Pathogenesis of Graves' disease

INTRODUCTION — Graves' disease is a syndrome that may consist of hyperthyroidism, goiter, thyroid eye disease (Graves' orbitopathy), and occasionally a dermopathy referred to as pretibial or localized myxedema (PTM). The terms Graves' disease and hyperthyroidism are not synonymous, because some patients may have orbitopathy but no hyperthyroidism, and there are other causes of hyperthyroidism in addition to Graves' disease.

Hyperthyroidism is the most common feature of Graves' disease, affecting nearly all patients, and is caused by autoantibodies to the thyrotropin receptor (TRAb) that activate the receptor, thereby stimulating thyroid hormone synthesis and secretion as well as thyroid growth (causing a diffuse goiter). The presence of TRAb in serum and an orbitopathy on clinical examination distinguishes the disorder from other causes of hyperthyroidism. Other causes of an overactive thyroid gland are discussed separately. (See "Disorders that cause hyperthyroidism".)

This topic will review the immune pathogenesis of Graves' thyroid disease, with emphasis on the role of B and T cells in the production of the TRAb that are responsible for the thyroid stimulation and growth. The pathogenesis of Graves' orbitopathy and dermopathy are reviewed separately. (See "Clinical features and diagnosis of thyroid eye disease" and "Pretibial myxedema (thyroid dermopathy) in autoimmune thyroid disease".)

THE THYROID GLAND IN GRAVES' DISEASE — The thyroid is usually, but not always, diffusely enlarged. The histology of the thyroid gland in patients with Graves' hyperthyroidism is characterized by follicular hyperplasia, intracellular colloid droplets, cell scalloping, a reduction in follicular colloid, and a patchy (multifocal) lymphocytic infiltration. Only rarely are lymphoid germinal centers seen. The histologic picture may be greatly influenced by pretreatment with antithyroid drugs, causing an underestimation of the degree of lymphocytic infiltration (picture 1). The majority of intrathyroidal lymphocytes are T cells but plenty of B cells may be present, though nothing like that seen in chronic autoimmune thyroiditis (Hashimoto's disease). In some areas, thyroid epithelial cell size correlates with the intensity of the lymphocytic infiltrate, suggesting thyroid-cell stimulation by local B cells secreting stimulating TRAb [1]. In addition, scattered, small areas of thyroid cell apoptosis may be found if stained for appropriately [2,3].

THE TSH RECEPTOR — In Graves' disease, the main autoantigen is the thyroid-stimulating hormone (TSH) receptor (TSHR), which is expressed primarily in the thyroid but also in adipocytes, fibroblasts, bone cells, and a variety of additional sites [4,5]. This antigen has been reviewed extensively elsewhere (figure 1) [4,5]. The TSHR is a G-protein coupled receptor with seven transmembrane-spanning domains responsible for signal transduction. TSH, acting via the TSHR, regulates thyroid growth and thyroid hormone production and secretion. The TSHR undergoes complex post-translational processing involving dimerization and intramolecular cleavage; the latter modification leaves a two-subunit structural form of the receptor. Data suggest that there is eventual shedding or degradation of the TSHR ectodomain [6], although this has not been demonstrated in vivo. Each of these post-translational events may influence the antigenicity of the receptor and the shed ectodomain fragment is considered the most antigenic region of the receptor. However, factors that contribute to TSHR presentation as a target for the immune system in humans are not well understood but are considered to be primarily factors that build on a state of enhanced genetic susceptibility combined with a failure of immune tolerance. Such susceptibility may be translated by variable expression of the TSHR on thymic epithelial cells, which is of great importance in determining self-tolerance [7,8].

THYROID AUTOANTIBODIES — It is well known that lymphocytes from Graves' thyroid tissue spontaneously secrete thyroid autoantibodies, including thyrotropin receptor antibodies (TRAb), in vitro, providing evidence of their activated state [9]. Additional evidence for their presence and activated state comes from the decline in serum thyroid autoantibody concentrations after antithyroid drug treatment, after thyroidectomy, and late after radioactive iodine therapy. (See 'Influence of radioiodine on TRAb' below.)

