Graves’ Disease
Graves’ disease is now universally classified among the autoimmune organ-specific diseases because it fulfills all the criteria required for this definition (Table 9-1). The main pathogenic mechanism is stimulation of growth and function of the thyroid gland by circulating antibodies directed against the thyroid-stimulating hormone (TSH) receptor (TSHR), thereby mimicking the effects of TSH. Thus TSHR is the major autoantigen of Graves’ disease.
Table 9-1
Criteria for Organ-Specific Autoimmune Diseases and Their Presence in Graves’ Disease
| Criteria | Present in Graves’ Disease |
| Lymphocytic infiltration of the target organ | Yes |
| Identification of the specific antigen(s) | Yes |
| Production of humoral and/or cellular autoimmune responses (or both) in animals sensitized by autologous antigen | Yes |
| Presence of organ-specific lesions in autosensitized animals | Yes |
| Association with other autoimmune diseases | Yes |
Historical Notes
Graves’ disease is the eponym by which a syndrome characterized by diffuse goiter and hyperthyroidism is recognized in English-speaking countries. Robert James Graves (Fig. 9-1) was a brilliant and productive Irish physician who contributed in many ways to the development of medical science of his time.1 Credit for his prominent position is probably due to his description in 1835 of “… three cases of violent and long palpitations in females, in each of which the same peculiarity presented … enlargement of the thyroid gland”, which was the first report of toxic diffuse goiter.2 However, Caleb Hillier Parry, a less renowned physician of Bath, England, had described a similar syndrome earlier, in 1825: “There is one malady which I have in five cases seen coincident with what appears to be enlargement of the heart, and which, so far as I know has not been noticed in that connection by medical writers. This malady to which I allude is enlargement of the thyroid gland”.3 He also described protrusion of the eyes as a feature of the syndrome. Even earlier than that, in 1805, the Italian Giuseppe Flajani in Rome had reported two cases of diffuse swelling of the neck accompanied by palpitations.4 He failed to recognize the thyroidal origin of the swelling and named it “bronchocele.” In 1840, in Germany, Carl A. von Basedow described “Exophthalmos durch Hypertrophie del Zellgewebes in der Augenhohle,” or exophthalmos caused by hypertrophy of the cellular tissue of the orbit.5 This was in fact the first description of the complete syndrome that included the triad exophthalmos, goiter, and palpitations. Von Basedow was struck by the prominence of the eye changes and made exophthalmos the hallmark of the disease. His descriptions were widely disseminated at the time, so that in most non-English-speaking European countries, the disease is still called Basedow’s disease. In 1880, Ludwig Rehn performed the first thyroidectomy for toxic diffuse goiter, and in 1909, Kocher was awarded the Nobel Prize for his innovations in thyroid surgery.6 In 1886, Moebius proposed that exophthalmic goiter was due to an excessive function of the thyroid gland.6 In 1911, Marine proposed treatment of Graves’ disease with iodine in the form of Lugol’s solution.7 In the early 1940s, the antithyroid drugs thioureas were described,8 and Astwood introduced them into clinical use for the control of thyrotoxicosis.9 At the same time, physicists and physicians in Boston and in Berkeley started to treat thyrotoxic patients with radioiodine (131I).10 In the span of just a few years, the two mainstays of modern treatment of Graves’ disease were initiated. The following decade was marked by the discovery in 1956 of the long-acting thyroid stimulator (LATS) by Adams and Purves11 and by the subsequent identification of this stimulator as an antibody, thereby forming the basis for our current understanding of the pathogenic mechanisms of Graves’ disease. Cloning of TSHR12,13 is only the most recent and will certainly not be the last memorable event in the uncovering of a disease that has paralleled the development of modern medicine across 2 centuries.
