsequently, the Vβ3⁺Vβ8⁺ JTag transfectants coexpressing either CD3γWT-GFP or CD3γLLAA-GFP were stimulated with plate-bound anti-human Vβ8 mAb for various times and analyzed for the expression of Vβ8 and Vβ3. In Vβ3⁺Vβ8⁺ transfectants coexpressing CD3γWT-GFP, both engaged and nonengaged TCR were down-regulated (Fig. 5,B). However, the engaged TCR were down-regulated with faster kinetics and to a greater extent than the nonengaged TCR. In Vβ3⁺Vβ8⁺ transfectants coexpressing CD3γLLAA-GFP, down-regulation of engaged TCR was partially inhibited whereas comodulation of nonengaged TCR was almost completely inhibited (Fig. 5 C). These results indicated that comodulation is strongly dependent on the LL-based motif of CD3γ and further confirm previous studies indicating that efficient ligand-mediated TCR down-regulation is dependent on this motif (18).

In this study, we found that TCR ligation with mAb and peptide-MHC complexes induced down-regulation of nonengaged TCR in Ag-specific mouse T cell hybridomas. This comodulation was dependent on the affinity of the stimulating Abs and on the activity of PKC. Furthermore, comodulation was dependent on PTK activity when the cells were stimulated with mAb having low affinity similar to the affinity of peptide-MHC complexes. Finally, we found that the LL-based motif in CD3γ was required for TCR internalization of nonengaged receptors.

There has been some disagreement in the literature concerning TCR comodulation. In some studies comodulation were observed (28, 30, 31, 32), whereas other studies failed to detect comodulation (3, 27, 33). Hou et al. (3) examined for comodulation following 20 h of stimulation with a high-affinity mAb and found no comodulation. However, we found re-expression of comodulated receptors following 2 h of stimulation with a high-affinity mAb, and it is therefore likely that Hou et al. (3) did not observe any comodulation because they analyzed for comodulation too late. Furthermore, some degree of comodulation is actually seen in the studies by Stotz et al. (33) and Valitutti et al. (27), although the contrary was reported by these authors.

We found that comodulation was highly dependent on the affinity of the stimulating ligand. Following stimulation with a high-affinity ligand, a rapid down-regulation of engaged as well as nonengaged receptors was observed. Whereas the expression level of engaged TCR remained low during the experiments, comodulation was reversible in that nonengaged receptors became re-expressed at the cell surface after 30 min of stimulation. In contrast, when T cells were stimulated with low-affinity ligands, the engaged and nonengaged receptors followed approximately the same kinetics for down-regulation. In agreement with these results, a recent study observed recycling of nonengaged receptors following stimulation with a high-affinity ligand and, furthermore, found approximately the same kinetics for engaged and nonengaged receptors following stimulation with a low-affinity ligand (32). The observation that comodulation is reversible argues against the suggestion that physical association between engaged and nonengaged TCR may be the mechanism behind comodulation as proposed by others (28, 29) based on studies suggesting that each TCR complex comprises two Tiαβ dimers. According to these studies, reversible comodulation of Vβ2 might be explained by down-regulation of Vβ2/Vβ8-containing TCR complexes and the isolated reappearance of Vβ2/Vβ2-containing TCR complexes. However, we do not think that the reversible comodulation of Vβ2 observed following stimulation with the high-affinity anti-Vβ8 mAb can be explained by accumulation of Vβ2-Vβ2 complexes at the cell surface during the rather short time interval of the experiments. It is known that TCR down-regulated by direct triggering do not recycle but are degraded. If the observed Vβ2 down-regulation was caused by down-regulation of Vβ2-Vβ8 complexes, the reappearance of Vβ2 should be caused solely by new synthesis of Vβ2. The rate constant for TCR new synthesis is low (∼0.001 min⁻¹, C.G., unpublished data) and it has been shown that re-establishment of normal TCR levels following TCR triggering with anti-TCR mAb takes >24 h (44). However, in our experiments normal levels of Vβ2 are observed ∼90 min after maximal comodulation. This could easily be explained by recycling of internalized Vβ2, as the recycling rate constant is ∼50 times higher (0.05 min⁻¹) than the rate constant for new synthesis (0.001 min⁻¹) (13). The question is then how could the comodulation of Vβ2 be reversible. Stimulation with the high-affinity anti-Vβ8 mAb causes a rapid and profound down-regulation of Vβ8. It may be suggested that TCR signaling through Vβ8 is significantly reduced after ∼1 h due to the induced low levels of Vβ8. Whereas the directly triggered Vβ8 are degraded, the comodulated Vβ2 quickly recycles back to the surface when PKC activity returns to basic prestimulation levels (19).

