Autoimmunity and Autoimmune Disease

Published on 18/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1486 times

Chapter 20 Autoimmunity and Autoimmune Disease

• Autoimmunity is associated with disease. Autoimmune mechanisms underlie many diseases, some organ-specific, others systemic in distribution, and autoimmune disorders can overlap – an individual may have more than one organ-specific disorder, or more than one systemic disease.

• Genetic factors play a role in the development of autoimmune diseases. Twin studies show that there is a heritable component to autoimmunity. The vast majority of autoimmune diseases are polygenic but HLA genes are particularly important.

• Self-reactive B and T cells persist even in normal subjects. Autoreactive B and T cells persist in normal subjects, but in disease are selected by autoantigen in the production of autoimmune responses.

• Controls on the development of autoimmunity can be bypassed. Microbial cross-reacting antigens and cytokine dysregulation can lead to autoimmunity.

• In most diseases associated with autoimmunity, the autoimmune process produces the lesions. The pathogenic role of autoimmunity can be demonstrated in experimental models. Human autoantibodies can be directly pathogenic. Immune complexes are often associated with systemic autoimmune disease. Autoantibody tests are valuable for diagnosis and sometimes for prognosis.

• Treatment of autoimmune disease has a variety of aims. Treatment of organ-specific diseases usually involves metabolic control. Treatment of systemic diseases includes the use of anti-inflammatory and immunosuppressive drugs. Biological therapies using monoclonal antibodies against pro-inflammatory cytokines have revolutionized the treatment of the autoimmune rheumatic diseases. B cell-directed therapies have proved highly effective in many autoimmune diseases.

Like a highly trained army, the immune system has evolved to recognize and destroy foreign invading forces. Sometimes immune recognition fails resulting in ‘friendly fire’ against the body’s own tissue. For example, ‘friendly fire’ directed against synovial tissue causes rheumatoid arthritis; whereas an attack on cells within the pancreas results in diabetes mellitus.

Non-specific inflammation (e.g. in response to infection) invariably leads to some degree of ‘collateral’ damage, but because of efficient clearance of the inciting pathogen and negative feedback loops, this is usually self-limited. In contrast, once initiated autoimmune reactions usually persist because the inciting self-antigen cannot be cleared without complete destruction of the target tissue. Furthermore, the tissue destruction resulting from the autoimmune attack may expose previously hidden antigens, leading to further autoantibody production. This phenomenon is known as epitope spread.

Autoimmunity and autoimmune disease

Because the repertoire of specificities expressed by the B cells and T cells is generated randomly, it includes many that are specific for self components. The body therefore requires self-tolerance mechanisms to distinguish between self and non-self determinants to avoid autoreactivity (see Chapter 19). However, such mechanisms may fail and a number of diseases have been identified in which there is autoimmunity, with copious production of autoantibodies and autoreactive T cells. The targets of the autoimmune ‘attacks’ vary widely as we shall see.

Not every ‘autoimmune event’ leads to clinically overt disease. For example, anti-nuclear antibodies may be observed in the healthy relatives of patients with systemic lupus eythematosus, as well as a small number of unrelated, healthy individuals.

Autoimmunity strictly refers to an inappropriate adaptive immune response, i.e. a loss of self-tolerance. Although the terms autoinflammatory and autoimmune are often used interchangeably, the two are not synonymous. Autoinflammatory diseases can occur without autoimmunity. For example, the periodic fever syndromes, which include familial Mediterranean fever and tumor necrosis factor receptor associated periodic syndrome (TRAPS), are caused by dysregulation of the innate immune system, without any adaptive immune response against ‘self’. Conversely, autoimmune diseases need not be inflammatory; examples include immune thrombocytopenia, hemolytic anemia and myasthenia gravis.

Autoimmune conditions present a spectrum between organ-specific and systemic disease

Hashimoto’s thyroiditis is highly organ-specific

One of the earliest examples in which the production of autoantibodies was associated with disease in a given organ is Hashimoto’s thyroiditis. (Thyroiditis is a condition that is most common in middle-aged women and often leads to the formation of a goiter or hypothyroidism.) The gland is infiltrated, sometimes to an extraordinary extent, with inflammatory lymphoid cells. These are predominantly mononuclear phagocytes, lymphocytes and plasma cells, and secondary lymphoid follicles are common (Fig. 20.1). The gland often also has regenerating thyroid follicles.

Fig. 20.1 Pathological changes in Hashimoto’s thyroiditis

In the normal thyroid gland (1), the acinar epithelial cells (aec) line the colloid space (cs) into which they secrete thyroglobulin, which is broken down on demand to provide thyroid hormones (cap, capillaries containing red blood cells). In the Hashimoto gland (2), the normal architecture is virtually destroyed and replaced by invading cells (ic), which consist essentially of lymphocytes, macrophages, and plasma cells. A secondary lymphoid follicle (sf), with a germinal center (gc) and a mantle of small lymphocytes (m), is present. H&E stain. × 80. (3) In contrast to the red color and soft texture of the normal thyroid, the pale and firm gross appearance of the Hashimoto gland reflects the loss of colloid and heavy infiltration with inflammatory cells.

((2) Reproduced from Woolf N. Pathology: basic and systemic. London: WB Saunders; 1998.)

The serum of patients with Hashimoto’s disease usually contains antibodies to thyroglobulin. These antibodies are demonstrable by agglutination and by precipitin reactions when present in high titer. Most patients also have antibodies directed against a cytoplasmic or microsomal antigen, also present on the apical surface of the follicular epithelial cells (Fig. 20.2), and now known to be thyroid peroxidase, the enzyme that iodinates thyroglobulin. The antibodies associated with Hashimoto’s thyroiditis react only with the thyroid, so the resulting lesion is highly localized.

Fig. 20.2 Autoantibodies to thyroid

Healthy, unfixed human thyroid sections were treated with patients’ serum, and then with fluoresceinated rabbit anti-human immunoglobulin. (1) Some residual thyroglobulin in the colloid (RTg) and the acinar epithelial cells (AEC) of the follicles, particularly the apical surface, are stained by antibodies from a patient with Hashimoto’s disease, which react with the cells’ cytoplasm but not the nuclei (N). (2) In contrast, serum from a patient with systemic lupus erythematosus (SLE) contains antibodies that react only with the nuclei of acinar epithelial cells and leave the cytoplasm unstained.

(Courtesy of Mr G Swana.)

SLE is a systemic autoimmune disease

By contrast, the serum from patients with diseases such as systemic lupus erythematosus (SLE) reacts with many, if not all, of the tissues in the body. In SLE, one of the dominant antibodies is directed against the cell nucleus (see Fig. 20.2).

Hashimoto’s thyroiditis and SLE represent the extremes of the autoimmune spectrum (Fig. 20.3):

• the common target organs in organ-specific disease include the thyroid, adrenal, stomach, and pancreas;

• the non-organ-specific diseases, often termed systemic autoimmune diseases, which include rheumatological disorders, characteristically involve the skin, kidney, joints, and muscle (Fig. 20.4).

Fig. 20.3 The spectrum of autoimmune diseases

Autoimmune diseases may be classified as organ-specific or non-organ-specific depending on whether the response is primarily against antigens localized to particular organs or against widespread antigens.

Fig. 20.4 Two types of autoimmune disease

Although the non-organ-specific diseases characteristically produce symptoms in the skin, joints, kidney, and muscle, individual organs are more markedly affected by particular diseases, for example the kidney in SLE and the joints in rheumatoid arthritis.

The location of the antigen determines where a disease lies in the spectrum

The autoantigen in organ-specific autoimmune disease is expressed only in that tissue. In contrast, autoantigens in systemic autoimmune disease are typically present in multiple tissues (if not ubiquitously) throughout the body. Examples include tRNA synthetases in myositis (which in spite of its name may also involve the skin, joints, lungs, and heart), small nuclear ribonucleoproteins (snRNPs) in SLE, and topoisomerases in systemic sclerosis. These small proteins play a vital role in the cellular ‘machinery’. tRNA synthetases are involved in protein translation, snRNPs in mRNA splicing, and topoisomerases in DNA replication. Thus it is not difficult to appreciate the myriad clinical manifestations of an autoimmune ‘attack’ on these targets.

An individual may have more than one autoimmune disease

Autoimmune diseases ‘hunt in packs’; in patients with an autoimmune disease, the chance of developing an additional autoimmune disease is significantly elevated. This holds true for both organ-specific and systemic autoimmune disease.

• Thyroid antibodies occur with a high frequency in patients with pernicious anemia who have gastric autoimmunity, and these patients have a higher incidence of thyroid autoimmune disease than the normal population. Similarly, patients with thyroid autoimmunity have a high incidence of stomach autoantibodies and to a lesser extent the clinical disease itself (pernicious anemia).

• Around 15% of patients with primary Sjögren’s syndrome have concomitant autoimmune hypothyroidism, and up to 40% have thyroid antibodies. Around 10% of SLE patients have thyroid antibodies in their serum.

• 30% of patients with SLE and Sjögren’s syndrome have a second, third or even fourth autoimmune disease.

