B-cell and T-cell receptor

During the course of an infection, B cells can further alter the specificity of the antibody they produce. When a mature B cell meets an antigen that its B-cell receptor recognises (the B-cell receptor comprises the antibody the cell produces anchored on the cell surface) then the B cell can undergo a process called somatic hypermutation. Here an enzyme called activation-induced cytidine deaminase (AID) makes random mutations in the antibody variable region genes. If the mutations result in an antibody that more strongly binds to their targets then these B cells will survive and may differentiate into antibody-producing plasma cells with the new specificity.

B and T lymphocytes are the primary effectors of adaptive immune responses. In mammals B and T cells develop in the bone marrow and thymus, respectively, and upon completing maturation circulate through blood, tissues, and lymphatics. B and T cells produce immune responses by interacting with each other, antigen, and antigen-presenting dendritic cells in secondary lymphoid organs such as spleen and lymph nodes.

There are three essential features of adaptive immune responses: sufficient diversity to deal with a universe of antigens, discrimination of self from non-self, and long-lasting immunologic memory. To explain these phenomena, Burnet and Talmage developed a theory of clonal (or cellular) selection. In their model, antigen would select specific clones from a vast array of immune cells via preexisting antigen receptors. This important idea differed from other theoretical models in that all of the other models proposed that antigen was in some way used as a template to induce antibody specificity. In Burnet's model memory would be provided by expanding the size of an antigen-specific clone, and random mutation would be allowed to enhance affinity. Furthermore, the model proposed that cells with self-reactive receptors would be clonally deleted during development.

It is well known that receptor diversity is generated during lymphocyte development by random combinatorial joining of antigen receptor gene fragments. This random rearrangement of V, D, and J gene segments can produce large numbers of potentially useful receptors, but frequently the receptors are nonfunctional because they are out-of-frame, or occasionally they are self-reactive. Quality control is achieved by feedback signals from developing receptors. Cells that fail to produce an in-frame receptor are never allowed to complete maturation, and until recently it was believed that lymphocytes that develop autoreactive receptors are either deleted or rendered unable to respond to antigen (anergic). Early experiments and work with transgenic mice supported Burnet's idea that clones of cells that develop self-reactive receptors are deleted and that occasional clones that survive become anergic. However, recent results from a number of laboratories suggest that the immune system practices molecular selection of receptors in addition to clonal selection of lymphocytes. This minireview will deal with this remarkable newly appreciated feature of immune cells, the capacity to revise their antigen receptors prior to cellular selection.

Evidence for Receptor Selection in Developing Lymphocytes

Immunoglobulin receptor selection was first detected in the bone marrow of transgenic mice that carry self-reactive antibody. Instead of the expected clonal deletion of all self-reactive cells in these mice, occasionally B lymphocytes were found that had undergone receptor editing: these B cells had deleted their autoreactive receptors and developed entirely new receptors by V(D)J recombination. One interpretation of these findings is that B cells with self-reactive receptors are induced to undergo V(D)J recombination by interacting with antigen. In this model, B cells expressing autoreactive receptors would survive by altering their receptors to produce non-self-reactive antibody. This idea is a serious challenge to Burnet's theory; he stated that “The clonal selection hypothesis could be disproved by showing that cells of a pure clone could, by appropriate manipulations, be induced to produce any one of a variety of antibodies.” . During receptor editing, B cells are “manipulated” by self-antigen to produce “any one of a variety of antibodies” and the non-self-reactive receptor is selected. However, it was difficult to place these early transgenic experiments in a physiologic context because the antibody transgenes were integrated outside their normal genomic locus. Thus, the emergence of non-self-reactive B cells would require transgene inactivation by an undetermined mechanism as well as new receptor assembly. Furthermore, it could be argued that this phenomenon might be due to the selection of randomly occurring mutations and that these mutant clones were the substrates for cellular selection as proposed by Burnet.

Figure 1Receptor Selection Can Precede Clonal Selection in Both Developing and Mature B Cells

Two recent observations suggest that random mutation and selection as envisioned by Burnet cannot entirely account for B cell receptor editing. First, in developing B cells, antigen receptor binding to self-antigen induces new V(D)J recombination. Second, binding to self-antigen induces replacement of the autoreactive receptor genes by non-self-reactive receptors. The latter observation was made using targeted mutation of antibody genes to create strains of mice in which pre-rearranged antigen receptors of known specificity were placed in the appropriate genomic context. In these experiments, antigen-induced recombination in autoreactive B cells resulted in high-efficiency replacement of targeted immunoglobulin heavy and light chain V regions with newly recombined non-self-reactive V genes. Developing self-reactive B cells induced to recombine their antibody genes emerged from the bone marrow with a new set of non-self-reactive receptors.

