Autoimmune Thyroid Disease Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Thyroid Autoimmunity Wilmar M. Wiersinga Department of Endocrinology and Metabolism, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
Abstract
Autoimmune thyroid disease (AITD) is a multifactorial or so-called ‘complex’ disease in which autoimmunity against thyroid antigens develops against a particular genetic background facilitated by exposure to environmental factors. Clinically relevant thyroid antigens are thyroid peroxidase (TPO), thyroglobulin (TG) and thyroid-stimulating hormone receptor (TSHR). Autoimmunity against these self-antigens gives rise to thyroid antibodies (Abs). TPO-Ab and TG-Ab are the hallmark of chronic lymphocytic thyroiditis (also called chronic autoimmune thyroiditis or
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Autoimmune thyroid disease (AITD) is a multifactorial disease in which autoimmunity against thyroid antigens develops against a particular genetic background facilitated by exposure to environmental factors. Immunogenicity of the major thyroid antigens thyroid peroxidase, thyroglobulin (TG) and thyrotropin receptor (TSHR) is increased by genetic polymorphisms, a high number of antigenic peptides available for binding to human leukocyte antigen (HLA), and a high degree of glycosylation. Antigens bound to HLA are presented by antigen-presenting cells to T cell receptors. Further interaction between both cells is required via binding of CD40 ligand to CD40 and of B7-1/2 to CD28 for activation of T cells. Complex regulatory mechanisms serve to prevent an immune response directed against ‘self’-antigens in the thymus (central tolerance) and in peripheral tissues (peripheral tolerance) with the help of regulatory T cells. Breakdown of tolerance to thyroid antigens can result in thyroid autoimmunity, which may happen in subjects who have the wrong genes and who are exposed to the wrong environment. Polymorphisms in thyroid genes (TG, TSHR) and immunoregulatory genes (HLA, CTLA4, PTPN22, CD40, FCRL3, IL2RA, FOXP3) would contribute for about 70% to AITD, and environmental exposures (like iodine, smoking, infections, parity) for the remaining 30%. Thyroid-infiltrating activated T cells may lead to cell-mediated immunity, thyroid injury and eventually hypothyroidism, whereas humoral immunity via TSHR-stimulating antibodies may give rise to hyperthyroidism. Pediatric Hashimoto’s and Graves’ disease are less prevalent than in adults, and the female preponderance is also less marked in children. The discrepancy is probably due to relatively less involvement of environmental insults in children, whereas the prevalence of risk alleles in AITD © 2014 S. Karger AG, Basel children is higher than in AITD adults.
Hashimoto’s thyroiditis) which may lead to Hashimoto’s hypothyroidism although most patients will remain euthyroid. TSHR-Abs are the hallmark of Graves’ disease resulting in Graves’ hyperthyroidism. Hashimoto’s hypothyroidism and Graves’ hyperthyroidism can be viewed as the opposite ends of a continuous spectrum of AITD, but in reality there is substantial overlap between Hashimoto’s and Graves’ diseases. The rare entity of hashitoxicosis refers to patients with high concentrations of TPO-Ab and TG-Ab who initially present with hyperthyroidism, but in whom hyperthyroidism is replaced rather soon by hypothyroidism; this sequence of events is frequently related to a change from stimulating to blocking TSHR-Abs [1]. Whereas almost all Hashimoto patients have TPO-Ab and/or TG-Ab, these Abs are also present in up to 70% of patients with Graves’ hyperthyroidism. Hypothyroidism due to chronic autoimmune thyroiditis occurs in the long run in up to 20% of patients with Graves’ hyperthyroidism who have entered remission after a course of antithyroid drugs; blocking TSHR-Ab contributed to the hypothyroidism in one third of the cases, and chronic autoimmune thyroiditis in the remaining two thirds [2]. In this review we will discuss the characteristics of thyroid antigens that contribute to their immunogenicity, the essential role of the immunological synapse in which antigens are presented to T cells, and how tolerance to self-antigens can be broken by genes and facilitated by environmental factors. We also discuss what is known on the early stages of AITD immunopathogenesis in humans and on the causes of differences in prevalence and gender distribution between childhood AITD and adulthood AITD.
The 3 major thyroid autoantigens involved in AITD are TPO, TG and TSHR. There is insufficient evidence that autoimmunity against other thyroid proteins plays an important role in the initiation or maintenance of thyroid autoimmunity [3]. Abs against the sodium/iodide symporter expressed on the basolateral membrane of thyrocytes have been detected in sera of AITD patients albeit in low frequency, but these Abs are rarely capable of modulating sodium/iodide symporter activity [4, 5]. Abs against pendrin (located at the apical membrane of thyrocytes) were present in 81% of AITD patients in Japan [6], but only in 8.8% in the UK when a more advanced radioligand binding assay was employed [7]. Autoimmunity against the insulin-like growth factor 1 receptor is thought to play an additional role in Graves’ ophthalmopathy (GO) [8], but Abs against it in sera of patients with GO are as common as in controls (10 vs. 11%) and thus lack specificity [9]. Table 1 lists the characteristics of the 3 major thyroid autoantigens [10]. TPO is a membrane-bound homodimer of two 107-kDa subunits. Its enzymatic activity is essential for thyroid hormone synthesis [11]. TG is the largest and most abundant thyroid antigen, consisting of two 330-kDa monomers. Coupling and iodination of
140
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Thyroid Antigens
Table 1. Characteristics of the 3 major thyroid autoantigens Characteristic
TPO
TG
TSHR
Gene chromosome Gene polymorphism Thyroid localization Thyroid concentration Protein size, kDa Protein type Number of peptides Glycosylation Binding to mannose receptor Immunogenicity
2p25 no apical membrane ++ 2 × 107 membrane bound ++ 10% no some
8q24 yes mainly colloid ++++ 2 × 330 soluble ++++ 12% yes high
14q31 yes basolateral membrane + 84.5 (approx. 60 TSHR A subunit) soluble (TSHR A subunit) + approx. 40% yes some
Antigen characteristics in bold contribute to the degree of immunogenicity. Table adapted and modified from McLachlan and Rapoport [10].
