Central tolerance to self revealed by the autoimmune regulator

Abstract

The autoimmune regulator (Aire) was initially identified as the gene causing multiorgan system autoimmunity in humans, and deletion of this gene in mice also resulted in organ-specific autoimmunity. Aire regulates the expression of tissue-specific antigens (TSAs) in medullary thymic epithelial cells (mTECs), which play a critical role in the negative selection of autoreactive T cells and the generation of regulatory T cells. More recently, the role of Aire in the development of mTECs have helped elucidate its ability to present the spectrum of TSAs needed to prevent autoimmunity. Molecular characterization of the functional domains of Aire have revealed multiple binding partners that assist Aire’s function in altering gene transcription and chromatin remodeling. These recent advances have further highlighted the importance of Aire in central tolerance.

Introduction

The immune system has evolved a complex set of responses to fight off unwanted infections through the generation of a T cell repertoire that can recognize a broad array of antigens. The unfortunate consequence of such diversity is the potential for the generation of an autoreactive T cell that recognizes a self-antigen. To combat this problem, the immune system has evolved elegant tolerance mechanisms to prevent their generation and to keep them in check when present.

Over the last decade, there have been rapid advances in our understanding of critical pathways that regulate immune tolerance, and many of these recent insights have relied, in part, on genetics and rare Mendelian forms of autoimmunity. Study of immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome, a rare monogenic disorder due to Foxp3 mutations, helped establish the critical role of T regulatory (T_reg) cells in preventing organ-specific autoimmunity.¹ Defects in innate sensing pathways also play a critical role in tolerance, as shown by studies of patients with Aicardi-Goutières syndrome, a disease characterized by neurological dysfunction and lupus-like features.² Among these findings has been the discovery of the function of the autoimmune regulator (Aire) gene. It has become clear that Aire plays a key role in establishing T cell tolerance to self, particularly in the thymus. Here, we review some of the recent progress in elucidating the function of Aire and its effects on the immune system.

Autoimmune polyglandular syndrome type 1: a Mendelian organ-specific autoimmune syndrome

The role of Aire in immune tolerance was first made evident with its identification as the causative gene for mediating multiorgan system autoimmunity, termed autoimmune polyglandular syndrome type 1 (APS-1) or autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED).^3,4 It is primarily considered an autosomal recessive disorder that occurs with variable frequency, depending on ethnic background. Studies have now identified autosomal dominant mutations that cluster in plant homeodomain 1 (PHD1) and one unique mutation (G228W) in the SAND domain.^5–7 These patients typically develop the classic constellation of autoimmune adrenocortical failure (i.e., Addison’s disease), autoimmune hypoparathyroidism, and chronic mucocutaneous candidiasis, as well as other autoimmune manifestations.^8–11 The disease usually starts in childhood, evolving over time and having heterogeneity in its organ involvement. These patients develop a spectrum of autoantibodies to organ tissue mediating immune destruction.¹² More recently, it was found that patients develop autoantibodies to cytokines, such as type I interferons (IFN-ω and IFN-α) and T helper (T_H)17-related cytokines (IL-22, IL-17F, and IL-17A).^13–15 Almost all patients with Aire deficiency develop type I interferon autoantibody, which serves as an excellent biomarker for this disease. Interestingly, the expected clinical consequences of increased viral susceptibility due to the presence of these autoantibodies do not occur. On the other hand, autoantibodies to the T_H17-related cytokines do correlate with candidiasis, supporting the role of the T_H17 pathway in providing immunity to candidal infection.^14,15

It is now appreciated that, in addition to genetic defects, acquired defects in Aire activity may occur. For example, reduced Aire expression can be present in a type of thymic tumor known as a thymoma, whose presence is often associated with autoimmunity.¹⁶ Some of these patients interestingly develop myasthenia gravis with autoantibodies such as anti-acetylcholine receptors. They rarely show the symptoms that are seen in APS-1 patients.¹⁷ These two disease processes stemming from a deficiency in Aire both result in autoimmunity, suggesting a clear role of this gene in breaking immune tolerance. To better understand the underlying mechanism, four research groups have generated Aire-deficient mouse models that shared features similar to human APS-1 patients.^18–21,22 These mice develop lymphocytic infiltrates similar to those in humans, but in different organs such as the salivary gland, eye, reproductive organs, and liver.^19,23 They also develop tissue-specific autoantibodies, as do APS-1 patients.^19,23 Taken together, Aire deficiency, whether in humans or mice, leads to autoimmunity, clearly highlighting the importance of this gene in immune tolerance.

