NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Janeway CA Jr, Travers P, Walport M, et al. Immunobiology: The Immune System in Health and Disease. 5th edition. New York: Garland Science; 2001.

  • By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Cover of Immunobiology

Immunobiology: The Immune System in Health and Disease. 5th edition.

Show details

Self-tolerance and its loss

Tolerance to self is acquired by clonal deletion or inactivation of developing lymphocytes. Tolerance to antigens expressed by grafted tissues can be induced artificially, but it is very difficult to establish once a full repertoire of functional B and T lymphocytes has been produced, which occurs during fetal life in humans and around the time of birth in mice. We have already discussed the two important mechanisms of self-tolerance—clonal deletion by ubiquitous self antigens and clonal inactivation by tissue-specific antigens presented in the absence of co-stimulatory signals (see Chapters 7-8). These processes were first discovered by studying tolerance to nonself, where the absence of tolerance could be studied in the form of graft rejection. In this section, we will consider tolerance to self and tolerance to nonself as two aspects of the same basic mechanisms. These mechanisms consist of direct induction of tolerance in the periphery, either by deletion or by anergy. There is also a state referred to as immunological ignorance, in which T cells or B cells coexist with antigen without being affected by it. Finally, there are mechanisms of tolerance that involve T-cell-T-cell interactions, known variously as immune deviation or immune suppression. In an attempt to understand the related phenomena of autoimmunity and graft rejection, we also examine instances where tolerance to self is lost.

13-25. Many autoantigens are not so abundantly expressed that they induce clonal deletion or anergy but are not so rare as to escape recognition entirely

We saw in Chapter 7 that clonal deletion removes immature T cells that recognize ubiquitous self antigens and in Chapter 8 that antigens expressed abundantly in the periphery induce anergy or clonal deletion in lymphocytes that encounter them on tissue cells. Most self proteins are expressed at levels that are too low to serve as targets for T-cell recognition and thus cannot serve as autoantigens. It is likely that very few self proteins contain peptides that are presented by a given MHC molecule at a level sufficiently high to be recognized by effector T cells but too low to induce tolerance. T cells able to recognize these rare antigens will be present in the individual but will not normally be activated. This is because their receptors only bind self peptides with very low affinity, or because they are exposed to levels of self peptide that are too low to deliver any signal to the T cell. Such T cells are said to be in a state of immunological ignorance. This state has been demonstrated experimentally using transgenic animals in which ovalbumin was expressed at high or very low concentrations in the pancreas. Lymphocytes reactive to ovalbumin were transferred to these animals. The lymphocytes transferred to animals expressing high levels of ovalbumin proliferated in response to ovalbumin presented by antigen-presenting cells and then died. In contrast, the lymphocytes transferred to animals expressing very low levels of pancreatic ovalbumin did not divide but persisted and could be stimulated normally when exposed to high levels of ovalbumin in vitro (Fig. 13.31).

Figure 13.31. T cells are ignorant of very low levels of autoantigen.

Figure 13.31

T cells are ignorant of very low levels of autoantigen. . Transgenic mice were developed that expressed ovalbumin in the pancreas at high or very low levels. CD8 lymphocytes specific for ovalbumin were injected into these mice. After 3 days, the regional lymph (more...)

In the organ-specific autoimmune diseases such as type I IDDM and Hashimoto's thyroiditis, autoimmunity is unlikely to reflect a general failure of the main mechanisms of tolerance—clonal deletion and clonal inactivation. For example, clonal deletion of developing lymphocytes mediates tolerance to self MHC molecules. If such tolerance were not induced, the reactions to self tissues would be similar to those seen in graft-versus-host disease (see Section 13-21). To estimate the impact of clonal deletion on the developing T-cell repertoire, we should remember that the frequency of T cells able to respond to any set of nonself MHC molecules can be as high as 10% (see Section 5-14), yet responses to self MHC antigens are not detected in naturally self-tolerant individuals. Moreover, mice given bone marrow cells from a foreign donor at birth, before significant numbers of mature T cells have appeared, can be rendered fully and permanently tolerant to the bone marrow donor's tissues, provided that the bone marrow donor's cells continue to be produced so as to induce tolerance in each new cohort of developing T cells (Fig. 13.32). This experiment, performed by Medawar, validated Burnet's prediction that developing lymphocytes collectively carrying a complete repertoire of receptors must be purged of self-reactive cells before they achieve functional maturity; it won them a Nobel Prize.

