It is also possible that the TCR may retain a residual specificity for the self peptide on which it was positively selected. Marrack et al demonstrated vestigial reactivity among mature T cells to the ligand presented on the original selecting thymic epithelial cells, which leads to the question as to whether certain low affinity self peptides that contribute to positive selection act as potential targets during subsequent autoimmune disease.
In these ways, autoimmunity may be better described as a ‘side-step’ of self tolerance rather than a complete breakdown: it is necessary that self antigen is used to select the thymocytes for use in the immune repertoire but it is also necessary that such selected cells do not react against self. The use of fuzzy logic by Leng and Bentwich emphasises this nature of thymocyte selection as a continuum rather than a Boolean system; exemplarily the same peptide of LCMV can induce positive or negative selection of self reactive CD8+ T cells depending on the concentration of Ag. If the T cell affinity for the Ag falls between the two extremities of selection and deletion and it is subsequently activated, or if it retains an affinity for the original selecting ligand, then self tolerance will be compromised.
Cryptic epitopes and autoimmunity
In order to dominate an immune response to a protein antigen a peptide must be readily generated by an antigen presenting cell and expressed at the cell surface bound to a MHC molecule. Pircher et al showed that the epitope binding avidity of a T cell required for induction of intrathymic tolerance is much less than the epitope binding avidity capable of activating a mature T cell. So, in theory, an epitope that precludes even the induction of tolerance should not be able to then elicit an autoimmune response by activating T cells.
However, if avidity of epitope binding is considered as a product of both affinity and frequency then it is possible to compensate for a low epitope binding affinity by increasing the density of presentation at the antigen presenting cell surface which could, for example, be increased by immunization. If these cryptic self determinants, absent or present at subthreshold levels during T cell selection are then presented at a higher concentration then the T cells may become autoaggressive. There are a number of possible routes as to how these cryptic epitopes can become visible to the immune system:
Processing can be increased by receptor down regulation. CD4 can be downregulated by HIV-1 gp120, especially if it is cross linked by α-gp120 Ab. The resultant presentation of cryptic CD4 epitopes on activated T cells (which act as professional antigen presenting cells to prime CD4 T cells) results in CD4 T cell attack of activated T cells and production of α-CD4 Ab by B cells.
Processing can be altered by ligand binding. Ab can modulate the processing of Ag so that the production of some epitopes increases by a factor of 10 to 100. Ab binding can suppress generation of some epitopes and boost processing of others in ‘footprinted’ domains. Such epitopes are only semi cryptic as they can be generated at lower stimulatory levels in the absence of Ab. In this way B cells can produce the boosting Ab, which then binds to the Ag ad produces increased levels of epitope, resulting in T cell activation. As this T cell activation further increases B cell production of Ab, the reaction is dominant and self sustaining.
Cytokines and dominant immune suppression
The balance of the Th1/Th2 immune response is as critical in the progression of autoimmunity as in the non-harmful immune response. The Th1 response propagates autoimmune damage; TNFα mice develop spontaneous inflammatory arthritis, while the Th2 response can be thought of as protective; IL-10-/- mice develop chronic enterocolitis. The use of skin graft mouse models has shown that the production of IL-4, IL-10 and TGFβ by T suppressor cells can ameliorate the autoimmune response. This ‘innocent bystander suppression’ has been shown to prevent EAE even when induced with myelin (Weiner et al). If such specific suppression could be elicited at will, autoimmunity would not be a problem. However, it appears that such suppression already naturally slows the progression of autoimmune disease and attempts to supplement its effect in humans, by feeding with antigen, have shown no benefit. Thus the immune system has developed its own system for the control of thymocytes which as a necessary side effect of selection fall within the fuzzy continuum. When this control mechanism fails, autoimmunity results.
Genetic associations in autoimmunity
A complex combination of susceptibility genes and environmental factors contribute to the development of autoimmunity. As such, many autoimmune diseases have been linked with a specific genetic locus, e.g. HLA-B27 and Ankylosing spondylitis. Wicker et al produced congenic mice on a NOD background that had little or no disease, showing that a genetic defect may dispose toward autoimmune reactions in general rather than limit the response to specific diseases. Involvement has concentrated on, understandably, the possession of MHC and TCR alleles. The MHC bcg gene that has been identified as being involved in IDDM susceptibility is also involved in the resistance/susceptibility to infections such as tuberculosis, leismaniasis and salmonellosis. In EAE the Vβ8 allelic variation of the TCR engages MBP to cause demyelination and subsequent paralysis. Thus a genetic predisposition can aid the development of autoimmune disease but it appears an important factor of the environmental component is that of infection.