Immunological functions of Aire

We have now come to appreciate that Aire plays a key role in thymic tolerance (Fig. 1). Aire is primarily expressed in the thymic medulla, suggesting a possible role in T cell tolerance, and was shown to express various tissue-specific genes within the thymus in cells known as medullary thymic epithelial cells (mTECs). Aire knockout mice show reduced expression of a subset of tissue-specific genes, which correlate with the specific organ involvement and associated autoantibodies. Taken together, the functions of Aire clearly affect thymic tolerance.

Figure 1.

The role of Aire in central tolerance. Central tolerance primarily occurs in the medullary compartment of the thymus where T cells develop (T cell progenitor to double-negative (DN) T cells to double-positive (DP) T cells and single-positive (SP) T cells). Aire is expressed in medullary thymic epithelial cells (mTECs) derived from TEC progenitors located in the corticomedullary junction of the thymus. These Aire⁺ mTECs express tissue-specific antigens (TSAs) and can hand-off these antigens to thymic dendritic cells (DCs). Post-Aire cells also express TSAs but at a lower level and are found near Hassall’s corpuscles. Expression of TSAs in the medulla affect T_reg cell generation and negative selection of T cells, which are critical for maintaining central tolerance.

The role of Aire in negative selection

The role of Aire in negative selection (i.e., deletion of autoreactive cells) was first illustrated using a transgenic mouse system bearing a neoantigen and a specific T cell receptor recognizing the neoantigen. The absence of Aire allows these autoreactive T cells to escape into the periphery where, if Aire was present, these autoreactive cells would be deleted,^23,24 which is also seen if there is a TCR transgene specific for self-antigens.^25,26 In addition, endogenous T cells directed to a specific self-antigen can be detected in the periphery in the absence of Aire, and their presence is correlated with disease.²⁶ Deletion is primarily mediated by Aire expression in mTECs, and recent data suggest that thymic B cells may also express low levels of Aire and participate in T cell selection.²⁷

Aire in T_reg cell selection

Initial data did not strongly support a role for Aire in T_reg cell generation. T_reg cell numbers and frequency are generally similar between Aire-deficient and wild-type mice, and the transfer of T_reg cells did not prevent or rescue autoimmunity in Aire-deficient mice.^23,28 However, patients with AIRE mutations show decreased T_reg cell numbers, and in vitro suppression assays revealed defects suggesting a role of Aire in T_reg cells.^29,30

To further explore the effects of Aire on T_reg cell biology, a transgenic system expressing hemagglutinin (HA) on mTECs using the Aire promoter provided initial insight that T_reg cells can be selected by specific Aire-driven antigen on mTECs.³¹ These data were further strengthened in the context of an endogenous antigen that was Aire dependent; in this model, antigen-specific T_reg cells failed to develop in the absence of Aire, clearly illustrating the role of Aire in the selection of specific T_reg cells.³²

Given how Aire affects specific T_reg cell selection, one could speculate that Aire would likely affect the T_reg TCR repertoire. To evaluate this possibility, the T_reg TCR repertoire was examined in mice with fixed TCRb with limited TCRa arrangements. It was found that the T_reg TCR repertoire was similar between Aire-sufficient and Aire-deficient mice;³³ however, this limited TCR repertoire may have excluded specific TCRs that are Aire dependent. With new and improved sequencing methodology, a broader and more natural TCR repertoire analysis was performed on mice with a fixed TCRβ chain with the endogenous Tcra locus. In this model, TCR repertoire differences were revealed, and Aire specifically affected those T_reg TCRs that were lower in frequency, which may explain why such differences were not initially seen in a limited TCR repertoire model analysis.³⁴