tyrosines into iodothyronines occurs within the soluble TG molecule. The TSHR belongs to the family of G-protein-coupled receptors with 7 transmembrane-spanning domains. Ligation of the TSHR with TSH stimulates thyroid hormone synthesis and release as well as thyroid growth. The TSHR consists of an extracellular domain, linked by a hinge region to the transmembrane and intracellular domain. At the surface of the thyrocyte, the holoreceptor undergoes cleavage in the hinge region due to breakage of disulfide bonds. As a result of this posttranslational modification, the extracellular TSHR A subunit is shed. The shed TSHR A subunit, rather than the holoreceptor, is apparently the autoantigen in Graves’ disease [12]. Interestingly, the related luteinizing hormone and follicle-stimulating hormone receptors do not undergo intracellular cleavage with shedding of the ectodomain, and autoimmunity against these gonadotropin receptors is not observed in humans. The TSHR A subunit is a soluble protein of about 60 kDa, and heavily glycosylated. Immunogenicity of antigens is expected to be higher in case of (a) genetic polymorphisms, (b) a high number of peptides available for binding to MHC (major histocompatibility complex) molecules on antigen-presenting cells (APCs), (c) membrane-bound antigens and (d) a high degree of glycosylation which facilitates antigen binding to cell surface mannose receptors on APCs [10]. According to these features, the immunogenicity of TG is higher than that of TPO and TSHR (table 1).
The immunological synapse refers to the interaction between APCs and T lymphocytes (T cells). Macrophages (Mφs), dendritic cells (DCs) but also B lymphocytes can act as professional APCs. Antigens are taken up by APCs and processed to peptides
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
141
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
The Immunological Synapse
142
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
that bind to MHC molecules on the surface of APCs. Antigens bound to human leukocyte antigen (HLA) class I or class II molecules are thus presented to T cells. The complex of MHC and peptide epitopes may be recognized by heterodimeric T cell antigen receptors (TCRs, T cell receptors), composed of α- and β-chains. T cell activation requires – next to recognition of the antigen – costimulation of T cells [10]. Interaction between MHC-bound antigen and TCR leads to induction of CD40 ligand on T cells. Binding of CD40 ligand to constitutively expressed CD40 on APCs results in induction of B7-1/2 molecules on the APC. B7-1/2 then binds to constitutively expressed CD28 on T cells. Costimulation is completed and T cell activation is initiated [10]. This process also induces expression of the CTLA4 (cytotoxic T-lymphocyteassociated protein 4) molecule, which reduces interactions between APCs and T cells. Engagement with CTLA4 on T cells terminates the immune response, resulting in anergy. Activation of T cells results in generation of cytokines. Proliferation, differentiation and maturation of T cells leads to infiltrating CD4+ T cells and cytotoxic CD8+ T cells (CD4 being the receptor for MHC class II and CD8 being the coreceptor for MHC class I molecules). B cells may turn into plasma cells secreting Abs. Some B cells remain as memory B cells or function as APCs by their cell surface immunoglobulins that function as specific antigen receptor [10]. These specific antigens are then internalized, processed and presented to T cells. Mounting an immune response against exogenous antigens (e.g. from invading microorganisms) has obvious advantages, but presentation of endogenous ‘self’-antigens to T cells may result in autoimmunity. A number of complex regulatory mechanisms serves to prevent an immune response directed against self-antigens. The process of self-tolerance is enacted in the thymus (central tolerance) and in peripheral tissues. Immature T cells originating from the bone marrow enter the thymus, where they undergo a process of selection and finally exit as CD4+ or CD8+ T cells depleted of high-affinity binding sites of self-peptides [10]. Central tolerance is accomplished by negative selection of autoreactive T cells in the thymic medulla. These self-reactive T cells emerge during the random recombination of gene segments that encode variable parts of the TCR for the antigen [13]. Thymic medullary epithelial cells express peptides from self-proteins, which in cooperation with dendritic cells are presented to immature T cells. T cells that recognize these self-peptides with high affinity are deleted. T cells that have moderate affinities for self-peptides are positively selected and leave the thymus to become mature T cells. In general, the higher the concentration of autoantigen in the thymus is, the greater the degree of self-tolerance will be. An important factor in controlling the expression of self-proteins in the thymus is the autoimmune regulator gene (AIRE). In the absence of AIRE, the levels of some self-proteins in thymic medullary epithelial cells are reduced, and autoimmunity develops spontaneously [10]. The rare autoimmune polyendocrinopathy/candidiasis/ectodermal dystrophy (APECED) syndrome in humans, also known as autoimmune polyglandular syndrome type 1, is a monogenetic disease due to AIRE mutations, which becomes manifest in infancy or youth, presenting itself (next to mucocutaneous can-
didiasis) with hypoparathyroidism or Addison’s disease; TG-Ab, TPO-Ab and autoimmune hypothyroidism can be part of the syndrome as well [14]. AIRE mutations are seldom observed in adult AITD patients, occurring in about 0.3–0.6% of patients with Graves’ disease or autoimmune hypothyroidism. AIRE is not considered as a susceptibility gene for common autoimmune endocrinopathies [15, 16]. Animal experiments have shown that intrathymic expression of TSHR, TG and especially of TPO are all AIRE dependent. The spontaneous development of Abs against these thyroid antigens is not always directly related to their level of intrathymic expression (which is higher for TG than for TPO and TSHR). Central tolerance seems less important for thyroid autoimmunity than thyroid antigen immunogenicity (table 1) [10]. Central tolerance may not eliminate all self-reactive T cells. Self-antigen presentation in the thymus generates regulatory T cells (Treg) that can inhibit in the periphery those self-reactive T cells that escaped negative selection in the thymus [13]. Treg cells can be natural (constitutive, developing in the thymus) or inducible (involved in the adaptive immune response) [10]. Both natural and inducible Treg cells are characterized by the expression of CD4, CD25 [the interleukin (IL)-2 receptor α-chain], and the transcription factor FOXP3 (forkhead box P3 protein). Mutations in FOXP3 result in the immunodysregulation/polyendocrinopathy/enteropathy/X-linked (IPEX) syndrome, a fatal condition characterized by neonatal onset of autoimmune diseases. FOXP3+ Treg cells suppress immune responses of self-reactive T cells, thereby contributing to peripheral tolerance. Another subset of Treg cells express CD8 and CD122 (IL-2 receptor β-chain); they also control autoreactive T cells in the periphery [10]. B cells with highaffinity receptors for self-antigens are tolerated by clonal deletion, anergy (functional inactivation) and receptor editing. In the case of receptor editing, B cells which express immunoglobulins on their surface that have specificity as an autoantigen receptor can edit and replace these receptors with different Ab gene rearrangements [10]. Tolerance mechanisms in B cells may be regarded as a secondary or ‘fail-safe’ mechanism.