Epidemiology
Graves’ disease is a relatively prevalent disorder, and it is the most frequent cause of thyrotoxicosis in iodine-sufficient countries.14
Several studies have attempted estimating the exact frequency of Graves’ disease in the general population. However, comparison of surveys is difficult because of the use of different criteria in population sampling, because of ethnic differences, and because diagnostic tools have changed over the years. In the United States, a large survey performed in the 1970s estimated the prevalence of Graves’ disease to be 0.4%.15 A similar prevalence (0.6%) was found in the Pescopagano study in Italy.16 The Whickham survey in the United Kingdom suggested a prevalence of 1.1% to 1.6% (i.e., about threefold to fourfold higher) for thyrotoxicosis of all causes, of which Graves’ disease was presumably the most frequent.17,18 A recent study performed in Sweden has shown an incidence of Graves’ disease of approximately 25 cases/100,000 per year, which is relatively high compared with other studies, a possible interpretation of which is a high iodine intake of the selected population (see later discussion).19 Overall, a meta-analysis of various studies has estimated the general prevalence of the disorder to be about 1%,20 which makes it one of the most frequent clinically relevant autoimmune disorders.
The dietary iodine supply appears to be a major factor in determining the frequency of Graves’ disease.21–28 For example, iodine supplementation of previously iodine-deficient Tasmania induced a threefold increase in the incidence of hyperthyroidism in 3 years.21 Although the increase was mainly due to iodine-induced thyrotoxicosis in patients with autonomously functioning nodules or goiters, it was also shown that LATSs or LATS protectors were present in a number of cases of thyrotoxicosis in the early supplementation period, thus suggesting that some of the iodine-induced cases of hyperthyroidism might have been due to Graves’ disease. Since the Tasmanian report, outbursts of iodine-induced thyrotoxicosis have been reported in many countries after the implementation of iodine supplementation programs.22–28 Thyrotoxicosis occurred primarily in older people with preexisting nodular goiter. However, in a study performed in Switzerland, a slight and transient increase in the incidence of Graves’ disease was noted after stepwise, full iodine supplementation (Fig. 9-2).23 Similar increases in the incidence of thyrotoxicosis have been reported in Sweden (16.6/100,000),24 New Zealand (15/100,000),25 Britain (23/100,000),26 and Denmark27, in the latter population especially in younger age groups.28 Population-based studies also show differences in the incidence of Graves’ disease in populations with different but relatively constant iodine intake. Nevertheless, this corresponds to a higher frequency of nonautoimmune thyrotoxicosis in iodine-deficient areas. Thus, by comparing two genetically similar populations that differed in terms of iodine intake (iodine-sufficient Iceland versus iodine-deficient East Jutland in Denmark), it was found that the incidence of Graves’ hyperthyroidism was slightly higher in the iodine-sufficient (20/100,000 inhabitants/year) than in the iodine-deficient population (15/100,000/year), but that the incidence of thyrotoxicosis of all causes was greater in the iodine-deficient (39/100,000/year) than in the iodine-sufficient (23/100,000/year) population.14 Clearly, these findings show that fear of an increased incidence of Graves’ disease should not prevent the implementation of iodine supplementation programs.
FIGURE 9-2 Increased incidence of hyperthyroidism from Graves’ disease and toxic nodular goiter in Switzerland after the introduction of iodine supplementation. (Modified from Burgi H, Kohler M, Morselli B: Thyrotoxicosis incidence in Switzerland and benefit of improved iodine supply. Lancet 352:1034, 1998.)
Although ethnic differences in susceptibility to Graves’ disease are likely to exist, they have not been consistently investigated in comparative studies. As with many other autoimmune disorders, Graves’ disease is about fivefold more prevalent in women than men. The reasons for this observation are understood only in part, but some hypotheses will be discussed later. The annual incidence is clearly and consistently related to age, with peaks in the fourth to sixth decades of life,28 although Graves’ disease can be observed in people of any age, including children.
