Update in Endocrine Autoimmunity

Abstract

Context: The endocrine system is a common target in pathogenic autoimmune responses, and there has been recent progress in our understanding, diagnosis, and treatment of autoimmune endocrine diseases.

Synthesis: Rapid progress has recently been made in our understanding of the genetic factors involved in endocrine autoimmune diseases. Studies on monogenic autoimmune diseases that include endocrine phenotypes like autoimmune polyglandular syndrome type 1 and immune dysregulation, polyendocrinopathy, enteropathy, X-linked have helped reveal the role of key regulators in the maintenance of immune tolerance. Highly powered genetic studies have found and confirmed many new genes outside of the established role of the human leukocyte antigen locus with these diseases, and indicate an essential role of immune response pathways in these diseases. Progress has also been made in identifying new autoantigens and the development of new animal models for the study of endocrine autoimmunity. Finally, although hormone replacement therapy is still likely to be a mainstay of treatment in these disorders, there are new agents being tested for potentially treating and reversing the underlying autoimmune process.

Conclusion: Although autoimmune endocrine disorders are complex in etiology, these recent advances should help contribute to improved outcomes for patients with, or at risk for, these disorders.

Recent progress made in understanding the genetics, diagnosis, and treatment of endocrine autoimmune diseases is reviewed.

Autoimmune diseases represent a significant health burden in the developed world afflicting 5–10% of the population (1), and a sizable percentage of these diseases involve an untoward immune response against an endocrine organ. Virtually any endocrine organ can be targeted by the immune system as part of an autoimmune response, and frequently responses to multiple organs can occur in the same patient as part of a polyglandular autoimmune syndrome. More common endocrine autoimmune syndromes include Hashimoto’s thyroiditis, Graves’ disease, and type 1 diabetes, whereas more rare syndromes include Addison’s disease, oophoritis, lymphocytic hypophysitis, and hypoparathyroidism. For years, the etiology and pathogenesis of these disorders have remained obscure, but the diseases are generally thought to involve a cellular and humoral immune response that pathologically targets the affected organ(s). This is evidenced by a wide number of observations, including the presence of autoantibodies in affected patients, improvement of some diseases by immunosuppressive drugs, and the demonstration of lymphocytic infiltrates in the targeted organs. Over the last few years, rapid progress in our understanding of these diseases has come through a number of efforts, particularly in genetics. In this review, I will highlight some of the recent advances in our understanding, diagnosis, and treatment of endocrine autoimmune diseases.

Genetics

There is good evidence that most autoimmune endocrine diseases have a genetic component to their etiology. Some of the best evidence comes from familial inheritance studies on type 1 diabetes and thyroiditis (2,3). In the case of type 1 diabetes, the lifetime concordance rate for disease in monogenic twins is around 50% and for siblings is around 3–4%. This shows significant risk when compared with the general population risk of around 0.3%. These data also show that there is a significant genetic contribution to disease risk and that other factors (i.e. environmental) are also involved in disease pathogenesis. For several decades, the major genetic association of autoimmune endocrine diseases with polymorphisms in the human leukocyte antigen (HLA) region has been recognized. The HLA is a genetic region on chromosome 6 that encodes a large number of immune response genes, and in most cases disease risk maps to polymorphisms in the major histocompatibility complex (MHC) class II genes DR and DQ. The MHC class II gene products along with antigenic peptides are part of the ligand complex for CD4+ T-cell receptors, and the association likely highlights the importance of T cells in these diseases (4). Interestingly, it remains to be determined how these risk polymorphisms lead to increased susceptibility to autoimmunity. Some investigators have proposed promiscuous peptide binding by MHC risk alleles as a potential mechanism, but more definitive data are needed (5). It is also important to note that in most cases, subjects harboring a MHC risk allele are more likely not to develop autoimmunity except in rare isolated incidents (6), thus, these risk alleles should be thought of as being necessary but not sufficient for the development of disease. Recently, significant progress has been made in expanding our understanding of genetic disease risk beyond the MHC, particularly with informative monogenic forms of endocrine autoimmunity and in highly powered genetic studies that include genome-wide association (GWA) efforts.

