Autoimmune diseases occur when a sustained, specific, adaptive immune response is generated against self-components, and results in tissue damage or dysfunction. They probably affect more than 3% of Western populations, more commonly women than men, and have peak incidence in the third to sixth decades.
Aetiology and pathogenesis
These can usefully be described in terms of (1) susceptibility—inherited or acquired defects in pathways required to maintain tolerance to self antigens render the individual susceptible to disease initiation; (2) initiation of autoimmunity—interaction between susceptibility genes and environmental events initiate an immune response directed at self antigens; (3) propagation—specific immune response to self antigens causes damage of tissues, with the release of more antigens that further drive the immune response; (4) transition—targets of immune response change and are amplified, with clinical symptoms assuming a recognizable phenotype.
Although a single immune effector pathway may predominate in generating tissue dysfunction and damage in some autoimmune diseases, it is much commoner for multiple effector pathways to participate in generating the final phenotype. Those pathways which generate tissue damage or dysfunction include autoantibody binding to target cells, immune complex-mediated activation of complement and Fc receptor (FcR) pathways, cytokine pathways, as well as lymphocyte-mediated cytotoxicity of target cells. The nature and sites of tissue damage determine the pathological and clinical features of specific diseases.
Tissue-specific autoimmune diseases
Immune-mediated damage is restricted to a particular tissue or organ that specifically expresses the targeted antigen, e.g. (1) Graves’ disease—autoantibodies bind to and stimulate the TSH receptor, resulting in thyrotoxicosis; (2) myasthenia gravis—autoantibodies target the acetylcholine receptor at the neuromuscular junction, resulting in muscular weakness and fatigue due to the inefficient transmission of the acetylcholine signal; (3) type 1 diabetes—a cytotoxic T-cell response to the β cells of the pancreatic islets results in destruction of the insulin-producing cells.
Systemic autoimmune diseases
Typically characterized by simultaneous damage in multiple tissues, e.g. kidney, lung, skeletal muscle, nervous system, and skin. Unlike autoantibodies in tissue-specific autoimmune diseases, which target tissue-specific antigens, the autoantibodies in systemic autoimmune diseases are frequently directed against molecules expressed ubiquitously in multiple tissues, e.g. (1) aminoacyl-tRNA synthetases—targeted in autoimmune myositis and associated interstitial lung disease; (2) small nuclear ribonucleoproteins (snRNPs)—targeted in systemic lupus erythematosus (SLE); (3) toposiomerase-1—targeted in scleroderma. Each of these molecules is expressed in all cells, where they play critical roles in essential cellular processes, e.g. protein translation, mRNA splicing, and DNA replication and remodelling (respectively).
Nonsustained autoimmune diseases
Organ or tissue damage and dysfunction tend to be self-limited and resolve after the first attack, and are very unlikely to recur, e.g. epidemic Guillain–Barré syndrome. These diseases typically occur in the setting of infection, and are associated with cross-reactive antibody responses that recognize both components of the infecting organism as well as the target tissue.
The effector mechanisms that the immune system utilizes to destroy extracellular pathogens, or host cells that harbour intracellular foreigners (e.g. mycobacteria or viruses) must be appropriately targeted if indiscriminate damage to normal host tissue is to be avoided. Under most inflammatory circumstances, some bystander tissue damage is unavoidable. In most situations, this damage is self-limited, due to efficient clearance of the exogenous antigen source and appropriate down-modulation of the immune response. Tissue damage in autoimmune diseases differs fundamentally from bystander damage, in that the host immune system is specifically activated and driven by self-components, focusing damaging immune effector pathways on host tissues expressing those components, in an autoamplifying and self-sustaining way. The danger inherent in initiating a self-sustaining, specific immune response directed against components of self-tissues is intuitively apparent, since antigen clearance under these circumstances is necessarily associated with complete tissue destruction.
It is now clear that an autoimmune component is a feature of many human diseases. Indeed, there are some estimates that autoimmune diseases afflict more than 3% of Western populations, and imposes a significant personal and economic burden on individuals and nations. This chapter will illustrate many of the principles unifying various autoimmune states, and will present a conceptual framework within which to understand their aetiology, pathogenesis, and pathology. The rapid advances in knowledge being made in this group of disorders predict that disease mechanisms will soon be more clearly understood, and will greatly impact therapeutics.
Autoimmune diseases may affect individuals at all stages of life. In general, diseases have a predilection for beginning after the second decade, with peak incidence in the third to sixth decades. In many instances, there is a preference for the female gender, with the magnitude of this sex difference varying amongst the different diseases. Thus, for the systemic autoimmune diseases (e.g. systemic lupus eythematosus (SLE), rheumatoid arthritis, Sjögren’s syndrome, scleroderma, and autoimmune myositis), and autoimmune thyroid disease, the female:male (F:M) ratio is approximately 4–9:1, while for type 1 diabetes, multiple sclerosis, and myasthenia gravis, the female predominance is much less prominent (F:M ratio <2:1). The exact mechanisms underlying this female predominance remain unknown, but this striking biological difference provides a major clue to pathways underlying susceptibility to autoimmunity. Recent studies describing possible gender-related differences in Toll-like receptor (TLR) expression may be relevant in this regard (see below).
An important theme related to the development of various autoimmune diseases has emerged in recent years. One of the most unexpected observations came from studies of patient populations in whom blood samples had been stored for a period of years prior to the onset of clinical disease (military cohorts or stored blood bank samples), allowing investigators to address whether autoantibodies are first generated coincident with clinical disease, or precede this. In a landmark study by Harley and colleagues in SLE, clear evidence was obtained showing that the relatively ‘nonspecific’ antinuclear autoantibodies and antiphospholipid antibodies generally precede the diagnosis of SLE, often by a period of several years. These investigators also showed that phenotype-specific autoantibodies in SLE (e.g. anti-Sm, anti-RNP) occurred around the time of onset of clinical disease, suggesting that different autoantibody specificities were marking different phases in the disease. Similar observations have been made in rheumatoid arthritis, where anti-cyclic citrullinated peptide antibodies predate clinical symptoms, whereas more specific antibodies (e.g. anti-vimentin) only occur when disease becomes established.