In addition to its effects on T_reg cell selection, Aire may also play a role in the generation of unique T_reg cell subsets that are generated during the perinatal period.³⁵ These perinatal T_reg cells have been shown to be better suppressors in vitro and may operate more efficiently in vivo in the adoptive transfer setting. How Aire influences this unique population of T_reg cells remains open to question, as the exact antigen specificities for potential Aire-dependent T_reg cells in this setting have not been determined. Previous work by two independent studies have shown that grafting of Aire wild-type and knockout thymic lobes into a single mouse was not sufficient to prevent autoimmunity.^20,23 If T_reg cell selection was the primary problem in the Aire-deficient model, the result of these latter experiments would be predicted to be protection against autoimmunity. Taken together, it does appear that Aire can influence the T_reg cell repertoire; however, the degree to which this contributes to autoimmunity in the Aire-deficient setting remains open to debate.

The role of Aire in peripheral tolerance

Aire also potentially plays a role in peripheral tolerance. In humans, Aire has been found in lymph nodes at both the transcript and protein levels,^4,36 and in mice, the detection of peripheral Aire had been more challenging until the generation of an Aire-reporter mouse line. Aire expression was then found to be present in secondary lymphoid organs, such as peripheral lymph nodes and spleen.³⁷ These Aire-expressing cells are bone marrow derived and are similar to dendritic cells (DCs) in cell morphology, expression of Zbtb46, and presence of certain markers such as CD45 and CD11c, although at a lower level than DCs. Expression profiling of these cells revealed a distinct pattern that is different from both plasmacytoid and conventional DCs.^37,38 These characteristics suggest a potentially new cell population that has been termed extrathymic Aire-expressing cells (eTACs).

Functionally, eTACs may play a role in maintaining peripheral T cell tolerance. When eTACs were engineered to express a specific antigen, the antigen-specific CD8⁺ T cells were deleted.³⁷ In contrast, antigen-specific CD4⁺ T cells were rendered anergic. These antigen-specific CD4⁺ T cells had elevated PD-1 levels and other markers associated with tolerance, and they were unable to produce IFN-γ. TCR signaling in these CD4⁺ T cells are impaired in the ability to increase calcium and activate the Ras–MAPK pathway by phosphorylating ERK. The eTACs were found to be unresponsive to inflammatory signals that thus did not upregulate costimulatory markers, such as CD80, supporting their ability to lead to T cell anergy.³⁸ Altogether, the data suggest eTACs as a potential new cell population that may play a role in peripheral T cell tolerance, complementing the important tolerogenic role of Aire in the thymus.

In addition to eTACs, two other cell types have also recently been observed to express Aire. Through the use of an Aire-GFP BAC transgenic reporter, thymic B cells can, to some extent, acquire Aire expression.²⁷ Interestingly, this expression appears to occur through a CD40 signaling pathway and requires cognate help interactions with developing thymocytes. Additionally, these thymic B cells can contribute to thymic selection in transgenic models. It remains unclear how Aire is functioning in these cells, given that these Aire-expressing thymic B cells were not enriched for tissue-specific antigens (TSAs). In addition to thymic B cells, keratinocytes were also shown to express Aire under an inflammatory condition in an oncogenic-driven tumor model.³⁹ In this setting, Aire itself may be a contributor to the transformation process, but precisely how this happens and how widespread this process is in other tumors remains to be determined.