The systemic autoimmune rheumatic diseases show considerable overlap. This is typified by the so called ‘overlap syndromes’ in which patients exhibit mixed features of SLE, myositis, and scleroderma. Features of rheumatoid arthritis, notably an erosive arthritis, are present in 5% of patients with SLE. In these diseases immune complexes are deposited systemically, particularly in the kidney, joints, and skin, giving rise to widespread lesions.

The mechanisms of immunopathological damage vary depending on where the disease lies in the spectrum:

• where the antigen is localized in a particular organ, type II hypersensitivity (e.g. autoimmune hemolytic anemia) and type IV cell-mediated reactions, as in type 1 insulin-dependent diabetes, are most important (see Chapters 24–26);

• in systemic disorders such as SLE, type III immune complex deposition leads to inflammation through a variety of mechanisms, including complement activation and phagocyte recruitment. In rheumatoid arthritis immune complexes directly stimulate production of cytokines such as TNFα.

Genetic factors in autoimmunity

There is an undoubted familial incidence of autoimmunity. This is largely genetic rather than environmental, as may be seen from studies of identical and non-identical twins. Thus if one of a pair of identical twins gets SLE, there is a 25% chance the other will develop it. If the twin is not identical the concordance rate is only 2–3%.

Within families, clustering of distinct autoimmune diseases has been reported. A large population-based survey found that families with a proband with rheumatoid arthritis were more likely to manifest other autoimmune disorders. This finding holds true for relatives of patients with other autoimmune diseases, such as multiple sclerosis, which suggests the presence of shared pathogenic factors across the autoimmune diseases. However, the majority of individuals with autoimmune disease will not have an affected first degree relative. Thus whilst genetic factors are important in the pathogenesis of autoimmunity, they are usually not sufficient to cause disease without additional environmental influences.

Within the families of patients with organ-specific autoimmunity, not only is there a general predisposition to develop organ-specific antibodies, other genetically controlled factors tend to select the organ that is mainly affected. Thus, although relatives of patients with Hashimoto’s thyroiditis and families of patients with pernicious anemia both have a higher than normal incidence and titer of thyroid autoantibodies, the relatives of patients with pernicious anemia have a far higher frequency of gastric autoantibodies (Fig. 20.5), indicating that there are genetic factors that differentially select the stomach as the target within these families.

Fig. 20.5 Overlap between thyroid and gastric autoimmunity

Genetic factors affect the predisposition to autoimmunity and selection of the target organ in the relatives of patients with Hashimoto’s thyroiditis and pernicious anemia.

(Data adapted from Doniach D, Roitt IM, Taylor KB. Ann NY Acad Sci 1965;124:605.)

Usually, several genes underlie susceptibility to autoimmunity

Clearly, the vast majority of autoimmune diseases are not single gene disorders. Rather, they occur as a result of the complex interplay of multiple genetic and environmental factors. Evidence from genome-wide association studies has demonstrated that many genes contribute to disease susceptibility, i.e. autoimmunity is usually polygenic. Thus the effect of variation in any one gene is by itself typically small.

There is emerging interest in how epigenetic modifications of DNA and associated histones may provide explanations of the molecular basis of complex polygenic disease traits such as autoimmunity. Epigenetic changes may occur as a result of environmental stimuli, but may also be partly heritable.

Certain HLA haplotypes predispose to autoimmunity

Further evidence for the operation of genetic factors in autoimmune disease comes from their tendency to be associated with particular HLA specificities (Fig. 20.6). For most autoimmune diseases, the MHC region which is located on the short arm of chromosome 6, provides the strongest genetic component to disease susceptibility. Recent high-density genome mapping in multiple autoimmune diseases has demonstrated complex, multilocus effects that span the entire region, with both shared and unique loci across diseases.

Fig. 20.6 HLA associations in autoimmune disease

The relative risk is a measure of the increased chance of contracting the disease for individuals bearing the HLA antigen, relative to those lacking it. Virtually all autoimmune diseases studied have shown an association with some HLA specificity. The greater relative risk for Addison’s disease associated with HLA-DR3 as compared with HLA-B8 suggests that DR3 is closely linked to or even identical with the ‘disease susceptibility gene’. In this case it is not surprising that B8 has a relative risk greater than 1, because it is known to occur with DR3 more often than expected by chance in the general population, a phenomenon termed linkage disequilibrium. Both DQ2 and DQ8 are also associated with type 1 diabetes mellitus, and are usually found in conjunction with the extended haplotype containing the DR3/4 disease-associated alleles. Rheumatoid arthritis is linked to a pentamer sequence in DR1 and certain subtypes of DR4, but not to any HLA-A or HLA-B alleles.

In some autoimmune disorders single HLA genes appear to determine disease susceptibility (e.g. HLA-B27 and ankylosing spondylitis). In others, such as rheumatoid arthritis, complex interactions between alleles at multiple genes within the HLA are involved. Rheumatoid arthritis shows no associations with the HLA-A and -B loci haplotypes, but is associated with a nucleotide sequence known as the ‘shared epitope’ (encoding amino acids 70–74 in the DRβ chain) that is common to DR1 and major subtypes of DR4; the MHC association with rheumatoid arthritis is restricted to patients who are positive for antibodies to citrullinated peptides. The nucleotide sequence of the shared epitope is also present in the dnaJ heat-shock proteins of various bacilli and EBV gp110 proteins, suggesting an interesting possibility for the induction of autoimmunity by a microbial cross-reacting epitope (see below). The plot gets deeper, though, with the realization that HLA-DR molecules bearing this sequence can bind to another bacterial heat-shock protein, dnaK, and to the human analog, namely hsp73, which targets selected proteins to lysosomes for antigen processing.

The haplotype B8, DR3 is common in both the organ-specific diseases and systemic autoimmune diseases like SLE and myositis, though Hashimoto’s thyroiditis tends to be associated more with DR5. Interestingly, for type 1 diabetes mellitus, DQ2/8 heterozygotes have a greatly increased risk of developing the disease (see Fig. 20.6).

Genes of the MHC including HLA-A1, B8, and DR3 have been linked to SLE, although over a dozen other genes (including those regulating interferon α) have shown a stronger linkage.

Genes outside the HLA region also confer susceptibility to autoimmunity

Although HLA risk factors tend to dominate, autoimmune disorders are genetically complex and genome-wide searches for mapping the genetic intervals containing genes for predisposition to disease also reveal a plethora of non-HLA genes (Fig. 20.7) affecting:

• loss of tolerance;

• lymphocyte activation (receptor signaling pathways and co-stimulation);

• microbial recognition;

• cytokines and cytokine receptors;

• end-organ targeting.

Fig. 20.7 Some genetic loci associated with autoimmune disease

CRP, C reactive protein; CTLA4, cytotoxic T-lymphocyte-associated protein 4; FCGR, Fc gamma receptor; IL, interleukin; MHC, major histocompatibility complex; PARP, poly-ADP-ribose polymerase; PDCD1, programmed cell death 1; PTPN22, protein tyrosine phosphatase, non-receptor type 22; TLR5,: toll-like receptor 5; TNF-α, tumor necrosis factor α; AAV, ANCA-associated vasculitis; AS, ankylosing spondylitis; AT, autoimmune thyroiditis; IBD, inflammatory bowel disease; MS, multiple sclerosis; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; T1D, type 1 diabetes mellitus.

One of the commonest genes outside of the HLA region to be associated with autoimmune disease is the protein tyrosine phosphatase gene (PTPN22), which is expressed in lymphocytes. The minor allele of PTPN22 (Trp620) is associated with type 1 diabetes, rheumatoid arthritis, autoimmune thyroiditis and SLE. Crohn’s disease, in contrast, is linked to the more common allele (Arg620). The minor allele results in gain of function and results in inhibition of B and T cell activation and homozygous individuals have a profound defect in lymphocyte receptor signaling. The mechanisms by which this leads to autoimmunity are not clear, but possibilities include failure to delete autoreactive T cells in the thymus, impaired regulatory T cell function, and ineffective clearance of pathogens. It is interesting to note that two alleles with contrasting effects on lymphocyte receptor signaling are associated with distinct autoimmune diseases.

Autoimmunity is associated with genes that control lymphocyte activation

Autoimmunity is associated with a multitude of genes encoding molecules expressed by lymphocytes which modulate co-stimulation signals. These include CD2/CD58, TNFSF15, and ISOSLG. A notable example is a single nucleotide polymorphism (SNP) linked to CTLA-4 (Fig. 20.w1), a molecule present on Tregs (see Chapter 19).

Fig. 20.w1 CTLA-4-linked single nucleotide polymorphism associated with autoimmune disease

The CT60 SNP variant in the non-coding 6.1 kb 3′ region of CTLA-4 is responsible for the association of autoimmune disease with lower messenger RNA levels of the soluble variant splice isoforms of CTLA-4.

(Data adapted from Ueda H, et al. Nature 2003;423:506.)

Mutations in the transcriptional autoimmune regulator (Aire) gene are associated with the multiorgan disorder known as autoimmune polyglandular syndrome type 1, characterized by Addison’s disease, hypoparathyroidism, and type 1 diabetes mellitus. Deficiency in Aire prejudices the establishment of self tolerance to a number of organ-specific antigens such as glutamic acid decarboxylase, GAD65, one of the key autoantigens in type 1 diabetes mellitus, by preventing its expression in the thymus.