It is therefore operationally useful to divide autoimmune diseases into separate kinetic phases: (1) susceptibility—predisease, in which inherited or acquired defects in pathways required to maintain tolerance to self-antigens render the individual susceptible to disease initiation); (2) initiation of autoimmunity—the interface of susceptibility genes and unique environmental events, which initiate an immune response directed at self antigens—this phase may be prolonged, and is generally not accompanied by clinical symptoms; (3) propagation—a self-amplifying phase in which the specific immune response to self-antigens causes damage of tissues, with the release of more antigens, which further drive the immune response. It is important to note that this last, amplified phase does not manifest initially fully developed, but rather evolves over time towards the diagnostic phenotype. This generally occurs subacutely over weeks to months, and frequently begins with nonspecific symptoms and signs. Examples include the fatigue and constitutional symptoms that predate diagnosis of SLE and rheumatoid arthritis. This period—during which the targets of immune response change and are amplified, and during which clinical symptoms assume a recognizable phenotype—can be viewed as a transition phase, and its recognition is of importance in terms of diagnosis and early intervention.
Both genetic and environmental factors play important roles in initiation and propagation of autoimmune diseases. They probably play their central roles by regulating the activation, function, and targets of the host immune system. There is also evidence that stochastic processes play an important role in disease initiation, greatly complicating studies to define the causes and mechanisms of autoimmune disease (see below).
Although autoimmune diseases in humans are genetically complex, significant advances in understanding have occurred over the past several years. In some cases, advances have come from the study of autoimmunity with mendelian patterns of inheritance (e.g. APECED, IPEX, C1q deficiency—see definitions below). Advances have also come from genetic association studies of various autoimmune phenotypes (e.g. rheumatoid arthritis, SLE, type 1 diabetes). Together, the studies stress that multiple genes interact in rendering an individual susceptible to autoimmunity, and highlight a critical role for pathways of tolerance induction, immunoregulation, and setpoints/thresholds for immune signalling in avoiding emergence of autoimmunity. A reciprocal role of target tissue pathways (e.g. antigen structure/expression) in regulating autoimmunity has also been recognized. The following are some general principles regarding the genetics of autoimmunity that have emerged in recent years.
Certain MHC class II alleles are associated with disease susceptibility
One of the most striking genetic associations with autoimmunity resides in the major histocompatibility complex (MHC), an area on chromosome 6 in humans which is highly enriched in genes that participate directly and indirectly in the immune response. The strength of association of different autoimmune phenotypes with MHC class II genes in this area is very robust (odds ratios in the 3–8 range). For example, patients with rheumatoid arthritis have an increased frequency of HLA DR4. HLA DR4 (initially defined serologically) encompasses numerous different alleles that have been defined by sequencing. Interestingly, not all subtypes of HLA DR4 are associated with an increased frequency of rheumatoid arthritis, but those alleles that are associated with rheumatoid arthritis share a short amino acid sequence (QKRAA) at positions 70 to 74 of the β chain of the HLA DR molecule. This sequence, termed the ‘shared epitope’ is located along the peptide-binding groove of HLA DR4 which presents peptides to the antigen receptor of T cells. Interestingly, this same ‘shared epitope’ is present in many HLA DR1-positive individuals with rheumatoid arthritis. A similar principle appears to hold for patients with type 1 diabetes, where there is a strong association of disease with specific DQβ genotype. Since MHC class II molecules function as a scaffold for presentation of specific peptides to CD4T cells (see below), it is possible that this MHC-encoded susceptibility to disease reflects the ability of these alleles to present unique self peptides to autoreactive T cells. The presence of significant linkage disequilibirum within the MHC region (i.e. large stretches of DNA do not undergo recombination, generating functional cassettes of associated genes) also creates the potential for the disease-association of particular MHC alleles to be influenced by additional genes on the extended haplotype in affected individuals. Studies to define these other genes, and the mechanisms whereby they influence development of autoimmunity, are challenging and are ongoing in many diseases.
Incomplete thymic tolerance induction predisposes to autoimmunity
Significant insights into basic mechanisms can derive from the study of rare human phenotypes. This has been true for autoimmunity, where several monogenic disorders have defined important pathogenic principles. Autoimmune polyendocrine syndrome type I (APS1; OMIM 240300), also called autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED), is a rare disease in which patients develop multiple autoimmune diseases, often beginning in childhood. The syndrome is characterized by striking autoimmunity directed against multiple different target tissues, including parathyroids, adrenals, pancreatic β cells, parietal cells, thyroid, liver, and gonads. Numerous autoantigens have been defined as targets of autoimmunity in APS1, and include enzymes specifically expressed in various endocrine tissues (e.g. steroid 21-hydroxylase—specific for adrenal cortex; steroid 17a-hydroxylase—found in adrenal cortex and gonads, GAD65—found in pancreatic islets, and thyroid peroxidase). The genetic basis of APS1 was mapped to a gene on chromosome 21q22.3, subsequently termed AIRE (for autoimmune regulator). AIRE expression is highest in the thymus, where it is expressed in medullary thymic epithelial cells. Significant evidence has now been obtained that AIRE is a transcriptional regulator, which regulates expression in thymic epithelial cells of various peripheral autoantigens normally expressed exclusively in endocrine target tissues. Thus, AIRE appears to regulate the ectopic expression in the thymus of tissue-restricted autoantigens, and provide an antigen source against which to establish central tolerance. Several AIRE-deficient mouse models were subsequently generated; these animals developed various autoimmune endocrine phenotypes, resembling those found in human APS1.