Etiology
Most of the pathogenic mechanisms of Graves’ disease have been clarified since the first description of LATS.11 It is now well established that Graves’ disease is an organ-specific autoimmune disorder, with involvement of both T-cell and B-cell–mediated immunity against thyroid antigens. TSHR is the main antigen involved, and circulating autoantibodies against TSHR (TSH receptor antibodies [TRAbs]) that are capable of stimulating the receptor are responsible for the most distinctive features of the disease, namely hyperthyroidism and goiter. Nevertheless, in spite of advancements in understanding the pathogenic mechanisms of Graves’ disease, the ultimate cause of the disease remains elusive. The majority of investigators share the opinion that Graves’ disease is a multifactorial disease caused by a complex interplay of genetic, hormonal, and environmental influences that lead to the loss of immune tolerance to thyroid antigens and to the initiation of a sustained autoimmune reaction.
Genetics of Graves’ Disease
The strongest data in support of a genetic predisposition to Graves’ disease comes from twin studies.29,30 Dizygotic twins share on average 50% of their genome, whereas monozygotic twins share 100%. Moreover, twins are likely to share environmental factors more than any other kind of siblings. Several large twin studies have reported greater concordance rate of Graves’ disease in monozygotic than in dizygotic twins.29 Data obtained with modern diagnostic tools have shown a relatively low (~17% to 35%), but still significant, concordance in monozygotic twins.29,30 These findings clearly show a genetic influence, possibly characterized by low penetrance of the genes involved.
Another tool widely used to establish the existence of a genetic predisposition to any condition is family studies, in which the prevalence of the disease in relatives of index cases is compared with the prevalence in the general population. Early family studies showed a high prevalence of Graves’ disease and other thyroid abnormalities in first-degree relatives of patients with Graves’ disease and Hashimoto’s thyroiditis.31,32 The prevalence of circulating thyroid autoantibodies in siblings of patients was as high as 56% in some studies,31 which suggested a dominant mode of inheritance. These observations have been consistently confirmed in highly selected populations,33 but the results may not be applicable to the general population because of ascertainment bias. In an extensive segregation analysis with randomly ascertained probands, circulating antibodies were found in only 25% of the offspring of positive parents and in 14% of the offspring of negative parents, thus suggesting a multigenic model with less than 100% penetrance for the antibody trait.34 With the exception of very early studies, the prevalence of overt Graves’ disease has been found to be relatively low in siblings of patients.33,36 However, initial abnormalities in thyroid function compatible with subclinical hyperthyroidism or hypothyroidism have been reported.33,35 Villanueva et al. found that 36% of Graves’ patients with ophthalmopathy have a family history of either Graves’ disease or autoimmune thyroiditis, which in 23% of the cases affected first-degree relatives.37 Autoimmune thyroiditis is frequently observed in siblings of probands with Graves’ disease, as well as the contrary,33,37 suggesting that the two diseases may share some susceptibility genes that predispose to thyroid autoimmunity, but that the full expression of the phenotype depends on other genes and/or on environmental factors. Other organ-specific, non-thyroid-related autoimmune diseases may also be more prevalent in relatives of patients with Graves’ disease.33
These data and those obtained in twin studies are indicative of a complex multigenic pattern of inheritance of Graves’ disease. Some of the components of the phenotype, such as the presence of circulating antithyroglobulin and antithyroperoxidase antibodies, may be inherited in a dominant fashion with high penetrance.33 However, these genetic determinants do not appear to be sufficient for full expression of the disease. Clearly, other genes must be involved, which is in line with the complexity of the inheritance observed. Also, it appears from epidemiologic and experimental data that environmental factors (reviewed later in this chapter) play an important role by modulating the effect of an inherited predisposition. Based on the above evidence, a number of genes or loci have been investigated as candidates for predisposing factors (Table 9-2).33,38