SLE is associated with multiple gene loci

Genome wide-association studies have identified multiple genetic loci associated with SLE (see Fig. 20.7). Null alleles which cause a deficiency of one of the early complement components (C1q, C2, or C4) strongly predispose individuals to a lupus-like disease, although there is a lower incidence compared to ‘classic’ SLE of certain clinical features, notably glomerulonephritis and serological characteristics (ANA and anti-dsDNA antibodies) in these individuals.

The IL2RA locus (which encodes the high-affinity IL-2 receptor) is associated with SLE, type 1 diabetes, multiple sclerosis, Grave’s disease, and ANCA-associated vasculitis.

Serum interferon alpha activity is a heritable risk factor for SLE and there is a strong linkage between single-nucleotide polymorphisms in the genes encoding tyrosine kinase 2 (TYK2) and interferon regulatory factor 5 (IRIF5). TYK2 binds to the type I IFN receptor complex. IRF5 is induces type I IFN gene expression and is downstream of pattern recognition receptor signaling. Microarray studies of peripheral blood mononuclear cells from patients with SLE show a strong type 1 interferon gene expression. The link between SLE and genes controlling interferon suggests the host response to viruses may be important in the pathogenesis of this disorder.

Using a mouse model of lupus, Wakeland and colleagues showed that some linked genes have a cumulative effect leading to autoimmune disease. Thus mice with the sle1 mutation alone show loss of tolerance to chromatin and develop antinuclear antibodies. Mice with sle3 have polyclonal T and B cell activation, with only mild glomerulonephritis. However, mice with both sle1 and sle3 have a full autoimmune response with antibodies to nuclear antigens, splenomegaly and fatal glomerulonephritis.

Autoimmunity and autoimmune disease

Despite the complex selection mechanisms operating to establish self tolerance during lymphocyte development, the body contains large numbers of lymphocytes, which are potentially autoreactive.

Thus, many autoantigens, when injected with adjuvants, make autoantibodies in normal animals, demonstrating the presence of autoreactive B cells, and it is possible to identify a small number of autoreactive B cells (e.g. anti-thyroglobulin) in the normal population.

Autoreactive T cells are also present in normal individuals, as shown by the fact that it is possible to produce autoimmune lines of T cells by stimulation of normal circulating T cells with the appropriate autoantigen (e.g. myelin basic protein [MBP]) and IL-2.

Autoantibody production alone does not equal autoimmune disease

Healthy individuals may have autoantibodies without clinical disease. For example, 13% of healthy individuals have significant levels of anti-nuclear antibodies (ANA) and their prevalence in the general population aged over 60 rises to 20–30%. Transiently positive ANA may occur following infection. In patients who eventually develop autoimmune disease, autoantibody production may predate clinical disease by years – in one study autoantibodies preceded the onset of clinical manifestations by up to 9 years.

The autoantibodies typically appear in a stereotyped order, providing insight into the sequential pathogenic changes that occur as SLE develops. Antinuclear antibodies, antibodies to Ro, and antibodies to beta2 glycoprotein 1 appear first; anti-dsDNA antibodies typically appear 1–2 years before symptom onset; antibodies to Sm and RNP appear in the months immediately preceding symptoms. Similarly rheumatoid factors and anti-CCP antibodies have been identified in the serum of individuals years before they develop clinically overt rheumatoid arthritis.

Progression to autoimmune disease occurs in stages

These findings allow us to conceptualize autoimmune disease as a multistep process. The first stage is predisposition of an individual to autoimmunity by his or her genes, and other factors such as female hormones. The second phase is initiated by an event, probably stochastic or perhaps caused by an environmental trigger such as infection or ultraviolet radiation, leading to loss of self-tolerance and autoantibody production. This however is not alone sufficient to cause disease. A further step is required before progression to a third phase involving tissue damage by the autoimmune attack. This autoimmune attack leads to further release of self-antigens, which are not removed in the normal efficient manner (see below), and propagation of the autoimmune response, resulting in the clinical manifestations of disease. The earliest clinical features of systemic autoimmune diseases such as SLE or RA are usually non-specific such as fatigue or constitutional symptoms. This prodrome typically precedes the development of the classic disease phenotype by weeks to months.

During the propagation phase, not only is there an autoimmune response to an increasing number of autoantigens (demonstrated by the sequential development of multiple autoantibodies in patients with SLE), but also to more epitopes within each antigen – a phenomenon termed epitope spread. Epitope spread can be involve multiple epitopes on the same molecule (intramolecular spread), or epitopes on different molecules associated as part of a macromolecular complex (intermolecular spread). The latter provides a mechanism for how antibodies to non-protein self antigens such as DNA and phospholipid can occur.

Autoimmunity results from antigen-driven self-reactive lymphocytes

Given that autoreactive B cells exist, the question remains whether they are stimulated to proliferate and produce autoantibodies by interaction with autoantigens or by some other means, such as non-specific polyclonal activators or idiotypic interactions (see below and Fig. 20.8).

Fig. 20.8 Possible model of T cell help via processing of intermolecular complexes in the induction of autoimmunity

An immune complex consisting of autoantigen (e.g. DNA) and a naturally occurring (germline) autoantibody is taken up by an antigen-presenting cell (APC), and peptides derived by processing of the idiotypic segment of the antibody (Id) are presented to TH cells. B cells that express the ‘pathogenic’ autoantibody can capture the complex and so can receive T cell help via presentation of the processed Id to the TH cell. Similarly, an anti-DNA-specific B cell that had endocytosed a histone–DNA complex could be stimulated to autoantibody production by histone-specific TH cells.

Evidence that B cells are selected by antigen comes from the existence of high-affinity autoantibodies, which arise through somatic mutation, a process that requires both T cells and autoantigen. In addition, patients’ serum usually contains autoantibodies directed to epitope clusters occurring on the same autoantigenic molecule. Apart from the presence of autoantigen itself, it is very difficult to envisage a mechanism that could account for the co-existence of antibody responses to different epitopes on the same molecule. A similar argument applies to the induction, in a single individual, of autoantibodies to organelles (e.g. nucleosomes and spliceosomes, which appear as blebs on the surface of apoptotic cells) or antigens linked within the same organ (e.g. thyroglobulin and thyroid peroxidase).

The most direct evidence for autoimmunity being antigen driven comes from studies of the Obese strain chicken, which spontaneously develops thyroid autoimmunity. If the thyroid gland (the source of antigen) is removed at birth, the chickens mature without developing thyroid autoantibodies (Fig. 20.9). Furthermore, once thyroid autoimmunity has developed, later removal of the thyroid leads to a gross decline of thyroid autoantibodies, usually to undetectable levels.

Fig. 20.9 Effect of neonatal thyroidectomy on Obese chickens

Because removal of the thyroid at birth prevents the development of thyroid autoantibodies, it would appear that the autoimmune process is driven by the autoantigen in the thyroid gland.

(Based on data from de Carvalho et al. J Exp Med 1982;155:1255.)

Comparable experiments have been carried out in the non-obese diabetic (NOD) mouse, which models human autoimmune diabetes – chemical destruction of the β cells leads to decline in pancreatic autoantibodies.

In organ-specific disorders, there is ample evidence for T cells responding to antigens present in the organs under attack. But in non-organ-specific autoimmunity, identification of the antigens recognized by T cells is often inadequate. However, histone-specific T cells are generated in patients with SLE and histone could play a ‘piggyback’ role in the formation of anti-DNA antibodies by substituting for natural antibody in the mechanism outlined in Figure 20.8.

Another possibility is that the T cells do not ‘see’ conventional peptide antigen (possibly true of anti-DNA responses), but instead recognize an antibody’s idiotype (an antigenic determinant on the V region of antibody).

In this view SLE, for example, might sometimes be initiated as an ‘idiotypic disease’, following the model shown in Figure 20.8. In this scheme, autoantibodies are produced normally at low levels by B cells using germline genes. If these then form complexes with the autoantigen, the complexes can be taken up by APCs (including B cells) and components of the complex, including the antibody idiotype, presented to T cells. Idiotype-specific T cells would then provide help to the autoantibody-producing B cells.

Evidence for the induction of anti-DNA and glomerulonephritis by immunization of mice with the idiotype of germline ‘natural’ anti-DNA autoantibody lends credence to this hypothesis.

The ‘waste disposal’ hypothesis of SLE

Antibodies to nuclear components are the serological hallmark of SLE. The question arising from this observation is how nuclear components normally hidden are detected by the immune system as antigen. The answer appears to lie with apoptosis. There is strong evidence that SLE (and possibly other autoimmune diseases) are diseases of failure of clearance of apoptotic cells, i.e. due to decreased macrophage ‘scavenger’ function. When a cell undergoes apoptosis (programmed cell death), blebs of cellular material are formed on the cell surface. Antigens normally buried deep within the cell (and so not detected by the immune system) are exposed on the cell surface. In healthy individuals these apoptotic cells are efficiently cleared. However, in SLE apoptosis is defective; it has been demonstrated that scavenging of apoptotic debris in vitro by macrophages from lupus patients is less efficient than by macrophages from healthy controls. Thus the antigens contained within the apoptotic blebs may trigger an autoimmune response. Apoptotic blebs vary in size; larger blebs contain antigens including Sm, Mi-2, Ro-60, and La, whilst smaller blebs contain fodrin, Jo-1, Ro-52, and ribosomal P. Antibodies against all of these targets can be found in systemic autoimmune diseases such as SLE, Sjögren’s syndrome, and myositis (Fig. 20.10).