Impaired clearance and tolerance induction by apoptotic cells: susceptibility defect in systemic autoimmunity
Although little is known in humans about the thymic pathways of tolerance induction to ubiquitously expressed autoantigens, there is accumulating evidence to suggest that in the periphery, apoptotic cells play an important role in providing a source of autoantigens against which the organism becomes tolerant. Apoptotic cells are generally very efficiently cleared by phagocytic cells; these events are normally associated with the production of anti-inflammatory cytokines and result in tolerance induction. Interestingly, early components of the classical complement pathway (e.g. C1q and C4) and C-reative protein (CRP) are required for efficient apoptotic cell clearance, with production of interleukin (IL)-10 and transforming growth factor β (TGFβ). It is of particular note, therefore, that homozygous C1q deficiency is associated with a striking susceptibility to SLE, suggesting that rapid, efficient, tolerance-inducing clearance of apoptotic cells may play a similar role to AIRE expression in the thymus in preventing subsequent emergence of autoimmunity to ubiquitously expressed autoantigens. Additional support for this model comes from recent studies of milk fat globule-EGF factor 8 (MFG-E8), a glycoprotein secreted from macrophages that is required for the efficient attachment and clearance of apoptotic cells by macrophages and immature dendritic cells. MFG-E8 is also expressed in tingible-body macrophages at the germinal centres of secondary lymphoid tissues. Interestingly, many unengulfed apoptotic cells are present in the germinal centres of the spleen in MFG-E8-deficient mice, which develop a striking lupus-like phenotype. Other examples exist in which defects in clearance of apoptotic cells are associated with development of systemic autoimmunity (e.g. Mer deficiency). Together, the data strongly suggest that efficient, anti-inflammatory clearance of apoptotic cells plays a central role in tolerance induction and prevention of autoimmunity.
Defective production of regulatory T cells
Although pathways exist that (1) regulate autoantigen expression at sites of tolerance induction, and (2) guide autoantigens towards tolerance-inducing outcomes, these pathways alone are clearly insufficient to prevent the emergence of autoimmune disease. This fact is highlighted by the emergence of autoimmunity when regulatory T-cell differentiation is abnormal in humans with IPEX syndrome (immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome; OMIM 304790). IPEX is a rare X-linked recessive disorder, which is characterized by type 1 diabetes, thyroiditis, atopic dermatitis, and inflammatory bowel disease, and is caused by mutations in the FOXP3 gene. FOXP3 is a member of the forkhead family of transcription factors, and is essential for the development of regulatory T cells (Tregs), which regulate the activation and differentiation of effector T cells at many different levels. It is therefore likely that induction of tolerance is incomplete under most circumstances, and that self-sustaining autoimmunity is normally limited by Treg function.
Signalling thresholds and susceptibility to autoimmunity
Several modulators of T-cell signalling have also been defined as important susceptibility determinants in autoimmunity. For example, CTLA4 polymorphisms are associated with increased risk of a variety of autoimmune diseases, including type 1 diabetes, Graves’ disease, SLE, and rheumatoid arthritis. Similarly, a functional polymorphism in PTPN22 has been identified as a major risk factor for several human autoimmune diseases, including SLE, rheumatoid arthritis, and type 1 diabetes. Although the exact mechanisms underlying susceptibility to autoimmunity remain unclear, in both cases the polymorphisms appear to regulate the balance of stimulatory and inhibitory signalling in effector and regulatory T cells, favouring effector T-cell activation.
The genetic studies in autoiommunity therefore highlight that there are many barriers to the development of autoimmunity, including effective tolerance induction in the thymus and periphery, tightly regulated immune signalling, and homeostatic pathways of immunoregulation to limit self responses should these occur. There are also cassettes of immune response genes encoded in the MHC which appear to be more likely to capture specific self-antigens and generate a response to them. It is likely that the genetic susceptibility to autoimmunity in outbred humans represents an integrated threshold involving genes that regulate these various pathways, upon which environmental and stochastic events act to accomplish disease initiation and propagation.
Twin studies in human autoimmune diseases showing that individuals with an identical genotype may be variably affected by disease (concordance rates vary widely, from 15 to 50%, in identical twins with SLE or rheumatoid arthritis) demonstrate that environmental insults and stochastic events likely play a significant role in the development of autoimmunity. Many potential environmental insults have been suggested to play a role in autoimmune diseases. These include infections, irradiation, and exposure to drugs and toxins. For example, exacerbations of SLE can follow sunlight exposure, and there are numerous reports that disease initiation may have a similar association with ultraviolet irradiation in rare patients. Numerous infections have been postulated to play a role in disease initiation across the spectrum of human autoimmune diseases. In rare cases, the association between antecedent infection and subsequent development of disease is evident (e.g. coxsackievirus infection-induced autoimmune myocarditis, acute rheumatic fever following streptococcal infection, Epstein–Barr virus infection and childhood SLE). In the majority of autoimmune diseases, however, it has not been possible to confirm this environmental connection with any certainty. This does not imply that a causal connection does not exist in these instances, but rather reflects several features of the diseases that greatly complicate the firm establishment of such a connection: (1) kinetic complexity of the autoimmune diseases—since development of autoimmunity occurs in several distinct phases, and once the propagation phase begins, establishment of a recognizable disease phenotype often takes months, evidence of the initiating insult may have disappeared by the time the environmental component is sought for the first time; (2) several different environmental insults may induce a similar response; (3) the environmental force may be extremely frequent in the population, but may only induce autoimmune disease in a unique subset of individuals with appropriate susceptibility genes.
How environmental forces influence initiation of autoimmune diseases is not yet known for most autoimmune diseases, but several plausible mechanisms have been advanced. These include (1) the disruption of cell and tissue barriers, allowing previously sequestered antigens access to a previously ignorant immune system (see below); (2) inducing novel pathways of antigen presentation, (3) alteration of the structure of self antigens, and (4) molecular mimicry. Some of these mechanisms are dealt with in more detail below.
Although extraordinarily complex in detail, the adaptive immune response operates by a set of relatively simple principles: (1) the immune system has the capacity to discern molecular structure in extremely fine detail; (2) it has a uniquely adapted set of signalling systems that computes the amount of antigen; (3) it responds in a binary way to contextual information, i.e. seeing an antigen in the setting of a dangerous context (e.g. infection) initiates an immune response, whereas seeing the antigen in the absence of such co-stimulatory signals leads to tolerance. Numerous studies over the past two decades have underscored that the sustained autoimmune response is extremely similar to adaptive immune responses directed against foreign pathogens, except that the driving antigens in autoimmune disease are self molecules. For example, autoantibodies in most autoimmune diseases display evidence of isotype switching (e.g. from IgM to IgG or IgA), and show features of having undergone affinity maturation through somatic hypermutation. These properties of autoantibodies require the activity of antigen-specific CD4+ T cells, and have therefore focused much attention on defining the mechanisms whereby self-reactive T cells are activated in autoimmunity. Since this is such a central issue in the understanding of autoimmunity, and since there are numerous mechanisms employed by the normal individual to prevent activation of autoreactive T cells, it is important to review briefly the mechanisms that the normal immune system uses to maintain tolerance against self proteins.