Autoantibodies to the TSH receptor — Over 50 years ago, serum from patients with Graves' hyperthyroidism was found to contain a long-acting thyroid stimulator (LATS) [10]. LATS proved to be an immunoglobulin and inhibited the binding of radiolabeled TSH to thyroid membranes, suggesting that such activity was due to the presence of an antibody to the TSH receptor (now called TRAb) [11].

The first proof that TRAb stimulated the thyroid gland in humans came from infusion experiments in human volunteers and later from the transient detection of TRAb in the serum of hyperthyroid neonates of mothers with Graves' disease and persisting TRAb [12,13].

Stimulatory TRAb have several other characteristics:

●They are specific for autoimmune thyroid disease, especially Graves' disease, in contrast to antibodies to thyroglobulin (Tg) and thyroid peroxidase (TPO) which occur in the "normal" population. Almost all patients with Graves' hyperthyroidism have detectable TRAb when measured by sensitive assays [14-16]. TRAb can also be detected in approximately 10 percent of Hashimoto's thyroiditis patients. These antibodies are unique to humans; no animals develop Graves' disease, although it can be induced in rodents by immunization with native TSH receptor (TSHR) [17-20].

●TRAb are usually of the immunoglobulin G1 (IgG1) subclass, which suggests that they are oligoclonal [21], in contrast to antibodies to TPO and Tg, which are polyclonal.

●The serum concentrations of TRAb are relatively low to begin with and tend to decline in patients treated with an antithyroid drug. If high concentrations persist, the patient is likely to become hyperthyroid again when the drug is discontinued. However, such a reaction assumes the thyroid is capable of secreting excess thyroid hormones once again and has not been damaged by ongoing thyroiditis or the patient has iodine deficiency [22-24]. Measuring serum TRAb in these patients can be helpful, but only when the result is positive. A significant number of patients negative for TRAb after a course of antithyroid drugs will still have a recurrence [25]. In practice, we always measure serum TRAb at the time of planned cessation of drug therapy and continue the drugs if TRAb remain detectable [22,23].

●Stimulating TRAb, like TSH, stimulate the synthesis and activity of the sodium-iodide symporter, explaining the increased uptake of iodide by thyroid tissue in Graves' disease in the absence of TSH [26].

●TRAb, like TSH, stimulate different subtypes of G proteins (primarily Gs-alpha), resulting especially in increased protein kinase A (PKA) and thyroid adenylate cyclase activity, which leads to increased thyroid hormone synthesis, secretion, and cell survival. High levels of TRAb also stimulate Gq and then the protein kinase C (PKC) pathway leading to modulation of cell proliferation [27,28].

Influence of radioiodine on TRAb — In patients treated with radioactive iodine, the serum thyrotropin receptor antibodies (TRAb) concentrations initially rise, reaching a peak three to five months after treatment, and then gradually decline [29]. The initial increase in serum antibody concentrations after radioiodine therapy may explain why, in some patients, Graves' orbitopathy may first appear or may transiently worsen afterwards (see "Treatment of thyroid eye disease"). The TRAb may then either gradually decline or, more commonly, persist for many years after radioiodine treatment [30]. Although it is theoretically possible for extrathyroidal TSHRs to act as antigenic stimuli in the absence of the thyroid, TRAb gradually fall after thyroidectomy and disappear in 70 to 80 percent of patients after 18 months [30].

Different types of TRAb — Not all thyrotropin receptor antibodies (TRAb) are stimulatory. Some, including those found in the serum of patients with Hashimoto's thyroiditis, block the binding and action of TSH and, therefore, can cause hypothyroidism (figure 2). Blocking TRAb can be found in 10 to 15 percent of Hashimoto's patients. However, some patients with Graves' disease have a mixture of TRAb, both stimulating and blocking, and the clinical presentation may depend upon a balance between these different antibodies. A third group of TRAb used to be called neutral TRAb, and they bind to the hinge region of the receptor, which connects the ectodomain to the transmembrane domain, and they do not influence TSH binding. These antibodies can in fact have cell-signaling capability, inducing thyroid cell stress and even apoptosis. They are better referred to as hinge-region or cleavage-region TRAb, reflecting the region on the TSHR to which they are directed but remain of uncertain clinical significance [31,32]. Evidence of their potential role comes from a mouse model where such neutral antibodies induced thyroid cell stress and apoptosis [33].