Monogenic diseases

Autoimmune polyglandular syndrome type 1 (APS1) is a rare monogenic autosomal recessive disorder characterized by a panoply of autoimmune syndromes in the same patient, many of which are directed against endocrine organs. Prominent clinical features are hypoparathyroidism, Addison’s disease, and mucocutaneous candidiasis (7). More variable endocrine features also include Hashimoto’s thyroiditis, oophoritis, type 1 diabetes, and lymphocytic hypophysitis. Through a positional cloning effort, the defective gene was identified in 1997 by two independent groups and termed autoimmune regulator (Aire) (8,9). Since its identification, much has been learned about the function of Aire in promoting immune tolerance and has been accelerated by the generation of a mouse model by knocking out the murine orthologue of the gene (10,11). Aire appears to function as a transcription factor and is mainly expressed in a specialized subset of cells in the thymus called medullary epithelial cells (mTECs). Within mTECs, Aire helps promote the transcription of many self-antigen genes, including the insulin gene (a known endocrine autoantigen) (11). A consequence of this self-antigen expression within the thymus is that it promotes the negative selection (or deletion) of autoreactive thymocytes that naturally develop in the thymus (12,13,14). Thus, in the absence of Aire, there is a failure to delete autoreactive T cells within the thymus, which then leads to a predisposition to widespread multi-organ autoimmunity (Fig. 1). Mouse studies have confirmed that the thymic defect is sufficient to induce the autoimmune syndrome associated with disease (11), and recent studies in humans have suggested that the long-known association of thymomas with the autoimmune syndrome myasthenia gravis may be attributable to the loss of AIRE expression in this thymic tumor (15). In addition, there is a developing picture that similar mechanisms are in play for more common endocrine autoimmune syndromes, like type 1 diabetes, in which a polymorphism in the insulin gene has been demonstrated to control thymic expression levels and correlates with disease risk (i.e. high thymic expression alleles have lower disease risk) (16,17,18). Recent associations with variation in the thyroglobulin gene and thyroiditis (3,19) could involve a similar mechanism, but this has yet to be tested. An autosomal dominant allele of AIRE has also been recently associated with Hashimoto’s thyroiditis (20), and recently the susceptibility has been shown to be due to a quantitative effect on self-antigen expression within the thymus (21). Together, these recent advances on Aire have helped establish a critical relationship between thymic expression of self-antigens and the prevention of autoimmune endocrine syndromes.

Figure 1.

Model of the function of Aire in the thymus. A, Aire appears to help mediate the transcription of many self-antigens in mTECs in the thymus. B, Impact of Aire on T-cell selection. These self-antigens are then presented in the thymus to developing thymocytes (blue-colored cells) in the medulla, and this results in the deletion of self-antigen specific thymocytes in this compartment. In the absence of Aire, the self-antigens fail to be generated by these mTECs, and self-antigen specific T cells mature and escape the thymus and migrate into the periphery and promote autoimmune responses.

Another monogenic autoimmune syndrome that has brought new mechanistic insights to immune tolerance is immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX). This is an X-linked disorder that is characterized by a severe autoimmunity syndrome in which most affected subjects usually die before the age of 2 yr if they do not receive bone marrow transplantation. Common autoimmune endocrine syndromes in these patients include type 1 diabetes and thyroiditis (22). The defective gene in this disorder has been mapped to the transcription factor FoxP3, and recent studies have established that FoxP3 plays a critical role in the function of a special T-cell subset called regulatory T cells (Tregs) (23,24,25). Tregs are CD4+CD25+ T cells that have the remarkable capability to suppress effector T-cell responses, including those directed at self (Fig. 2) (26). These cells develop within the thymus and are thought to have a preferential specificity for self-antigens, perhaps at least in part due to Aire-dependent mechanisms (27). Preferential depletion (28) or loss of function of these cells (through knocking out FoxP3) has been demonstrated in animal models to lead to catastrophic autoimmunity similar to that in IPEX patients. FoxP3 likely plays a number of critical functions in allowing the suppressor activity of these cells to be promoted, but the exact details of the suppression mechanism remain unclear, especially in vivo (29). Interestingly, Tregs have been used as a tool to suppress and reverse type 1 diabetes in animal models (30,31), and this has important future clinical implications. This is because the suppression mechanism in vivo appears to be dependent on the antigenic specificity of the Treg population that is used. Thus, it may someday be possible to induce antigen or organ-specific tolerance by treatment with clonal populations of Tregs as a method to cure or reverse a given autoimmune disease without conferring the risk of global immunosuppression.

Figure 2.