Breakdown of Tolerance to Self-Antigens
Genetic Factors Thyroid-Stimulating Hormone Receptor Single-nucleotide polymorphisms (SNPs) in TSHR have been associated with Graves’ disease. The functional consequences of these intronic SNPs are not entirely clear.
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
143
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Tolerance to self-antigens is usually accomplished effectively at the level of the thymus (central tolerance). In case some autoreactive T cells escape central tolerance, tolerance may be effectuated also in the peripheral tissues (peripheral tolerance). If tolerance to self-antigens fails, autoimmunity develops. Breaking tolerance to self-antigens can happen in subjects who have the wrong genes and who are exposed to the wrong environment (fig. 1) [17].
Generation of immune repertoires in thymus or bone marrow
Central tolerance Anti-self lymphocytes deleted by apoptosis Wrong genes
Some anti-self lymphocytes escape central tolerance
Wrong environment
Peripheral tolerance • Ignorance • Anergy • Homeostasis – CTLA4 • Regulation
Failure of tolerance resulting in autoimmunity
Fig. 1. Interaction of genes and environment on the development of autoimmune disease. Reproduced with permission from Weetman [17].
They give rise to RNA splice variants, which may increase the level of potentially autoantigenic TSHR A subunits [18]. Subjects homozygous or heterozygous for TSHR SNP rs179247 have fewer thymic TSHR mRNA transcripts than subjects who do not carry the SNP [19]. SNP carriers may thus have decreased central tolerance to TSHR, which puts them at risk for autoimmunity against TSHR. Thyroglobulin Multiple SNPs in TG have been associated with both Graves’ and Hashimoto’s diseases. SNPs were located in exons in Caucasians and in introns in Japanese [20, 21]. The susceptible haplotype would enhance antigen presentation.
144
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Human Leukocyte Antigen MHC molecules bind antigenic peptides and present them to T cells. MHC genes are associated with thyroid autoimmunity. For instance, a single amino acid variation in the peptide-binding cleft of HLA-DR3 results in an arginine at position 74 of the β-chain, which is strongly associated with Graves’ disease. This variant interacts with the SNP in exon 33 of TG, suggesting that HLA-DRβ-Arg74 presents the disease-associated TG SNP alleles more efficiently to T cells [20, 22]. This may happen both in the thymus and in the periphery.
CD40 Genotype The CD40 genotype associated with Graves’ disease increases the expression of CD40, also on nonimmune cells like thyrocytes [20]. Intrathyroidal CD40 expression enhances TSHR-Ab responses in experimental animals, but it is unknown whether CD40 expression reduces the efficacy of deleting autoreactive T cells in the thymus [10]. Cytotoxic T-Lymphocyte-Associated Protein 4 The association of CTLA4 polymorphisms with Graves’ disease is well known. CTLA4 reduces the interaction between APCs and T cells, and CTLA4 polymorphisms reduce the efficacy of deleting autoreactive T cells in the thymus [20]. Protein Tyrosine Phosphatase Nonreceptor Type 22 Lymphoid tyrosine phosphatase is encoded by PTPN22 (protein tyrosine phosphatase nonreceptor type 22); it is a powerful inhibitor of T cell activation. SNPs in PTPN22 are linked with both Graves’ and Hashimoto’s disease. Because the SNP makes the protein an even stronger inhibitor of T cells (as it is a gain-of-function mutation), it is not understood how the SNP contributes to the breakdown of tolerance [20]. Fc Receptor-Like 3 Gene The Fc receptor-like 3 gene (FCRL3) shares significant structural homology to classical receptors for immunoglobulin constant chains (Fc receptors). The gene encodes for a member of the immunoglobulin receptor superfamily, expressed particularly strongly on the surface of B cells but also on T cells. FCRL3 polymorphisms predispose for Graves’ disease by downregulation of B cell receptor-mediated signaling, incomplete induction of anergy and deletion in autoreactive B cells, and finally to breakdown of B cell tolerance; the presence of FCRL3 on a subset of Treg cells is linked to reduced suppressor activity [23].
Forkhead Box P3 Polymorphisms in FOXP3, the definitive marker of Treg cells, have been associated with AITD in Caucasians (especially with Graves’ disease below the age of 30 years) but not in Japanese [26, 27]. The location of FOXP3 on the X chromosome might contribute to the female preponderance in AITD.