Similarly, experiments with T cell receptor (TCR) transgenic mice and T cell lines showed continued TCRα gene rearrangements and suggested that receptor selection might also occur during T cell development in the thymus. Continuing TCRα gene rearrangement has now been shown to be an important element in establishing the developing T cell repertoire. In transgenic mice expressing a TCR that recognizes a pigeon cytochrome c peptide, expression of pigeon cytochrome c peptide as a self-antigen results in deletion of transgene-expressing thymocytes. This loss of transgenic anti-self-specific T cells produces a decrease in the number of thymocytes, which is entirely consistent with clonal (cellular) selection as it was originally proposed. However, as with all transgenes, the anti-pigeon cytochrome c TCR is randomly integrated in the genome, and therefore the transgenic TCR cannot be deleted by secondary V(D)J recombination. In contrast to the anti-pigeon cytochrome c TCR transgene, there is no loss of thymocytes when pigeon cytochrome c is expressed as a self-antigen in mice that carry the anti-pigeon cytochrome c TCRα chain introduced into the TCRα locus by gene targeting. Instead of antigen-driven cellular deletion, there appears to be receptor deletion by V(D)J recombination, and the pigeon cytochrome c reactive TCRα chain is replaced by another TCRα chain. The result is a normal thymus in which mature T cells do not express the self-reactive TCRα chain. In the bone marrow, editing appears to be limited to a specific stage in B cell development, and it can be predicted that a corresponding restriction may be found in developing T cells in the thymus. However, in contrast to B cells, the role of antigen in TCR editing is not entirely clear. Exposure to self-antigen may simply arrest T cells at a stage that precedes cellular selection and in which there are continuous TCRα rearrangements.

Although editing and receptor selection were not part of Burnet's model, the clonal selection theory could certainly accommodate receptor editing if receptor selection occurs before cellular selection. Indeed, antigen-driven receptor selection was proposed by Jerne in modeling a mechanism for repertoire diversification. Jerne's model comes very close to predicting receptor editing in suggesting that self-antigen drives clonal diversification during lymphocyte development. In this model clonal (cellular) selection occurs only after molecular selection produces a diversified non-self-reactive group of antigen receptors.

V(D)J Recombination in Peripheral B Lymphocytes

Until recently, receptor editing was thought to be limited to developing B cells in the bone marrow, but new data from several laboratories suggests that antigen receptors can be revised in mature B cells and T cells

The possibility that there is ongoing V(D)J recombination in mature B cells was suggested by the finding that germinal center (GC) B cells express the recombinase activating genes, RAG1 and RAG2, which are required for V(D)J rearrangement. The GC is the anatomical site where antibody-producing cells are clonally expanded, and as predicted by Burnet, GC B cells undergo random somatic point mutation in their antibody genes. Burnet proposed mutation and selective cellular expansion to explain the properties of adjuvants and immunologic memory. “If we have a clone with a reactive site not quite appropriate to determinant D, but sufficiently close for contact to provoke activation and proliferation, occasionally, cells of this clone will multiply more rapidly on the average than other clones. Mutation within the clone will be more likely to occur and any favorable mutation will automatically be strongly favored …” 

Ongoing recombination in mature B cells was confirmed by showing that they contain intermediates of V(D)J recombination and that mature B cells stimulated to express RAGs can change their antibody genes. However, not all GCs contain B cells that express RAGs, and RAG expression is heterogeneous even in RAG-positive GCs. In addition, the number of peripheral B cells that express RAGs has not been determined, and it is not clear whether there are specific subsets of peripheral B cells that undergo V(D)J recombination or whether peripheral V(D)J recombination is limited to the GC.

Finding RAG expression in GCs suggested that V(D)J recombination might be a second mechanism for producing random changes in antibody specificity in expanding clones of B cells. If recombination in mature B cells is random and not regulated by antigen it would be entirely consistent with Burnet's theory. However, three independent lines of evidence suggest that V(D)J recombination in GC B cells is regulated by antigen. First, RAGs are preferentially expressed in GC B cells that have receptors with low affinity for antigen. When GC cells responding to antigen were sorted into subsets expressing high- and low-affinity receptors, RAGs were found in the low-affinity group. Thus, RAGs are expressed in cells with receptors that could be improved by further V(D)J recombination. Second, binding of a low-affinity antigen to immunoglobulin transgenic B cells induced RAG expression and V(D)J recombination in vivo. In contrast, high-affinity antigens did not have this effect. This experiment established that regulation of secondary recombination in the periphery differs from regulation of RAG expression in the developing B cell where high-affinity self-antigen binding induces editing. Third, cross-linking GC B cell antigen receptors turns off RAG expression. Together these three sets of observations suggest a model whereby RAGs are expressed in peripheral B cells exposed to antigens to which their receptors bind at low affinity, and that recombination is turned off by high-affinity antigen binding.

