T cell receptor specificity for major histocompatibility complex proteins

The ligands for alpha beta T cell receptors (αβTCRs) are usually major histocompatibility complex (MHC) proteins bound to peptides. Although there is evidence that T cell receptor variable regions have been selected evolutionarily to bind MHC, the rules governing this interaction have not previously been apparent. However, recent solved structures of T cell receptors with related variable regions bound to MHC plus peptides suggest that some amino acids in variable region CDR1 and CDR2s almost always react in a consistent way with MHC. These amino acids may therefore have been selected evolutionarily to predispose T cell receptors toward recognition of MHC ligands.

Introduction

Proteins have been evolutionarily selected to react specifically with particular targets, other proteins, nucleic acids, substrates for enzymes, and so on. This phenomenon is not difficult to understand for most proteins, which have well-defined sequences and a relatively small collection of ligands that are usually created directly or indirectly by the genome of their host. However, the phenomenon is more complicated for the antigen-recognizing proteins of the specific immune system, which are germline encoded, have related and very variable sequences but a huge range of unpredictable ligands that are not encoded by their host. For antibody proteins, the explanation is quite straightforward, the variability created during the assembly of immunoglobulin genes coupled with selection by antigen and somatic mutation during response picks out antibody proteins that can bind virtually any antigen. The situation for alpha beta T cell receptors (αβTCRs) is less easily understood. The variability is created only during gene rearrangement, not somatic mutation, and the repertoire of αβTCRs on mature T cells is tremendously biased toward recognition of antigens in the form of peptides bound to major histocompatibility complex (MHC) proteins.

How is this bias created? There are two explanations. One possibility is that the specificities of αβTCRs, like those of immunoglobulins, are intrinsically random. T cell receptors could react with any ligand. However, in order to turn into mature T cells, developing thymocytes must react with MHC/peptides in the thymus. Thus, the bias for MHC reactivity on mature T cells could be caused entirely by positive selection from pool of random TCRs. Plenty of proteins are expressed on thymus cortical epithelial cells, the major cells participating in positive selection. Of all of these, why are TCRs selected for reaction with MHC? Perhaps positive selection requires co-engagement of CD4 or CD8 [1, 2], and these proteins react with MHC, not other proteins on thymus epithelial cells. Alternatively thymocytes may be selected by reaction with all the other proteins on cortical epithelial cells, but then all those with high avidity/affinity are deleted by negative selection. In this case some mature T cells should bear αβTCRs with low affinity for proteins such as ICAM-1. These T cells will be invisible because they will never encounter a peripheral variant of their target ligand for which they have high affinity. However, to some extent a test of this idea has been performed, searching for T cells specific for the MHC-related protein, DM, and such cells were not found [3].

Finally, αβTCRs may react with MHC because they have a built-in affinity for these proteins [4]. If that happens, then probably it will operate via the CDR1 and 2 sequences of the TCR α and β chains since these are germline encoded, and since, in the solved structures of TCRs bound to MHC, these portions of the TCR contact MHC helix [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17••, 18, 19••]. Against this idea is the fact that every species contains many TCR α and β variable regions and, although these fall into families, there is no close resemblance between the different groups [20]. Also, until recently, the solved structures of TCRs bound to MHC/peptide have failed to reveal any pattern whereby particular V region amino acids contact MHC in a reproducible way [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17••, 18, 19••]. In favor of the idea is the fact that the TCR V region sequences are relatively conserved from man to mouse [20, 21], indicating functions for the various V families, and the fact that TCRs nearly always engage MHC diagonally [22^•], suggesting either that docking of CD4 and CD8 requires the diagonal (see above) or that some conserved feature of TCRs and MHC drives the phenomenon.

With the increasing numbers of solved structures of TCRs bound to MHC, evidence for conserved interactions is accumulating.