Binding sites for TRAb — Thyrotropin receptor antibodies (TRAb) bind to a complex conformational epitope on the extracellular domain of the TSHR, principally in the region of the receptor to which TSH binds, which is a binding pocket encompassing the leucine-rich repeat region (figure 1) [34]. The differences in functional activity of TRAb (stimulating, blocking, or "neutral") relate to their variable molecular binding characteristics, resulting in different cell-signaling initiatives. Stimulating TRAb only have affinity for a conformational epitope in the ectodomain with approximately a dozen binding sites and will not interact with an unfolded/reduced receptor molecule [4]. Shedding of the subunit of the extracellular domain into the extracellular fluid may be an important stimulus for TRAb formation. Only a minority of TRAb bind to linear epitopes, and these are likely to be of the blocking or neutral variety [35-38].

Relationship with autoimmune thyroiditis — Lymphocytic infiltration of the thyroid and anti-Tg and anti-TPO antibodies in the serum occur in both Graves' disease and chronic autoimmune (Hashimoto's) thyroiditis, suggesting that the two disorders are related in fundamental ways and consistent with their appearance within the same family and their sharing of a number of susceptibility genes including human leukocyte antigen (HLA) (see "Pathogenesis of Hashimoto's thyroiditis (chronic autoimmune thyroiditis)"). Therefore, it appears that Graves' disease may develop on a background of thyroiditis. Several other observations are compatible with this hypothesis:

●Areas of cellular apoptosis may be seen even in Graves' thyroid glands [2,3].

●The presence of antibodies that bind to the TSHR in both disorders.

●Progression from Graves' hyperthyroidism to chronic autoimmune thyroiditis and hypothyroidism is well recognized [39]. The converse also occurs [40], and there are patients who have hypothyroidism one year, Graves' hyperthyroidism another, and hypothyroidism again later [41]. This is sometimes called Graves' alternans.

●In families of patients, some members may have Graves' disease and others may have chronic autoimmune thyroiditis [42].

Modifying the B cell repertoire — Since thyroid autoantibodies are the hallmark of Graves' disease, it is clear that B cells must be important in their immunopathogenesis. Use of a monoclonal antibody to the CD34 antigen on the surface of B cells has demonstrated that changing the B cell repertoire can have profound influences on Graves' disease [43]. This has been exemplified by the effective use of rituximab (monoclonal anti-CD20) in patients with active Graves' orbitopathy [43,44]. These results indicate that B cells not only have a role in hyperthyroidism of Graves' disease but also Graves' orbitopathy and suggest that TRAb interactions with retroorbital TSHRs expressed on fibroblasts and adipocytes may be important in disease etiology. B cells are excellent presenters of antigen to T cells, and it is likely that a primary function of thyroid autoantibodies on the surface of B cells is to home the B cell to the thyroid gland, where it obtains thyroid antigens and presents them to the T cells.

T CELLS IN GRAVES' DISEASE — T cells are present in the immune repertoire of patients with Graves' disease that react with appropriately processed peptides derived from all thyroid autoantigens. These activated T cells in turn release a variety of cytokines and increase the secretion of thyroid-specific autoantibodies from B cells.

The current concept is that thyroid-specific T cells in Graves' disease primarily act as helper (CD4+ Th1) cells. However, distinct subsets of T cells have been identified that are distinguished most easily by the cytokines that they produce (table 1):

●CD4+ Th1 cells – When activated, these cells secrete interleukin-2 (IL-2), interferon gamma (IFN-gamma), and tumor necrosis factor (TNF)-alpha, which in turn activate cytotoxic (CD8+) cells and may induce thyroid cell apoptosis.