Table 9-2
Genetic Determinants Associated With Graves’ Disease
| Gene | Possible Mechanism | Evidence in Favor of an Association |
| HLA-DR | Altered antigen presentation | Fair |
| CD40 | Altered antigen presentation | Fair |
| CTLA-4 | Altered antigen presentation | Good |
| PTPN22 | Altered T-cell activation | Fair |
| Thyroglobulin | Loss of tolerance | Good |
| TSH-R | Loss of tolerance | Poor |
Genes Predisposing to Graves’ Disease
In the last 20 years, the impressive advancement of biomedical research has allowed a remarkable expansion in genetics, which has led to the identification of several genes involved in the predisposition to Graves’ disease.33,38 Linkage, association, and candidate genes analyses, as well as whole-genome screening, have all been used to accomplish this goal. Based on the data available, genetic susceptibility to thyroid autoimmunity in general and to Graves’ disease in particular seems to be determined by a combination of a number of genes, some which most likely still remain to be identified.33,38,39
1:: The HLA Complex The HLA complex, which is located on the short arm of human chromosome 6, contains the sequence encoding about a hundred genes, most of which are involved in regulation of the immune response.40–42 The HLA genes are classically grouped into three major classes. Class I includes histocompatibility genes expressed on the surface of most cells (HLA-A, HLA-B, and HLA-C). Class II includes histocompatibility genes expressed exclusively on the surface of leukocytes and immune cells (HLA-DR). Class III includes a heterogeneous group of genes encoding molecules involved in the immune response, such as some complement factors, cytokines, and lymphocyte surface molecules. Other genes in this class are not clearly related with immunity. Most genes of the HLA complex are highly polymorphic, which makes them excellent candidates for disease susceptibility.
Experimental thyroiditis in the mouse was the first autoimmune disease to be associated with HLA.43,44 Early population studies in humans indicated an association of Graves’ disease with HLA-B8 and a relative risk of 3.9 in white patients.45 Subsequent studies also suggested an influence of that haplotype on the clinical course of the disease.33,44 However, HLA-DR3 (HLA-DRB1*03) was later shown to increase the risk to a greater extent and was considered to be the true determinant of the disease because it was in linkage disequilibrium with the B8 allele.33 Among Caucasians, HLA-DQA1*0501 was found to confer a relatively high risk within DR3 itself.33,44 This allele is in linkage disequilibrium with both B8 and DR3 and gives a relative risk of 3 to 4 for Graves’ disease in the white population. Sequencing of the DRβ-1 chain of HLA-DR3 allowed the identification of Arg74 as the critical amino acid conferring susceptibility to Graves’ disease.33,44
Different haplotypes seem to be involved in ethnic groups other than Caucasians: DQ3 in patients of African descent and Bw46 in those of Asian descent, although the data available are limited and have not always been reproducible.33,44
In general, HLA associations have been shown to confer a relatively low risk, even with alleles that have a high prevalence in the general population. Thus linkage analysis, a powerful tool for mapping essential predisposition genes, has been negative when the HLA region was examined using different polymorphisms at the same locus.33,44 Overall, it seems that the HLA locus explains a small fraction of the total genetic predisposition, but it is neither the major nor the only determinant, although it represents an established risk-increasing factor.
2:: CD40 CD40, a member of the tumor necrosis factor receptor family, is expressed in B cells and other antigen-presenting cells and is involved in B-cell activation and proliferation, antibody secretion, immunoglobulin class switching, affinity maturation and generation of memory cells.46 Linkage studies have shown an association of the CD40 gene with Graves’ disease and the subsequent sequencing of the gene led to the identification of a C/T polymorphism at the 5′ untranslated region of CD40 strongly associated with Graves’ disease.33 This polymorphism influences the translational efficiency of the gene, which may have functional consequences in the CD40 protein.