Model of Treg function. Tregs expressing the FoxP3 gene play a key role in dampening responses by effector T cells (Teff), including autoreactive T cells specific for organ-specific antigens. This suppression is essential because the loss of Treg function has been demonstrated to lead to catastrophic autoimmunity like that in patients with the IPEX syndrome. The suppression by these cells in vivo also appears to be antigen specific and raises the possibility that these cells could be harnessed to induce antigen-specific immune tolerance in the future.

GWA studies

Rapid advances in human genetics have afforded the opportunity to identify new risk alleles associated with common diseases, like type 1 diabetes and thyroiditis, that have previously been elusive. This has been due to a number of factors, including the completion of the human genome sequence, the development of a catalog of common genetic variation (i.e. the haplotype map), affordable technologies for high-density/high-throughput genotyping, and adequately powered sample sizes of cases and controls (32,33). In this regard, the most progress has been made with studies on type 1 diabetes and thyroiditis, in which adequately powered sample collections have been amassed to detect common variants using GWA and confirm previously established associations. Studies with type 1 diabetes samples have established a large number of genes associated with risk outside of the HLA region. Before the advent of GWA, the insulin (17,34), PTPN22 (35), CTLA4 (36), and interleukin-2 receptor α-chain (also known as CD25) (37) genes were established to be associated with disease, and have also been confirmed with GWA. With the advent of large GWA studies on type 1 diabetes, MDA5 (38), KIAA0350 (a C-type lectin of unknown function) (39,40,41), and several loci harboring other genes have been associated with disease (41). Although Hashimoto’s thyroiditis and Graves’ disease are distinct in their clinical presentations, they likely share many commonalities in their pathogenesis. Most large genetic studies on autoimmune thyroid disease have used large Graves’ collections, and there has been difficulty in detecting loci when Graves’ and Hashimoto’s patients are pooled together (42). In fact, a very recent study on Hashimoto’s thyroiditis patients has demonstrated different HLA class II associations when compared with Graves’ (43). To date, established genes outside of HLA for Graves’ include the TSH receptor (44,45), PTPN22 (46,47), CTLA4 (36), and FCRL3 (a Fc receptor family member) (45). Beyond these recent findings, it should be noted that there is an extensive body of literature examining candidate gene associations with thyroiditis, type 1 diabetes, and Addison’s disease. These reported associations may hold true associations but have yet to be replicated in these large collection studies for thyroiditis and type 1 diabetes. This may be due to many factors, but caution is warranted given the likely bias for reporting false-positive results in such studies, especially those that may be underpowered or may have unrecognized population stratification (48). The NALP1 gene, a likely regulator in the innate immune system, was also recently shown to have an association with multiple autoimmune diseases in families with vitiligo (49). In this study, families with two or more members with vitiligo and at least one with an autoimmune condition that included but was not limited to type 1 diabetes, Addison’s disease, and thyroiditis were collected, and convincing linkage was demonstrated to this gene.

In terms of the non-HLA genes outlined previously, the risk conferred by them, with few exceptions, is relatively small, with most having an odds ratio less than 1.5. In addition, the biological mechanisms by which these common alleles confer genetic risk still remain to be completely elucidated (Table 1). Despite this, when these findings are put into the context of what we know about autoimmunity and immune tolerance mechanisms, a picture is starting to emerge. First, there appears to be at least a set of genes that generally increase autoimmune disease risk, like PTPN22, CTLA4, NALP1, and FCRL3, which have established risk for many autoimmune diseases. For example, PTPN22 has been established as a risk gene for rheumatoid arthritis, systemic lupus erythematosus, juvenile rheumatoid arthritis, and myasthenia gravis, in addition to its established association with thyroiditis and type 1 diabetes. Second, some disease risk genes fit into context with established pathways related to immune tolerance. For example, CTLA4 (which is highly expressed in T cells) is known to play a critical role in dampening and suppressing T-cell responses in biological studies (50), and its association with multiple autoimmune diseases makes good sense. PTPN22 encodes a signaling phosphatase expressed in T cells that likely controls T-cell signaling, and the risk variant encodes an amino acid change that likely confers biological activity in T-cell activation pathways. Third, there are associations with emerging immune tolerance pathways. For instance, the association with CD25 may have a relationship with the function and activity of CD4+CD25+Tregs. The association of innate immune response genes like MDA5 and NALP1 may help explain the bridge between environmental triggers and activation of autoimmune responses. The association of the TSH receptor with Graves’ may also have a relationship with thymic expression of self-antigens, but making these links will need more study. Finally, there are some associations that are not completely clear, like KIAA0350, which may help identify unexpected pathways associated with disease.