Figure 13.32. Tolerance to allogeneic skin can be established in bone marrow chimeric mice.

Figure 13.32

Tolerance to allogeneic skin can be established in bone marrow chimeric mice. If mice are injected with allogeneic bone marrow at birth (top panel) before they achieve immune competence, they become chimeric, with T cells and antigen-presenting cells (more...)

Clonal deletion reliably removes all T cells that can mount aggressive responses against self MHC molecules; the organ-specific autoimmune diseases, which involve rare T-cell responses to a particular self peptide bound to a self MHC molecule, are therefore unlikely to reflect a general failure in clonal deletion; nor are they likely to be caused by a random failure in the mechanisms responsible for anergy. Rather, the lymphocytes that mediate autoimmune responses seem not to be subject to clonal deletion or inactivation. Such autoreactive cells are present in all of us, but they do not normally cause autoimmunity because they are activated only under special circumstances.

A striking demonstration that autoreactive T cells can be present in healthy individuals comes from a strain of mice carrying transgenes encoding an autoreactive T-cell receptor specific for a peptide of myelin basic protein bound to self MHC class II molecules. The autoreactive receptor is present on every T cell, yet the mice are healthy unless their T cells are activated. As the level of the specific peptide:MHC class II complex is low except in the central nervous system, a site not visited by naive T cells, the autoreactive T cells remain in a state of immunological ignorance. When these T cells are activated, for example, by deliberate immunization with myelin basic protein, as in EAE, they migrate into all tissues, including the central nervous system, where they recognize their myelin basic protein:MHC class II ligand. Recognition triggers cytokine production by the activated T cells, causing inflammation in the brain and the destruction of myelin and neurons that ultimately causes the paralysis in EAE.

It is likely that only a small fraction of proteins will be able to serve as autoantigens. An autoantigen must be presented by an MHC molecule at a level sufficient for the antigen to be recognized by effector T cells, but must not be presented to naive T cells at a level sufficient to induce tolerance. Many self proteins are expressed at levels too low to be detected even by effector T cells. It has been estimated that we can make approximately 105 proteins of average length 300 amino acids, capable of generating about 3 × 107 distinct self peptides. As there are rarely more than 105 MHC molecules per cell, and as the MHC molecules on a cell must bind 10–100 identical peptides for T-cell recognition to occur, fewer than 10,000 self peptides (<1 in 3000) can be presented by any given antigen-presenting cell at levels detectable by T cells. Thus, most peptides will be presented at levels that are insufficient to engage effector T cells, whereas many of the peptides that can be detected by T cells will be presented at a high enough level to induce clonal deletion or anergy. However, as shown in Fig. 13.33, a few peptides may fail to induce tolerance yet be present at high enough levels to be recognized by effector T cells. Autoreactivity probably arises most frequently when the antigen is expressed selectively in a tissue, as is the case of insulin in the pancreas, rather than ubiquitously, because tissue-specific antigens are less likely to induce clonal deletion of developing T cells in the thymus. The nature of such peptides will vary depending on the MHC genotype of the individual, because MHC polymorphism profoundly affects peptide binding (see Section 5-13). This argument leaves aside the crucial issue of how T cells specific for such autoantigens are activated to become effector T cells, which we will consider later.

Figure 13.33. An autoantigenic peptide will be presented at different levels on different MHC molecules.

Figure 13.33

An autoantigenic peptide will be presented at different levels on different MHC molecules. Peptides bind to different MHC molecules with varying affinities; in this example, a self peptide binds well to MHCa, less well to MHCb, and poorly to all other (more...)

If it is true that only a few peptides can act as autoantigens, then it is not surprising that there are relatively few distinct autoimmune syndromes, and that all individuals with a particular autoimmune disease tend to recognize the same antigens. If all antigens could give rise to autoimmunity, one would expect that different individuals with the same disease might recognize different antigens on the target tissue, which does not seem to be the case. Finally, because the level of autoantigenic peptide presented is determined by polymorphic residues in MHC molecules that govern the affinity of peptide binding, this idea could also explain the association of autoimmune diseases with particular MHC genotypes (see Fig. 13.3).