●CD4+ Th2 cells – These cells secrete IL-4 and IL-5 (but not IFN-gamma) and activate antibody production, amongst other actions.

●CD4+ Th17 cells – This proinflammatory subset of cells secretes IL-17 under the influence of IL-23.

●CD4+ Treg cells – These are the anti-inflammatory regulatory T cells (Tregs) (CD4+CD25+), which diminish the activity of Th1 and Th2 cells [45]. One characteristic of Treg cells is the expression of the transcription factor Foxp3.

●CD8+ cytotoxic cells – These act primarily as destructive T cells under the control of Th2 and Treg cells.

While all types of T cells are found in the thyroid glands of patients with Graves' hyperthyroidism, it is helpful to consider Graves' disease as a mixture of Th1 and Th2 autoimmune responses. As far as Graves' disease is concerned, the primary pathophysiology is related to thyrotropin receptor antibodies (TRAb) of the immunoglobulin G1 (IgG1) subclass, which are driven by IFN-gamma [46,47]. Moreover, the Th1 and Th2 types of T helper cells interact with each other so that a predominance of, for example, Th1 cells does not necessarily mean that the predominant result is apoptosis. T cells may induce target cell death directly, and there is some evidence for a minor degree of thyroid-cell apoptosis in Graves' disease [2,3]. In the past, we have identified a clone of T cells that specifically lysed autologous thyroid cells from a patient with goitrous autoimmune thyroiditis (Hashimoto's disease), but we were unable to identify similar cells in patients with Graves' hyperthyroidism [48]. This suggests that thyroid antibodies may be involved in any thyroid cell death that occurs in Graves' thyroid, and we have suggested that the neutral region TRAb may be responsible [32]. In contrast, there are extensive data on the importance of apoptosis in the thyroid destruction of Hashimoto's disease. (See "Pathogenesis of Hashimoto's thyroiditis (chronic autoimmune thyroiditis)".)

Mechanisms of T cell activation — The T cell receptor sees antigen in the context of human leukocyte antigen (HLA). This means that the T cell receptor complexes with an HLA molecule on the surface of an antigen-presenting cell (figure 3); CD8+ cells with HLA class I molecules and CD4+ cells with HLA class II molecules. This complex forms only when the appropriate antigenic peptide (for example from the TSH receptor [TSHR]) is present in the binding pocket of the HLA molecule. Data show that binding pocket residue Arg74 is important for the binding of thyroid-related peptides, and helps explain the HLA association with Graves' disease [49]. Once this complex is formed, the T cell requires an additional stimulus to proliferate and secrete cytokines. This additional stimulus is called "costimulation" and is provided by costimulatory molecules on the same T cell and antigen-presenting cells (table 2) [50]. If no costimulation occurs, the T cell may become anergic or even apoptotic. Thyroid cells express major histocompatibility complex (MHC) molecules in autoimmune thyroid disease and may express costimulatory molecules (such as CD40), aiding in intrathyroidal T cell activation.

Changing the T cell population — Perturbing the T cell repertoire, in particular, disrupting the Treg cells, can result in Graves' hyperthyroidism in susceptible patients by facilitating production of TRAb. In a group of 27 patients with multiple sclerosis (a Th1-predominant disease) treated with a monoclonal antibody to T cells, peripheral blood CD4+ and CD8+ T cell counts fell to less than 20 percent of normal for at least 18 months and multiple sclerosis disease activity decreased in all patients, but nine developed Graves' hyperthyroidism 6 to 31 months after treatment [51]. Graves' hyperthyroidism has also occasionally been a complication of interferon alfa treatment in patients with hepatitis C, again thought to be on the basis of changes in T cell repertoire [52]. The same explanation is given for the improvement in Graves' disease by the onset of pregnancy where the action of Treg cells is enhanced [53]. Similarly, the improvement in Graves' orbitopathy following corticosterone therapy has been suggested as secondary to changes in the Th17/Treg ratio [54].