Table 1.

Autoimmune endocrine disease susceptibility genes identified or confirmed in recent high-powered genetic studies (see text for references)

Gene	Associated autoimmune endocrine disease	Putative role of gene variant
HLA-DR,DQ (MHC class II)	T1D, GD, HT	Antigen presentation to CD4+ T cells
HLA-B (MHC class I)	T1D	Antigen presentation to CD8+ T cells
HLA-C (MHC class I)	GD	Antigen presentation to CD8+ T cells
Insulin	T1D	Thymic expression to promote negative selection
TSH receptor	GD	? Antigen recognition, ? thymic expression
CTLA4	T1D, GD, HT	Inhibitory T-cell signaling
PTPN22	T1D, GD, HT	? T-cell signaling
CD25	T1D	? Treg activity and function
MDA5	T1D	Innate immune response signaling
FCRL3	GD	Unknown
KIAA0350	T1D	Unknown

AD, Addison’s disease; GD, Graves’ disase; HT, Hashimoto’s thyroiditis; T1D, type 1 diabetes; ?, possible but not clearly established. 

Another general emerging set of findings with large case control collections has been a more thorough analysis of the HLA region with high-density marker genotyping. The HLA poses a particular challenge to geneticists because it is such a polymorphic and gene-rich region. This makes identifying true risk associations more difficult because the identified risk may be in linkage disequilibrium with the true risk variant. In type 1 diabetes, recent new data have emerged that have extended our growing knowledge of MHC class II alleles associated with disease risk and protection (51), and also in identifying additional disease risk (albeit lower) associated with MHC class I alleles (52). Additional studies have identified MHC haplotypes that provide extreme risk for the development of type 1 diabetes (6), which likely contain several synergistic loci. Likewise, a recent study on Graves’ patients has demonstrated disease risk attributable to MHC class I (53). Together, these findings reveal the rich complexity of the HLA region, and clearly a more detailed study of the region will be needed to unravel completely the risk associated with this locus.

Diagnostics

Autoantibodies

Autoantibodies are a key tool in the diagnosis of patients with autoimmune endocrine diseases and those at risk for these diseases. As outlined earlier, a major clinical phenotype of patients with the APS1 disorder is the presence of hypoparathyroidism, which is presumably autoimmune in origin, and a recent study has identified a parathyroid autoantigen called NACHT leucine-rich-repeat protein 5 (NALP5) (54). Interestingly, NALP5 is highly expressed in both the parathyroid and ovary, and autoreactivity to NALP5 may explain both the hypoparathyroidism and oophoritis associated with the APS1 disorder. However, it still remains to be determined if NALP5 is expressed in the thymus under the control of AIRE. A similar set of studies searching for pituitary autoantibodies has revealed tudor domain containing protein 6 as a pituitary autoantigen in APS1 subjects (55). The autoantigen is quite prevalent in APS1 subjects, but its direct correlation with pituitary autoimmunity in APS1 or in isolated lymphocytic hypophysitis remains to be established. Another set of recent studies has found that autoantibodies to type 1 interferons are generally predictive of the APS1 disorder (56,57,58). The clinical meaning of these autoantibodies currently remains unclear but may have some relationship to the candidiasis commonly observed in APS1 subjects. The specificity of this test for APS1 also appears to be on par with gene sequencing of AIRE in the initial studies, and raises the possibility that this assay may be of utility in patients and those at risk for the disorder. Recently, a new autoantigen has also been established for subjects with type 1 diabetes (59). ZnT8 is an islet-specific zinc transporter for which a large number of subjects with type 1 diabetes have reactive autoantibodies. The marker may prove particularly useful in subjects who test negative for other established autoantibodies to glutamate decarboxylase, insulin, and I-A2.