New insights into mTEC development and turnover

Aire plays several roles in the immune system to prevent the development of autoimmunity. On a cellular level, it functions to upregulate expression of TSA, and it plays a role in the differentiation of a specialized subset of the thymus known as mTECs (Fig. 1).There is also some speculation that TSA expression may be linked to maturation of these cells.⁴⁰

Aire controls TSA expression

It is well established that mTECs express a range of TSAs that play a role in negative selection and T_reg cell generation. Previously, it was estimated that roughly 3000 genes were expressed in these cell types and play a role in the selection of T cells.⁴¹ However, this covers only a fraction of all known genes, suggesting incomplete self-antigen display for T cell selection, resulting in potential escape of autoreactive T cells. Recently, through the use of singlecell sequencing, a broader expression profile was identified, revealing that roughly 90% of all known genes are indeed expressed in TECs, including Aire⁺ mTECs.⁴² Many genes were previously masked because of low levels of expression in the bulk population.⁴² It was also recently by a comparison of expression profiles between cells that there is coordinated expression of gene clusters, which is speculated to be attributed to chromatin remodeling. While there is this coordinated expression of gene clusters on a single-cell level, the establishment of the gene cluster appears to be stochastic since there is variation on an individual level.^43,44 Furthermore, the TECs can also hand-off these antigens to other thymic antigen presenting cells, such as DCs.^34,45 Collectively, these cells allow for a wide breath of antigen presentation, resulting in improved T cell selection and self-tolerance.

Aire in medullary thymic epithelial cells

The role of Aire in mTEC development has been illustrated using an in vivo ablation system. Using an Aire-containing BAC transgenic model, diphtheria toxin receptor (DTR) was expressed on Aire⁺ mTECs to mark them for ablation, and deletion was controlled temporally with diphtheria toxin (DT) administration. Loss of these cells did result in decreased negative selection and T_reg cell production as anticipated. Surprisingly, these ablated cells were able to recover after 5 days.⁴⁶ Bromodeoxyuridine (BrdU) labeling experiments as well as fate mapping studies revealed that these Aire⁺ cells have a half-life of 7–14 days.^46–48 Taken together, the data suggest a progenitor cell population that replenishes the Aire⁺ mTEC population.

Recent studies have identified this progenitor cell population using different methods (Fig. 1).^49–51 One research group identified this progenitor cell line by identifying a rare cell population derived from adult mouse thymic stromal cells that formed thymospheres, a property of progenitor cells. They were Sca1 positive, and negative for CD24, CD45, and EpCam.⁴⁹ The other group identified this progenitor population by subsetting adult thymic stromal cells using Sca1 and alpha-6 integrin, both of which are markers found in other epithelial progenitor cells⁵⁰ and are localized to the corticomedullary junction. Both research groups found that these cells had low mitotic activity, self-renewal capacity, and an ability to give rise to cTEC and mTEC lineages, including Aire⁺ cells. These cells were also found in FoxN1-deficient mice, an athymic mouse line, suggesting that they are not FoxN1 dependent.^49,50 Furthermore, it was demonstrated that these progenitor cells, specifically Cld3,4^hi, could prevent the development of autoimmunity in the aly/aly mouse model.⁵² The identification of this progenitor cell identifies the cellular source for replenishing the mTECs that undergo rapid turnover, which helps ensure a near complete presentation of the self-antigen repertoire of mTECs, such that autoimmunity can be prevented through proper negative selection and T_reg cell generation.

Preventing autoimmunity can also negatively affect the ability of the immune system to fend off tumors since the effector T cells against these tumor antigens may have been deleted or suppressed. Thus, temporary blockade of mTECs could provide a small window of opportunity for tumor-specific T cells to escape tolerance mechanisms. Studies have previously shown that RANK–RANKL and CD40–CD40L play a role in the development of Aire⁺ mTECs.^53–56 RANK signaling blockade in vivo is able to deplete Aire⁺ mTECs,^46,57 leading to alterations in negative selection by preventing thymic deletion and to loss of T_reg cell development in both a restricted and polyclonal T cell repertorie.⁵⁷ Anti-RANKL treatment in a B16 melanoma model allowed temporary escape of these self-antigen–specific T cells to target the tumor and improve survival.⁵⁷ Modulation of the mTEC compartment could potentially be part of the immunotherapy regimen for treating cancers.