Intrathyroidal T cell receptor V gene repertoire — As discussed above, T cells are activated by the binding of complexes of HLA (major histocompatibility molecules) and antigenic peptides processed from proteins by antigen-processing cells. These complexes bind to antigen receptors on the surface of T cells. These receptors consist of two noncovalently linked chains (alpha and beta), each with variable (V), diversity (D), and junctional (J) regions and common constant regions. The V, D, and J genes code for the sites on the receptors that recognize the HLA-antigen complex, affording antigenic specificity. In addition to the many V (>100) and J (>50) genes present in the genome, random nucleotide additions and deletions to the D region add immense complexity to the T cell antigen receptor repertoire, causing this region to be the major site of antigen recognition [55]. Evidence for an etiologic role for T cells in Graves' hyperthyroidism is the finding that the antigen receptors of T cells isolated from thyroid tissue are the products of a limited number of V gene families [56-58]. This observation suggests that the thyroid tissue of these patients attracts and activates T cells with particular types of antigen receptors, rather than nonspecifically.

In support of this concept is evidence for clonally expanded T cell populations within the thyroid gland in Graves' disease. These data have been obtained by direct sequencing of the genes for T cell antigen receptors from intrathyroidal T cells [58-60]. These results indicate limited T cell heterogeneity in Graves' disease and point to the primacy of T cells in disease etiology. The findings are similar to those in synovial tissue from patients with rheumatoid arthritis and central nervous system plaques from patients with multiple sclerosis [61]. However, as the pathologic process progresses, there is likely to be a less restricted response [62].

The role of suppressor effects of T cells — We now understand that the immune system exerts some of its overall control via Treg (CD4+CD25+Foxp3+) cells which exert "suppression" by cytokine secretion and cell-cell contact. This function may be diminished in Graves' disease, although not all studies have found this [63,64]; caution is needed since T regulatory function may be decreased by high thyroid hormone levels [65]. The ratio of Th17/Treg is often used as an assessment of Treg function since absolute numbers vary quite widely between individuals. Even if there were only subtle defects in regulatory T cell function in patients with Graves' hyperthyroidism, the phenomena of deletion and anergy will also contribute to antigen-specific tolerance [66,67]:

●Deletion of immune cells via apoptosis occurs when immature T and B cells bind antigens in the absence of costimulatory molecules.

●Anergy can also occur when mature immune cells bind antigen in the absence of co-stimulatory molecules, leading to desensitization rather than deletion.

IMMUNE MECHANISMS OF DISEASE — A variety of immune mechanisms may be involved in the pathogenesis of Graves' hyperthyroidism. The major mechanisms for which there is some evidence are molecular mimicry (specificity crossover), thyroid-cell expression of HLA (human leukocyte-associated) molecules (antigens), and bystander activation.

Molecular mimicry — Molecular mimicry implies structural similarity between some infectious or other exogenous agent and human proteins, such that antibodies and T cells activated in response to the exogenous agent react with the human protein, in this instance, one or more thyroid proteins. As an example, in an analysis of 600 monoclonal antibodies raised against a large variety of viruses, 4 percent of the monoclonal antibodies cross-reacted with uninfected tissues [68]. With respect to Graves' hyperthyroidism, there is no strong evidence that molecular mimicry plays a role. The suggestive evidence is:

●The serum of some patients contains antibodies that react with antigens derived from Yersinia enterocolitica [69]. Furthermore, serum from some patients recovering from Yersinia infections blocks the binding of TSH to its receptors. In addition, in a report of twins discordant for Graves' disease, the twin with Graves' disease had an increased odds ratio of prior Yersinia infection [70]. However, patients who have or have had infections with these organisms do not have thyroid dysfunction.

●Structural similarities between retroviral sequences and the TSH receptor (TSHR) have been detected [71].

●Bacterial heat shock proteins can elicit antibody and T cell responses, which may cross-react with host heat shock proteins. Heat shock protein 72 can be detected in thyroid tissue from patients with Graves' hyperthyroidism but not in thyroid tissue from normal subjects [72].