3:: CTLA-4 CTLA-4 is a T-lymphocyte surface protein with a major role in down-regulation of the immune response.47 Several studies have provided evidence that CTLA-4 is linked to Graves’ disease, autoimmune thyroiditis, and to the production of autoantibodies against thyroid antigens.33,44–48 Several variants of the CTLA-4 gene have been implicated in its possible causative role in Graves’ disease, among which a CTLA-4 polymorphism at position 60 was found to be the most suitable candidate in a large comprehensive analysis, the functional relevance of the polymorphism being possibly due to a reduced mRNA expression encoding the soluble form of the molecule.47 Nevertheless, a subsequent study did not confirm these data.48 Although CTLA-4 seems to be a genetic determinant of Graves’ disease, the causative variant remains to be identified with certainty, and it is possible that a haplotype consisting of more than one variant is responsible for the association.33
4:: The Protein Tyrosine Phosphatase 22 (PTPN22) Gene PTPN22 is a powerful inhibitor of T-cell activation.49 A single nucleotide polymorphism at codon 620 associated with other autoimmune diseases was found to be associated with both Graves’ disease and autoimmune thyroiditis, with significant ethnic differences in the association.49,50
5:: Thyroglobulin Thyroglobulin is the precursor of thyroid hormones and a major autoantigen in thyroid autoimmunity. Recent wide genome screens have provided evidence for a strong linkage between a locus on chromosome 8q24, where the thyroglobulin gene is located, and autoimmune thyroid diseases.33,51 Sequence analysis of the thyroglobulin gene has shown numerous single nucleotide missense polymorphism, suggesting that amino acid variants in the Tg proteins may contribute the pathogenesis of autoimmune thyroid diseases, including Graves’ disease.33,52
6:: TSH-R Despite the central role of TSH-R in the pathogenesis of Graves’ disease, the association of the disease with the gene encoding the receptor remains controversial.33 Although three common missense single nucleotide polymorphisms have been found to be associated with Graves’ disease, these associations were not always confirmed by other studies. The extent of the contribution of the TSH-R gene remains to be established.33
7:: Other Genes In the search for genetic determinants of Graves’ disease, several other genes have been studied over the years, namely genes involved in the immune response. The immunoglobulin genes were studied extensively, but conflicting results were observed in association studies.33 Other candidate immunoregulatory genes that have been studied include interleukin 1 (IL-1), IL-1 receptor antagonist, tumor necrosis factor receptor 2 (TNF-2), and interferon γ (INF-γ). None of these genes showed significant associations with Graves’ disease.33 Additional loci have been linked in families with Graves’ disease, including one on chromosome 14q31 (GD-1), one on chromosome 20 (GD-2), and one on chromosome Xq21-22 (GD-3).33
Environmental Factors and Graves’ Disease
The relative low penetrance in twins and first-degree relatives of patients with Graves’ disease suggests that environmental factors must play a major role in inducing the disease in genetically susceptible individuals.33,53,54 Several studies have shown that various nongenetic factors may in fact contribute to the development of Graves’ disease.
Infections
Over the years, both experimental and epidemiologic evidence has suggested that infections could play a role in the pathogenesis of Graves’ disease.54,55 Seasonal and geographic variations in the incidence of the disease have been reported,56,57 although seasonal variations have not been confirmed in other studies.58 Blood group nonsecretors, who are more prone than secretors to infections, are more frequently found among patients with Graves’ disease than in controls.59 This observation has been interpreted as indirect evidence that infectious pathogens may be involved in the etiology of Graves’ disease, although a direct genetic effect of the secretor status could also explain these results. Evidence of a recent viral infection has been reported in a high percentage of patients with Graves’ disease.54,55
Molecular mimicry has been invoked to explain the association between infections and Graves’ disease.60 Molecular mimicry is based on the hypothesis that cross-reactions of some microbial antigens with a self-antigen may cause an immune response to autoantigens. In Graves’ disease, the pathogen Yersinia enterocolitica has been thoroughly studied after reports of association of this microbe with the disease. A high prevalence of circulating antibodies against Y. enterocolitica has been observed in patients with Graves’ disease, and Yersinia antibodies were found to interact with thyroid structures.61–63 In a recent study from Denmark, it was shown that the occurrence of IgA and IgG antibodies against Yersinia not only is more frequent in Graves’ patients than in case controls but also in twins with Graves’ disease compared with their discordant twins.64 Saturable binding sites for TSH have been found in Yersinia and were also recognized by TRAbs from patients with Graves’ disease.65,66 In animals immunized with Yersinia proteins, antibodies developed against human thyroid epithelial cells and TSHR.67,68 Overall, the affinity of these cross-reactive antibodies to the thyroid was low, and immune responses were transient. Low-affinity binding sites for TSH have also been found in other bacteria, including some species of Leishmania and Mycoplasma.53,54 However, it must be noted that thyroid autoimmunity does not develop in most patients with Yersinia infections,69 so the evidence in favor of Yersinia infections as a precipitating cause of Graves’ disease awaits confirmation.