Post-Aire cells

Initial data suggested that Aire expression reflects the terminal mTEC state;⁴⁷ however, new data have delineated a post-Aire state, using various fate-mapping mouse models.^{35,46,48,58–60} The post-Aire cell population is regulated by the lymphotoxin signaling pathway,⁵⁸ whereas RANK and CD40L signaling may not play key roles as seen in the development of Aire⁺ mTECs.⁶⁰ Furthermore, these post-Aire cells express intermediate levels of TSAs;⁴⁶ are positive for involucin, a terminal maturation marker for keratinocytes; and have been associated with Hassall’s corpuscles structures found in human thymus.⁶¹ Hassall’s corpuscles have been suggested to play a role in T_reg cell induction on the basis of identification of TSLP⁺ DCs in these structures and their generation of human T_reg cells in an in vitro system. ⁶² Thus, it is tantalizing to speculate that these cells could be mediating this role in vivo.

Molecular mechanisms of Aire

Aire was initially hypothesized to behave as a transcription factor that turned on TSAs to facilitate negative selection of autoreactive thymocytes. However, several aspects of Aire and its gene expression pattern was not conducive to features of classical transcription factors. First, Aire regulates the expression of thousands of TSAs and thus it is unlikely that there is a consensus DNA sequence that Aire binds.⁶³ Second, the transcriptional start sites were different from the endogenous start site, and Aire uses different transcriptional machinery.⁶⁴ Third, the pattern of TSA expression varied between cell types.^37,65–67 Taken together, the data show that Aire functions differently from classical transcription factors, given the diverse expression pattern.

Structural and functional domains of Aire

Insight into the molecular function of Aire arose from characterization of its structure and functional domains. Aire contains a caspase-recruitment domain (CARD), a SAND domain, two PHD fingers, and LXXLL motifs (Fig. 2). The CARD domain participates in oligomerization of Aire, and the PHD1 domain in Aire binds unmethylated histone H3 at lysine-4 (H3K4m0), which is associated with repressed areas of transcription.^68–71 Recognition of this histone mark directs Aire to help in the expression of Aire-dependent TSAs, but it is not the primary mechanism for directing Aire to these target genes.⁷² The SAND domain binds other proteins, rather than DNA. The LXXLL motifs are thought to recruit Aire-binding partners. Taken together, the properties of Aire transcription and its link to epigenetics through the H3K4Me0 recognition motif suggest that Aire may use a fascinating array of mechanisms to target and promote TSA expression. Gaining insight on these pathways has been challenging because of limited cell numbers in vivo, but inroads have been made using transfection models.

Figure 2.

Aire protein domains. Aire contains multiple functional domains, including caspase-recruitment domain (CARD), SAND (Sp100, AIRE-1, NucP41/75, and DEAF-1) domain, two plant homeodomain (PHD) fingers, proline rich region (PRR), and four LXXLL motifs. These domains bind different Aire-binding partners and epigenetic marks that result in Aire’s unique functions.

Aire in transcription

Several Aire-binding partners play a role in directing transcription by elongation and potentially altering splicing (Fig. 3). The first Aire-binding protein identified was cyclic AMP response element–binding protein (CBP),⁷³ which is a common transcriptional coactivator that colocalizes with Aire in the nucleus and together can drive expression of reporters and endogenous genes.⁷⁴ The common Finnish mutation, R257X, disrupts the binding of Aire to CBP. CBP was shown to acetylate Aire in the SAND domain, which can alter the expression pattern for Aire-dependent TSAs.⁷⁵

Figure 3.

Aire-binding partners. Aire binds to multiple protein partners, enabling an influence on both transcription and chromatin remodeling.