13-26. The induction of a tissue-specific response requires the presentation of antigen by antigen-presenting cells with co-stimulatory activity

As we learned in Chapter 8, only antigen-presenting cells that express co-stimulatory activity can initiate clonal expansion of T cells—an essential step in all adaptive immune responses, including graft rejection and, presumably, autoimmunity. In tissue grafts, it is the donor antigen-presenting cells in the graft that initially stimulate host T cells, starting the direct allorecognition response that leads to graft rejection. Antigen-presenting cells bearing both graft antigens and co-stimulatory activity travel to regional lymph nodes. Here they are examined by large numbers of naive host T cells and can activate those that bear specific receptors (see Fig. 13.26). Grafts depleted of antigen-presenting cells are tolerated for long periods, but are eventually rejected. This rejection is due to the recipient's T cells responding to graft antigens, both MHC and minor H antigens, after they have been processed and presented by recipient antigen-presenting cells (see Fig. 13.25).

The ability of the recipient's antigen-presenting cells to pick up antigens in tissues and initiate graft rejection may be relevant to the initiation of autoimmune tissue damage as well. Transplantation experiments show that host antigen-presenting cells can stimulate both cytotoxic T-cell responses and inflammatory TH1 responses against the transplanted tissue; thus, tissue antigens can be taken up and presented in conjunction with both MHC class I and class II molecules by antigen-presenting cells. In autoimmunity, tissues may be similarly attacked by MHC class I-restricted cytotoxic T cells or injured by inflammatory damage mediated by TH1 cells, as a consequence of the uptake and presentation of tissue antigens by such antigen-presenting cells.

To induce a response to tissue antigens, the antigen-presenting cell must express co-stimulatory activity. As we saw in Chapters 2 and 8, the expression of co-stimulatory molecules in antigen-presenting cells is regulated to occur in response to infection. Transient autoimmune responses are seen in the context of such events, and it is thought that one trigger for autoimmunity is infection. This is discussed further in Section 13-31.

13-27. In the absence of co-stimulation, tolerance is induced

Activation of naive T cells requires interaction with cells expressing both the appropriate peptide:MHC complex and co-stimulatory molecules; in the absence of full co-stimulation, specific antigen recognition leads to partial T-cell activation, leading to T-cell anergy or deletion (see Section 8-11). Tissue cells are not known to express B7 or other co-stimulatory molecules, and can therefore induce tolerance. Experiments with transgenes show that the expression of foreign antigens in peripheral tissues can in some cases induce tolerance, whereas in other cases the foreign antigen seems not to be presented to naive T cells at a sufficient level and is ignored (see Fig. 13.31). Autoimmunity can be induced by coexpression of a foreign antigen and B7 in the same target tissue, but as B7 expression on peripheral tissue cells is not by itself a sufficient stimulus for autoimmunity, it is clear that the loss of tolerance to self tissues requires the coexpression of both a suitable target antigen and co-stimulatory molecules. As discussed in Section 13-25, antigens that are unable to induce clonal anergy or deletion, but that can nonetheless act as targets for effector T cells, can serve as autoantigens; these antigens are likely to be tissue-specific and relatively few.

By analogy with graft rejection, it seems likely that autoimmunity is initiated when a professional antigen-presenting cell picks up a tissue-specific autoantigen and migrates to the regional lymph node, where it is induced to express co-stimulatory activity. Once an autoantigen is expressed on a cell with co-stimulatory potential, naive ignorant T cells specific for the autoantigen can become activated and can home to the tissues, where they interact with their target antigens. At this stage, the absence of co-stimulatory molecules on tissue cells that present the autoantigen can again limit the response. Armed effector T cells kill only a limited number of antigen-expressing tissue cells if these lack co-stimulatory activity; after killing a few targets, the effector cell dies. Thus, not only can responses not be initiated in the absence of co-stimulatory activity, they also cannot be sustained. Therefore, in addition to the question of how autoimmunity is avoided, we must ask: Why does it ever occur? How are responses to self initiated, and how they are sustained?

13-28. Dominant immune suppression can be demonstrated in models of tolerance and can affect the course of autoimmune disease

In some models of tolerance, it can be shown that specific T cells actively suppress the actions of other T cells that can cause tissue damage. Tolerance in these cases is dominant in that it can be transferred by T cells, which are usually called suppressor T cells or regulatory T cells. Furthermore, depletion of the suppressor T cells leads to aggravated responses to self or graft antigens. Although it is clear that immune suppression exists, the mechanisms responsible are highly controversial. Here, we will examine the phenomenon in several animal models.

