Autoimmunity represents the failure of self-tolerance. Before proceeding, the reader is recommended to glance back at Fig. 22, which summarizes the mechanisms by which the immune system normally safeguards its lymphocytes against self-reactivity. This is essentially a problem for the adaptive immune system, since both B and T cells generate their antigen-binding receptors by random gene rearrangement (see Figs 12 and 13) and receptors recognizing self antigens are bound to be generated in the process.
The main mechanisms by which these are prevented from causing harm are shown in Fig. 22. The figure above highlights some of the points at which they can break down or be induced to fail. These are numerous, but two influences are particularly significant: genetics and infection. Identical twins show concordance rates around 30% for many autoimmune diseases (concordance is the frequency of disease in one twin occurring in the other). The association of autoimmune diseases with individual HLA genes, especially class II, implies a crucial role for CD4+ T cells, although the association is fairly weak (relative risk 4–14; relative risk is the chance of developing the disease compared with people without the gene) except for ankylosing spondylitis, where the very strong link (over 90) is with a class I gene, B 27. The role of infection in autoimmunity is suggestive but seldom clear- cut: autoimmune disease frequently follows infection, but no autoimmune disease has yet been convincingly shown to be due to a specific pathogen. The killing of virus-infected cells by cytotoxic T cells could be regarded as an exception, but here the autodestruction is a beneficial part of recovery, although it may cause excessive damage, e.g. hepatitis B and the myocarditis of coxsackie virus infection.
It is important to realize that autoimmunity (centre of figure) does not necessarily mean autoimmune disease (right), the latter term being restricted to conditions where there is reasonable evidence that the symptoms are in fact due to autoantibodies and/or autoreactive T cells (see opposite page). The finding of autoantibodies in the absence of obvious disease, or even in healthy people, emphasizes the fact that the precise aetiology of most autoimmune diseases is still not fully understood.
Tolerance The mechanisms responsible for making sure that lymphocytes do not generally react to self antigens (self-tolerance) are explored in Fig. 22. However, in some cases tolerance is not complete. This can result from incomplete clonal deletion, or a breakdown in peripheral tolerance. Deficiency in the TREG subpopulation has been reported in several autoimmune diseases, including diabetes, rheumatoid arthritis and SLE. Expression of class II MHC antigens on thyroid epithelial cells, or pancreatic beta cells, perhaps as a result of infection, may also contribute to breakdown of tolerance. Sometimes, tissue injury or infection can allow antigens that are usually screened from the immune system (e.g. in the eye) to become accessible.
Macrophages have a key role in many autoimmune diseases, by releasing cytokines that cause local inflammation, enzymes and reactive chemicals (free radicals) that damage the tissue. Antibodies against TNF-α, a key macrophage-derived inflammatory cytokine, are very effective in treatment of rheumatoid arthritis, psoriasis and Crohn’s disease. Macrophage activation is dependent on autoreactive TH cells that release IFNγ and IL-17.
Cytotoxic T cells (TC) in killing virus-infected cells, may damage normal tissues. Liver damage in hepatitis B is a classic example. In other cases, however, autoreactive TC ‘break tolerance’ and target specific autoantigens in organs such as the thyroid or the pancreas.
Drugs frequently bind to blood cells, either directly (e.g. sedormid to platelets; penicillin to red cells) or as complexes with antibody (e.g. quinidine). Alpha methyldopa can induce antibodies against Rhesus blood group antigens, towards which B-cell tolerance is particularly unstable.
Cross-reacting antigens shared between microbe and host may stimulate T help for otherwise silent self-reactive B cells – the ‘T-cell bypass’. Cardiac damage in streptococcal infections and Chagas’ disease appear to be examples of this.
Polyclonal activation Many microbial products (e.g. endotoxins, DNA) can stimulate B cells, including self-reactive ones. The EB virus infects B cells themselves and can make them proliferate continuously.