DNA-dependent protein kinase (DNA–PK) was identified in two different screens searching for Aire-binding partners.^76,77 It has been speculated that the associated DNA–PK complex induces double-stranded DNA breaks and leads to removal of nucleosomes, allowing for expression of repressed genes.⁷⁸ Recently, a patient with mutations in DNA–PK developed autoimmune symptoms; studies suggested an inability to express Aire-dependent TSAs, whereas Aire-independent TSAs were unaffected.⁷⁹

Positive transcription elongation factor b (P-TEFb) is another protein that interacts with Aire and is positioned in areas of stalled transcription.⁸⁰ P-TEFb phosphorylates RNA polymerase II, DRB sensitivity-inducing factor (DSIF), and negative elongation factor (NELF). Phosphorylation of these proteins leads to release of stalled RNA polymerase II, allowing for expression of TSA transcripts.⁸¹ Patients with a mutation causing a deletion of the C-terminus leads to an inability to bind P-TEFb and a block in TSA expression.⁸² RNAi screens identified additional factors that further support the role of P-TEFb in mediating Aire-directed TSA expression.⁸³

Aire can also interact with proteins involved in splicing⁷⁷ and can lead to increased splicing of mRNA.⁸² More recently, a reanalysis of microarray and RNA-Seq data from murine mTECs showed that the presence of Aire was associated with increased alternative spliced isoforms in TSAs, suggesting that the actions of Aire lead to increased exposure to alternative TSA isoforms.⁸⁴

Recently, another protein, sirtuin-1 (SIRT1), was shown to affect TSA gene expression in Aire⁺ mTECs, and its deficiency led to organ-specific autoimmunity.⁸⁵ Sirt1 is highly expressed in mTECs and in vitro assays suggest association with Aire. It deacetylates proteins such as Aire (acetylated by CBP), which is speculated to affect TSA expression in an Aire-dependent manner.⁸⁵

Aire in chromatin remodeling

Given the ability of Aire to express a wide spectrum of TSAs, it has been speculated that Aire participates in chromatin remodeling, which opens up areas of repressed transcription, allowing transcription of a broad array of genes (Fig. 3). The genes regulated by Aire are clustered but do not have any specific chromosomal preferences, and the pattern is stochastic, suggesting a role for Aire in epigenetic mechanisms.^64,86 Also, as mentioned above, the PHD1 domain recognizes histone mark, which lends further support for a role in chromatin remodeling.

One protein partner suggesting a role in this process is DAXX, which was identified from a yeast two-hybrid system using the HSR/CARD domain as bait. The W78R mutation, a known mutation present in APS-1 patients, disrupts binding of this protein to Aire. DAXX is known to bind histones and recruit histone deacetylases and also binds and recruits other proteins (Pax3, Pax5, p53, glucocorticoid receptor, and HDAC). Together, the actions of DAXX and its protein complex implicates Aire in epigenetics and can alter its accessibility to other transcriptional targets.⁸⁷

Recently, our research group identified two new Aire-binding partners that provide additional evidence for Aire in chromatin remodeling. The MBD1–ATF7ip complex targets Aire to repressed areas of transcription with methylated CpG dinucleotides. In particular, the SAND domain, which was previously thought to be a potential DNA-binding domain but lacks the canonical sequence, interacts with ATF7ip. The dominant disease-causing G228W mutation is located in the SAND domain and disrupts the interaction. Findings on this mechanism help elucidate how Aire can drive TSA expression in different cell types due to different methylation patterns based on the cellular environment.⁸⁸ Taken together, it now appears that Aire likely exploits several different types of binding partners and methods to activate transcription of TSAs, and there may be other similar properties that remain to be identified.

Conclusions

Advancing our understanding of the function of Aire has provided insight into self-tolerance mechanisms and the development of autoimmunity, as well as the development cancer. More research is needed to further elucidate the molecular mechanisms by which Aire promotes TSA expression. It also remains to be determined if there are other factors in the thymus that may carry out similar activities to those of Aire, as there remains significant TSA expression in the thymus of Aire-deficient animals. Finally, recent work has demonstrated the potential for selective manipulation of thymic Aire-expressing cells, which could open the door to developing a potential therapy for diseases such as malignancy.