Neonatal rats can be rendered tolerant to allogeneic skin grafts by prior injection of allogeneic bone marrow. This tolerance is highly specific and can be transferred to normal adult recipient rats. This shows that tolerance in this model is dominant and active, as the transferred cells prevent the lymphocytes of the recipient from mediating graft rejection. In order to transfer this tolerance, cells of both the allogeneic graft donor and the neonatal tolerized host must be transferred. Removal of either cell type abolishes the transfer of tolerance.

This finding is reminiscent of Medawar's studies of tolerance in neonatal bone marrow chimeric mice discussed in Section 13-25. In both cases, even injection of massive numbers of normal syngeneic lymphocytes, which would react vigorously against the foreign cells in the normal environment of the cell donor, did not break tolerance. Tolerance could be broken only by alloreactive cells from an animal that had been immunized with cells from the allogeneic donor before transfer; such cells probably break tolerance by killing the allogeneic donor cells. Thus, an active host response prevents graft rejection in this model. The tolerance is specific for cells of the original donor, and so the suppression must also be specific.

When mice transgenic for a T-cell receptor specific for myelin basic protein are crossed with RAG-/- mice, they spontaneously develop EAE. TCR transgenic mice that have functional RAG genes are able to rearrange their endogenous TCRα chain genes. Since TCRα chain expression is not allelically excluded (see Section 7-16), many T cells in such TCR transgenic mice nevertheless express receptors containing endogenous TCRα chains and have a diverse repertoire. In the RAG-/- mice, no such rearrangements can occur, and the only T-cell receptor expressed is encoded by the transgenes. The observation that, when the background population of diverse T cells is lost, the mice develop an autoimmune disease suggests that this population contains cells normally capable of suppressing the autoimmune disease. Such cells have been shown in an increasing number of autoimmune diseases, and their isolation as cloned T-cell lines is a major goal for people who study the induction of autoimmunity. The reason for this renewed interest in T cell-mediated regulation is that the intentional induction of such cells could be a major advance in the prevention of autoimmune disease.

The mechanisms of tolerance in these animal models are not fully understood. There is evidence for the existence of CD4-positive regulatory cells that can inhibit autoimmune disease. These cells have an activated phenotype and can be identified by the expression of CD25, the IL-2 receptor α chain. Depletion of these cells from the peripheral immune system of normal mice leads to the development of insulin-dependent diabetes, thyroiditis, and gastritis. Similarly, if thymocytes depleted of CD25+ CD4 cells are adoptively transferred into athymic recipients, autoimmune disease results. It is not known how these cells mediate their effects, but one possibility is by secreting cytokines that inhibit other lymphocytes.

In the NOD mouse model for type I IDDM, transfer of a particular insulin-specific T-cell clone can prevent the destruction of pancreatic β cells by autoreactive T cells. This suggests that the insulin-specific T cells can suppress the activity of other autoaggressive T cells in an antigen-dependent manner. They do this by homing to the islet, where they react with insulin peptides presented on the macrophages or dendritic cells. This stimulates the secretion of cytokines, prominent among which is TGF-β, a known immunosuppressive cytokine. There are interesting hints that such cells naturally affect the course of the autoimmune response that causes human diabetes; β-cell destruction in humans occurs over a period of several years before diabetes is manifest, yet when new islets are transplanted from an identical twin into his or her diabetic sibling, they are destroyed within weeks. This suggests that, in the normal course of the disease, specific T cells protect the β-cells from attack by effector T cells and the disease therefore progresses slowly. It might be that after the host islets have been destroyed, these protective mechanisms decline in activity but the effector cells responsible for β-cell destruction do not.

If specific suppression of autoimmune responses could be elicited at will, autoimmunity would not be a problem. Feeding with specific antigen is known to elicit a local immune response in the intestinal mucosa, and responses to the same antigen given subsequently by a systemic route are suppressed (see Section 10-20). This response has been exploited in experimental autoimmune diseases by feeding proteins from target tissues to mice. Mice fed with insulin are protected from diabetes; mice fed with myelin basic protein are resistant to EAE (Fig. 13.34).

Figure 13.34. Antigen given orally can lead to protection against autoimmune disease.