Central and peripheral tolerance
To prevent the survival of lymphocytes that will likely encounter their cognate antigens in healthy self tissues, with potential autoimmune destruction of tissues, the immune system spends significant energy on testing the specificity of all receptors generated during antigen-independent development of lymphocytes initially in the thymus, and subsequently in the periphery. When the T-cell receptor generated through somatic recombination recognizes a peptide–MHC complex in the thymus with high affinity/avidity, cells expressing this receptor are negatively selected (since they are likely self-reactive, and will recognize their cognate antigens at additional peripheral sites). These self-reactive cells undergo apoptosis in the thymus, and never make it into the periphery. In contrast, those T-cell receptors that have some affinity for the selecting MHC molecule, but not for the peptide contained in the groove, likely will recognize foreign peptides, and are positively selected. This process of establishing tolerance to self-proteins in the thymus is called ‘central tolerance’. T cells exiting the thymus therefore include cells that can recognize peptides within the scaffold of the MHC molecule used to select that T cell, but have not encountered their specific peptide in the thymus. Since not all self-antigens are expressed in the thymus, there is still a chance that these T cells will encounter a self peptide–MHC complex in the periphery for which they have high affinity. Since cells that have left the thymus no longer have the developmental context that likely denotes a self peptide (i.e. recognition with high affinity of a peptide–MHC complex during development in the thymus), peripheral T cells utilize another binary system to define whether a high affinity interaction should lead to activation or inactivation. This binary system uses additional cell surface molecules (called costimulatory molecules) to denote context. Thus, when peripheral T cells recognize an MHC–peptide complex with high affinity in the absence of costimulation (through ligation of CD28 by surface CD80 or CD86 on the antigen-presenting cell), T cells are inactivated or tolerized. This is known as ‘peripheral tolerance’. In contrast, when peripheral T cells recognize an MHC–peptide complex with high affinity in the presence of costimulation, these T cells are activated.
In addition to T-cell tolerance, B-cell tolerance to self components is also actively maintained. Thus, if B cells encounter either soluble or membrane-bound antigen during development in the bone-marrow, these cells are either deleted (tolerance) or inactivated such that they become refractory to specific stimulation by their antigen (anergy).
Mechanisms allowing an immune response to be directed against self antigens
Although tolerance to self molecules is stringently maintained at the T and B cell levels, reactivity against self molecules may still be possible for several reasons. These include:
There are numerous mechanisms used to establish and maintain T and B cell tolerance. There is accumulating evidence that defects in regulation of these pathways may result in the failure to eliminate autoreactive lymphocytes, or an altered activation threshold for lymphocytes. Examples include defects in the Fas/Fas-ligand system, a receptor–ligand pair which is required for removal of activated, self-reactive lymphocytes. Mice or humans with defects in this pathway manifest profound lymphadenopathy and a spectrum of autoimmunity. Similarly, defects in regulatory molecules which normally function to dampen the immune response (e.g. CTLA-4, the inhibitory T cell receptor for the costimulatory molecules B7-1 and B7-2) may result in profound autoimmune responses. Mice lacking CTLA-4 develop fatal autoimmunity, with widespread T-cell infiltrates, and CTLA-4 polymorphisms are associated with autoimmunity in humans. It should be remembered that the immune system is a highly complex system, with interdependent regulation present at numerous levels. It is likely that many of the other susceptibility genes in human autoimmunity impinge on these immunoregulatory pathways.
Existence of sites of immune privilege
Strict sequestration of tissue-specific antigens behind anatomical and immunological barriers prevents the development of tolerance to molecules expressed preferentially at these sites. Events (e.g. penetrating trauma) which breach this tight boundary may allow initiation of an immune response to these previously hidden self molecules. Relevant examples include antigens within the eye, testis, and central nervous system. In the eye, for instance, penetrating injury to one eye may be followed by development of inflammation in the contralateral eye (sympathetic ophthalmitis) approximately 1 to 2 weeks after injury. Several mechanisms have been proposed to be responsible for maintaining the immune-privileged status of these tissues. One powerful mechanism appears to involve the constitutive expression of Fas ligand in the relevant tissue (e.g. eye). When this molecule binds to and activates its receptor on lymphocytes, these cells undergo apoptotic death, and are prevented from entering the tissue.
Immunodominance and cryptic determinants
Not all regions of a molecule are equally immunogenic. Regions of the molecule that are well-captured by class II MHC molecules during natural processing of self-antigens are able to tolerize T cells (these determinants have been termed ‘immunodominant’ by Sercarz and colleagues). In contrast, regions of self molecules that are not generated in significant amount during natural antigen processing (so-called ‘cryptic determinants’) cannot effectively tolerize T cells, since they are never seen by these cells either in the thymus or peripherally. This immunodominance appears to be influenced by the intrinsic affinity of the peptide for MHC class II, as well as by neighbouring structural determinants on the antigen that may influence its binding to the peptide-binding groove. On self molecules, two sets of determinants can therefore be functionally defined:
- ◆ those that are easily processed and presented (comprising the dominant self), which readily tolerize developing T cells
- ◆ those that are not presented in appreciable amounts after natural processing (comprising the cryptic self), which do not tolerize
There are unusual circumstances in which natural processing of self-antigens is altered from the default pathway. Examples include novel proteolysis of autoantigen (which destroys the dominant epitope or generates a new dominant epitope) prior to entry into the processing pathway, as well as high-affinity binding to specific receptors or antibodies, which can hinder access of the dominant epitope to the antigen-binding groove of MHC class II molecules, or optimize the loading of a previously cryptic epitope. Since T cells recognizing these cryptic peptides have not previously been tolerized, such ‘autoreactive’ T cells can now be activated.