Animal Models

Animal models have proven to be invaluable in furthering our understanding of autoimmunity, given the inherit complexity of these diseases. Both the Aire knockout and FoxP3 knockout lines of mice have been valuable in unraveling the function of Aire and FoxP3 as outlined previously, but there have also been other recent advances with other animal models. A broad concept worth mentioning with animal models is segregating these models into those that have spontaneous development of autoimmune disease vs. those that are induced (i.e. immunizing with organ extract or antigen in the context of a strong adjuvant). Although induced models may be of some value, they are also hampered in identifying precipitating factors for disease because this is likely bypassed by the immunization process. Certainly, one of the most widely used spontaneous models in autoimmune endocrine disease research is the nonobese diabetic mouse strain model of autoimmune diabetes, which shows defects in multiple pathways of immune tolerance (60). This mouse strain has proven to be valuable in dissecting out the role of various immune cell populations and immune pathways in their contribution to the autoimmune diabetes process. In addition, it should also be noted that this strain has been shown to have an increased susceptibility to spontaneous autoimmune thyroiditis when its MHC locus is replaced in a congenic fashion (61) or when crossed to a dominant point mutation in Aire (21). A spontaneous thyroiditis model was also recently described using a T-cell receptor transgenic approach and emphasizes the importance again of T cells in driving this autoimmune disease (62). Another interesting development in animal models is the recent demonstration of genetic susceptibility loci in Portuguese water dogs for Addison’s disease (63). This dog breed shows a relatively high predisposition to acquired adrenal insufficiency with estimates around 1.5% of these dogs being affected [compared with approximately 0.01% in the human population (64)]. With recent advances in the genetic study of dogs and excellent pedigree records for this breed, Chase et al. (63) were able to demonstrate significant linkage for Addison’s disease to two loci in the dog genome. One locus was in the region of the dog MHC, and the second was in a genetic region rich for immune response-related genes, which includes CTLA4. Further work will be needed in this system to unravel the exact genes and polymorphisms responsible for the Addison’s disease risk, but this unique animal model may bring new mechanistic insights for this disorder. These recent findings also suggest that further work in nonrodent models of autoimmune endocrine conditions may be genetically tractable given the rapid advances in whole genome sequencing.

Environmental Effects

Recent epidemiological evidence has suggested that there is an increasing incidence of many autoimmune conditions, including type 1 diabetes (65,66). A prevailing hypothesis for the increase in these recent trends is the “hygiene hypothesis,” whereby the relative decrease in childhood infections from improved living conditions and increased immunizations may be a factor (67). Along these lines, Kondrashova et al. (68) recently examined the prevalence of thyroid autoimmunity in two geographically adjacent regions in Russia and Finland that share similar genetic ancestry. In this study an increased prevalence of antithyroid peroxidase and antithyroglobulin antibodies was observed in Finnish children over Russian children. The authors go on to suggest that the increased rate in Finland could be due to socioeconomic factors that include a lower rate of childhood infections.

Treatment

The mainstay of treatment for most autoimmune endocrine disorders is of course replacement therapy with the exception of Graves’ disease. To date, the main area for which some progress has been made in reversing or treating the underlying autoimmune process has been in type 1 diabetes, and this was recently reviewed in this series (69). One developing area of immunotherapy outside of type 1 diabetes worth mentioning involves the B-cell depleting agent rituximab (anti-CD20). This drug has been demonstrated to have efficacy in the treatment of several autoimmune diseases (70) with a relatively good side effect profile, and initial case reports suggested that it may have some efficacy in the treatment of Graves’ disease (71,72). Because a pathogenic autoantibody is responsible for this disorder, this is a rational treatment, however, it should be noted that plasma cells do not express CD20, and depletion of mature anti-TSH receptor antibody producing cells may be intractable to this approach. Recently, two controlled pilot studies for the treatment of Graves’ with anti-CD20 showed less encouraging results but some efficacy in patients with low anti-TSH receptor antibody levels (73,74). There has also been a case report of ulcerative colitis being associated with the treatment of a Graves’ patient in a similar trial (75) and brings into question the need for this therapy over established treatments. Despite this, rituximab may prove to be worthwhile in unique circumstances such as in the prevention of severe ophthalmopathy in those patients receiving thyroid ablation.

Conclusion

The endocrine system is commonly pathologically targeted by the immune system and can often lead to clinical disease through complete destruction of the organ. For years, our main genetic understanding of these disorders has been that the MHC genetic region encodes a significant degree of risk. Recent, rapid advances in genetics have shed new light on immune pathways and mechanisms that are involved in the pathogenesis of these diseases. These pathways include those revealed by monogenic autoimmune diseases, like APS1 and IPEX, which reveal the importance of thymic selection and Tregs in maintaining tolerance. In addition, rigorously powered genetic studies have reinforced the notion that T-cell response genes are involved in disease pathogenesis and that many autoimmune endocrine diseases share similar genetic risk. In addition, to our advancing knowledge in genetics, there have also been recent strides in identifying new diagnostic markers and new treatments for these diseases. Despite these advances, much work remains to be done, including addressing the fundamental question of why the endocrine system is so commonly targeted by autoimmune responses.