Of note, this use of the term molecular mimicry should not be confused with the relationship between hyperthyroidism and orbitopathy in Graves' disease, where it is likely that the two tissues contain the same antigen, the TSHR, so that a crossover immune reaction against the thyroid antigen affects the retroorbital tissues. (See "Clinical features and diagnosis of thyroid eye disease".)

Thyroid cell expression of HLA molecules — Thyroid epithelial cells from patients with autoimmune thyroid disease (including Graves' disease), but not normal subjects, express major histocompatibility complex (MHC) human leukocyte antigen (HLA) class II molecules, notably HLA-DR molecules (picture 1) in addition to enhanced expression of HLA class I. This expression could be the direct result of viral or other infections of thyroid epithelial cells, or it may be induced by cytokines such as interferon gamma (IFN-gamma) produced by T cells that have been attracted to the gland either by an infection or directly because of the presence of thyroid antigens [73]. The potentially important role of interferon alpha, which may be induced by viral infection, further supports the concept of viral involvement [74].

Class II molecule expression provides a mechanism for presentation of thyroid antigens to autoreactive T cells resulting in their activation and with the potential for persistence of thyroid disease. Several experimental observations provide support for this hypothesis:

●Induction of class II molecules on thyroid epithelial cells by interferon gamma can induce autoimmune thyroiditis in susceptible mice [75].

●Viruses can directly induce class II molecule expression on thyroid cells, independent of cytokine secretion [76,77].

●Thyroid epithelial cells expressing class II molecules can present viral peptide antigens to cloned T cells [78]. Thyroid antigen-specific T cell clones in normal rats react specifically with cloned autologous thyroid cells in the absence of more conventional antigen-presenting cells [79].

●An animal model of Graves' disease induced by cells expressing the TSHR is only effective when the cells also express MHC class II antigens [17,18].

These findings strongly support the view that an insult, such as infection, may induce class II molecule expression on human thyroid cells and that these cells then may act as antigen-presenting cells to initiate an autoimmune response. The expression of a T cell costimulator molecule, CD40, on thyroid epithelial cells indicates that costimulatory molecules are available for this action. In addition, intrathyroidal dendritic cells and B cells may also serve as potent antigen-presenting cells [50]. The description of hyperthyroidism in mice immunized with fibroblasts coexpressing class II molecules and human TSHRs provides further evidence that cells need not be "professional" antigen-presenting cells to present antigen so long as they can acquire the ability to express class II molecules [17].

However, the context in which HLA class II is induced by cytokines is of prime importance. As an example, transgenic mice with thyroid cells which secrete IFN-gamma showed limited thyroiditis even in susceptible strains, and this may be due to the too high local concentration of IFN-gamma [62].

Bystander activation — In order for this model of HLA class II antigen expression and presentation of antigens to be realized, there must be a local insult to initiate the responses. As mentioned above, this may take the form of a direct insult to the thyroid by a viral infection of the thyroid cells or of immune cells. Even the arrival of activated T cells within the thyroid gland may perhaps initiate such a series of events in a susceptible subject with the appropriate immune repertoire. Evidence has mounted that such bystander activation of local T cells, which may not be thyroid specific, may exert via cytokines a marked activation effect on resident thyroid-specific T cells. Evidence for such bystander effects has been obtained in an animal model of viral-induced autoimmune insulitis [80] and in experimental autoimmune thyroiditis [81]. The attractiveness of this sequence of events is that many different types of infections would lead to the same clinical disease phenotype.

PRECIPITATING AND PREDISPOSING FACTORS — Several factors that predispose to Graves' hyperthyroidism have been proposed (table 3).

Genetic susceptibility — There is abundant epidemiologic evidence for genetic susceptibility to Graves' hyperthyroidism and chronic autoimmune thyroiditis (table 4) [82-84].

●The diseases cluster in families and are more common in females.

●The concordance rate in monozygotic twins is 20 to 40 percent.

●The sibling recurrence rate for Graves' disease exceeds 10.0 [85].