There are several clear demonstrations that autoreactive T cells recognizing cryptic epitopes can be activated in vivo through altered processing of self molecules to reveal these previously immunocryptic epitopes. For instance, high-affinity binding of the HIV surface protein gp120 to CD4 alters the processing of CD4, and activates T cells which recognize epitopes of CD4 not generated during normal antigen processing. This mechanism may account for the autoimmune response to CD4 seen during HIV infection. Similarly, although intact mouse cytochrome c is not immunogenic in mice, cleavage of the molecule into smaller peptides induces a robust T-cell response to cryptic areas of cytochrome c, which were never previously presented by the natural processing pathway, and therefore did not induce tolerance.
The revelation of cryptic epitopes in self-antigens is likely to be a highly relevant mechanism in many human autoimmune diseases, but the studies to demonstrate the importance of this mechanism have only recently begun in earnest. Since the structure of autoantigens influences the hierarchy of dominant and cryptic and determinants generated when the molecule is processed, unique processes which alter the structure of molecules may play critical roles in initiation of autoimmune diseases. These unique events likely do not occur during normal homeostasis, but may occur preferentially during infectious or other pro-immune events occurring at the host-environment interface. Relevant examples include:
- ◆ Activation of unique proteolytic pathways that specifically alter the structure of autoantigens during immune effector pathways. It has recently been observed that the majority of autoantigens targeted across the spectrum of human autoimmune diseases are specifically cleaved by granzyme B during killing of infected target cells by cytotoxic lymphocytes. This cleavage generates unique molecular fragments never generated in the organism during development or homeostasis. Interestingly, this cleavage is a unique feature of autoantigens, and does not affect non-autoantigens. Although it has been proposed that these cleavage events allow the efficient presentation of previously cryptic epitopes, this remains to be formally demonstrated.
- ◆ Additional post-translational modifications that alter conformation of antigens, and modify their subsequent processing. It is noteworthy that numerous post-translational modifications of autoantigens occur, and that in some cases the autoimmune response is strictly dependent on the occurrence of these modifications. Examples include phosphorylation, acetylation, deimination, and isoaspartyl formation, amongst others.
- ◆ Formation of high-affinity complexes between autoantigens and other viral or self-proteins.
In all these examples, it should be remembered that the initiating event in autoimmunity requires that, on the background of appropriate susceptibility genes, several stringent criteria needed to initiate a primary immune response must be simultaneously satisfied. These include the generation of suprathreshold concentrations of self molecules that have a structure not previously tolerized by the immune system, and the presentation of these unique molecular forms to T lymphocytes in the presence of costimulation (i.e. in a proimmune context).
Foreign antigens, which clearly differ from their homologous self-antigens in some areas, may nevertheless bear significant structural similarity to self-antigens in other regions. Initiation of an immune response to the foreign antigen may generate a cross-reactive antibody response that also recognizes the self-protein (molecular mimicry). When the antigen is a cell-surface molecule, antibody-mediated effector pathways can lead to host tissue damage. Although the antibody response is cross-reactive with self molecules, the T cells that drive this response are directed at the foreign antigen (see below). Diseases involving this sort of antigen mimicry therefore tend to be self-limited. It is important to realize that molecular mimicry alone cannot explain self-sustaining autoimmune diseases, which are driven by self-antigens and autoreactive T cells. In these cases, there is a requirement for overcoming T cell tolerance to the self protein. The simultaneous liberation of self-antigen in the presence of the cross-reactive antibody response likely play critical roles in this regard (see below).
Mechanistic insights into molecular mimicry
Although a number of microbial and viral antigens have regions of high homology with various human autoantigens, a causal link between exposure to these foreign antigens and the onset or exacerbation of autoimmune disease has been extremely difficult to establish. There are, however, clear examples that suggest the existence of ‘one-shot’ autoimmune processes, in which cross-reactive antibodies directed against surface self-antigens are generated following infection, and result in tissue damage. This persists until infection is cleared, and the immune response wanes. Although the mechanistic details of this scheme are difficult to prove in vivo, several pertinent examples exist. One of these is a seasonal epidemic form of Guillain–Barré syndrome seen in northern China, which follows Campylobacter jejuni infection. Affected patients make antibodies recognizing gangliosides, and the disease has a self-limited course, which rarely recurs. The anti-ganglioside antibodies generated are likely responsible for the pathological findings of acute motor axonal neuropathy. Another plausible example of this mechanism (although with meager in vivo evidence) is immune thrombocytopenia (ITP) in children. This process characteristically (1) follows an infectious process; (2) demonstrates anti-platelet antibodies, and (3) frequently shows durable remissions. The mechanistic details of this process have been difficult to prove in vivo, and cross-reactive epitopes on potentially initiating pathogens have not yet been defined.
The single episodes of tissue damage in the setting of a cross-reactive immune response following infection must be contrasted to the sustained, autoamplifying disease frequently seen in other autoimmune syndromes. The central issues in this regard are (1) how T-cell tolerance to self antigens might initially be broken, and (2) once this has occurred, why these antigens continue to drive the immune response to self. Examination of tolerance to cytochrome c, a ubiquitous protein that has regions of homology and divergence across different species, has been very useful in understanding molecular mimicry of cross-reactive epitopes. Mouse cytochrome c shares significant homology with human cytochrome c, although the proteins are entirely different in other areas. When Mamula and colleagues used mouse cytochrome c to immunize mice, no T-cell or antibody response to the murine protein was observed. When human cytochrome c was similarly used to immunize mice, strong T-cell epitopes on the foreign cytochrome c were able to induce a strong antibody response to the foreign protein. The antibodies induced recognized both the murine and the human forms of cytochrome c, i.e. cross-reactive antibodies that recognize the self protein were produced. The T cell response to cytochrome c was, however, directed entirely against the foreign (human) form of the protein, and no T cells against the murine protein could be found. These cross-reactive antibodies disappear as the immune response to the foreign protein wanes.