Fig. 20.10 Induction of surface blebs during apoptosis

Apoptosis of keratinocytes exposed to ultraviolet light is illustrated. The different constituents of developing small and large surface blebs during apoptosis are shown. PARP denotes poly-ADP-ribose polymerase.

(Based on Rahman A, Isenberg DA: Systemic lupus erythematosus. N Engl J Med, 2008; 358:929–939.)

A defective complement pathway may also contribute to ineffective clearance of apoptotic cells as C1q binds to cell debris, allowing macrophages with C1q receptors to engulf the apoptotic cells. Complement deficiency in SLE is usually attributed to consumption as a secondary consequence of immune complex formation. However, it is clear that in a very small number of patients complement deficiency is the cause rather than the effect of SLE. In patients with these rare genetic disorders of deficiency of complement components (including C1q, C2 and C4, as discussed previously), there is a hugely increased risk of developing a lupus-like disease. Reduced clearance of immune complexes by the spleen has been demonstrated in a patient with C2 deficiency and SLE. This problem was correctable with transfusions of fresh-frozen plasma containing C2. The C1q knock-out mouse also provides evidence for the role of complement in clearance of apoptotic cells. This mouse develops a lupus-like glomerulonephritis, and renal biopsy reveals multiple apoptotic fragments.

The clustering of the lupus autoantigens in apoptotic blebs provides an elegant explanation for how epitope spread occurs in SLE. Intermolecular epitope spread occurs when molecules in close physical proximity are taken up by an APC. Uptake of an apoptotic cell by an APC will result in presentation of the antigens from the surface of the apoptotic cell. Thus the apoptotic cell acts as a cellular platform or ‘scaffold’ physically linking the multiple autoantigens targeted by the immune response in SLE.

Defective clearance of apoptotic cells may induce autoimmunity

In mice deletion of some of the molecules involved in the clearance of apoptotic cells leads to autoantibody production and lupus-like disease. However, not all deletions resulting in defective clearance of apoptotic cells result in autoimmunity. Furthermore, immunization of mice with apoptotic cells does not produce autoimmune disease. This may be accounted for by the immunosuppressive capabilities of apoptotic cells, especially through release of anti-inflammatory cytokines from phagocytic cells. SLE-related autoantibodies and glomerulonephritis resembling human lupus nephritis can however be induced in mice by co-immunization with an apoptotic cell binding protein (human beta2 glycoprotein1) and the powerful adjuvant, LPS. Not only did the mice produce high titres of IgG anti-beta2 glycoprotein1 antibodies, but also antibodies to Ro, La, dsDNA, Sm, and nRNP in sequential order, recapitulating human disease. Knockout of CD28 abrogated these effects, indicating that CD28-mediated co-stimulation of T cells is required. This mouse model provides evidence of intermolecular epitope spread, and suggests that long-lived memory T cells responding to a single protein (beta2 glycoprotein1) are capable of providing help to other auto-antigen specific B cells if the protein interacts with a ‘scaffold’ containing the other autoantigens. Of note, immunization with either LPS or beta2 glycoprotein1 alone did not result in lupus autoantibody production.

Induction of autoimmunity

Normally, naive autoreactive T cells recognizing cryptic self epitopes are not switched on because the antigen is presented on ‘non-professional’ APCs such as pancreatic β-islet cells or thyroid epithelial cells which lack co-stimulator molecules, or because it is presented only at low concentrations on ‘professional’ APCs. However the normal mechanisms that maintain self-tolerance may be bypassed.

Molecular mimicry by cross-reactive microbial antigens can stimulate autoreactive lymphocytes

Infection with a microbe bearing antigens that cross-react with the cryptic self epitopes (i.e. have shared epitopes) will load the professional APCs with levels of processed peptides that are sufficient to activate the naive autoreactive T cells. Once primed, these T cells are able to recognize and react with the self epitope on the non-professional APCs because they:

• no longer require a co-stimulatory signal; and

• have a higher avidity for the target, due to upregulation of accessory adhesion molecules (Fig. 20.11).

Fig. 20.11 Cross-reactive antigens induce autoimmune TH cells

The inability of naive TH cells to recognize autoantigen on a tissue cell, whether because of low concentration or low affinity, can be circumvented by a cross-reacting microbial antigen at higher concentration or with higher innate affinity, together with a co-stimulator such as B7 on a ‘professional’ APC; this primes the TH cells (1). Due to increased expression of accessory molecules (e.g. LFA-1 and CD2) the primed TH cells now have high affinity and, because they do not require a co-stimulatory signal, they can interact with autoantigen on ‘non-professional’ APCs such as organ-specific epithelial cells to produce autoimmune disease (2).

Cross-reactive antigens that share B cell epitopes with self molecules can also break tolerance, but by a different mechanism. Many autoreactive B cells cannot be activated because the CD4⁺ TH cells they need are unresponsive, either because:

• these TH cells are tolerized at lower concentrations of autoantigens than the B cells; or

• because they recognize only cryptic epitopes. However, these ‘helpless’ B cells can be stimulated if the cross-reacting antigen bears a ‘foreign’ carrier epitope to which the T cells have not been tolerized (Fig. 20.12). The autoimmune process may persist after clearance of the foreign antigen if the activated B cells now focus the autoantigen on their surface receptors and present it to normally resting autoreactive T cells, which will then proliferate and act as helpers for fresh B cell stimulation.

Fig. 20.12 Induction of autoantibodies by cross-reactive antigens

The B cell recognizes an epitope present on autoantigen, but coincidentally present also on a foreign antigen. Normally the B cell presents the autoantigen, but receives no help from autoreactive TH cells, which are functionally deleted. If a cross-reacting foreign antigen is encountered, the B cell can present peptides of this molecule to non-autoreactive T cells and thus be driven to proliferate, differentiate, and secrete autoantibodies.

Molecular mimicry operates in rheumatic fever

An example of a disease in which such molecular mimicry may operate is rheumatic fever, in which autoantibodies to heart valve antigens can be detected. These develop in a small proportion of individuals several weeks after a streptococcal infection of the throat. Carbohydrate antigens on the streptococci cross-react with an antigen on heart valves, so the infection may bypass T cell self tolerance to heart valve antigens. Historically many concepts about autoimmunity arose because the early immunologists often had a background in infectious diseases, and brought with them ideas about cross-reactivity and molecular mimicry extrapolated from rheumatic fever. However, the evidence for molecular mimicry in most chronic autoimmune diseases is lacking or absent. A notable exception is post-infective polyneuropathy.

There is circumstantial evidence for molecular mimicry in anti-neutrophil cytoplasmic antibody (ANCA) associated vasculitis (AAV). Antibodies to lysosomal membrane protein-2 (LAMP-2) are a subtype of ANCA found in most patients with pauci-immune focal necrotizing glomerulonephritis (FNGN). The autoantibodies to LAMP-2 commonly recognize an epitope with 100% homology to the bacterial adhesion FimH. Rats injected with FimH develop antibodies to LAMP-2 and pauci-immune FNGN. Furthermore, many humans with pauci-immune FNGN have evidence of recent infection with fimbriated organisms.

Shared B cell epitopes between Yersinia enterolytica and the extracellular domain of the thyroid stimulating hormone (TSH) receptor have been described.

In some cases foreign antigen can directly stimulate autoreactive cells

Another mechanism to bypass the tolerant autoreactive TH cell is where antigen or another stimulator directly triggers the autoreactive effector cells.

For example, lipopolysaccharide or Epstein–Barr virus causes direct B cell stimulation and some of the clones of activated cells will produce autoantibodies, though in the absence of T cell help these are normally of low titer and affinity. However, it is conceivable that an activated B cell might pick up and process its cognate autoantigen and present it to a naive autoreactive T cell.

Infection may trigger relapse in autoimmune disease

Many autoimmune diseases have a relapsing-remitting phenotype, characterized by periods of disease activity punctuated by periods of quiescence. This relapsing-remitting phenotype suggests there may be some varying influence on the autoimmune process. Clinical observation suggests that infections appear to trigger increased autoimmune disease activity, although proving a causal link is difficult. In Wegener’s granulomatosis relapses are closely correlated with recent infection, and chronic Staphylococcus aureus nasal carriage is linked with more frequent relapses of upper respiratory tract disease. Whether infection increases the supply of autoantigen or causes a general stimulation of the immune response through cytokine stimulation is unclear.

Cytokine dysregulation, inappropriate MHC expression, and failure of suppression may induce autoimmunity

It appears that dysregulation of the cytokine network can also lead to activation of autoreactive T cells, and as noted earlier genes affecting cytokines and their receptors (e.g. IL2RA) are implicated in autoimmune disease by genome-wide association studies.

One experimental demonstration of this is the introduction of a transgene for interferon-γ (IFNγ) into pancreatic β-islet cells. If the transgene for IFNγ is fully expressed in the cells, MHC class II genes are upregulated and autoimmune destruction of the islet cells results. This is not simply a result of a non-specific chaotic IFNγ-induced local inflammatory milieu because normal islets grafted at a separate site are rejected, implying clearly that T cell autoreactivity to the pancreas has been established.

