The recent demonstration that T cells can recognize a single foreign pMHC complex suggests that TCR triggering does not require aggregation. Or does it?
These are interesting times for those investigating the mechanism of signal transduction through the T cell antigen receptor (TCR), hereafter referred to as TCR triggering. A major point of contention is whether, as with most other receptors that trigger tyrosine phosphorylation, binding-induced multimerization or aggregation is required for TCR triggering. An article published 4 years ago showing that soluble monomeric TCR ligand—peptide bound to a major histocompatibility complex protein (pMHC)—could trigger CD8+ T cells had seemed firm evidence that TCR triggering did not require aggregation1. However, two recent studies suggested that this may be an artifact, resulting from peptide transfer to and presentation by MHC class I molecules on the T cell themselves2,
3. Now a report from Mark Davis's laboratory showing that a single foreign (agonist) pMHC complex can trigger T cells4 seems to be conclusive evidence that aggregation is not required for TCR triggering. Or is it?
Despite intensive study, the mechanism of TCR triggering remains an enigma, and there are a number of competing models5. The key issue is to explain how the binding of agonist pMHC to the TCR leads to biochemical changes (such as tyrosine phosphorylation) in the cytoplasmic portions of the TCR-associated CD3 subunits. Perhaps the most popular models are those proposing that TCR binding to pMHC leads to aggregation of TCRs, which induces phosphorylation by bringing together TCR-associated tyrosine kinases and their substrates, which are regions of the cytoplasmic portions of CD3 subunits. Certainly, artificial aggregation of the TCR is sufficient for TCR triggering; several groups have used streptavidin tetramer technology to show that dimeric soluble agonist pMHC complexes are sufficient to trigger T cells. The original report that monomeric soluble pMHC seemed sufficient to trigger CD8+ T cells1 suggested that TCR aggregation was not necessary for triggering. The investigations that contradicted this result found that TCR triggering by monomeric pMHC class I required expression of the same MHC class I molecule on the T cells2,
3. These and other data led to the conclusion that the peptide was being transferred from the soluble pMHC class I to T cell surface MHC class I and then presented to responding T cells. Thus, the results of experiments that used soluble pMHC seem to be consistent with a requirement for aggregation in TCR triggering.
But what happens under physiological conditions, in which T cells must detect agonist pMHC on the surface of other cells? One requirement of aggregation models is that a sufficiently high density of pMHC ligand be available on the surface of the presenting or target cells to induce aggregation. Several attempts have been made to estimate the minimum number of pMHC complexes required to induce T cell activation, leading to estimates of ten or fewer per cell. A drawback of all these studies is that they measured the average surface density on many cells, leaving open the possibility that rare cells were present that had significantly higher amounts of pMHC. Neither could these studies exclude the possibility that agonist pMHC is heavily concentrated in certain areas of the cell surface, thereby facilitating aggregation.
Davis and colleagues have used a novel approach to measure directly the minimum number of pMHC complexes required to activate a T cell4. The key innovation was to take advantage of the extraordinary fluorescent brightness of the phycoerythrin (PE) protein, which enabled them to detect and count individual agonist pMHC complexes at the interface between living T cells and antigen-presenting cells (APCs). Cells were pulsed with biotinylated peptides and then incubated with saturating levels of PE-streptavidin. In careful control experiments, they showed that, under these conditions, essentially all the agonist pMHCs are labeled and that all labeled agonist pMHCs can be detected by fluorescent microscopy. They then counted the number of agonist pMHC molecules at T cell−APC interfaces and correlated this with increases in the intracellular Ca2+ concentration, a marker of TCR triggering. They showed that cells responded more often than not, even when only a single agonist pMHC was present in the contact interface4. They also showed that there was a correlation between the size of the Ca2+ response and the number of agonist pMHC complexes, with a maximal response observed with 10−20 agonist pMHC complexes4.
Given that it would seem to be impossible for a single agonist pMHC to induce TCR aggregation, these results appear to rule out conventional TCR aggregation as the mechanism of TCR signaling. There are, however, "unconventional" models of TCR aggregation in which aggregation is not driven by binding to two or more agonist pMHC ligands and which are, thus, compatible with these data. One possibility is that the engaged TCR undergoes a conformational change upon binding pMHC, which enables other unengaged TCRs to bind to it. The evidence originally cited in support of this model has been disputed, however, and structural evidence is strongly against binding-induced conformational changes in the TCR itself5. A second possibility is that TCR engagement of self-pMHC contributes to aggregation. Indeed, this is what Davis and colleagues propose in their pseudodimer model of TCR triggering4. This model was inspired in part by their finding that TCR triggering requires more agonist pMHCs in the presence of a blocking CD4 monoclonal antibody (mAb) and by the observation that CD4 and CD8 coreceptors engage MHC at an angle that precludes simultaneous direct interactions with the TCR (Fig. 1). The pseudodimer model proposes that, after a TCR binds to specific pMHC, CD4 associated with the same TCR engages a self-pMHC molecule on the same APC and that enables a second TCR to bind to the self-pMHC, resulting in TCR dimerization.
Figure 1. Coreceptor function and a proposed role for self-pMHC in TCR triggering.
Two models are consistent with the obsevation that the coreceptors CD4 and CD8 (not shown) bind pMHC at an angle that precludes direct association with the cognate TCR. The pseudodimer model proposes that CD4 associates directly with the TCRs that have engaged agonist pMHC but binds a different MHC which is likely to present self-peptide. The conventional model proposes that CD4 and TCR bind the same pMHC complex and associate with each other indirectly through intermediate molecules.
This is an ingenious model that accounts for many observations that are difficult to reconcile. But is it correct? One prediction of the model is that simultaneous recognition of self-pMHC is critical for TCR triggering. There is, indeed, growing evidence that TCR engagement of self-pMHC contributes to the function of peripheral T cells. In addition to a role in T cell homeostasis and survival, it has been reported that prior exposure to self-pMHC modulates T cell responses to foreign pMHC, although there are conflicting reports concerning whether self-pMHC engagement inhibits6,
7 or enhances8 T cell responses. There is little data on the more relevant question of whether simultaneous engagement of self-pMHC enhances recognition of foreign pMHC. This is a technically difficult experiment, but one group has devised a clever way of doing it9. They took advantage of transporter associated with antigen processing−deficient APCs, which have few self-pMHCs, and an agonist pMHC-specific mAb, to expose T cells to defined amounts of agonist pMHC in the context of high or low amounts of self-pMHC. They were unable to detect any difference in the T cell responses, indicating that simultaneous self-pMHC recognition made no contribution to TCR triggering in these cells. Although this experiment needs to be repeated in other systems and with CD4+ T cells, it does argue against self-pMHC recognition as a key element of TCR triggering.
One important feature of the pseudodimer model is that it is compatible with structural evidence that coreceptors cannot simultaneously bind directly to a TCR and pMHC in a TCR-pMHC complex (Fig. 1). However, these structural data are also compatible with a more conventional model in which coreceptors associate with the TCR indirectly, via CD3 and the tyrosine kinases ZAP-70 and Lck.
In conclusion, the demonstration by Davis and colleagues that a single pMHC is sufficient to induce TCR triggering is a technical tour de force and places an important constraint on models of TCR triggering. Aggregation models seem less plausible now, but are not completely ruled out because TCR engagement of self-pMHC may contribute to aggregation. In contrast, models that do not require TCR aggregation, including binding-induced changes in the conformation10 or mobility of the TCR-CD3 complex, now seem more likely to be correct5.