Interestingly, when mouse cytochrome c was included with human cytochrome c during the immunization, a T-cell response to human cytochrome c, and a humoral response to the human protein that cross-reacts with the murine protein, was induced. Within a few days, a strong helper T-cell response specific for murine cytochrome c was detected. This breaking of T-cell tolerance to murine cytochrome c was dependent on activated B cells specific for cytochrome c, which likely exert their effect through altering the processing of mouse cytochrome c, potentially uncovering previously cryptic epitopes in the self-protein. In the presence of continued release of self antigen, this response may become self-sustaining—self-antigen driving autoreactive T cells, providing help to autoantibody-producing B cells.
Molecular mimicry may therefore induce the production of cross-reactive antibodies, which, in the absence of liberation of significant amounts of self antigen, should disappear when the foreign pathogen is cleared. The form of epidemic motor axonopathy described above is likely representative of this scenario. Under highly novel conditions, the simultaneous liberation of significant amounts of self antigen in the setting of a cross-reactive antibody response may allow effective presentation of cryptic epitopes in the self-antigen to autoreactive T cells by activated cross-reactive B cells. If continued release of self antigen occurs, a specific, adaptive immune response to self will be sustained. Antigen release from tissues likely plays a critical role in driving this autoimmune process. Understanding the mechanisms of ongoing antigen release at sites of tissue damage in autoimmune disease (e.g. unique pathways of cell injury and death) is a high priority for future work, as it provides a novel target for therapy (see below).
It is clear from the above discussion that extraordinary complexity is operative in initiation of the human autoimmune diseases. The patient population is genetically heterogeneous, the human immune system is complex and extremely plastic, and it interacts with a plethora of environmental stimuli and stochastic events. The simultaneous confluence of susceptibility factors and initiation forces to set off the self-sustained and autoamplifying process is therefore an extremely rare occurrence. In contrast, once activation of autoreactive T cells has occurred, the ability of the immune system to vigorously respond to vanishingly low concentrations of antigen, to amplify the specific effector response to those antigens, and to spread the response to additional antigens in that tissue, greatly reduces the stringency that must be met to keep the process going.
Effector mechanisms in autoimmune diseases
The initiation phase of autoimmunity requires cooperation between many different cell types, including antigen-presenting cells, T cells and B cells, as well as numerous soluble mediators including antibodies, chemokines, and cytokines. The effector phase of autoimmunity uses the same immune and inflammatory effector mechanisms that the immune system has evolved for removing and destroying pathogens. These include activation of the complement cascade, which generates signals that effect inflammatory cell recruitment and activation. Similarly, ligation of activating Fc receptors on inflammatory cells by immune complexes activates macrophage and neutrophil effector function. Autoantibodies directed against cell surface antigens initiate antigen-dependent cellular cytotoxicity (ADCC), likely mediated by macrophages and natural killer (NK) cells. Cytokines and chemokines play a central role in inflammatory cell recruitment and activation in the target tissue. Tissue damage can also be effected by cytolytic lymphocytes. The pathology characteristic of each autoimmune disease reflects both the particular antigens targeted, as well as the predominant effector mechanisms activated.
One principle of central importance in the effector phase of autoimmunity is autoamplification, which appears to play a central role in the self-sustaining nature of the autoimmune process. Thus, immune effector pathways cause damage of cells in the target tissue, liberating antigen which further stimulates the immune response and effector pathways, thus liberating more antigen. Although this is likely an oversimplification, the view that the immune system plays a role in generating an ongoing supply of autoantigen is useful therapeutically, since it focuses attention on controlling both the supply of antigen as well as immune effector pathways (see below).
Principles of amplification
One of the central features of human autoimmunity is the tendency of the process to amplify progressively with the accumulation of significant immune-mediated tissue damage. Furthermore, in the vast majority of cases, once such amplification begins, the process is very unlikely to resolve spontaneously. Properties of autoantigens themselves may be very important in this phase, in terms both of acquisition of adjuvant properties, and of regulation of expression. The essential features of amplification are a substrate cycle, in which antigen expression and adjuvant properties induce an immune response, which induces increased antigen expression and tissue damage—and further drive the immune response. The importance of tissue-specific autoantigen expression in focusing such immune responses is only beginning to be recognized.
Acquisition of adjuvant properties by disease-specific autoantigens
In recent years there have been dramatic advances in the understanding of the mechanisms whereby specific molecules are selected as antigens in the various autoimmune syndromes. In spite of the fact that tens of thousands of molecules could be targeted by the immune system in autoimmunity, the number of molecules that are frequently targeted in the different phenotypes are markedly restricted—limited perhaps to 100 or so molecules. This has led to the proposal that frequently targeted autoantigens may themselves have properties that make them proimmune. This was first suggested by Plotz and colleagues. who observed that the autoantigenic histidyl aminoacyl tRNA synthetase which is targeted in autoimmune myositis (but not non-auto-antigenic lysyl- and aspartyl-aminoacyl tRNA synthetases) is chemoattractant to immature dendritic cells and other leucocytes. The authors suggested that the selection of a self molecule as a target for an autoantibody response may be a consequence of proinflammatory properties of the molecule itself. They further suggested that modification of autoantigen structure during processes of cell damage or death may be critical in recruiting these additional functions of autoantigens.
Toll-like receptors (TLRs)
One of the most likely receptor systems to sense and transduce the pro-inflammatory properties of autoantigens is the Toll-like receptor (TLR) family, which is the primary innate immune system transducer of pathogen-associated molecular patterns. Ligands for TLRs include both microbial and endogenous molecules, the latter group being particularly relevant to autoimmunity (see below). Microbial ligands include components of Gram-positive bacteria, Gram-negative bacteria, yeast, and protozoans. Although viral and bacterial nucleic acids are the most likely ligands for TLRs, accumulating data demonstrates that complexes containing endogenous nucleic acids are also able to signal through TLRs. Although the exact nature and source of endogenous ligands for TLRs in vivo remains unclear, recent studies have demonstrated that components from stressed, injured and dying cells may play critical roles. Working in several models, numerous investigators have now provided evidence that the targeting of frequently targeted nucleoprotein autoantigens (which contain DNA or RNA) results from the ability of these nucleic acid components to ligate TLRs both in vitro and in vivo. For example, when TLR9-deficiency is bred on to MRL-lpr mice—which are an excellent model of SLE—animals no longer get autoantibody responses to chromatin. Similarly, when mice are rendered TLR7-deficient, the autoantibody response to Sm is markedly inhibited, and severity of the SLE phenotype is improved. These data confirm that autoantigens frequently selected in different autoimmune phenotypes likely have the dual property of being able to simultaneously activate the innate and adaptive immune systems, and that the ability to coligate TLRs plays a critical role. The likelihood that the TLR-autoantigen interface will be therapeutically relevant in autoimmune processes is very high.