Figure 13.34

Antigen given orally can lead to protection against autoimmune disease. Experimental allergic encephalomyelitis (EAE) is induced in mice by immunization with spinal cord homogenate in complete Freund's adjuvant (upper panels); the disease is mediated (more...)

EAE is usually caused by TH1 cells that produce IFN-γ in response to myelin basic protein. In mice fed this protein, CD4 T cells found in the brain instead produce cytokines such as TGF-β and IL-4 (see Fig. 13.34). TGF-β, in particular, suppresses the function of inflammatory TH1 lymphocytes. In these cases, the protection seems to be tissue-specific rather than antigen-specific. Thus, feeding with myelin basic protein will protect against EAE elicited by immunization with other brain antigens. Feeding with antigen might induce the production of T cells producing TGF-β and IL-4 because these cytokines are also required for IgA production against antigens ingested in food. If feeding with antigen works as a clinical therapy, it would have the advantage over treatments with immunosuppressive drugs that it does not alter the general immune competence of the host. Unfortunately, early studies of this approach to treatment in humans with multiple sclerosis or rheumatoid arthritis have shown minimal, if any, benefit. These strategies may prove more successful in preventing the onset of disease than reversing established disease. However, this approach requires the identification of patients at the very onset of disease or those who are at extremely high risk of developing disease and adds impetus to studies to identify the disease susceptibility genes for the development of autoimmune diseases.

Another strategy for controlling immune responses is to manipulate the cytokine balance that determines whether a CD4 T-cell response is predominantly TH1 or TH2. It is possible experimentally to switch TH1 to TH2 responses by administration of IL-4, and vice versa using IFN-γ. This is known as immune deviation and is discussed further in Chapter 14.

Unlike human diabetes, which follows a chronic progressive course in humans, multiple sclerosis is a chronic relapsing disease with acute episodes followed by periods of quiescence. This suggests a balance between autoimmune and protective T cells that can alter at different stages of the disease. However, it remains to be proven whether the specific suppressive cells discussed in this section exist naturally and contribute to self-tolerance, or whether they arise only upon artificial stimulation or in response to autoimmune attack. Nevertheless, because they can play an active, dominant part in self-tolerance, they are particularly attractive potential mediators for immunotherapy of autoimmune disease.

13-29. Antigens in immunologically privileged sites do not induce immune attack but can serve as targets

Tissue grafts placed in some sites in the body do not elicit immune responses. For instance, the brain and the anterior chamber of the eye are sites in which tissues can be grafted without inducing rejection. Such locations are termed immunologically privileged sites (Fig. 13.35). It was originally believed that immunological privilege arose from the failure of antigens to leave privileged sites and induce immune responses. Subsequent studies have shown, however, that antigens do leave immunologically privileged sites, and that these antigens do interact with T cells; but instead of eliciting a destructive immune response, they induce tolerance or a response that is not destructive to the tissue. Immunologically privileged sites seem to be unusual in three ways. First, the communication between the privileged site and the body is atypical in that extracellular fluid in these sites does not pass through conventional lymphatics, although proteins placed in these sites do leave them and can have immunological effects. Naive lymphocytes, similarly, may be excluded by the tissue barriers of privileged sites, such as the blood-brain barrier. Second, humoral factors, presumably cytokines, that affect the course of an immune response are produced in privileged sites and leave them together with antigens. The anti-inflammatory cytokine TGF-β seems to be particularly important in this regard: antigens mixed with TGF-β seem to induce mainly T-cell responses that do not damage tissues, such as TH2 rather than TH1 responses. Third, the expression of Fas ligand by the tissues of immunologically privileged sites may provide a further level of protection by inducing apoptosis of Fas-bearing lymphocytes that enter these sites. This mechanism of protection is not fully understood, as it appears that under some circumstances the expression of Fas ligand by tissues may trigger an inflammatory response by neutrophils.

Figure 13.35. Some body sites are immunologically privileged.

Figure 13.35

Some body sites are immunologically privileged. Tissue grafts placed in these sites often last indefinitely, and antigens placed in these sites do not elicit destructive immune responses.