Infectious agents and the break of self tolerance
There has long been debate about microbial involvement in the establishment of autoimmune disease. The classical experiment that showed IL-2-/- mice succumbed to irritable bowel syndrome (IBS) whilst their sterile IL-2-/- counterparts were immune suggested that infection was an important environmental factor in the development of autoimmunity. Indeed, the presence of M.tuberculosis in Freund’s adjuvant is absolutely required for the induction of EAE; this acts via the infection of antigen presenting cells, inducing presentation of co-stimulatory molecules and the activation of T cells without inducing anergy.
Bacterial superantigen binding to a polyclonal T cell could also overcome clonal anergy. Molecular mimicry can trigger autoimmune reactions; microbial antigens can elicit antibody responses that react not only with the pathogen but also with host antigens that are similar in structure, e.g. the response against Streptococcus pyogenes M protein can cross react with kidney, joint and heart antigens to produce rheumatic fever. However, such responses are usually transient as the helper T cells are specific for the microbe and not self proteins. Autoreactive B cell escape can occur if host proteins form a complex with bacteria, thus allowing these cells to receive inappropriate T cell help.
Determinant spreading amplifies failures in tolerance
It is often seen in autoimmune disease that after an initial response to self the autoimmune repertoire expands as a result of diversification of T and B cell responses to other antigens at the site of inflammation, some of which may be self antigens. The immune response initially targets a peptide epitope on the target antigen, and then expands to target other peptide antigens on that target antigen (known as intramolecular spreading) and then spreads to different epitopes on different proteins (intermolecular spreading).
Our understanding of this phenomenon has been aided by the study of the NOD mouse model of insulin dependant Diabetes mellitus. Initially the immune response it directed against a limited region of the GAD molecule. It then spreads to additional determinants and then on to heat shock proteins, e.g. HSP60 (tolerisation to HSP60 reverses the disease process in NOD mice). This emphasises the conserved nature of many structures between non self and self and confuses the distinction made between an effective immune response and the breakdown of self tolerance.
Failure of apoptosis and autoimmunity
Takahashi et al used MRL/lpr and gld mice to show a link between aberration of apoptosis and autoimmune disease. These mice are defective in the production of fas and fas ligand respectively which results in the failure of apoptosis of auto reactive B cells. Decreased clearance of apoptotic debris by complement also leads to the transfer of antigen to follicular dendritic cells and other antigen presenting cells, and hence the breakdown of self tolerance. This is seen in the human disease, Systemic lupus erythematosus, in which there is a deficiency of the classical pathway in the shape of C1q. These aberrations result in a direct failure of one of the central mechanisms of self tolerance-clonal inactivation.
Conclusion
The state of self tolerance includes those subsets of autoreactive cells that are not subject to clonal deletion or inactivation. The existence of such immunologically silent cells are the necessary result of the process of thymocyte selection that occurs during development and normally do not cause autoimmunity. Only a few of such peptides can act as autoantigens (shown by the fact that there are relatively few distinct autoimmune diseases and that all individuals with a particular autoimmune disease tend to recognize the same antigens) and the immune system has developed regulatory mechanisms (T suppressor cells) to ensure that if these cells do become recognised they will be eliminated. However, the consequence of infection, particularly if coupled with a genetic susceptibility, can increase the activation of these silent cells, overcoming the defence of the suppressor T cells and leading to autoimmunity.
The failure of clonal inactivation, as seen in the failure of apoptosis can be considered as a true breakdown of self tolerance. The activation of the immunologically silent cells that constitute a normal part of the immune system as a result of infection can also be considered as a breakdown, but the presence of such cells themselves should not be considered a failure of self tolerance.
Bibliography
-
Eaves IA, Wicker LS, Ghandour G, Lyons PA, Peterson LB, Todd JA, Glynne RJ. (2002) Genome Res. 12(2):232-243
- Fairchild P et al. (1996) Immunol. Today 14:203-208
- Lanzavecchia (1995) J. Exp. Med. 181:1597-1601
- Leng Q and Bentwich Z. (2002) Scand J Immunol. 56(3):224-232
-
(1998) PNAS. 95(8):4522-4526
-
(1991)
Nature. 351(6326):482-485
-
(1988) J. Neuroimmunol. 18(4):341-351
- Steinman (1995) Cell 80:7-10
- Wucherpfennig et al. (1995) J. Exp. Med. 181:1945-1948