One of the major pathways downstream of TLR ligation in autoimmunity appears to centre on a relatively rare class of immature dendritic cells (plasmacytoid dendritic cells or pDCs), which can secrete large amounts of type I interferons upon TLR ligation, and which express TLR7 and TLR9 at high levels. Ronnblom and colleagues have demonstrated that, when added to material from apoptotic or necrotic cells, autoantibodies from SLE and Sjögren’s syndrome patients with specificity for DNA or RNA autoantigens induce striking interferon secretion. Type I interferons have a broad set of functions which likely contribute to the feed forward, propagation phase of systemic autoimmune diseases. For example, they (1) promote the differentiation of monocytes into mature DCs, which drive autoreactive T and B cell responses; (2) increase target cell sensitivity to killing pathways; (3) up-regulate cytotoxic effector pathways; and (4) up-regulate expression of autoantigens. Targeting interferon pathways in systemic autoimmunity may therefore have important therapeutic potential.
|Table 1 Autoantigens targeted in several tissue-specific autoimmune diseases|
|Disease||Tissue target||Prominent autoantigen(s)||Proposed disease mechanisms||Clinical features|
|Autoimmune haemolytic anaemia||Erythrocyte surface||Components of the Rh antigen, Band 3.1, glycophorin, and several unidentified molecules||Antibody-mediated destruction and clearance of erythrocytes||Anemia|
||Multiple||Mild to severe chronic hepatic dysfunction in young women|
|Epidemic Guillain–Barré syndrome (N. China)||Motor axons||Axonal gangliosides||Infection with Campylobacter jejuni induces cross-reactive antibody, which mediates axonal damage||
|Grave’s disease||Thyroid gland||TSH receptor||Antibody-mediated stimulation of TSH receptor, leading to excessive thyroid hormone secretion||
|Diabetes (type 1)||β cells of the islets of Langerhans||
||Cytotoxic lymphocyte-mediated destruction of islet cells||Insulin deficiency and diabetes|
|Idiopathic thrombocytopenia||Platelet surface||Platelet integrins||Antibody-mediated platelet destruction and phagocytosis||Thrombocytopenia, bleeding|
|Inflammatory bowel disease||Gastrointestinal tract||
||Cytokine and lymphocyte-mediated epithelial damage and dysfunction||Chronic intestinal inflammation marked by remission and relapse|
|Multiple sclerosis||Myelinated nerve fibers||
||Demyelinating disorder primarily affecting young adults: Protean clinical manifestations depending on location and size of classic plaques|
|Myasthenia gravis||Neuromuscular junction||Nicotinic AChR||Antibody-induced blockade and down-regulation of AChR||Striated muscle fatigue and weakness|
||Infection with coxsackievirus induces myocardial damage and immunization with cardiac autoantigens||Subacute congestive heart failure|
|Pemphigus vulgaris||Hemidesmosome junctions||Desmoglein-3||Antibody-mediated disruption of epithelial cell junctions, with epidermal cell detachment||Blistering skin lesions|
|Rasmussen’s encephalitis||Inhibitory neurons||Type 3 glutamate receptor||Antibody-mediated blockade of inhibitory neurotransmitter signalling||Severe epileptic seizures, progressive degeneration of a single cerebral hemisphere|
|Stiff man syndrome||GABA-ergic neurons modulating spinal cord reflexes||
||Blockade of inhibitory neurotransmitter signalling, possible autoantibody-mediated||Rare disease characterized by severe, progressive stiffness with superimposed episodic muscle spasms, may be associated with autoimmune disease or malignancy|
|Vitiligo||Melanocytes||Tyrosinase, TRP-1||Cytotoxic lymphocyte-mediated damage of melanocytes||Skin depigmentation|
AChR acetylcholine receptor; ASCA, anti-Saccharomyces cerevisiae antibodies; MOG, myelin/oligodendrocyte glycoprotein; PLP, proteolipid protein.
|Table 2 Systemic autoimmune diseases|
|Disease||Prominent tissue target||Prominent autoantigen(s)||Proposed disease mechanisms||Clinical features|
|Rheumatoid arthritis||Synovial joints||
||Symmetric, erosive polyarthritis|
|Scleroderma||Skin, lung, GI, kidney, heart||
||Blood vessel damage by activated effector lymphocytes and autoantibodies||
|Sjögren’s syndrome||Exocrine glandular epithelial tissue||Ro/SS-A; La/SS-B||Epithelial cell death induced by cytotoxic lymphocytes and other immune effector pathways||Keratoconjunctivitis sicca|
|Systemic lupus erythematosus||Numerous, including skin, kidney, joints, haematologic elements, nervous system||
|Wegener’s granulomatosis||Numerous, including upper airways, lungs, kidneys and skin||
||c-ANCA binding to neutrophil surface induces degranulation in the vessel wall with consequent damage||Multisystem inflammatory vascular disease with predominance of sinuses, middle ear, lung, and renal involvement|
DM, dermatomyositis; PM, polymyositis.
Autoantigen expression in the target tissue in autoimmunity
Another important component of the amplification cycle is the target tissue itself, and particularly the amounts and forms of autoantigens expressed at these sites. Unfortunately, very little is currently known about such parameters in vivo in relevant target tissues, in either normal or pathological circumstances. Insights from studies on human autoimmune myopathies have begun to provide important insights into this problem. Thus, myositis-specific autoantigens are expressed at very low levels in normal muscle, but at high levels in myositis tissue, where antigen expression is at highest levels in regenerating muscle cells. These data suggest that enhanced autoantigen expression in the target tissue may be a feature of disease propagation, and that antigen expression during tissue repair may provide an ongoing antigen source to sustain and amplify tissue damage. In this regard, the regulation of antigen expression (rather than exclusively pathways of immune-mediated damage) may have important therapeutic potential.