The surface expression of MHC class II in itself is not sufficient to activate the naive autoreactive T cells, but it may be necessary to allow a cell to act as a target for the primed autoreactive TH cells. It was therefore most exciting when cells taken from the glands of patients with Graves’ disease were found to be actively synthesizing class II MHC molecules (Fig. 20.13) and so were able to be recognized by CD4⁺ T cells.

Fig. 20.13 Human thyroid sections stained for MHC class II

(1) Normal thyroid with unstained follicular cells, and an isolated dendritic cell that is strongly positive for MHC class II. (2) Thyrotoxic (Graves’ disease) thyroid with abundant MHC class II molecules in the cytoplasm, indicating that rapid synthesis of MHC class II molecules is occurring.

In this context it is interesting that isolated cells from several animal strains that are susceptible to autoimmunity are also more readily induced by IFNγ to express MHC class II molecules than cells from non-susceptible strains.

The argument that imbalanced cytokine production may also contribute to autoimmunity receives further support from the unexpected finding that tumor necrosis factor (TNF; introduced by means of a TNF transgene) ameliorates the spontaneous SLE-like disease of F1 (NZB × NZW) mice. Furthermore, serological (and less commonly clinical) features of SLE have developed in humans treated with TNF blockade.

Aside from the normal ‘ignorance’ of cryptic self epitopes, other factors that normally restrain potentially autoreactive cells may include:

• regulatory T cells;

• hormones (e.g. steroids);

• cytokines (e.g. transforming growth factor-β [TGFβ]); and

• products of macrophages.

Deficiencies in any of these factors may increase susceptibility to autoimmunity.

The feedback loop on TH cells and macrophages through the pituitary–adrenal axis is particularly interesting because defects at different stages in the loop turn up in a variety of autoimmune disorders (Fig. 20.14).

Fig. 20.14 Defects in the cytokine/hypothalamic-pituitary-adrenal feedback loop in autoimmunity

Production of IL-1 is defective in the NOD mouse (1) and diabetes-prone BB rat (2); the disease can be corrected by injection of the cytokine. The same is true for the production of TNFα by the NZB × W lupus mouse (3). Patients with rheumatoid arthritis have a poor hypothalamic response to IL-1 and IL-6 (4). The hypothalamic–pituitary axis is defective in the Obese strain chicken and in the Lewis rat, which is prone to the development of Freund’s adjuvant-mediated experimental autoimmune disease (5). (CRH, corticotropin releasing hormone; MØ, macrophage.)

For example, patients with rheumatoid arthritis have low circulating corticosteroid levels compared with controls. After surgery, although they produce copious amounts of IL-1 and IL-6, a defect in the hypothalamic paraventricular nucleus prevents the expected increase in adrenocorticotropin (ACTH) and adrenal steroid output.

There is currently intense interest focused on the role of Tregs. Patients with rheumatoid arthritis, for example, reveal a deficiency of Treg function (see below and Fig. 20.w5).

A subset of CD4 regulatory cells present in young healthy mice of the NOD strain, which spontaneously develop IDDM, can prevent the transfer of disease provoked by injection of spleen cells from diabetic animals into NOD mice congenic for the severe combined immunodeficiency trait; this regulatory subset is lost in older mice.

Pre-existing defects in the target organ may increase susceptibility to autoimmunity

We have already noted the sensitivity of target cells to upregulation of MHC class II molecules by IFNγ in animals susceptible to certain autoimmune diseases. Other evidence also favors the view that there may be a pre-existing defect in the target organ.

In the Obese strain chicken model of spontaneous thyroid autoimmunity, not only is there a low threshold of IFNγ induction of MHC class II expression by thymocytes, but when endogenous TSH is suppressed by thyroxine treatment, the uptake of iodine into the thyroid glands is far higher in the Obese strain than in a variety of normal strains. Furthermore, this is not due to any stimulating effect of the autoimmunity because immunosuppressed animals show even higher uptakes of iodine (Fig. 20.15).

Fig. 20.15 Thyroid ¹³¹I uptake in TSH-suppressed chickens

Thyroid ¹³¹l uptake in Obese strain chickens and in the related Cornell strain is abnormally high compared with that in normal strains. Endogenous TSH production was suppressed by administration of thyroxine; therefore the experiment measured TSH-independent ¹³¹l uptake. Values were far higher than normal in Obese strain chickens, which spontaneously develop thyroid autoimmunity, and even higher in the non-autoimmune Cornell strain from which the Obese strain was bred. This abnormality was not due to immune mechanisms because immunosuppression actually increased ¹³¹l uptake into the thyroid gland.

Interestingly, the Cornell strain (from which the Obese strain was derived by breeding) shows even higher uptakes of iodine, yet these animals do not develop spontaneous thyroiditis. This could be indicative of a type of abnormal thyroid behavior, which in itself is insufficient to induce autoimmune disease, but does contribute to susceptibility in the Obese strain.

Other situations in which the production of autoantigen is affected are:

• diabetes mellitus, in which one of the genetic risk factors is linked to a transcription factor controlling the rate of insulin production; and

• rheumatoid arthritis, in which the agalacto IgG glycoform is abnormally abundant.

The post-translational modification of arginine to citrulline, producing a new autoantigen in rheumatoid arthritis, represents yet another mechanism by which autoimmunity can be evoked.

Q. What evidence from rheumatoid arthritis supports the view that autoimmune diseases have a multifactorial etiology?

A. Defects in hypothalamic feedback, Tregs and post-translational modification of antigens (glycosylation and citrulline formation) as well as a strong MHC association have all been identified.

Autoimmune processes and pathology

Autoimmune processes are often pathogenic. When autoantibodies are found in association with a particular disease there are three possible inferences:

• the autoimmunity is responsible for producing the lesions of the disease;

• there is a disease process that, through the production of tissue damage, leads to the development of autoantibodies;

• there is a factor that produces both the lesions and the autoimmunity.

Autoantibodies secondary to a lesion (the second possibility) are sometimes found. For example, cardiac autoantibodies may develop after myocardial infarction.

However, sustained production of autoantibodies rarely follows the release of autoantigens by simple trauma. In most diseases associated with autoimmunity, the evidence supports the first possibility, that the autoimmune process produces the lesions.

The pathogenic role of autoimmunity can be demonstrated in experimental models

The most direct test of whether autoimmunity is responsible for the lesions of disease is to induce autoimmunity deliberately in an experimental animal and see if this leads to the production of the pathological lesions.

Autoimmunity can be induced in experimental animals by injecting autoantigen (self antigen) together with complete Freund’s adjuvant, and this does indeed produce organ-specific disease in certain organs. For example:

• thyroglobulin injection can induce an inflammatory disease of the thyroid;

• Myelin Basic Protein (MBP) can cause encephalomyelitis.

In the case of thyroglobulin-injected animals, not only are thyroid autoantibodies produced, but the gland becomes infiltrated with mononuclear cells and the acinar architecture crumbles, closely resembling the histology of Hashimoto’s thyroiditis.

While these animal models show that experimentally-induced autoimmunity can produce pathological lesions, caution needs to be taken when extrapolating evidence from such models to human disease. It is important to distinguish animal models in which autoimmune disease is artificially induced by large doses of mycobacterial adjuvant, from inbred animal strains which spontaneously develop autoimmune disease (e.g. F1 NZB × NZW mice, discussed below). The former probably have very limited relevance to the pathogenic mechanisms in spontaneous autoimmune disease in humans, not least because they are often of short duration. This may account for the failure of many drugs which showed promise in animal models to translate into effective therapeutics in clinical practice.

TH cells are essential for induction of autoimmunity in EAE

Susceptibility of rats and mice to MBP-induced experimental allergic encephalomyelitis (EAE) depends on a small number of gene loci, of which the most important are the MHC class II genes.

MBP-induced encephalomyelitis can be induced in susceptible strains by injecting T cells specific for MBP. These pathogenic T cells belong to the CD4/TH1 subset. It has been found that induction of disease can be prevented by treating the recipients with antibody to CD4 just before the expected time of disease onset, blocking the interaction of the TH cells’ CD4 with the class II MHC molecules of antigen-presenting target cells (see Chapter 8).

The results indicate the importance of class II-restricted autoreactive TH cells in the development of these conditions, and emphasize the prominent role of the MHC.

Examples of spontaneous autoimmunity

It has proved possible to breed strains of animals that are genetically programmed to develop autoimmune diseases closely resembling their human counterparts.

One well-established example mentioned above is the Obese strain chicken, which parallels human autoimmune thyroid disease in terms of:

• the lesion in the gland;

• the production of antibodies to different components in the thyroid; and

• the overlap with gastric autoimmunity.

So it is of interest that, when the immunological status of Obese strain chickens is altered, quite dramatic effects are seen on the outcome of the disease. For example:

• removal of the thymus at birth appears to exacerbate the thyroiditis, suggesting that the thymus exerts a controlling effect on the disease through Tregs; but

• if the entire T cell population is abrogated by combining thymectomy with massive injections of anti-chick T cell serum, both autoantibody production and the attack on thyroid are completely inhibited.