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
145
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Interleukin-2 Receptor α-Chain CD25 is a marker for the IL-2 receptor α-chain present predominantly on CD25+ T cells. The gene is a susceptibility locus for Graves’ disease, involved with Treg cells [24]. First- and second-degree female euthyroid relatives of AITD patients (who by definition are at an increased risk to develop AITD themselves) had a lower proportion of CD4+ CD25+ T cells and lower serum concentrations of soluble IL-2 receptor [25]. The data are interpreted as a sign of poor capability to preserve tolerance in AITD relatives, present already prior to the occurrence of thyroid Abs and an abnormal TSH.
Environmental Factors Iodine TPO-Ab, TG-Ab and autoimmune hypothyroidism are more common in iodinesufficient than in iodine-deficient regions. Recent population-based studies in Denmark provide further proof for this modulating effect of iodine intake: after mandatory iodization of salt, the prevalence of TPO-Ab and TG-Ab increased (odds ratios were 1.80 and 1.49, respectively) as did the incidence rate of hypothyroidism (relative risk 1.23, 95% CI 1.07–1.42) [29, 30]. The effect is also seen in childhood. In second-grade primary school students, the mean urinary iodine excretion in children with Hashimoto’s thyroiditis was higher than in healthy controls (132 vs. 81 μg/l, p < 0.001) [31]. A cross-sectional epidemiological study in children and adolescents aged between 1 and 16 years in Spain observed thyroid autoimmunity (positive TPO-Ab and/or TG-Ab) in 3.7% [32]. Grouped according to iodine intake as assessed by the median urinary iodine concentration, the prevalence of thyroid autoimmunity was 2.6% in iodine deficiency (300 μg/l). It is not completely understood how iodine affects thyroid autoimmunity. Animal experiments have shown that incorporating dietary iodine into thyroglobulin increases its immunogenicity. TG is iodinated in the thyroid. Central tolerance is likely induced by noniodinated TG, whereas iodinated TG may act as a neoantigen. Iodination of TG may alter recognition of the antigen by Abs, e.g. by unmasking a cryptic (B domain) epitope on TG recognized by human TG-Ab. Iodine is essential for human T cell recognition of human TG [33, 34]. It is less
146
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Whenever studied, the role of the above-mentioned susceptibility genes is similar in childhood and adulthood AITD. Polymorphisms in thyroid-specific genes may promote Graves’ disease or Hashimoto’s disease, or both. Polymorphisms in immune-regulatory genes may promote AITD, but are not specific for AITD as they are associated with other autoimmune diseases as well; it explains why various autoimmune diseases may occur in the same patient. It must be acknowledged that the mechanism of action of many susceptibility loci is incompletely understood. For instance, why are so many SNPs located in noncoding parts of the gene? The mechanisms are apparently more complex than previously thought, and gene-gene or geneenvironment interactions have hardly been evaluated. Genome-wide association studies continue to detect additional genes and loci conferring risk for AITD [28]. The odds ratio of each locus for AITD is rather low in the order of 1.5–2.0, with slightly higher odds of 2.0–4.0 for HLA. A recent Chinese study found that taking together the effects of 4 HLA loci and 5 non-HLA loci accounted for only 9.3% of heritability of Graves’ disease [28]. The data strongly suggest there must be many more still undetected susceptibility genes, each variant contributing just a little to the development of AITD.
likely that iodide could induce de novo thyroid autoimmunity in humans. Available data rather suggest that increased iodine intake enhances ongoing autoimmune responses [10].
Infections Infection with Yersinia enterocolitica has long been implicated in the pathogenesis of Graves’ disease because IgG from Graves’ patients inhibits binding of TSH to Y. enterocolitica outer membrane proteins, and conversely IgG from patients with Y. enterocolitica infection inhibit binding of TSH to thyroid membranes. There is crossreactivity between Y. enterocolitica outer membrane proteins and TSHR-Abs [38, 39]. Clinical evidence supporting a role of Y. enterocolitica infection in the pathogenesis of AITD is, however, limited. In the Amsterdam AITD cohort, the only prospective study in this area, no association was found between Yersinia outer membrane protein IgA or IgG status and de novo occurrence of thyroid Abs or development of overt autoimmune hypo- or hyperthyroidism [40]. Hepatitis C virus is the only infectious agent that is clearly associated with an increased risk for autoimmune thyroiditis [41]. It is able to infect human thyrocytes resulting in the production of proinflammatory cytokines which may enhance autoimmune responses [42]. Thyroid Abs are more frequent in children with untreated hepatitis C virus infection than in controls [43]. Enhancement of autoimmune responses by microorganisms likely occurs as a result of bystander activity [10]. Autoreactive T cells may bypass negative selection during infection and respond to self-antigen stimulation [44]. In this context, the hygiene hypothesis should be mentioned, which postulates that subjects regularly exposed to
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
147
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Smoking Smoking is a well-established risk factor for Graves’ disease. The odds ratio for Graves’ hyperthyroidism is 3.30 (95% CI 2.09–5.22) in current smokers as compared to never smokers. The odds for GO is even higher: 4.40 (95% CI 2.88–6.73) in ever smokers versus never smokers [35]. Data on the effect of smoking in children are limited. According to a European questionnaire study, out of 1,963 patients with juvenile Graves’ hyperthyroidism, 33% had GO. The prevalence of GO among juvenile patients with Graves’ hyperthyroidism was 37, 27 and 26% in countries in which the smoking prevalence among teenagers was ≥25, 20–25 and
Thyroid Autoimmunity Wilmar M. Wiersinga Department of Endocrinology and Metabolism, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