The clinical features of the different autoimmune diseases are extremely diverse, and reflect the specific tissue dysfunction which results from activity of immune effector pathways. Almost all tissues may be affected, including prominent involvement of endocrine organs, nervous system, eye, bone marrow elements, kidney, muscle, skin, liver and gastrointestinal tract, blood vessels, lung, and joints. For tissue-specific autoimmune processes (e.g. type 1 diabetes, ITP, autoimmune haemolytic anaemia (AIHA)—see Table 5.4.1), symptoms may relate to tissue hypofunction resulting from (1) target cell destruction (for type 1 diabetes, destruction of the β cells of the pancreatic islets; for ITP and AIHA, destruction and phagocytosis of platelets and erythrocytes); (2) antibody-mediated interference with function or down-regulation of autoantigen expression (e.g. myasthenia gravis, bullous pemphigoid). In other cases, symptoms may arise from tissue hyperfunction (e.g. Graves’ disease) due to activating effects of antibody binding (where antibodies to the TSH receptor induce nonphysiological secretion of thyroid hormone).
In the case of systemic autoimmune processes (Table 5.4.2), symptoms frequently result both from localized target tissue destruction (e.g. skeletal muscle in polymyositis, skin disease in SLE) as well as from the more general activities of inflammatory effector pathways. The latter result from (1) immune-complex deposition at multiple sensitive sites (e.g. joints, kidney, skin, and blood vessel walls) with activation of the complement cascade and recruitment and activation of myelomonocytic cells; (2) ongoing secretion of proinflammatory cytokines. In this regard, the profoundly positive effects of TNF inhibition recently observed on the inflammatory symptoms and joint destruction in rheumatoid arthritis underscore the central role of these general inflammatory mediators in generation and maintenance of the disease phenotype in systemic autoimmune diseases.
Although the barriers that need to be overcome in terms of initiating an autoimmune disease are stringent, and are difficult to satisfy even in the setting of appropriate susceptibility genes, the immune system is equipped with a powerful memory. The mechanisms of this memory are still incompletely defined, but include the generation of a population of memory cells specific for the antigen that initiated the response, which respond vigorously (both in terms of clonal expansion as well as effector function) to very low concentrations of antigen if they encounter it again. Since the autoimmune diseases are disorders driven by the ongoing release of self-antigen, this immunological memory constitutes a major barrier to complete cure. Autoimmune diseases therefore tend to be self-sustaining over long periods, and are often punctuated by clinical exacerbations (flares), which are likely due to re-exposure of the primed immune system to antigen (e.g. SLE, autoimmune myositis, rheumatoid arthritis). The possibility of disease recurring, even after long clinical remission, remains present in most of the autoimmune diseases. Tissue-specific autoimmune diseases may result in the complete destruction of the target tissue over time, with loss of function of that tissue accompanied by a waning immune response (e.g. type 1 diabetes). Interestingly, in cases where immune-mediated tissue pathology results from effector pathways being driven by a cross-reactive T-cell response to a foreign antigen (e.g. epidemic Guillain–Barré syndrome), disease has a finite duration, and generally does not recur.
It is not possible to discuss the therapy of this broad group of disorders in any detail in this chapter, but a few principles that underlie current approaches to therapy are discussed. Autoimmune diseases cause significant tissue dysfunction through (1) inflammation, (2) tissue destruction with loss of functional units, (3) the consequences of healing, and (4) functional disturbances (e.g. interference with acetylcholine signalling by autoantibody to the acetylcholine receptor and inducing receptor down-modulation in myasthenia gravis). Therapeutic interventions in autoimmune diseases are generally focused on controlling immune and inflammatory pathways, and at replacing or accommodating lost function.
Control of immune and inflammatory pathways responsible for ongoing damage
Since in most instances the critical autoantigens and effector pathways responsible for unique diseases have not been defined, this goal is frequently extremely challenging. Thus, frequent use is made of anti-inflammatory and immunosuppressive therapies which broadly target many aspects of the immune response (e.g. steroids, azathioprine, cyclophosphamide, methotrexate, mycophenolate). Since a robust immune response is required to protect the host from a myriad of infectious threats, this nontargeted suppression of the immune system can have deleterious consequences in terms of increased susceptibility to infection, with its attendant high morbidity and mortality. In this regard, therapeutic targeting of specific inflammatory pathways is extremely attractive, and there are recent examples in which this approach has been highly successful. In rheumatoid arthritis, the maintenance of chronic inflammatory joint pathology appears to be dependent on the activity of tumour necrosis factor (TNF). Specific inhibition of TNF through the use of either soluble TNF receptors or humanized monoclonal antibodies has led to an astonishing effect on disease activity in rheumatoid arthritis, with abolition of systemic symptoms, and a striking decrease in the rate of joint destruction. These positive effects were associated with only a minimal increase in susceptibility to infection, although this risk is certainly present. This therapy also served as a model demonstrating that the use of injectable forms of biological therapies (monoclonal antibodies or receptors) as therapeutic agents in the general population was feasible. However, the significant expense associated with such therapies will be a major challenge to health care systems worldwide, particularly as additional effective and powerful agents have been discovered. Although the combinatorial use of agents targeting different pathways in autoimmune disease is attractive, the effects are unpredictable, and great caution will need to be exercised. As noted above, modulation of additional immune effector pathways, TLR signalling, or autoantigen expression in the target tissue are also important targets for novel therapy in the autoimmune diseases.
Another example of specific targeting of proinflammatory pathways is that of intravenous immunoglobulin (IVIG). This is prepared from pooled serum and its major component is immunoglobulin G (IgG). IVIG therapy has been used as a treatment of several autoimmune diseases, including ITP, autoimmune myositis, and acute demyelinating polyneuropathy, but is only available at prohibitive cost. Recent data from mice has demonstrated that IVIG induces surface expression of the inhibitory Fcγ receptor (Fcγ RIIB) on macrophages, and shifts the balance of signalling through Fc receptors towards inhibition, down regulating the pro-inflammatory response to immune complexes. It is likely that continued identification of additional agents that precisely modulate specific inflammatory pathways will have a major therapeutic impact on this group of diseases.