●There are associations with a number of immune-related genes which have also been found with many other autoimmune diseases and presumably underpin the inherited susceptibility to autoimmunity; for example, with certain alleles of cytotoxic T lymphocyte-associated (CTLA)-4 [84,86]. As an example, in one study of 379 patients with Graves' hyperthyroidism in the United Kingdom, 42 percent had a particular allele (G allele) of the cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) gene, as compared with 32 percent of 363 normal subjects [84,86].

●In keeping with an immune-related susceptibility seen in almost all autoimmune diseases, there is a well-known association with certain alleles of human leukocyte antigen (HLA) on chromosome 6 [82]. As an example, a study of White patients in North America found that HLA-DRB1*08 and DRB3*0202 were associated with the disease and that DRB1*07 was protective [87]. Detailed studies have shown convincingly that the presence of Arg74 is the important peptide in the HLA DR binding pocket rather than just the HLA subtype [49,88]. As far as thyroid-specific gene associations are concerned, there is now evidence of increased risks associated with polymorphisms of intron 1 in the TSH receptor (TSHR) gene [89-92] and the thyroglobulin (Tg) gene [93]. The data suggest that the influence of HLA and Tg polymorphisms is more than additive.

●However, the associated risks with these gene associations are all relatively low so that their assessment is not currently clinically useful.

Infection — Autoimmune thyroiditis can be induced in experimental animals by certain viral infections. If infection were the direct cause of Graves' hyperthyroidism, an identifiable agent should be present in the majority of patients, and it should be possible to induce the disease by transferring the agent. Possible infections of the thyroid gland itself (eg, subacute thyroiditis, congenital rubella) have been associated with thyroid autoimmune disease and could initiate class II molecule expression. Hepatitis C infection is a well-recognized precipitator of autoimmune thyroid disease when treated with interferon therapy, although, most commonly, a thyroiditis develops rather than Graves' disease [94]. A report of retroviral sequences in the thyroid glands of patients with Graves' disease was not confirmed [95,96]. With the coronavirus 2019 (COVID-19) pandemic, there have been a flurry of reports trying to associate infection with autoimmune thyroid disease [97], and for Graves' disease, only a few case reports have emerged. While there may be a weak association with subacute thyroiditis, there is no good evidence of a connection with Graves' disease other than chance. There is also no evidence that any other infections or exposures lead directly to autoimmune thyroid disease in the majority of patients [98], although examination of thyroid tissue from patients with Graves' disease continues to yield suspicious evidence [99].

Stress — As compared with normal subjects or patients with toxic nodular goiter, patients with Graves' hyperthyroidism more often give a history of some type of psychologic stress, in particular negative life events such as loss of a spouse or a road traffic accident, before the onset of their hyperthyroidism [100-102]. In general, stress appears to induce a state of immune suppression, possibly mediated by the actions of cortisol on immune cells. Stress-induced suppression may be followed by rebound immunologic hyperactivity. Such a response could precipitate autoimmune thyroid disease in genetically susceptible subjects.

Biological sex — More females develop Graves' hyperthyroidism than males, with a ratio of approximately 4:1, an effect that is often said to be mediated in some way by more estrogen or less testosterone. There is a large body of evidence that moderate amounts of estrogen enhance immunologic reactivity [103-105]. However, it is just as likely that the extra X chromosome is the source of the enhanced susceptibility rather than sex steroids since the susceptibility continues after the menopause. For example, X chromosome inactivation has been associated with autoimmune thyroid disease [106].

Microbiota — The gut microbiome is required for normal immune system maturation, especially induction of tolerance. Alterations in bacterial function and diversity likely contributes to autoimmune diseases. Studies in mice have demonstrated that early life microbial exposures determine sex hormone levels and modify progression to autoimmunity suggesting that the gut microbiota may contribute to autoimmune thyroid disease susceptibility [107]. Data on the gut microbiome are being accumulated in patients with Graves' disease with and without orbitopathy but it is still too early for definitive conclusions [108,109].

Smoking — Smoking is a risk factor for Graves' hyperthyroidism (relative risk approximately 2.0) and an even stronger risk factor for Graves' orbitopathy [110-112]. The mechanism remains uncertain, other than the obvious effect of irritation [113].