Abstract
Autoimmune thyroid disease (AITD) is a multifactorial or so-called ‘complex’ disease in which autoimmunity against thyroid antigens develops against a particular genetic background facilitated by exposure to environmental factors. Clinically relevant thyroid antigens are thyroid peroxidase (TPO), thyroglobulin (TG) and thyroid-stimulating hormone receptor (TSHR). Autoimmunity against these self-antigens gives rise to thyroid antibodies (Abs). TPO-Ab and TG-Ab are the hallmark of chronic lymphocytic thyroiditis (also called chronic autoimmune thyroiditis or
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Autoimmune thyroid disease (AITD) is a multifactorial disease in which autoimmunity against thyroid antigens develops against a particular genetic background facilitated by exposure to environmental factors. Immunogenicity of the major thyroid antigens thyroid peroxidase, thyroglobulin (TG) and thyrotropin receptor (TSHR) is increased by genetic polymorphisms, a high number of antigenic peptides available for binding to human leukocyte antigen (HLA), and a high degree of glycosylation. Antigens bound to HLA are presented by antigen-presenting cells to T cell receptors. Further interaction between both cells is required via binding of CD40 ligand to CD40 and of B7-1/2 to CD28 for activation of T cells. Complex regulatory mechanisms serve to prevent an immune response directed against ‘self’-antigens in the thymus (central tolerance) and in peripheral tissues (peripheral tolerance) with the help of regulatory T cells. Breakdown of tolerance to thyroid antigens can result in thyroid autoimmunity, which may happen in subjects who have the wrong genes and who are exposed to the wrong environment. Polymorphisms in thyroid genes (TG, TSHR) and immunoregulatory genes (HLA, CTLA4, PTPN22, CD40, FCRL3, IL2RA, FOXP3) would contribute for about 70% to AITD, and environmental exposures (like iodine, smoking, infections, parity) for the remaining 30%. Thyroid-infiltrating activated T cells may lead to cell-mediated immunity, thyroid injury and eventually hypothyroidism, whereas humoral immunity via TSHR-stimulating antibodies may give rise to hyperthyroidism. Pediatric Hashimoto’s and Graves’ disease are less prevalent than in adults, and the female preponderance is also less marked in children. The discrepancy is probably due to relatively less involvement of environmental insults in children, whereas the prevalence of risk alleles in AITD © 2014 S. Karger AG, Basel children is higher than in AITD adults.
Hashimoto’s thyroiditis) which may lead to Hashimoto’s hypothyroidism although most patients will remain euthyroid. TSHR-Abs are the hallmark of Graves’ disease resulting in Graves’ hyperthyroidism. Hashimoto’s hypothyroidism and Graves’ hyperthyroidism can be viewed as the opposite ends of a continuous spectrum of AITD, but in reality there is substantial overlap between Hashimoto’s and Graves’ diseases. The rare entity of hashitoxicosis refers to patients with high concentrations of TPO-Ab and TG-Ab who initially present with hyperthyroidism, but in whom hyperthyroidism is replaced rather soon by hypothyroidism; this sequence of events is frequently related to a change from stimulating to blocking TSHR-Abs [1]. Whereas almost all Hashimoto patients have TPO-Ab and/or TG-Ab, these Abs are also present in up to 70% of patients with Graves’ hyperthyroidism. Hypothyroidism due to chronic autoimmune thyroiditis occurs in the long run in up to 20% of patients with Graves’ hyperthyroidism who have entered remission after a course of antithyroid drugs; blocking TSHR-Ab contributed to the hypothyroidism in one third of the cases, and chronic autoimmune thyroiditis in the remaining two thirds [2]. In this review we will discuss the characteristics of thyroid antigens that contribute to their immunogenicity, the essential role of the immunological synapse in which antigens are presented to T cells, and how tolerance to self-antigens can be broken by genes and facilitated by environmental factors. We also discuss what is known on the early stages of AITD immunopathogenesis in humans and on the causes of differences in prevalence and gender distribution between childhood AITD and adulthood AITD.
The 3 major thyroid autoantigens involved in AITD are TPO, TG and TSHR. There is insufficient evidence that autoimmunity against other thyroid proteins plays an important role in the initiation or maintenance of thyroid autoimmunity [3]. Abs against the sodium/iodide symporter expressed on the basolateral membrane of thyrocytes have been detected in sera of AITD patients albeit in low frequency, but these Abs are rarely capable of modulating sodium/iodide symporter activity [4, 5]. Abs against pendrin (located at the apical membrane of thyrocytes) were present in 81% of AITD patients in Japan [6], but only in 8.8% in the UK when a more advanced radioligand binding assay was employed [7]. Autoimmunity against the insulin-like growth factor 1 receptor is thought to play an additional role in Graves’ ophthalmopathy (GO) [8], but Abs against it in sera of patients with GO are as common as in controls (10 vs. 11%) and thus lack specificity [9]. Table 1 lists the characteristics of the 3 major thyroid autoantigens [10]. TPO is a membrane-bound homodimer of two 107-kDa subunits. Its enzymatic activity is essential for thyroid hormone synthesis [11]. TG is the largest and most abundant thyroid antigen, consisting of two 330-kDa monomers. Coupling and iodination of
140
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Thyroid Antigens
Table 1. Characteristics of the 3 major thyroid autoantigens Characteristic
TPO
TG
TSHR
Gene chromosome Gene polymorphism Thyroid localization Thyroid concentration Protein size, kDa Protein type Number of peptides Glycosylation Binding to mannose receptor Immunogenicity
2p25 no apical membrane ++ 2 × 107 membrane bound ++ 10% no some
8q24 yes mainly colloid ++++ 2 × 330 soluble ++++ 12% yes high
14q31 yes basolateral membrane + 84.5 (approx. 60 TSHR A subunit) soluble (TSHR A subunit) + approx. 40% yes some
Antigen characteristics in bold contribute to the degree of immunogenicity. Table adapted and modified from McLachlan and Rapoport [10].
