How the T Cell Repertoire Becomes Peptide and MHC Specific (original) (raw)

cells are usually cytotoxic, whereas CD4 + , class II-4 Department of Medicine restricted T cells often act to help the responses of other 5 Department of Pharmacology leukocytes (Goldrath and Bevan, 1999; Starr et al., University of Colorado Health Sciences Center 2002; von Boehmer et al., 2003). Denver, Colorado 80262 These observations have led to two longstanding problems in immunology. First, how can a relatively limited collection of TCR Vα, Jα, Vβ, Dβ, and Jβ segments Summary produce a set of TCRs that, in any individual, is quite specific, both for peptide and the alleles of MHC ex-T cells bearing ␣␤ T cell receptors (TCRs) recognize pressed in that individual? This property is usually antigens in the form of peptides bound to class I or thought to be the result of MHC-driven positive selecclass II major histocompatibility proteins (MHC). tion, a phenomenon that occurs during thymocyte de-TCRs on mature T cells are usually very specific for velopment. Positive selection allows conversion to maboth peptide and MHC class and allele. They are ture T cells of only those thymocytes that bear TCRs picked out from a precursor population in the thymus with low but appreciable affinity/avidity for MHC + pepby MHC-driven positive and negative selection. Here tide combinations present in the thymus (Goldrath and we show that the pool of T cells initially positively Bevan, 1999; Starr et al., 2002; von Boehmer et al., selected in the thymus contains many T cells that are 2003). However, two other events that occur in the thyvery crossreactive for peptide and MHC and that submus might bear on the specificity of mature T cells: (1) sequent negative selection establishes the MHCdeath by neglect, an event whereby thymocytes with restriction and peptide specificity of peripheral T cells. no appreciable affinity/avidity for MHC + peptide com-Our results also suggest that germline-encoded TCR binations in the thymus die (Berg et al., 1989; Scott et variable elements have an inherent predisposition to al., 1989), and (2) negative selection, an event whereby react with features shared by all MHC proteins. thymocytes bearing TCRs with high affinity/avidity for ligands in the thymus also die (Kappler et al., 1987; Ki-Introduction sielow et al., 1988). Negative selection prevents the peripheralization of most autoreactive cells (Mathis and T cells bearing αβ T cell receptors (TCRs) recognize Benoist, 2004). Because of the demands of positive and invading organisms in the form of peptides derived negative selection, only a few of all possible TCRs are from the invader bound to major histocompatibility proexpressed on mature T cells in any given animal. teins (MHC) of their host (Townsend and Bodmer, 1989; The second question deals with the problem of how Unanue, 1984). The genes coding for TCRs are prothe relatively conserved set of TCR polypeptides can duced by gene rearrangement during T cell developcreate TCRs that can react with all the different alleles ment in the thymus. This process creates receptors that of MHC that might be present in different individuals of are very different in sequence from one T cell to anthe species and, most dramatically, with MHC proteins other. However, all TCRs include a Vα, Jα, Vβ, Dβ, and of both types, class I and class II. The explanation for Jβ segment, each selected from a set of germline the class I versus class II problem might lie partly in the genes. Only the junction regions between each of these coexpression of CD8 versus CD4 by T cells since the segments are not germline encoded (Davis and Bjorkbinding of CD8 to class I and CD4 to class II certainly man, 1988). On the whole, these genes do not vary draincreases the avidity with which an individual T cell matically between different members of a species. reacts with target cells bearing class I or class II. Thus, each individual has the ability to produce T cells However, this cannot be the only solution to the probbearing approximately the same (albeit very large) collem since T cells are known which can react with their lection of TCRs. MHC + peptide target independently of their coexpres-In contrast, the MHC proteins to which these TCRs sion of CD8 or CD4 (Goldrath and Bevan, 1999; Starr et bind vary dramatically in sequence from one individual al., 2002; von Boehmer et al., 2003). Likewise, for some of a species to another. For example, in man, more than rare T cells, expression of CD4 or CD8 is not concordant with the class of MHC with which they react (Kirberg et al., 1994). In extreme examples of this unpre-*Correspondence: marrackp@njc.org Cell 248