Autoantibodies are found in every individual but rarely cause disease. In some diseases, raised autoantibody levels are clearly effect rather than cause (e.g. cardiolipin antibodies in syphilis). But in some dis- eases they are the first, major or only detectable abnormality and can cause damage in a variety of ways. They can attach to tissues and activate the complement system (see Fig. 6) leading to inflammation. They can react with specific receptors blocking important hormone or neurotransmitter signals. Or they can react with target autoantigens in blood, forming large complexes (see Figs 20 and 36), which accumulate in skin, lung or kidneys causing inflammation and organ damage.
The precise mechanisms that give rise to autoimmune diseases remain incompletely understood. Much of our current knowledge comes from the study of animal models, such as experimental allergic encephalitis and collagen-induced arthritis, in which autoimmunity is induced by direct immunization with self-proteins. These models have taught us much about how tolerance may be broken, but important differences remain between the corresponding animal and human diseases.
Genetics of autoimmunity Most autoimmune diseases have a genetic component and much effort is being devoted to identifying the genetic ‘risk factors’ associated with particular autoimmune diseases. The strongest associations are those with specific alleles of the MHC class
II genes (see opposite page), confirming that CD4+ T cells have an important role in the aetiology of these diseases. However, there are at least 20 other loci that contribute to an individual’s propensity to develop a particular autoimmune disease. Some of these appear to control the level of cytokines, others affect signalling pathways in immune cells while yet others affect non-immunological steps in tissue damage.
Haemolytic anaemia and thrombocytopenia, although they can be caused by drugs, are more often idiopathic. The correlation between autoantibody levels and red cell destruction is not always very close, suggesting another pathological process at work.
Thyroiditis is one of the best candidates for ‘primary’ autoimmunity. There may be stimulation (thyrotoxicosis) by antibody against the receptor for pituitary TSH, or inhibition (myxoedema) by cell destruc- tion, probably mediated by cytotoxic T cells and autoantibody.
Pernicious anaemia results from a deficiency of gastric intrinsic factor, the normal carrier for vitamin B12. This can be caused both by autoimmune destruction of the parietal cells (atrophic gastritis) and by autoantibodies to intrinsic factor itself.
Diabetes, Addison’s disease (adrenal hypofunction) and other endocrine diseases are often found together in patients or families, suggesting an underlying genetic predisposition. The actual damage is probably mainly T-cell mediated, against pancreatic β cells and the adrenal cortex, respectively.
Myasthenia gravis, in which neuromuscular transmission is intermittently defective, is associated with autoantibodies to, and destruction of, the postsynaptic acetylcholine receptors. There are often thymic abnormalities and thymectomy may be curative, although it is not really clear why.
Rheumatoid arthritis is characterized by autoantibody against IgG (rheumatoid factor) although not in every case. Joint damage may be partly mediated via immune complexes, and injections of antibodies against CD20, which result in depletion of B cells, is an effective treatment in a proportion of patients. T-cell-dependent activation of macrophages (type IV hypersensitivity) may also contribute. In either case the cytokines TNF-α and IL-1 cause the main pathology, by activating degradation of cartilage.
SLE In systemic lupus erythematosus the autoantibodies are against nuclear antigens, including DNA, RNA and nucleic acid binding proteins. The resulting immune complex deposition is widespread throughout the vascular system, giving rise to a ‘non-organ-specific’ pattern of disease. A localizing role for the antigen itself may explain why different complexes damage different organs. Patients with SLE also have very high levels of type I interferons, perhaps resulting from innate responses to circulating nucleic acids (see Fig. 5), which con- tribute to a generalized inflammation.
Treatment of autoimmunity
No cures exist for most autoimmune diseases, and treatment is symptomatic; examples are anti-inflammatory drugs for rheumatoid arthritis, insulin for type I diabetes, anti-thyroid drugs for thyrotoxicosis. Where autoantibodies are to blame (e.g. in myasthenia) plasmapheresis to remove them can provide short-term benefit. Remarkable improvement in patients with rheumatoid arthritis and Crohn’s disease has been achieved by treatment with a high-affinity antibody against TNF-α, which presumably blocks the inflammatory cascade within the affected tissue: this remains the best example of successful therapy using an anticytokine antibody. More antigen-specific approaches to immunomodulation, such as vaccination against particular families of T-cell receptors, or the simulation of specific TREG cells, are still at an experimental stage.