In this study, the Vβ2 chain used to transfect DO11.10 cells was obtained from the H-2K^b-specific KB5-C20 T cell line (38). With regard to the comodulation experiments using specific peptide/MHC stimulation, we cannot formally exclude the possibility that the TCR generated by pairing of the transfected Vβ2 chain and the endogenous DO11.10 TCRα chain recognized OVA_323–339/I-A^d although we find this unlikely.

We found that PTK is required for TCR down-regulation of both engaged and nonengaged receptors following stimulation with a low-affinity ligand. A previous study found that down-regulation of nonengaged receptors was dependent on PTK, whereas down-regulation of engaged receptors was independent of PTK (32). The divergence from our observations could be explained both by the use of a high-affinity ligand and by the fact that a TT ζ chimera was used as model for one of the TCR in this study (32). Indeed, we have data showing that stimulation with high-affinity ligands can induce TCR down-regulation independently of PTK (data not shown).

We have recently shown that efficient down-regulation of engaged TCR is dependent on PKC and the LL-based motif in CD3γ (18). In the present study, we elaborated on this observation and found that in addition to the requirement for efficient down-regulation of engaged TCR, PKC activity and the LL-based motif in CD3γ are required for comodulation of nonengaged TCR. In agreement with this, it has been shown that stimulation of the TCR leads to phosphorylation of CD3γ both in engaged and nonengaged receptors (30). Comodulation of Vβ3 in JTag cells cotransfected with CD3γLLAA-GFP was not completely abolished. This was probably due to the fact that although CD3γLLAA-GFP was overexpressed in the JTag cells they still expressed WT CD3γ. Thus, not all of the Vβ3⁺ TCR expressed at the cell surface contained CD3γLLAA-GFP but a small fraction contained the WT CD3γ and therefore maintained the ability to be comodulated. Inhibition of PKC activity led to down-regulation of engaged receptors with a slower kinetics compared with untreated cells. This indicated a role of PKC early in the TCR internalization. The requirements of PKC and the LL-based motif for down-regulation of engaged receptors could either be a direct effect or an indirect effect because it must be expected that a fraction of the Ab-specific receptors were comodulated rather than down-regulated by direct engagement.

The role of TCR down-regulation of both engaged and nonengaged TCR is still unknown. It could play a role in receptor revision and tolerance induction during T cell development as suggested by Fink and McMahan (45). Furthermore, it may serve as a protective mechanism to avoid overstimulation of the T cell or it may be a part of the signaling mechanism following TCR triggering. We have recently shown that PKC activation increases the endocytic rate of the TCR without affecting the exocytic rate (13). It is therefore likely that TCR triggering would adjust a new kinetic equilibrium for nonengaged TCR with reduced levels of TCR expressed on the surface and an increased pool of intracellular TCR. The recycling ability of this pool of nonengaged TCR allows directed exocytosis of nontriggered TCR, which may suggest that down-regulation of nonengaged receptors serves as a recruitment mechanism for TCR to the IS (Fig. 6). This could explain how a large proportion of the TCR are rapidly concentrated in the IS after contact between the T cell and the APC (24).

In conclusion, this study clearly demonstrates that TCR comodulation takes place following TCR triggering. At first sight, this indicates that the TCR serial triggering model is hereby falsified. However, due to the complicated trafficking and sorting of the TCR, including recycling of nonengaged receptors, the serial triggering model might still be valid. The only conclusion that can be made is that it is not correct to equate the number of down-regulated TCR with the number of triggered TCR. Therefore, it is an open question whether the number of triggered TCR outnumbers, equals, or is less than the number of down-regulated TCR.

The technical help of Bodil Nielsen is gratefully acknowledged.