T cells therefore play a variety of pivotal roles as mediators and regulators of this disease.

More directly, a diabetogenic CD4⁺ T cell clone can induce the chronic leukocytic infiltrate of T cells and macrophages, which damages the insulin-producing β cells of the pancreatic islets of Langerhans in the NOD murine model of type 1 diabetes mellitus (Fig. 20.w2).

Fig. 20.w2 Destruction of β cells in the pancreatic islets of Langerhans in the non-obese diabetic mouse by a diabetogenic T cell clone

NOD neonates were injected with the diabetogenic CD4⁺ T cell clone, BDC2.5. The cells infiltrated the pancreas and there was recruitment of other NOD cells. Of the neonates that receive this clone 80% become diabetic within 20 days. Snap-frozen pancreas sections taken from representative neonates killed 10 days after transfer of BDC2.5, before the onset of overt diabetes, were stained for insulin (red fluorescence) and infiltrating cells (green fluorescence). (1) Section from a normal NOD neonate showing an islet, but no infiltration of CD3⁺ T cells. Other sections show islets infiltrated with (2) macrophages (F4/80⁺), (3) B cells (B220⁺), (4) CD3⁺ T cells, (5) CD4⁺ T cells, and (6) CD8⁺ T cells, this section showing serious loss of insulin producing cells. × 400.

(Figures kindly provided by Drs Jenny Phillips and Anne Cooke.)

The role of the T cells in mediating this attack is further emphasized by the amelioration and prevention of disease by treatment of the mice with a non-depleting anti-CD4 monoclonal antibody, which in the presence of the pancreatic autoantigens, insulin, and glutamic acid decarboxylase (GAD) induces specific T cell anergy.

The dependence of yet another spontaneous model, the F₁ hybrid of New Zealand Black and White strains, on the operation of immunological processes is aptly revealed by the suppression of the murine SLE (associated with immune complex glomerulonephritis and anti-DNA autoantibodies), by treatment with anti-CD4 antibody or azathioprine (Fig. 20.w3).

Fig. 20.w3 Suppression of autoimmune disease

The New Zealand Black (NZB) mouse spontaneously develops autoimmune hemolytic anemia. The hybrid between this and the New Zealand White (NZW) strain develops DNA autoantibodies and immune complex glomerulonephritis, as in patients with SLE. Immunosuppression with monoclonal antibodies to the TH cell marker CD4 considerably reduced the severity of the glomerulonephritis and the titer of double-stranded DNA autoantibodies at 8 months of age, showing the relevance of the immune processes to the generation of the disease.

(Based on data from Wofsy D, Seaman WE. J Exp Med 1985;161:378.)

Human autoantibodies can be directly pathogenic

When investigating human autoimmunity directly rather than using animal models, it is of course more difficult to carry out experiments. Nevertheless, there is much evidence to suggest that autoantibodies may be important in pathogenesis, as discussed below.

Autoantibodies can give rise to a wide spectrum of clinical thyroid dysfunction

A number of diseases have been recognized in which autoantibodies to hormone receptors may actually mimic the function of the normal hormone and produce disease. Graves’ disease (thyrotoxicosis) was the first disorder in which such anti-receptor antibodies were clearly recognized.

The phenomenon of neonatal thyrotoxicosis provides us with a natural ‘passive transfer’ study, because the IgG antibodies from the thyrotoxic mother cross the placenta and react directly with the TSH receptor on the neonatal thyroid. Many babies born to thyrotoxic mothers and showing thyroid hyperactivity have been reported, but the problem spontaneously resolves as the antibodies derived from the mother are catabolized in the baby over several weeks.

Whereas autoantibodies to the TSH receptor may stimulate cell division and/or increase the production of thyroid hormones, others can bring about the opposite effect by inhibiting these functions, a phenomenon frequently observed in receptor responses to ligands that act as agonists or antagonists.

Different combinations of the various manifestations of thyroid autoimmune disease – chronic inflammatory cell destruction, and stimulation or inhibition of growth and thyroid hormone synthesis – can give rise to a wide spectrum of clinical thyroid dysfunction (Fig. 20.16).

Fig. 20.16 The spectrum of autoimmune thyroid disease

Responses involving thyroglobulin and the thyroid peroxidase (microsomal) surface microvillous antigen lead to tissue destruction, whereas autoantibodies to TSH (and other?) receptors can stimulate or block metabolic activity or thyroid cell division. ‘Hashitoxicosis’ is an unconventional term that describes a gland showing Hashimoto’s thyroiditis and Graves’ disease simultaneously.

A variety of other diseases are associated with autoantibodies

Myasthenia gravis provides an example of a disease where some of the autoantibodies can act as a receptor antagonist, blocking the acetyl choline receptor on the post-synaptic membrane of the neuromuscular junction, thus causing muscle weakness and fatigability. A parallel with neonatal hyperthyroidism has been observed – antibodies to acetylcholine receptors from mothers who have myasthenia gravis cross the placenta into the fetus and may cause transient muscle weakness in the newborn baby.

ANCA are thought to be directly pathogenic in ANCA-associated vasculitides by causing neutrophil degranulation leading to endothelial damage. ANCA directed against myeloperoxidase (MPO) are commonly found in microscopic polyangiitis. Knockout mice lacking murine MPO (muMPO) can be immunized with muMPO to produce antibodies to muMPO and passive transfer of these antibodies into suitable recipients produces necrotizing arteritis, focal necrotizing crescenteric glomerulonephritis and alveolar capillaritis and hemorrhage reminiscent of human disease.

Somewhat rarely, autoantibodies to insulin receptors and to α-adrenergic receptors can be found, the latter associated with bronchial asthma.

Neuromuscular defects can be elicited in mice injected with serum containing antibodies to presynaptic calcium channels from patients with the Lambert–Eaton syndrome, while sodium channel autoantibodies have been identified in the Guillain–Barré syndrome.

Yet another example of autoimmune disease is seen in rare cases of male infertility where antibodies to spermatozoa lead to clumping of spermatozoa, either by their heads or by their tails, in the semen.

In pernicious anemia an autoantibody interferes with the normal uptake of vitamin B₁₂

Vitamin B₁₂ is not absorbed directly, but must first associate with a protein called intrinsic factor; the vitamin–protein complex is then transported across the intestinal mucosa.

Early passive transfer studies demonstrated that serum from a patient with pernicious anemia (PA), if fed to a healthy individual together with intrinsic factor–B₁₂ complex, inhibited uptake of the vitamin.

Subsequently, the factor in the serum that blocked vitamin uptake was identified as antibody against intrinsic factor. It is now known that plasma cells in the gastric mucosa of patients with PA secrete this antibody into the lumen of the stomach (Fig. 20.17).

Fig. 20.17 Failure of vitamin B₁₂ absorption in pernicious anemia

Normally, dietary vitamin B₁₂ is absorbed by the small intestine complexed with intrinsic factor (IF), which is synthesized by parietal cells in gastric mucosa. In pernicious anemia, locally synthesized autoantibodies, specific for intrinsic factor, combine with intrinsic factor to inhibit its role as a carrier for vitamin B₁₂.

Antibodies to the glomerular capillary basement membrane cause Goodpasture’s disease

Goodpasture’s disease is characterized clinically by rapidly progressive glomerulonephritis and pulmonary haemorrhage. Patients with Goodpasture’s disease have circulating antibodies to the glomerular capillary basement membrane which bind to the kidney and lung (see Fig. 25.3). Evidence for the direct pathogenity of these antibodies was demonstrated by the passive transfer of antibodies eluted from renal biopsy specimens into primates (whose renal antigens were similar to humans). The injected monkeys subsequently died from glomerulonephritis. Subsequent work has shown that these antibodies bind to the several non-collagenous-1 (NC-1) domains of type IV collagen in the GBM. Moreover, immunization of animals with NC-1 domains induces glomerulonephritis, providing a causal link between autoantigen and antibody.

Blood and vascular disorders caused by autoantibodies include AHA and ITP

Autoimmune hemolytic anemia (AHA) and idiopathic thrombocytopenic purpura (ITP) result from the synthesis of autoantibodies to red cells and platelets, respectively.

The primary antiphospholipid syndrome characterized by recurrent thromboembolic phenomena and fetal loss is triggered by the reaction of autoantibodies with a complex of β₂-glycoprotein 1 and cardiolipin.

The β₂-glycoprotein is an abundant component of atherosclerotic plaques, and there is increasing attention to the idea that autoimmunity may initiate or exacerbate the process of lipid deposition and plaque formation in this disease, the two lead candidate antigens being heat-shock protein 60 (Fig. 20.18) and the low density lipoprotein, apoprotein B.

Fig. 20.18 Upregulation of heat-shock protein 60 in endothelial cells at a site of hemodynamic stress

Hsp60 expression (red) co-localized with intercellular adhesion molecule-1 (ICAM-1) expression (black) by endothelial cells and cells in the intima (macrophages) at the bifurcation of the carotid artery of a 5-month-old child. × 240.

(Photograph kindly provided by Professor G Wick.)