There are four possible outcomes of continuing recombination in mature B cells. Since V(D)J recombination is imprecise, the most likely result of any rearrangement is an out-of-frame gene that fails to encode a receptor, and loss of antibody expression in mature B cells leads to cell death. Therefore, the most frequent result of V(D)J recombination in mature B cells may be cell death. The next most likely outcome of new V(D)J recombination is assembly of a receptor of even lower affinity than the starting antibody. B cells with decreased affinity receptors would be predicted to continue to recombine. Occasionally, a B cell might develop a self-reactive receptor as a result of V gene replacement, and these cells would be expected to be deleted or made anergic if they are exposed to the self-antigen. Certainly, serendipitous assembly of a high-affinity receptor would be a rare event, but the frequency of this event may be similar to the frequency of a random somatic mutation that enhances antibody affinity. Should a high-affinity receptor develop in a cell undergoing peripheral editing, cross-linking by antigen would be expected to turn off recombination, thereby positively selecting the improved receptor. Any high-affinity clone developing by somatic mutation or editing would be expected to be preferentially expanded as suggested by Burnet.

Our understanding of V(D)J recombination in peripheral B cells is limited and several issues remain unresolved. These include: how is V(D)J recombination regulated in the periphery?; can B cells undergo one or several rounds of recombination?; does V(D)J recombination occur outside the selective environment of the GC, and if so what are the consequences of deregulated recombination?; finally and most importantly, what is the role of peripheral recombination in the antibody response?

V(D)J Recombination in Peripheral T Lymphocytes

McMahan and Fink have reported that mature CD4+ T cells in Vβ5 TCR transgenic mice can express RAGs and undergo TCRβ gene rearrangements. V(D)J recombination was not expected in mature T cells because repertoire selection in T cells is fundamentally different from that in B cells. T cells must be positively selected to express receptors that can recognize antigenic peptides bound to self major histocompatibility complex (MHC) molecules in the thymus. Only a small minority of all randomly assembled TCRs are of the right affinity to be positively selected in any given MHC background, and most developing TCRs are discarded. If TCR gene recombination in the periphery follows the same rules as it does in the thymus, then there must be a yet-to-be-defined peripheral mechanism for positive selection.

McMahan and Fink showed that the number of cells expressing transgenic Vβ5 TCR declines with age, whereas the Vβ5−CD4+ T cells that express endogenous Vβ genes increase. To account for the emergence of Vβ5− T cells that express endogenous Vβs, purified Vβ5− T cells were examined for RAG expression and ongoing V(D)J recombination. Vβ5− T cells showed RAG expression and they contained TCRβ chain–specific recombination intermediates, whereas Vβ5+ T cells did not. McMahan and Fink's observation that MHC class II–expressing B cells are required for the emergence of the Vβ5− T cells suggests that B cells could be involved in either induction of recombination or positive selection.

If V(D)J rearrangement in peripheral T cells is not limited to the Vβ5 TCR transgenic model, then this observation has profound implications for understanding immune responses. However, this first report raises many questions and it may be too early to try to fit peripheral T cell editing into a model of immune function. For example, what is the nature of the cells that express RAGs in the periphery? The origin of the Vβ5− T cells is not entirely clear. McMahan and Fink find that only Vβ5+ T cells show active recombination, but if the Vβ5− T cells arise from Vβ5+ T cells, then why aren't the Vβ5+ T cells positive for recombination and why is the emergence of Vβ5− T cells thymus-independent? If TCR gene editing is not limited to transgenic models like the one studied by McMahan and Fink, how is it regulated, and is T cell editing limited to a particular anatomic compartment like the GC or to a specific subset of T cells?

Summary

Secondary antigen receptor gene recombination occurs in developing lymphocytes, and in more mature T and B cells. In the developing lymphocyte, “editing” occurs in response to receptor ligation by autoantigens. As a result of receptor editing, anti-self-reactive cells are converted to non-self-reactive cells, and self-reactive clones are thereby salvaged before negative selection. This antigen-induced change in the receptor specificity was not foreseen in the clonal selection theory. However, editing could be incorporated into the clonal selection theory by limiting receptor selection to a specific stage in lymphocyte development that precedes cellular selection. We know much less about the origin, regulation, or function of V(D)J recombination in mature lymphocytes. Nevertheless, antigen-induced receptor selection is likely to play an important role in shaping immune responses.