Nature Immunology 2, 759 - 761 (2001)
Joerg Ermann & C. Garrison Fathman
Department of Medicine, Division of Immunology and Rheumatology, Stanford University School of Medicine, Stanford, CA 94305, USA. (http://www.nature.com/email_response/email.taf?address=cfathman%40leland.stanford.edu)
In this Overview, common themes of the accompanying News & Views on RA, SLE, IDDM, thyroiditis and MS are discussed. A unifying concept for the development of these and other autoimmune diseases should incorporate genetic predisposition, environmental factors and immune dysregulation.
The five accompanying News & Views articles in this issue of Nature Immunology review pathogenetic mechanisms in five clinical disorders that are generally regarded as being autoimmune in nature: rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), insulin-dependent diabetes mellitus (IDDM), autoimmune thyroid disease and multiple sclerosis (MS) (1-5). This article will explore three common themes that underlie the induction and perpetuation of these, and other, autoimmune diseases: genetic predisposition, environmental factors and immune regulation (Fig. 1).
Figure 1. Requirements for the development of an autoimmune disease.
The immune response of a genetically predisposed individual to an environmental pathogen, in association with defects in immunoregulatory mechanisms, can lead to the development of an autoimmune disease. The importance of the single components represented in this Venn diagram may vary between individuals and diseases. However, the appearance of an autoimmune disease requires the convergence of all three components. T, T cell; B, B cell; DC, dendritic cell.
What is an autoimmune disease?
The definition of an autoimmune disease is somewhat vague but includes the demonstration of autoimmune phenomena such as autoantibodies and/or autoreactive lymphocytes (antibodies or cells that react against self). Several autoimmune disorders—including Grave's disease (hyperthyroidism due to stimulating antibodies against the thyroid stimulating hormone receptor) (4), myasthenia gravis, pemphigus vulgaris and immune cytopenias—are mediated by pathogenic autoantibodies. In most cases, however, it is not clear what mechanistic role the autoimmune processes have in the pathogenesis of the disease. Autoimmune attack against "self" may be involved in the initiation and/or perpetuation of disease. The autoimmune processes seem to result, in certain instances, from a normal (or aberrant) immune reaction against an exogenous pathogen with subsequent "spreading" of the immune response to recognize self tissue; this reaction can continue in the apparent absence of the initiating pathogen. Most often, however, autoimmune phenomena are simply phenomenological events (for example, false-positive autoantibody tests) without pathogenetic relevance. As discussed in two News & Views (1, 5), the effector phase of some autoimmune diseases, which cause organ damage and clinically detectable disease, is mediated by nonimmune cells and events.
Genetic predisposition to autoimmune disease
A common feature of autoimmune diseases is their propensity to appear in families, which suggests an underlying genetic susceptibility. Not only humans with autoimmune diseases, but also their animal model counterparts, share this apparent genetic predisposition. The genetics of autoimmune diseases in humans and animal models are complex and apparently involve many genes (for a Review on non-MHC genes see Wakeland in this issue (6)). Only a few of the genes involved in the pathogenetic mechanisms that underlie autoimmune diseases are actually known. More commonly, allelic variants of chromosomal regions have been linked to an increased disease risk. Some of these "susceptibility regions" are similar in humans and rodents. More importantly, a number of the genetic loci relevant to at least four of the five diseases discussed in the accompanying News & Views articles are shared in some manner (6). It is not clear whether this "sharing" is due to the clustering of different, perhaps related, genes specific for individual diseases within these regions, or whether different diseases share a common set of genes that predispose them to autoimmune disease in general, while other loci determine the target organ. Any analysis of the genetic predisposition to develop an autoimmune disease is complicated by the existence of "protective genes" that may mask disease susceptibility and modify the risk imposed by "susceptibility genes"; this is clearly demonstrated in a mouse model for SLE (7). On the positive side, however, the identification of such "protective genes" may hold clues to new targets for therapeutic intervention.
One gene cluster stands out among all others in defining genetic susceptibility to all five of the autoimmune diseases described in the News & Views (as well as in their animal models): the region that encodes the major histocompatibility gene complex (MHC). The association between MHC products and autoimmune diseases has been known for more than 20 years and is one of the major arguments for a central role of T cells in the pathogenesis of these diseases. As Feldmann points out, RA (and other autoimmune diseases) can develop in the absence of the "disease-associated" MHC haplotype (1). In a disease as clinically heterogeneous as RA, indeed in most autoimmune diseases, it may be that disease heterogeneity obscures any absolute requirement for MHC identity among all diseased individuals. However, the MHC association is sufficiently strong in human IDDM to allow certain MHC alleles to be used as markers of genetic predisposition to the development of IDDM in models of disease prediction and intervention (see the News & Views by Eisenbarth (3)).