Immune complexes appear to be pathogenic in systemic autoimmunity

In the case of SLE, it can be shown that complement-fixing complexes of antibody with DNA and other nucleosome components such as histones are deposited in the kidney (see Fig. 25.3), skin, joints, and choroid plexus of patients, and must be presumed to produce type III hypersensitivity reactions as outlined in Chapter 25. A variety of different antibodies have been eluted from the kidney biopsies of patients with SLE. These include anti-dsDNA (nucleosomes), anti-Ro and anti-Sm/RNP. Whilst placing these antibodies at the scene of the crime, their mere presence does not prove they ‘pulled the trigger’. However, experiments using murine monoclonal anti-dsDNA antibodies in a rat kidney perfusion system, and other evidence from the use of human hybridoma derived anti-dsDNA antibodies in SCID mice, provides compelling evidence that some anti-dsDNA antibodies are genuinely pathogenic.

It has been proposed that anti-dsDNA/nucleosome antibodies bind to the negatively charged surface of the renal glomerulus via a histone (positively charged) ‘bridge’. The histone is part of the nucleosome complex. The formation of immune complexes at the glomerular surface membrane is thought to induce an inflammatory response leading to the kidney damage so frequently seen in patients with SLE, though the precise mechanisms have yet to be elucidated.

Autoantibodies to IgG provoke pathological damage in rheumatoid arthritis

The erosions of cartilage and bone in rheumatoid arthritis are mediated by macrophages and fibroblasts, which become stimulated by cytokines from activated T cells and immune complexes generated by a vigorous immunological reaction within the synovial tissue (Fig. 20.19).

Fig. 20.19 Pathology of rheumatoid arthritis

In the rheumatoid arthritis joint an inflammatory infiltrate is found in the synovial membrane, which hypertrophies forming a ‘pannus’ (1 and 2). This covers and eventually erodes the synovial cartilage and bone. Immune complexes and neutrophils (PMNs) are detectable in the joint space and in the extra-articular tissues where they may give rise to vasculitic lesions and subcutaneous nodules.

(Histological section reproduced from Woolf N. Pathology: basic and systemic. London: W.B. Saunders; 1998.)

The complexes can arise through the self-association of IgG rheumatoid factors specific for the Fcγ domains – a process facilitated by the striking deficiency of terminal galactose on the biantennary N-linked Fc oligosaccharides (Fig. 20.20). This agalacto glycoform of IgG in complexes can exacerbate inflammation through reaction with mannose-binding lectin and production of TNFα.

Fig. 20.20 Self-associated IgG rheumatoid factors forming immune complexes

The binding between the Fab on one IgG rheumatoid factor and the Fc of another involves the hypervariable region of the combining site. As it has been established that the Fab oligosaccharides, which occur on approximately one in three different immunoglobulin molecules, are not defective with respect to glycosylation in rheumatoid arthritis, a Fab galactose residue could become inserted in the Fc pocket left vacant by a galactose-deficient Cγ2 oligosaccharide, so increasing the strength of intermolecular binding. The stability and inflammatory potency of these complexes is increased by binding IgM rheumatoid factor and Clq.

Evidence for directly pathogenic T cells in human autoimmune disease is hard to obtain

Adoptive transfer studies have shown that TH1 cells are responsible for directly initiating the lesions in experimental models of organ-specific autoimmunity.

In humans, evidence for a pivotal role of T cells in the development of autoimmune disease includes:

• the production of high-affinity, somatically mutated IgG autoantibodies characteristic of T-dependent responses;

• the isolation of thyroid-specific T cell clones from the glands of patients with Graves’ disease;

• the beneficial effect of ciclosporin in prediabetic individuals; and

• the close associations with certain HLA haplotypes.

However, it is difficult to identify a role for the T cell as a pathogenic agent as distinct from a TH function in the organ-specific disorders.

Indirect evidence from circumstances showing that antibodies themselves do not cause disease, such as in babies born to mothers with type 1 diabetes mellitus, may be indicative.

The central role of TH1 cells in some autoimmune diseases has been challenged

Many autoimmune diseases such as rheumatoid arthritis were thought to be primarily driven by TH1 cells. For this reason, the proposal to use rituximab (a monoclonal antibody against the CD20 molecule which is expressed on B cells but not plasma cells) in rheumatoid arthritis was initially met with great skepticism. However, Rituximab’s success in the treatment of RA has emphasized that B cells play a key role its pathogenesis, and has led some to doubt that RA is ‘TH1 driven’. However, the precise mechanisms by which B cell depletion has a therapeutic effect are unclear. B cells are more than simply precursors to antibody producing plasma cells – they also act as antigen presenting cells, interacting with T cells in several ways. Thus rituximab’s efficacy in RA cannot be used as proof that T cells are not important in its pathogenesis.

TH17 cells – a new player in autoimmunity?

Recent interest has focused on the role of TH17 cells in auto-immunity. Their initial discovery came about when it was noted that deletion of key TH1 molecules such as IL12 and IFN-γR did not abrogate EAE and collagen induced arthritis (CIA) in mice. In fact, these animals showed enhanced susceptibility, casting doubt on the role of TH1 cells as the fundamental players in autoimmunity. This led to the identification of a new cytokine IL23 – knockout of IL23 is protective against experimentally induced autoimmunity and IL23 was subsequently shown to induce IL-17 from activated T cells. Such IL-23 driven TH cells show a unique pattern of gene expression (which differs from that of IL-12 driven TH1 cells), and are now known to be TH17 cells. Whilst TGFβ, IL1β, and IL6, and not IL23, are the key cytokines for the induction of TH17 cells, IL23 is important for their maintenance; IL-23 deficient mice show normal numbers of TH1 cells but a reduction in TH17 cells.

In humans, high levels of IL-17 and its receptor are found in the synovial fluid and tissue of patients with RA. However, the number of TH17 cells is not elevated in the synovial fluid or PBMCs with RA compared to healthy controls. Multiple SNPs in the IL23 receptor gene region, as well as other genes involved in the IL23/TH17 pathway, are associated with inflammatory bowel disease. The results of trials of anti-IL-17 monoclonal antibodies in patients with RA and MS will help clarify the role of TH17 cells in these diseases.

Autoantibodies for diagnosis, prognosis, and monitoring

Whatever the relationship of autoantibodies to the disease process, they frequently provide valuable markers for diagnostic purposes. A particularly good example is the test for mitochondrial antibodies, used in diagnosing primary biliary cirrhosis (Fig. 20.21). Historically, exploratory laparotomy was needed to obtain this diagnosis, and was often hazardous because of the age and condition of the patients concerned. Some other diagnostically useful antibodies are listed below:

• antibodies to cyclic citrullinatated peptides (CCP) have a higher specificity than rheumatoid factor in the diagnosis of RA;

• the presence of anti-nuclear antibodies is one of the revised ACR criteria for SLE, but is non-specific. In contrast anti-dsDNA antibodies (also one of the ACR criteria) are highly specific for SLE, being present in 70% of SLE patients but in less than 0.5% of controls. Anti-Sm antibodies are found in 10% of Caucasian and 30% of Afro-Caribbean SLE patients; like dsDNA antibodies they have a high specificity for SLE;

• autoantibodies often have predictive value. For instance, individuals testing positively for antibodies to both insulin and glutamic acid decarboxylase have a high risk of developing type 1 diabetes mellitus.

Fig. 20.21 Diagnostic value of anti-mitochondrial antibodies

Mitochondrial antibody tests using indirect immunofluorescence, together with percutaneous liver biopsy, can be used to assist in the differential diagnosis of these diseases. Many patients with primary biliary cirrhosis, but less than half of patients with cryptogenic cirrhosis or chronic active hepatitis, have anti-mitochondrial antibodies. The antibodies are rare in the other diseases.

Prognosis and disease subtype:

• in rheumatoid arthritis, CCP antibodies are associated with a poor prognosis and predict erosive disease;

• within the context of a patient with SLE the presence of specific autoantibodies are associated with specific disease manifestations. For example, anti-La antibodies are associated with features of Sjögren’s syndrome; anticardiolipin antibodies and anti-beta2 glycoprotein1 antibodies with thrombosis and miscarriage; anti-dsDNA antibodies with glomerulonephritis; anti-RNP with pulmonary hypertension, and anti-Ro with photosensitivity and the neonatal lupus syndrome.

Disease monitoring:

• anti-dsDNA antibodies can be used as a measure of disease activity in SLE. A rising titer of anti-dsDNA antibodies often heralds a disease flare, especially if accompanied by a falling C3 level, and should prompt the clinician to monitor the patient more frequently;

Treatment of autoimmune diseases

Many autoimmune diseases can be treated successfully. Often, in organ-specific autoimmune disorders, the symptoms can be corrected by metabolic control. For example:

• hypothyroidism can be controlled by administration of thyroxin;

• type 1 diabetes mellitus can be controlled by administration of insulin;

• in pernicious anemia, metabolic correction is achieved by injection of vitamin B₁₂;

• in myasthenia gravis metabolic correction is achieved by administration of cholinesterase inhibitors.

If the target organ is not completely destroyed, it may be possible to protect the surviving cells by transfection with FasL or TGFβ genes.

Where function is completely lost and cannot be substituted by hormones, as may occur in lupus nephritis or chronic rheumatoid arthritis, tissue grafts or mechanical substitutes may be appropriate. In the case of tissue grafts, protection from the immunological processes that necessitated the transplant may be required.