Viruses could theoretically trigger autoimmunity through several mechanisms, including interactions with autoantigens, permanent expression of viral proteins on the surface of epithelial cells, aberrant induction of HLA antigens on epithelial cells (see later), and molecular mimicry.60,70 In 1989, the presence of retroviral (HIV-1 glycosaminoglycan protein) sequences in the thyroid and peripheral mononuclear cells of patients with Graves’ disease was reported,71 but viral sequences were not found in control thyroids. This finding, however, remained isolated and was not confirmed in subsequent studies.72,73 Human foamy virus antigens were shown by immunofluorescence to be present in the thyroid of patients with Graves’ disease.74 Again, further studies using more specific and sensitive techniques failed to identify foamy virus DNA and antiviral antibodies in the blood of affected subjects.75,76 Homology between another HIV-1 protein (Nef) and human TSHR has also been reported, although sera from patients with Graves’ disease did not react with the peptide bearing the highest degree of homology.77 Another retroviral protein, p15E, has been isolated from the thyroids of patients with Graves’ disease but not from control glands.78 In this regard, it is worthwhile emphasizing that retroviral-like proteins, including p15E, are encoded by the normal human genome. Although their function is unclear, they may be expressed in many epithelial tissues under certain conditions such as inflammation and may modulate but not initiate the immune response.79 The finding of retroviral sequences or proteins in the glands of patients with Graves’ disease may therefore represent a secondary rather than a causative phenomenon. Circulating antibodies against another retroviral particle, namely HIAP-1, have been found in as many as 87.5% of patients with Graves’ disease as compared with 10% to 15% of controls,80 but HIAP-1 particles were not detected when human T cells were co-cultured with Graves’ thyrocytes.81
A highly speculative hypothesis has been raised that involves superantigens.55,60 Superantigens are endogenous or exogenous proteins, such as microbial proteins, capable of stimulating a strong immune response through molecular interactions with nonvariant parts of the T-cell receptor and the HLA class II proteins. Through this mechanism, superantigens are in theory capable of stimulating the expansion of autoreactive T cells and therefore of driving an autoimmune response.55,60 Such a mechanism has been suggested in rheumatoid arthritis, and a similar mechanism was proposed for Graves’ disease. In vitro superantigen stimulation of glands with autoimmune thyroid disease induced expression of HLA class II molecules on thyrocytes, and this phenomenon was IFN-γ dependent.82 The interpretation of this observation was that superantigen-reactive T cells exist among the lymphocytes infiltrating the thyroid in autoimmune thyroid disease, and that these lymphocytes may have been activated after exposure to extrinsic superantigens.
The most recent hypothesis, the so-called “hygiene hypothesis of autoimmunity,” implies that infections may protect from, rather than precipitate, autoimmune diseases. Exposure of the immune system to infective agents may somehow allow better control of autoimmune responses. In this regard, improved living standards have been associated with decreased exposure to infections and an increased risk of autoimmune diseases. Kondrashova et al. reported a much reduced prevalence of thyroid autoantibodies in the lower-economic population, which may suggest the hygiene hypothesis may apply to thyroid autoimmune diseases.83 Further studies are needed to investigate whether this is in fact the case.