The initial idea that the MHC class II gene product associated with IDDM was "altered" in some way and, therefore, represented a disease-specific gene product or a disease-associated mutation was dismissed when it was shown, by sequence analysis, that the disease-associated MHC gene seen in patients with IDDM was indistinguishable from the same gene in nondiseased people (8). However, it was found that in IDDM (and RA), different allelic variants of disease-associated MHC molecules, which increase the risk of disease, share certain structural features, as is described by Eisenbarth (3) and Feldmann (1) in this issue. In both human IDDM and the nonobese diabetic (NOD) mouse model of spontaneous IDDM, there is a substitution of a neutrally charged amino acid for the negatively charged amino acid aspartic acid and mouse MHC class II genes associated with diabetes susceptibility (7). In RA, HLA-DR variants that confer an increased risk of disease, or severity of disease, contain a short conserved amino acid sequence within a specific section of the -chain of the disease-associated MHC class II molecule, the "shared epitope". These data imply that it is the normal allelic variant of the MHC molecule itself, and not some other gene within the MHC complex, that confers the increased risk of developing autoimmune disease. In addition, because these structural features affect the peptide-binding characteristics of the MHC molecules, they point to a central role for antigen-presentation events in the pathogenesis of these diseases. However, the manner in which the MHC molecules affect predisposition to IDDM, RA or other autoimmune diseases is still not understood. MHC molecules serve both as "thymic selecting elements" to create the repertoire of naïve T cells, and then, in the periphery, present antigenic determinants of foreign proteins to the same T cells to prime them for antigen-specific immune responses. Thus the role of these predisposing MHC gene products could either be in selection of the repertoire during thymic development, in the presentation of (auto) antigenic peptides to the T cells in the periphery, or both. These alternatives have been discussed (9).
Environmental factors predisposing to autoimmune disease
Another disconcerting fact about genetic predisposition to autoimmune disease is the lack of concordance in identical twin pairs for any of the five diseases discussed in the News & Views, as well as other autoimmune diseases that have been studied. Autoimmune disease becomes manifest in less 50% of the twin siblings of an affected identical twin; this poses a major problem for any simple explanation of the genetic control of autoimmune disease development. To explain this low concordance rate among identical twin pairs, one has to either consider a certain stochastic element in disease development (for example, the creation and selection of the expressed T cell repertoire) or search for an initiating external event such as the response against an environmental pathogen.
Lyme arthritis represents an example for the potential evolution of an immune response against an infectious agent to an "autoimmune" response directed at a cross-reactive antigenic determinant of the pathogen that is shared with a self-protein. This disease was originally thought to represent a cluster of patients with juvenile RA but, through painstaking epidemiological analysis and diligent microbiology, was shown to represent initial infection with a tick-borne pathogen with a resultant autoimmune response to a self-protein, leukocyte function-associated antigen 1 (LFA-1, also known as CD11a and CD18). LFA-1 shares an antigenic determinant with the outer surface protein antigen of the inducing infectious agent, the spirochete Borellia burgdorferi (10). In about 10% of patients, antibiotic treatment does not resolve the disease, which suggests that the autoimmune process continues independently of pathogen persistence.
Similar mechanisms have been postulated for other diseases discussed here, some of which, like MS, show clear epidemiological features of an infectious disease (5). However, MS has defied multiple and prolonged conventional analyses to define any particular etiologic pathogen. Studies in animal models, in particular experimental autoimmune encephalomyelitis (EAE), have allowed a careful dissection of the kinetics of response and the evolution of autoimmune disease. However, because patients who state that they are going to develop an autoimmune disease in the near future are not encountered in the doctor's office, new and better methods of epidemiological evaluation must be employed to provide a pathophysiological link between the environment and the disease.
One such example of an epidemiological analysis of the potential relationship between the immune response to an infectious agent and the subsequent induction of an autoimmune disease is the previous infection with Epstein-Barr Virus (EBV) of SLE patients. This potential association highlights the inherent problems in assigning associations of previous infectious exposure to an autoimmune disease. Only by analyzing stored serum samples from military personnel who developed SLE, who had previously provided serum samples for an epidemiological analysis of another disease (AIDS), was one group able to propose that exposure to EBV may have triggered an immune response in SLE patients. They found that a single antigenic determinant on EBV was shared with one of the known SLE autoantigens (11). With such compelling associations between exposure to environmental antigens and resultant long-term autoimmune sequellae, it is likely that more such associations await discovery.