Conventional immunosuppressive therapy with anti-mitotic drugs at high doses can be used to damp down the immune response, but, because of the dangers involved, tends to be used only in organ or life-threatening disorders such as SLE, myositis and ANCA-associated vasculitis (AAV).

Advances in treatment have transformed the 5 year survival rate in severe systemic autoimmune diseases such as SLE and AAV from around 50% and less than 10% respectively in the mid 20th century to over 90% today. However, the costs of this success in controlling autoimmune disease activity are the adverse effects of immunosuppressants, especially glucocorticoids. Most of the early mortality in SLE or AAV is now due to infection secondary to immunosuppressive therapy rather than from uncontrolled autoimmune disease. Long term use of drugs such as cyclophosphamide and azathioprine also increases the risk of infertility and possibly malignancy. The challenge now is to minimize treatment toxicity. This is likely to be achieved through several means. Firstly, the rational design of targeted therapies (in contrast to the ‘shot-gun’ approach of non-specific cytotoxics such as cyclophosphamide) should reduce the toxicity of treatment regimes. Secondly, it should become possible to identify patients through means of biomarkers in whom immunosuppression can be safely reduced. Some such biomarkers have already been identified but require further validation before translation into clinical practice. For example, CD8⁺ T cell transcription signatures from blood samples taken at time of diagnosis in patients with AAV and SLE identify subgroups of patients with high and low risk of subsequent relapse.

The most recent treatments for autoimmune disease are antibodies that target individual elements of the immune system, including:

• TNFα and IL-6 in rheumatoid arthritis;

• B cells in several autoimmune diseases (RA, SLE, AAV);

• co-stimulatory interactions in rheumatoid arthritis;

• trans-endothelial migration in multiple sclerosis.

‘Biologics’ – a new paradigm in the treatment of autoimmune rheumatic disease

The treatment of patients with rheumatoid arthritis has been revolutionized by the introduction of biological agents. Infliximab, a murine-human chimeric antibody against TNFα, not only markedly alleviates the symptoms of RA such as joint pain, stiffness and swelling, but also halts the progression of joint destruction (Fig. 20.w4). A number of other anti-TNF agents have subsequently become available including adalimumab (a fully humanized monoclonal antibody against TNF), etanercept (a soluble TNF receptor), and certolizumab (a pegylated anti-TNF antibody). It is fascinating to record that the compromised regulatory T cell function in the patients is reversed by such therapy (Fig. 20.w5). Similarly impressive results in RA are seen with B-cell depletion using rituximab, an anti-CD20 monoclonal antibody, and tocilizumab, a monoclonal antibody against IL-6. Other biologics licensed for use in RA include abatacept, which inhibits the co-stimulatory interaction between T cells and APCs, and anakinra, an IL-1 blocker.

Fig. 20.w4 Synergy of anti-TNFα and methotrexate in the treatment of rheumatoid arthritis

(1) Infusions were given at times indicated by the arrows. Median joint scores were more effectively reduced by a combination of anti-TNFα with methotrexate, which (2) eliminated the anti-idiotypic response to infliximab (a humanized anti-TNFα monoclonal antibody).

(Data reproduced from Maini RN, et al. Arthritis Rheum 1998;41:1552, with permission of the authors and publishers.)

Fig. 20.w5 Reversal of compromised regulatory T cell function in rheumatoid arthritis by anti-TNFα therapy

CD4⁺ CD25^– cells isolated from peripheral blood were stimulated with a mixture of anti-CD3 and anti-CD28 alone or after adding back the CD4⁺ CD25⁺ putative regulatory T cells. Activation of the T cells was assessed by staining for intracellular TNFα. The lines join values for individual patients after adding back regulatory CD4⁺ CD25⁺ T cells. Although the regulatory T cells were dysfunctional in active rheumatoid arthritis (RA), they were able to suppress activation of the CD25^– population after successful anti-TNFα therapy.

(Data adapted from Ehrenstein MR, et al. J Exp Med 2004;200:277–285.)

Rituximab has shown notable efficacy across the spectrum of autoimmune disease from ITP to multiple sclerosis. Randomized controlled trials show that it has equivalent efficacy to cyclophosphamide in AAV. It is particularly useful in patients with AAV who are refractory to cyclophosphamide treatment. Rituximab treatment results in reduction of the titer of the pathogenic anti-neutrophil cytoplasmic antibodies. Despite initial encouraging results from uncontrolled series, two randomized controlled trials did not demonstrate benefit from rituximab in SLE. However, this may be accounted for by poor trial design. Belimumab, which targets the B-cell survival factor BAFF (also known as BLyS), has, in contrast to rituximab, demonstrated efficacy in SLE in two randomized trials. In the future, the combination of an anti-CD20 agent such as rituximab with anti-BLys therapy using belimumab may produce longer lasting disease remission.

Less well-established approaches to treatment may become practicable

As we understand more about the molecular mechanisms underlying autoimmunity, targeted therapy is becoming increasingly possible (Fig. 20.22):

• several centers are trying out autologous stem cell transplantation following hematoimmunoablation with cytotoxic drugs for patients with severe SLE and vasculitis;

• repeated injection of Cop 1 (a random copolymer of alanine, glutamic acid, lysine, and tyrosine) reduces relapse rate in relapsing–remitting multiple sclerosis. Cop 1 was originally designed to simulate the postulated ‘guilty’ autoantigen, MBP, and induce experimental autoimmune encephalitis; paradoxically it had the opposite effect. This suggests it is possible to achieve antigen-specific immune suppression.

• eculizumab, a monoclonal antibody against complement component C5, has been used successfully in paroxysmal nocturnal hemoglobinuria. It may prove effective in SLE;

• TH cell subsets show considerable plasticity in vivo. Manipulation of the balance between pro-inflammatory TH17 cells and anti-inflammatory regulatory T cells may prove fruitful. Therapeutic interventions to increase the numbers of Tregs might increase self-tolerance and ‘put the brakes’ on inflammatory cascades. Adoptive transfer of polyclonal Tregs can prevent the onset of diabetes in NOD mice. However, it has become clear that successful treatment of established autoimmune disease by adoptive transfer of Tregs is antigen-specific; only Tregs specific for pancreatic antigens (and not polyclonal T regs) can reverse diabetes in NOD mice. This presents difficulties for treatment in human disease, where the inciting antigen is rarely known;

• inhibition of the activation or differentiation of TH17 cells, and/or blockade of linked cytokines such as IL17A and IL-23 may prove valuable. Trials of anti-IL17 monoclonal antibodies are already underway;

• small molecule therapies have the advantage of being cheaper to manufacture than monoclonal antibodies, of particular relevance in an era of spiraling healthcare costs. Kinases are intracellular enzymes that play a crucial role in signal transduction pathways controlling a number of cellular functions. Kinase inhibitors have been successfully used in oncology, but may also prove effective in autoimmune diseases. For example, a splenic tyrosine kinase inhibitor, reduces severity of antibody-mediated glomerulonephritis in rats, and is being trialled in RA.

Fig. 20.22 Targeted therapeutic approaches in systemic lupus erythematosus

This simplified diagram, which is based on our increased understanding of the immunologic events thought to occur in lupus, indicates the targets of current therapeutic interventions. (APC, antigen-presenting cell; BlyS, B-lymphocyte stimulator; CAML, calcium modulator and cyclophilin ligand; CTLA-4-Ig, cytotoxic T-lymphocyte-associated protein 4 IgG1; TACl-Ig, transmembrane activator and CAML interactor immunoglobulin.)

(Based on Rahman A, Isenberg DA: Systemic lupus erythematosus. N Engl J Med 2008; 358:929–939.)

The prospects for these new treatments either individually or in combination with therapeutic antibodies is very positive.

Critical thinking: Autoimmunity and autoimmune disease (see p. 440 for explanations)

Miss Jacob, a 30-year-old Caribbean woman, was seen in a rheumatology clinic with stiff painful joints in her hands, which were worse first thing in the morning. Other symptoms included fatigue, a low-grade fever, a weight loss of 2 kg, and some mild chest pain. Miss Jacob had recently returned to the UK from a holiday in Jamaica and was also noted to be taking the combined oral contraceptive pill. Past medical history of note was a mild autoimmune hemolytic anemia 2 years previously.

On examination Miss Jacob had a non-specific maculopapular rash on her face and chest, and patchy alopecia (hair loss) over her scalp. Her mouth was tender and examination revealed an ulcer on the soft palate. She had moderately swollen and tender proximal interphalangeal joints. Her other joints were unaffected, but she had generalized muscle aches. The results of investigations are shown in the table.

Investigation	Result
radiograph of hands	soft tissue swelling, but no bone erosions
chest radiograph	a small pleural effusion at the right lung base
full blood count	a mild normocytic, normochromic anemia and mild lymphocytopenia
C-reactive protein levels	normal
erythrocyte sedimentation rate	raised
rheumatoid factor	negative
serum IgG levels	raised
anti-nuclear antibodies (ANA)	positive by immunofluorescence
anti-double-stranded DNA, anti-RNA, and anti-histone	positive by ELISA antibodies
complement (C3 and C4) levels	low
skin biopsy from an area unaffected by the rash	deposition of IgG and complement components at the junction between dermis and epidermis (lupus ‘band’ test)