tyrosines into iodothyronines occurs within the soluble TG molecule. The TSHR belongs to the family of G-protein-coupled receptors with 7 transmembrane-spanning domains. Ligation of the TSHR with TSH stimulates thyroid hormone synthesis and release as well as thyroid growth. The TSHR consists of an extracellular domain, linked by a hinge region to the transmembrane and intracellular domain. At the surface of the thyrocyte, the holoreceptor undergoes cleavage in the hinge region due to breakage of disulfide bonds. As a result of this posttranslational modification, the extracellular TSHR A subunit is shed. The shed TSHR A subunit, rather than the holoreceptor, is apparently the autoantigen in Graves’ disease [12]. Interestingly, the related luteinizing hormone and follicle-stimulating hormone receptors do not undergo intracellular cleavage with shedding of the ectodomain, and autoimmunity against these gonadotropin receptors is not observed in humans. The TSHR A subunit is a soluble protein of about 60 kDa, and heavily glycosylated. Immunogenicity of antigens is expected to be higher in case of (a) genetic polymorphisms, (b) a high number of peptides available for binding to MHC (major histocompatibility complex) molecules on antigen-presenting cells (APCs), (c) membrane-bound antigens and (d) a high degree of glycosylation which facilitates antigen binding to cell surface mannose receptors on APCs [10]. According to these features, the immunogenicity of TG is higher than that of TPO and TSHR (table 1).
The immunological synapse refers to the interaction between APCs and T lymphocytes (T cells). Macrophages (Mφs), dendritic cells (DCs) but also B lymphocytes can act as professional APCs. Antigens are taken up by APCs and processed to peptides
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
141
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
The Immunological Synapse
142
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
that bind to MHC molecules on the surface of APCs. Antigens bound to human leukocyte antigen (HLA) class I or class II molecules are thus presented to T cells. The complex of MHC and peptide epitopes may be recognized by heterodimeric T cell antigen receptors (TCRs, T cell receptors), composed of α- and β-chains. T cell activation requires – next to recognition of the antigen – costimulation of T cells [10]. Interaction between MHC-bound antigen and TCR leads to induction of CD40 ligand on T cells. Binding of CD40 ligand to constitutively expressed CD40 on APCs results in induction of B7-1/2 molecules on the APC. B7-1/2 then binds to constitutively expressed CD28 on T cells. Costimulation is completed and T cell activation is initiated [10]. This process also induces expression of the CTLA4 (cytotoxic T-lymphocyteassociated protein 4) molecule, which reduces interactions between APCs and T cells. Engagement with CTLA4 on T cells terminates the immune response, resulting in anergy. Activation of T cells results in generation of cytokines. Proliferation, differentiation and maturation of T cells leads to infiltrating CD4+ T cells and cytotoxic CD8+ T cells (CD4 being the receptor for MHC class II and CD8 being the coreceptor for MHC class I molecules). B cells may turn into plasma cells secreting Abs. Some B cells remain as memory B cells or function as APCs by their cell surface immunoglobulins that function as specific antigen receptor [10]. These specific antigens are then internalized, processed and presented to T cells. Mounting an immune response against exogenous antigens (e.g. from invading microorganisms) has obvious advantages, but presentation of endogenous ‘self’-antigens to T cells may result in autoimmunity. A number of complex regulatory mechanisms serves to prevent an immune response directed against self-antigens. The process of self-tolerance is enacted in the thymus (central tolerance) and in peripheral tissues. Immature T cells originating from the bone marrow enter the thymus, where they undergo a process of selection and finally exit as CD4+ or CD8+ T cells depleted of high-affinity binding sites of self-peptides [10]. Central tolerance is accomplished by negative selection of autoreactive T cells in the thymic medulla. These self-reactive T cells emerge during the random recombination of gene segments that encode variable parts of the TCR for the antigen [13]. Thymic medullary epithelial cells express peptides from self-proteins, which in cooperation with dendritic cells are presented to immature T cells. T cells that recognize these self-peptides with high affinity are deleted. T cells that have moderate affinities for self-peptides are positively selected and leave the thymus to become mature T cells. In general, the higher the concentration of autoantigen in the thymus is, the greater the degree of self-tolerance will be. An important factor in controlling the expression of self-proteins in the thymus is the autoimmune regulator gene (AIRE). In the absence of AIRE, the levels of some self-proteins in thymic medullary epithelial cells are reduced, and autoimmunity develops spontaneously [10]. The rare autoimmune polyendocrinopathy/candidiasis/ectodermal dystrophy (APECED) syndrome in humans, also known as autoimmune polyglandular syndrome type 1, is a monogenetic disease due to AIRE mutations, which becomes manifest in infancy or youth, presenting itself (next to mucocutaneous can-
didiasis) with hypoparathyroidism or Addison’s disease; TG-Ab, TPO-Ab and autoimmune hypothyroidism can be part of the syndrome as well [14]. AIRE mutations are seldom observed in adult AITD patients, occurring in about 0.3–0.6% of patients with Graves’ disease or autoimmune hypothyroidism. AIRE is not considered as a susceptibility gene for common autoimmune endocrinopathies [15, 16]. Animal experiments have shown that intrathymic expression of TSHR, TG and especially of TPO are all AIRE dependent. The spontaneous development of Abs against these thyroid antigens is not always directly related to their level of intrathymic expression (which is higher for TG than for TPO and TSHR). Central tolerance seems less important for thyroid autoimmunity than thyroid antigen immunogenicity (table 1) [10]. Central tolerance may not eliminate all self-reactive T cells. Self-antigen presentation in the thymus generates regulatory T cells (Treg) that can inhibit in the periphery those self-reactive T cells that escaped negative selection in the thymus [13]. Treg cells can be natural (constitutive, developing in the thymus) or inducible (involved in the adaptive immune response) [10]. Both natural and inducible Treg cells are characterized by the expression of CD4, CD25 [the interleukin (IL)-2 receptor α-chain], and the transcription factor FOXP3 (forkhead box P3 protein). Mutations in FOXP3 result in the immunodysregulation/polyendocrinopathy/enteropathy/X-linked (IPEX) syndrome, a fatal condition characterized by neonatal onset of autoimmune diseases. FOXP3+ Treg cells suppress immune responses of self-reactive T cells, thereby contributing to peripheral tolerance. Another subset of Treg cells express CD8 and CD122 (IL-2 receptor β-chain); they also control autoreactive T cells in the periphery [10]. B cells with highaffinity receptors for self-antigens are tolerated by clonal deletion, anergy (functional inactivation) and receptor editing. In the case of receptor editing, B cells which express immunoglobulins on their surface that have specificity as an autoantigen receptor can edit and replace these receptors with different Ab gene rearrangements [10]. Tolerance mechanisms in B cells may be regarded as a secondary or ‘fail-safe’ mechanism.