Paradoxically, the antigens sequestered in immunologically privileged sites are often the targets of autoimmune attack; for example, brain autoantigens such as myelin basic protein are targeted in multiple sclerosis. It is therefore clear that this antigen does not induce tolerance due to clonal deletion of the self-reactive T cells. Mice only become diseased when they are deliberately immunized with myelin basic protein, in which case they become acutely sick, show severe infiltration of the brain with specific TH1 cells, and often die.

Thus, at least some antigens expressed in immunologically privileged sites induce neither tolerance nor activation, but if activation is induced elsewhere they can become targets for autoimmune attack (see Section 13-25). It seems plausible that T cells specific for antigens sequestered in immunologically privileged sites are more likely to remain in a state of immunological ignorance. Further evidence for this comes from the eye disease sympathetic ophthalmia (Fig. 13.36). If one eye is ruptured by a blow or other trauma, an autoimmune response to eye proteins can occur, although this happens only rarely. Once the response is induced, it often attacks both eyes. Immuno-suppression and removal of the damaged eye, the source of antigen, is frequently required to preserve vision in the undamaged eye.

Figure 13.36. Damage to an immuno-logically privileged site can induce an autoimmune response.

Figure 13.36

Damage to an immuno-logically privileged site can induce an autoimmune response. In the disease sympathetic ophthalmia, trauma to one eye releases the sequestered eye antigens into the surrounding tissues, making them accessible to T cells. The effector (more...)

It is not surprising that effector T cells can enter immunologically privileged sites: such sites can become infected, and effector cells must be able to enter these sites during infection. As we learned in Chapter 10, effector T cells enter most or all tissues after activation, but accumulations of cells are seen only when antigen is recognized in the site, triggering the production of cytokines that alter tissue barriers.

13-30. B cells with receptors specific for peripheral autoantigens are held in check by a variety of mechanisms

During B-cell development in the bone marrow, B-cell antigen receptors specific for self molecules are produced as a consequence of the random generation of the repertoire. If a self molecule is expressed in the bone marrow in an appropriate form, clonal deletion and receptor editing can remove all of these self-reactive B cells while they are still immature (see Sections 7-9 and 7-10). There are, however, many self molecules available only in the periphery whose expression is restricted to particular organs. An example is thyroglobulin (the precursor of thyroxine), which is expressed only in the thyroid and at extremely low levels in plasma. Back-up mechanisms exist to ensure that B cells reactive to these self molecules do not cause autoimmune disease. When a mature B cell in the periphery encounters self molecules that bind its receptor, four proposed mechanisms could bring about the observed nonreactivity. Failure of any one of these mechanisms could lead to autoimmunity.

First, B cells that recognize a self antigen arrest their migration in the T-cell zone of peripheral lymphoid tissues (Fig. 13.37), just like B cells that bind a foreign antigen (see Chapter 9). However, in contrast to the response to foreign antigens, in which activated effector CD4 T cells are present, B cells binding self antigens will not be able to interact with helper CD4 T cells because normally no such cells exist for self antigens. This lack of interaction prevents the B cells from migrating out of the T-cell zones into the follicles; instead, the trapped self-reactive B cells undergo apoptosis.

Figure 13.37. Autoreactive B cells do not compete effectively in peripheral lymphoid tissue to enter primary lymphoid follicles.

Figure 13.37

Autoreactive B cells do not compete effectively in peripheral lymphoid tissue to enter primary lymphoid follicles. In the top panel, B cells are seen entering the T-cell zone of a lymph node through high endothelial venules (HEVs). Those with reactivity (more...)

A second mechanism is the induction of B-cell anergy, which is associated with downregulation of surface IgM expression and partial inhibition of the linked signaling pathways (Fig. 13.38). B-cell anergy can be induced by exposure to soluble circulating antigen (see Section 7-17); if mice are inoculated intravenously with protein solutions from which all protein aggregates have been rigorously removed to eliminate multivalent complexes, the peripheral B cells that bind these proteins can be inactivated. The lifespan of anergic B cells is short as they are usually eliminated after failing to enter the primary lymphoid follicles or after interacting with antigen-specific T cells, as described below. This form of B-cell tolerance can therefore be viewed as a form of delayed B-cell deletion, with the significant difference that there may be autoimmune diseases such as SLE in which anergic cells can be rescued.

Figure 13.38. Peripheral B-cell anergy.

Figure 13.38

Peripheral B-cell anergy. An autoreactive B cell encounters its soluble autoantigen in the periphery (first panel), which leads to the development of B-cell anergy (second panel). This is characterized by a reduction in both the expression of surface (more...)