Immune (dys)regulation in autoimmune disease
As pointed out by Lipsky in this issue (2), autoimmune disease is not the same thing as autoimmunity. Autoimmune phenomena can be shown in healthy human individuals, most frequently in the siblings of affected individuals: low-titer autoantibodies, for example, are a relatively common finding ("false-positive autoantibody tests"). As a matter of fact, autoreactivity is a built-in feature of the immune system. The T cell receptor repertoire is positively selected on MHC–self-peptide complexes in the thymus, and naïve T cells require contact with self-MHC molecules in the periphery for their survival and effector function (12). This means that all the T cells in the periphery are, by definition, autoreactive. T cells bearing receptors that encounter self-antigens with a high-avidity response during thymic development are negatively selected. However, this process is incomplete, as not all self-antigens can be sufficiently presented during the thymic selection processes. In addition, the demonstrable ability to induce autoimmune diseases in rodents with suitable immunization protocols (as in EAE or collagen-induced arthritis) is evidence that autoreactive lymphocytes with pathogenic potential exist in the periphery of normal animals and, by inference, in normal humans. A number of tolerance mechanisms exist in the periphery that keep these potentially dangerous self-reactive cells in check: clonal ignorance, deletion, anergy, immune deviation and suppression (13). However, it is still unclear how important any of these mechanisms are in preventing autoimmune disease.
Nevertheless, a number of examples from mouse models (and a few from patients) have shown that the disruption of distinct immunoregulatory pathways can lead to the development of disorders with autoimmune features. First, MRL lpr/lpr mice, which harbor a disruption of the gene that encodes Fas, spontaneously develop a multi-organ autoimmune disease with symptoms that are similar to SLE. The same phenotype is found in gld/gld mice, in which the gene that encodes the ligand of Fas (FasL) is disrupted. Fas-FasL interactions are thought to be important for the termination of immune responses through activation-induced cell death (AICD) of lymphocytes. In a small number of human subjects, a similar disease has been described as autoimmune lymphoproliferative syndrome; these patients have variable mutations in their Fas genes (14). Second, cytolytic T lymphocyte–associated antigen 4–deficient (CTLA-4-/-) mice succumb to a severe lymphoproliferative syndrome with organ infiltration within first 3–4 weeks of life. Third, mice with targeted mutations of interleukin 2 (IL-2) or CD25 (the chain of the high-affinity IL-2 receptor) develop a fatal disease characterized by lymphoproliferation, lymphocytic organ infiltration, colitis, autoantibody formation and anemia (15). Despite the fact that lymphoproliferation is not a prominent feature of all the diseases discussed in the News & Views, it is noteworthy that CTLA-4 has been mapped as a susceptibility gene in both human autoimmune thyroid disease (4) and IDDM (3), whereas IL-2 (as well as CTLA-4) are found in genetic regions linked to disease susceptibility in the NOD mouse (3).
The phenotype observed in CTLA-4-/-, IL-2-/- and CD25-/- mice has been explained as being the consequence of a lack of negative regulatory signals in the CTLA-4-/- mice and insufficient priming for AICD in the IL-2-/- and CD25-/- animals (15). More recently, a defect in CD4+CD25+ regulatory T cells has emerged as an interesting alternative hypothesis. These CD4+CD25+ regulatory "suppressor" T cells were originally described in mice. After neonatal thymectomy, adoptive transfer of CD4+CD25+ cells from adult animals can prevent multi-organ autoimmune syndrome from occurring in susceptible mouse strains. Removal of these cells has since been shown to lead to autoimmune diseases in a number of rodent models (reviewed by Powrie in this issue (16)). Interestingly, when analyzed ex vivo, CD4+CD25+ T cells have high intracellular concentrations of CTLA-4. Although controversial, it has been proposed that CTLA-4 signaling is in some way important for the function of these cells. In addition, CD4+CD25+ regulatory T cells are missing in mice that lack IL-2 or components of the IL-2 receptor signaling pathway and these mice have severe autoimmune phenotypes. This suggests that CD25 is not "just a marker" for this population of regulatory T cells, but that signals provided by IL-2 are essential for their generation and/or homeostasis. It is now clear that regulatory CD4+CD25+ T cells exist in humans and that they show in vitro characteristics that are similar to those seen in studies of their animal counterparts (17). Future investigations will show whether CD4+CD25+ T cells have any role in human autoimmune diseases, that is, whether quantitative or qualitative defects in these cells contribute to disease development.
A unifying concept for the development of an autoimmune disease needs to incorporate genetic predisposition, environmental factors and immune (dys)regulation (Fig. 1). Among the genetic markers of predisposition to autoimmune disease are specific sets of genes for MHC molecules that both shape and regulate the specificity of the adaptive immune response. In addition, variations in a number of other genes that are important in the regulation of immune responses have been associated with the development of autoimmune diseases. The genetic makeup of humans and mice determines not only how the immune system deals with antigenic challenges from the environment, but also how the immune system is regulated to remain tolerant towards self. Under certain environmental conditions, such as an infection, failure of regulatory mechanisms and/or an inappropriate immune response to cross-reactive self-antigens ensues, which leads to organ damage and/or dysfunction. Many of the steps involved in the pathogenesis of the autoimmune diseases discussed here await further studies. Future data will, hopefully, lead to a better understanding of the mechanisms that control the autoimmune "phenotype" and the development of new and better treatment strategies.
© 2001 Nature Publishing Group