The evolution of recognition systems that initiate destruction of ‘non-self’ material obviously brings with it the need for safeguards to prevent damage to ‘self’. This is a particularly acute problem for the adaptive immune system, because the production of T-cell and B-cell receptors involves an element of random gene rearrangement (see Figs 12 and 13), and therefore lymphocytes with receptors directed at ‘self’ will inevitably emerge in each individual. Furthermore, ‘self’ for one individual is not always the same as ‘self’ for another. For example, people of blood group A have red cells that carry antigen A but make antibodies to blood group B, and vice versa. The AB child of an A father and a B mother inherits the ability to make both anti-B and anti-A antibodies but must not make either, i.e. it must be tolerant to A and B. Adaptive immunity, both B and T cell, in fact protects itself against possible self-reactivity at several stages (as shown in the figure). It used to be assumed that elimination of potentially self-reactive clones (negative selection) was the basis of all unresponsiveness to self, but many other regulatory mechanisms are now recognized. Nevertheless, self-tolerance is not absolute, and in some cases failure may lead to self-destructive immune responses (see Fig. 38).
In certain circumstances, normally antigenic ‘non-self’ materials can trigger these safeguarding mechanisms, a state known as induced tolerance, which might be very undesirable in some infections but very useful in the case of an organ transplant. The mechanisms involved in induced tolerance are likely to be very similar to those that maintain self-tolerance. Note that tolerance is by definition antigen specific, and quite distinct from the non-specific unresponsiveness induced by damage to the immune system as a whole, which is instead described as immunodeficiency (see Fig. 33).
Clonal elimination A cornerstone of Burnet’s clonal selection theory (1959) was the prediction that lymphocytes were individually restricted in their recognition of antigen and that self-recognizing ones were eliminated early in life in the primary lymphoid organs. This is achieved for T cells by negative selection in the thymus (see Fig. 16), and for B cells in the bone marrow. Negative selection was first demonstrated convincingly for superantigens, such as those expressed by some mice endogenous retroviruses, because these delete a substantial proportion of T cells in the thymus. Neither B-cell nor T-cell deletion during development is complete, necessitating the existence of mechanisms of tolerance induction outside the primary lymphoid organs (peripheral tolerance).
Immunological ignorance Some antigens (e.g. those in the chamber of the eye) do not normally induce self-reactivity, simply because they never come into contact with cells of the normal immune system. This phenomenon is known as immunological ignorance. However, if the normal barriers are broken down, e.g. following injury or during a prolonged infection, these antigens can escape into the blood, and self-reactivity and damage of the tissue sometimes results.
Dendritic cells are thought to exist in both immature and mature states. Immature dendritic cells express MHC molecules but lack a full complement of costimulatory molecules such as CD80/86 or CD40 (see Fig. 18). Dendritic cells carry pattern recognition receptors (PRR; see Fig. 5), which recognize microbial products (such as the cell surface of bacteria) and trigger maturation. The processing and presentation of antigens, whether they be ‘self’ or ‘non-self’, by immature dendritic cells is thought to deliver a negative signal to T cells, and hence induce tolerance. In contrast, antigen presentation by mature dendritic cells results in full T-cell activation. The danger hypothesis postulates that both self-antigens and foreign antigens, administered in the absence of inflammation or pathogen-derived maturation stimuli, trigger tolerance. The hypothesis explains the old observation that soluble antigen is less immunogenic and more ‘tolerogenic’ than antigen administered in the presence of adjuvants, because it does not activate antigen-presenting cells to express the appropriate costimulatory molecules.
Negative signalling in T cells T cells express a number of molecules on their surface that transmit negative rather than activating signals. Engagement of these molecules (e.g. CTLA4, PD1) by ligands on the antigen-presenting cell surface serves to control and limit normal immune responses to prevent accidental collateral damage to self- tissues. However, this action may also limit the efficacy of an immune response, e.g. during chronic viral or bacterial infection or cancer. Antibodies to these molecules have shown promise for their ability to improve immune responses in these diseases, but the price may be the risk of some autoimmunity.
B-cell receptors (immunoglobulin) Exposure of B cells to high concentrations of antigen during their development leads to either clonal elimination (death of the B cell). B cells against self-antigens present at low concentrations (less than 10−5 mol/L) survive, but are never normally activated because they require help from T cells to trigger antibody secretion. This mechanism also guards against mature B cells that subsequently change their specificity because of somatic mutation of their V genes (see Figs 13 and 19) during an immune response. Thus, B-cell tolerance is determined by both ‘central’ tolerance (clonal deletion) and ‘peripheral’ tolerance (T-cell regulated).
T-cell receptors pass through an important selection process as they appear in the thymus (see Fig. 16), in which cells with receptors that have a sufficiently high affinity for self-peptides presented by thymic dendritic cells die by apoptosis and are therefore clonally deleted. Using transgenic technology, it is possible to create mice in which all B or T cells carry receptors of a single antigenic specificity. Despite the limitations of studying such artificial systems, these mice have been very important in clearly demonstrating clonal elimination and/ or clonal anergy.
Regulatory T cells (TREG, formerly known as suppressor T cells) TREG cells that inhibit self-reactive lymphocytes are believed to differentiate during thymic development, and are characterized by the expression of CD4, CD25 (one chain of the IL-2 receptor) and a transcription factor, FoxP3. Elimination of these subpopulations of cells, either experimentally or genetically, leads to the development of widespread autoimmunity, emphasizing the importance of these cells in maintain- ing normal ‘self’ tolerance. Other types of TREG can be induced, e.g. by administering antigens via the oral route, or by delivering repeated small doses of antigen. Regulatory or suppressive B cells have also been demonstrated. The mechanisms whereby regulatory cells inhibit their target (which is usually a TH) can include the release of the inhibitory cytokines IL-10 and TGF-β, but other less understood mechanisms probably contribute. The balance between TH and TREG probably determines the eventual outcome of most immune responses and there is enormous interest in trying to expand populations of antigen-specific TREG therapeutically so as to limit damaging autoimmune diseases (see Fig. 38).
Fetal (or neonatal) administration of antigen was the first method shown to induce tolerance. It probably operates by a combination of clonal elimination and deficient antigen presentation, due perhaps to antigen-presenting cell immaturity, although fetal B cells may also be particularly tolerizable because of differences in the way their Ig receptors are replaced (see above). There is some evidence that α-fetoprotein, a major serum protein in the fetus, can inhibit self-reactive T cells.
Oral route Antigens absorbed through the gut are first ‘seen’ by liver macrophages, which remove immunogenic aggregates, etc., leaving only soluble ‘tolerogen’. In addition, antigen-presenting cells in the gut may be specialized for tolerance induction, to prevent immune responses against food. The gut epithelium contains large numbers of TREG expressing suppressive cytokines such as IL-10 and TGF-β.
Antibody-induced tolerance Antibodies against some molecules on the surface of either T cells or antigen-presenting cells can help to induce a state of tolerance. Tolerance induced in this way is sometimes known as enhancement, from the ability to enhance the growth of tumours, transplants, etc. Antibodies to the CD4 molecule are particularly effective at inducing T-cell tolerance to antigens given at the same time.
High doses of antigen are usually more tolerogenic, although repeated low doses can also induce tolerance in T cells. As a rule, T-cell tolerance is easier to induce and lasts longer than B-cell tolerance.
Antigen suicide Antigens coupled to toxic drugs, radioisotopes, etc. may home in on specific B cells and kill them without exposing other cells to danger. A similar principle has been tried to eliminate tumour cells using toxins coupled to antibodies (see Fig. 42).