The third mechanism depends on the presence of T cells that are specific for the self antigen and express Fas ligand. In the rare instances when such an autoreactive T cell matures and is activated, it is able to kill autoreactive anergic B cells in a Fas-dependent manner (see Fig. 13.38). In the absence of the normal pathways of co-stimulation, anergized B lymphocytes that have been chronically exposed to self antigen show enhanced sensitivity to apoptosis after ligation of Fas by Fas ligand. They are therefore not subject to the stimulatory signals that oppose apoptosis in B cells whose surface receptors have just been ligated by foreign antigen. The importance of this mechanism is nicely illustrated by the consequences of mutation in the genes for Fas or Fas ligand. Mice deficient in Fas or Fas ligand develop severe autoimmune disease, similar to SLE, associated with the production of a similar spectrum of autoantibodies.

Finally, there is evidence for a distinct mechanism for dealing with B cells that develop self-reactive specificities as a result of somatic hypermutation during a response to a foreign antigen (Fig. 13.39). At a crucial phase at the height of the germinal center reaction, an encounter with a large dose of soluble antigen causes a wave of apoptosis in germinal center B cells within a few hours. Thus B cells that become reactive for abundant soluble self antigens could be removed.

Figure 13.39. Elimination of autoreactive B cells in germinal centers.

Figure 13.39

Elimination of autoreactive B cells in germinal centers. During somatic hypermutation in germinal centers (depicted in the top panel), B cells with autoreactive B-cell receptors can arise. Ligation of these receptors by soluble autoantigen induces apoptosis of (more...)

All these mechanisms reemphasize the fact that the mere existence in the body of some B lymphocytes with receptor specificities directed against self is not in itself harmful. Before an immune response can be initiated they need to receive effective help, the B-cell receptors must be ligated, and their intracellular signaling machinery must be set to respond normally.

13-31. Autoimmunity may sometimes be triggered by infection

Human autoimmune diseases often appear gradually, making it difficult to find out how the process is initiated. Nevertheless, there is a strong suspicion that infections can trigger autoimmune disease in genetically susceptible individuals. Indeed, many experimental autoimmune diseases are induced by mixing tissue cells with adjuvants that contain bacteria. For example, to induce EAE, the spinal cord or myelin basic protein used for immunization must be emulsified in complete Freund's adjuvant, which includes killed Mycobacterium tuberculosis (see Appendix I, Section A-4). When the mycobacteria are omitted from the adjuvant, not only is no disease elicited, but the animals become refractory to any subsequent attempt to induce the disease by antigen in complete Freund's adjuvant, and this resistance can be transferred to syngeneic recipients by T cells (Fig. 13.40). Infection is important in the induction of disease in several other known cases. For example, transgenic mice that express a T-cell receptor specific for myelin basic protein (see Section 13-25) often develop spontaneous autoimmune encephalo-myelitis if they become infected. One possible mechanism for this loss of tolerance is that the infectious agents induce co-stimulatory activity on antigen-presenting cells expressing low levels of peptides from myelin basic protein, thus activating the autoreactive T cells.

Figure 13.40. Bacterial adjuvants are required for induction of experimental autoimmune disease.

Figure 13.40

Bacterial adjuvants are required for induction of experimental autoimmune disease. Mice immunized with spinal cord homogenate in complete Freund's adjuvant, which contains large numbers of killed Mycobacterium tuberculosis, get experimental allergic encephalomyelitis (more...)

It has also been suggested that autoimmunity can be initiated by a mechanism known as molecular mimicry, in which antibodies or T cells generated in response to an infectious agent cross-react with self antigens. To show that infectious agents can trigger responses that can destroy tissues, mice were made transgenic for a viral nuclear protein driven by the insulin promoter, so that the protein was expressed only in pancreatic β cells. As the amount of viral protein expressed was low, the T cells that recognized this protein remained immunologically ignorant. That is, they were neither tolerant to the viral protein nor activated by it, and the animals showed no sign of disease. If they were infected with the live virus, however, they responded by making cytotoxic CD8 T cells specific for the viral protein, and these armed CD8 cytotoxic T cells could then destroy the β cells, causing diabetes (Fig. 13.41).

Figure 13.41. Virus infection can break tolerance to a transgenic viral protein expressed in pancreatic β cells.