Breakdown of Tolerance to Self-Antigens
Genetic Factors Thyroid-Stimulating Hormone Receptor Single-nucleotide polymorphisms (SNPs) in TSHR have been associated with Graves’ disease. The functional consequences of these intronic SNPs are not entirely clear.
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
143
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Tolerance to self-antigens is usually accomplished effectively at the level of the thymus (central tolerance). In case some autoreactive T cells escape central tolerance, tolerance may be effectuated also in the peripheral tissues (peripheral tolerance). If tolerance to self-antigens fails, autoimmunity develops. Breaking tolerance to self-antigens can happen in subjects who have the wrong genes and who are exposed to the wrong environment (fig. 1) [17].
Generation of immune repertoires in thymus or bone marrow
Central tolerance Anti-self lymphocytes deleted by apoptosis Wrong genes
Some anti-self lymphocytes escape central tolerance
Wrong environment
Peripheral tolerance • Ignorance • Anergy • Homeostasis – CTLA4 • Regulation
Failure of tolerance resulting in autoimmunity
Fig. 1. Interaction of genes and environment on the development of autoimmune disease. Reproduced with permission from Weetman [17].
They give rise to RNA splice variants, which may increase the level of potentially autoantigenic TSHR A subunits [18]. Subjects homozygous or heterozygous for TSHR SNP rs179247 have fewer thymic TSHR mRNA transcripts than subjects who do not carry the SNP [19]. SNP carriers may thus have decreased central tolerance to TSHR, which puts them at risk for autoimmunity against TSHR. Thyroglobulin Multiple SNPs in TG have been associated with both Graves’ and Hashimoto’s diseases. SNPs were located in exons in Caucasians and in introns in Japanese [20, 21]. The susceptible haplotype would enhance antigen presentation.
144
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Human Leukocyte Antigen MHC molecules bind antigenic peptides and present them to T cells. MHC genes are associated with thyroid autoimmunity. For instance, a single amino acid variation in the peptide-binding cleft of HLA-DR3 results in an arginine at position 74 of the β-chain, which is strongly associated with Graves’ disease. This variant interacts with the SNP in exon 33 of TG, suggesting that HLA-DRβ-Arg74 presents the disease-associated TG SNP alleles more efficiently to T cells [20, 22]. This may happen both in the thymus and in the periphery.
CD40 Genotype The CD40 genotype associated with Graves’ disease increases the expression of CD40, also on nonimmune cells like thyrocytes [20]. Intrathyroidal CD40 expression enhances TSHR-Ab responses in experimental animals, but it is unknown whether CD40 expression reduces the efficacy of deleting autoreactive T cells in the thymus [10]. Cytotoxic T-Lymphocyte-Associated Protein 4 The association of CTLA4 polymorphisms with Graves’ disease is well known. CTLA4 reduces the interaction between APCs and T cells, and CTLA4 polymorphisms reduce the efficacy of deleting autoreactive T cells in the thymus [20]. Protein Tyrosine Phosphatase Nonreceptor Type 22 Lymphoid tyrosine phosphatase is encoded by PTPN22 (protein tyrosine phosphatase nonreceptor type 22); it is a powerful inhibitor of T cell activation. SNPs in PTPN22 are linked with both Graves’ and Hashimoto’s disease. Because the SNP makes the protein an even stronger inhibitor of T cells (as it is a gain-of-function mutation), it is not understood how the SNP contributes to the breakdown of tolerance [20]. Fc Receptor-Like 3 Gene The Fc receptor-like 3 gene (FCRL3) shares significant structural homology to classical receptors for immunoglobulin constant chains (Fc receptors). The gene encodes for a member of the immunoglobulin receptor superfamily, expressed particularly strongly on the surface of B cells but also on T cells. FCRL3 polymorphisms predispose for Graves’ disease by downregulation of B cell receptor-mediated signaling, incomplete induction of anergy and deletion in autoreactive B cells, and finally to breakdown of B cell tolerance; the presence of FCRL3 on a subset of Treg cells is linked to reduced suppressor activity [23].
Forkhead Box P3 Polymorphisms in FOXP3, the definitive marker of Treg cells, have been associated with AITD in Caucasians (especially with Graves’ disease below the age of 30 years) but not in Japanese [26, 27]. The location of FOXP3 on the X chromosome might contribute to the female preponderance in AITD.