Pregnancy — Severe Graves' disease is uncommon during pregnancy because hyperthyroidism is associated with reduced fertility and increased pregnancy loss. When severe hyperthyroidism occurs, however, it can endanger both mother and fetus. Luckily, pregnancy is a time of immune tolerance so that the disease tends to improve as pregnancy progresses. During pregnancy, both T cell and B cell functions are diminished, while regulatory T cells increase, dampening the disease [53,114]. The slow rebound from this tolerant state after delivery results in enhanced immune reactivity, and this contributes to the development of the postpartum thyroid diseases, including the new onset or recurrence of Graves' disease [115].

It has also been suggested that fetal microchimerism (the presence of fetal cells in maternal tissue) might play a role in the development of pregnancy tolerance and postpartum autoimmune thyroid disease [116]. Up to 30 percent of young females give a history of pregnancy in the 12 months before the onset of Graves' disease [60], indicating that postpartum Graves' disease is a surprisingly common presentation and that pregnancy is a major risk factor in susceptible females.

Drugs — Iodine and iodine-containing drugs such as amiodarone and computed tomography (CT) scan contrast media may precipitate Graves' disease, or a recurrence of Graves' disease, in a susceptible individual [117]. Iodine is most likely to precipitate thyrotoxicosis in an iodine-deficient population simply by allowing the thyrotropin receptor antibodies (TRAb) to be effective in stimulating more thyroid hormone to be formed. Whether there is any other precipitating event is unclear. Iodine and amiodarone may also damage thyroid cells directly and release thyroid antigens to the immune system [118].

Interferon alfa treatment of patients with hepatitis C infection has been widely associated with the development of autoimmune thyroiditis but Graves' disease may also be precipitated presumably by influencing the immune repertoire [119]. Alemtuzumab, a monoclonal antibody against the T cell antigen CD52 used for treatment of multiple sclerosis, has been associated with a 10 to 15 percent incidence of new-onset Graves' disease [120]. The immune checkpoint inhibitors, such as CTLA-4 and programmed cell death 1 (PD-1)/PD-1 ligand 1 (PD-L1) negatively regulate the immune system, and autoimmune thyroid disease is one of their most common side effects [121]. However, these cases are mostly autoimmune thyroiditis, and Graves' disease is much less common. (See "Drug interactions with thyroid hormones", section on 'Checkpoint inhibitor immunotherapy'.)

SUMMARY

●Hyperthyroidism is the most common feature of Graves' disease, affecting nearly all patients and is caused by autoantibodies to the thyrotropin receptor (TRAb) that activate the receptor, thereby stimulating thyroid hormone synthesis and secretion and thyroid growth (causing a diffuse goiter). The presence of TRAb in the serum and orbitopathy on clinical examination distinguishes the disorder from other causes of hyperthyroidism. (See 'Introduction' above.)

●TRAb stimulate the thyroid gland and are specific for Graves' disease, in contrast to thyroglobulin (Tg) and thyroid peroxidase (TPO) antibodies. They bind mainly to the leucine-rich repeat region of the thyroid-stimulating hormone (TSH) receptor (TSHR) ectodomain to which TSH binds (figure 1). (See 'Autoantibodies to the TSH receptor' above.)

●T cells are present in patients with Graves' disease that react with appropriately processed peptides derived from all thyroid autoantigens but in particular the TSHR. These activated T cells in turn increase the secretion of thyroid-specific autoantibodies from B cells. Thyroid-specific T cells in Graves' disease primarily act as helper (CD4) rather than suppressor or cytotoxic (CD8) cells. (See 'T cells in Graves' disease' above.)

●A variety of immune mechanisms may be involved in the pathogenesis of Graves' hyperthyroidism. The major mechanisms for which there is reasonable evidence are thyroid cell expression of human leukocyte antigen (HLA)-associated molecules associated with bystander activation. (See 'Immune mechanisms of disease' above.)

●Possible precipitating and predisposing factors include genetic susceptibility, infection, stress, smoking, pregnancy, and iodine. (See 'Precipitating and predisposing factors' above.)