Figure 13.41

Virus infection can break tolerance to a transgenic viral protein expressed in pancreatic β cells. Mice made transgenic for the lymphocytic choriomeningitis virus (LCMV) nucleoprotein under the control of the rat insulin promoter express the nucleoprotein (more...)

It has long been known that molecular mimicry can operate in antibody-mediated autoimmunity; microbial antigens can elicit antibody responses that react not only with the antigens on the pathogen but also with host antigens of similar structure. This type of response occurs after infection with some Streptococcus species. These elicit antibodies that cross-react with kidney, joint, and heart antigens to produce rheumatic fever. Such responses are usually transient and do not lead to sustained autoantibody production, as the helper T cells involved are specific for the microbe and not for self proteins. Host proteins that form a complex with bacteria can induce a similar transient response; in this case, the antibody response is not cross-reactive but the bacterium is acting as a carrier, allowing B cells that express an autoreactive receptor to receive inappropriate T-cell help. These and some other mechanisms that could allow an infectious agent to break tolerance are summarized in Fig. 13.42. All of these mechanisms can be shown to act in experimental systems, and there is some evidence for their importance in human autoimmune disease as well.

Figure 13.42. There are several ways in which infectious agents could break self-tolerance.

Figure 13.42

There are several ways in which infectious agents could break self-tolerance. Because some antigens are sequestered from the circulation, either behind a tissue barrier or within the cell, an infection that breaks cell and tissue barriers might expose (more...)

The argument that some autoimmune diseases might be initiated by infection is strengthened by the fact that there are several human autoimmune diseases in which a prior infection with a specific agent or class of agents leads to a particular disease (Fig. 13.43). Disease susceptibility in these cases is determined largely by MHC genotype.

Figure 13.43. Association of infection with autoimmune diseases.

Figure 13.43

Association of infection with autoimmune diseases. Several autoimmune diseases occur after specific infections and are presumably triggered by the infection. The case of post-streptococcal disease is best known but is now rare because effective antibiotic (more...)

In most autoimmune diseases, however, there is still no firm evidence that a particular infectious agent is associated with disease onset. Furthermore, a number of animal models of autoimmunity show that infection is not necessary for certain diseases to develop. Indeed, in some cases, infection may prevent or delay disease onset. Murine models of SLE and diabetes show that these diseases develop in genetically predisposed animals housed in germ-free conditions. The prevalence of diabetes in nonobese diabetic (NOD) mice is higher in mice raised in pathogen-free environments than in colonies housed in environments where infectious diseases are not rigorously excluded.

In humans, SLE is a very rare disease in African populations living in West Africa. By contrast, SLE has a prevalence of as high as one in 500 African-American women living in the West Coast of the United States. It has been proposed that the high prevalence of infection, particularly by parasites, in West Africa may protect against the development of SLE. In support of this idea, infection of mice genetically predisposed to the development of SLE with murine strains of malaria delays the onset of SLE disease. The mechanism of this protective effect is not known.


Tolerance to self is a normal state that is maintained chiefly by clonal deletion of developing T and B cells and clonal deletion or inactivation of mature peripheral T and B cells. In addition, some antigens are ignored by the immune system, either because they are present at too low a level, or because they are present in immunologically privileged sites. When the state of self-tolerance is disrupted, autoimmunity can result. The process of clonal deletion influences the kinds of autoimmune disease that can occur. One group of autoantigens are those that do not trigger clonal deletion in the thymus, either because they are not abundant enough or because they are tissue-specific and not expressed in the thymus. Autoimmunity to these antigens, such as insulin, causes organ-specific autoimmune diseases such as type I diabetes mellitus. In the second category of autoimmune diseases, the systemic autoimmune diseases such as SLE, tolerance is broken to ubiquitous self antigens. Predisposition to these diseases may be due to inherited abnormalities in the regulation of immune responses and in the waste disposal mechanisms for removing dying cells at sites of inflammation. A third mechanism of self-tolerance—dominant suppression—has been noted in several experimental systems of autoimmunity and graft rejection; if this mechanism could be understood, it might be possible to use it to prevent both graft rejection and autoimmunity, which are closely related problems.

Image CH13F3
Image CH13F26
Image CH13F25

By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.

Copyright © 2001, Garland Science.
Bookshelf ID: NBK27174