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
145
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Interleukin-2 Receptor α-Chain CD25 is a marker for the IL-2 receptor α-chain present predominantly on CD25+ T cells. The gene is a susceptibility locus for Graves’ disease, involved with Treg cells [24]. First- and second-degree female euthyroid relatives of AITD patients (who by definition are at an increased risk to develop AITD themselves) had a lower proportion of CD4+ CD25+ T cells and lower serum concentrations of soluble IL-2 receptor [25]. The data are interpreted as a sign of poor capability to preserve tolerance in AITD relatives, present already prior to the occurrence of thyroid Abs and an abnormal TSH.
Environmental Factors Iodine TPO-Ab, TG-Ab and autoimmune hypothyroidism are more common in iodinesufficient than in iodine-deficient regions. Recent population-based studies in Denmark provide further proof for this modulating effect of iodine intake: after mandatory iodization of salt, the prevalence of TPO-Ab and TG-Ab increased (odds ratios were 1.80 and 1.49, respectively) as did the incidence rate of hypothyroidism (relative risk 1.23, 95% CI 1.07–1.42) [29, 30]. The effect is also seen in childhood. In second-grade primary school students, the mean urinary iodine excretion in children with Hashimoto’s thyroiditis was higher than in healthy controls (132 vs. 81 μg/l, p < 0.001) [31]. A cross-sectional epidemiological study in children and adolescents aged between 1 and 16 years in Spain observed thyroid autoimmunity (positive TPO-Ab and/or TG-Ab) in 3.7% [32]. Grouped according to iodine intake as assessed by the median urinary iodine concentration, the prevalence of thyroid autoimmunity was 2.6% in iodine deficiency (300 μg/l). It is not completely understood how iodine affects thyroid autoimmunity. Animal experiments have shown that incorporating dietary iodine into thyroglobulin increases its immunogenicity. TG is iodinated in the thyroid. Central tolerance is likely induced by noniodinated TG, whereas iodinated TG may act as a neoantigen. Iodination of TG may alter recognition of the antigen by Abs, e.g. by unmasking a cryptic (B domain) epitope on TG recognized by human TG-Ab. Iodine is essential for human T cell recognition of human TG [33, 34]. It is less
146
Wiersinga Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Whenever studied, the role of the above-mentioned susceptibility genes is similar in childhood and adulthood AITD. Polymorphisms in thyroid-specific genes may promote Graves’ disease or Hashimoto’s disease, or both. Polymorphisms in immune-regulatory genes may promote AITD, but are not specific for AITD as they are associated with other autoimmune diseases as well; it explains why various autoimmune diseases may occur in the same patient. It must be acknowledged that the mechanism of action of many susceptibility loci is incompletely understood. For instance, why are so many SNPs located in noncoding parts of the gene? The mechanisms are apparently more complex than previously thought, and gene-gene or geneenvironment interactions have hardly been evaluated. Genome-wide association studies continue to detect additional genes and loci conferring risk for AITD [28]. The odds ratio of each locus for AITD is rather low in the order of 1.5–2.0, with slightly higher odds of 2.0–4.0 for HLA. A recent Chinese study found that taking together the effects of 4 HLA loci and 5 non-HLA loci accounted for only 9.3% of heritability of Graves’ disease [28]. The data strongly suggest there must be many more still undetected susceptibility genes, each variant contributing just a little to the development of AITD.
likely that iodide could induce de novo thyroid autoimmunity in humans. Available data rather suggest that increased iodine intake enhances ongoing autoimmune responses [10].
Infections Infection with Yersinia enterocolitica has long been implicated in the pathogenesis of Graves’ disease because IgG from Graves’ patients inhibits binding of TSH to Y. enterocolitica outer membrane proteins, and conversely IgG from patients with Y. enterocolitica infection inhibit binding of TSH to thyroid membranes. There is crossreactivity between Y. enterocolitica outer membrane proteins and TSHR-Abs [38, 39]. Clinical evidence supporting a role of Y. enterocolitica infection in the pathogenesis of AITD is, however, limited. In the Amsterdam AITD cohort, the only prospective study in this area, no association was found between Yersinia outer membrane protein IgA or IgG status and de novo occurrence of thyroid Abs or development of overt autoimmune hypo- or hyperthyroidism [40]. Hepatitis C virus is the only infectious agent that is clearly associated with an increased risk for autoimmune thyroiditis [41]. It is able to infect human thyrocytes resulting in the production of proinflammatory cytokines which may enhance autoimmune responses [42]. Thyroid Abs are more frequent in children with untreated hepatitis C virus infection than in controls [43]. Enhancement of autoimmune responses by microorganisms likely occurs as a result of bystander activity [10]. Autoreactive T cells may bypass negative selection during infection and respond to self-antigen stimulation [44]. In this context, the hygiene hypothesis should be mentioned, which postulates that subjects regularly exposed to
Thyroid Autoimmunity Szinnai G (ed): Paediatric Thyroidology. Endocr Dev. Basel, Karger, 2014, vol 26, pp 139–157 (DOI: 10.1159/000363161)
147
Downloaded by: UCSF Library & CKM 169.230.243.252 - 3/1/2015 6:27:42 PM
Smoking Smoking is a well-established risk factor for Graves’ disease. The odds ratio for Graves’ hyperthyroidism is 3.30 (95% CI 2.09–5.22) in current smokers as compared to never smokers. The odds for GO is even higher: 4.40 (95% CI 2.88–6.73) in ever smokers versus never smokers [35]. Data on the effect of smoking in children are limited. According to a European questionnaire study, out of 1,963 patients with juvenile Graves’ hyperthyroidism, 33% had GO. The prevalence of GO among juvenile patients with Graves’ hyperthyroidism was 37, 27 and 26% in countries in which the smoking prevalence among teenagers was ≥25, 20–25 and
