Autoimmune regulator (AIRE) gene mutation is responsible for the development of organ-specific autoimmune disease with monogenic autosomal recessive inheritance. Although Aire has been considered to regulate the elimination of autoreactive T cells through transcriptional control of tissue-specific Ags in thymic epithelial cells, other mechanisms of AIRE-dependent tolerance remain to be investigated. We have established Aire-deficient mice and examined the mechanisms underlying the breakdown of self-tolerance. The production and/or function of immunoregulatory T cells were retained in the Aire-deficient mice. The mice developed Sjögren’s syndrome-like pathologic changes in the exocrine organs, and this was associated with autoimmunity against a ubiquitous protein, α-fodrin. Remarkably, transcriptional expression of α-fodrin was retained in the Aire-deficient thymus. These results suggest that Aire regulates the survival of autoreactive T cells beyond transcriptional control of self-protein expression in the thymus, at least against this ubiquitous protein. Rather, Aire may regulate the processing and/or presentation of self-proteins so that the maturing T cells can recognize the self-Ags in a form capable of efficiently triggering autoreactive T cells. With the use of inbred Aire-deficient mouse strains, we also demonstrate the presence of some additional factor(s) that determine the target-organ specificity of the autoimmune disease caused by Aire deficiency.
Autoimmune diseases are mediated by sustained adaptive immune responses specific for self-Ags through unknown mechanisms. Although breakdown of self-tolerance is considered to be the key event in the disease process, the mechanisms that allow the production of auto-Abs and/or autoreactive lymphocytes are largely enigmatic (1). The situation seems to have become more complicated due to the existence of multiple factors that influence the disease process, such as environmental factors, immune dysregulation, and genetic predisposition. In this regard, although only a small number of genes genetically relevant to the pathogenetic processes for the development of autoimmune diseases have been found so far (2), genetic engineering of such genes in mice should enable us to establish disease models and facilitate an understanding of the disease mechanisms to a large extent. One of these genes is the autoimmune regulator (AIRE)3 mutation, which is responsible for the development of autoimmune-polyendocrinopathy-candidiasis ectodermal dystrophy (APECED; Online Mendelian Inheritance in Man 240300) with autosomal recessive inheritance (3, 4, 5, 6).
The AIRE gene encodes a predicted 58-kDa protein carrying a conserved nuclear localization signal, two plant homeodomain (PHD)-type zinc fingers, four LXXLL motifs or nuclear receptor interaction domains, and the recently described homogeneously staining region (HSR) and SAND domains (3, 4); the HSR and SAND domains have been suggested to function in homodimerization and DNA binding, respectively (7, 8). Based on the fact that PHD resembles the RING finger, which can function as an E3 ubiquitin ligase, in both sequence and structure (9), we have recently found that AIRE acts as an E3 ubiquitin ligase through the N-terminal PHD domain (PHD1) (10). Because the ubiquitin-proteasome pathway plays an essential role in diverse cell functions such as cell cycle progression, signal transduction, cell differentiation, DNA repair and apoptosis (11, 12), we speculate that AIRE should play a fundamental role by facilitating polyubiquitinylation of the substrate(s) in yet undetermined processes. The significance of this finding was underscored by the fact that disease-causing missense mutations in PHD1 abolished its E3 ligase activity (10).
One important aspect of AIRE, in the context of autoimmunity, is its limited tissue expression in medullary thymic epithelial cells (mTEC) and cells of the monocyte-dendritic cell lineage of the thymus (13, 14). Both cell types are considered to play major roles in the establishment of self-tolerance by eliminating autoreactive T cells (negative selection) (1, 15) and/or by producing immunoregulatory T cells (Tregs), which prevent CD4+ T cell-mediated organ-specific autoimmune diseases (16, 17). For this purpose, thymic epithelial cells (TECs) have been postulated to express a set of self-Ags encompassing all of the self-Ags expressed by parenchymal organs. Supporting this hypothesis, analysis of gene expression in the thymic stroma has demonstrated that mTECs are a specialized cell type in which promiscuous expression of a broad range of peripheral tissue-specific genes is an autonomous property (18). Aire in TECs has been suggested to regulate this promiscuous gene expression (19).
Fundamental roles of Aire in the elimination of autoreactive T cells in vivo have been demonstrated by the use of a TCR-transgenic mouse system (20). Mice expressing hen egg lysozyme (HEL) in pancreatic β cells driven by the rat insulin promoter were crossed with mice expressing TCR specific for HEL, and the fate of HEL-specific T cells was monitored in either the presence or absence of Aire. Remarkably, Aire-deficient TCR-transgenic mice showed almost complete failure to delete the autoreactive (i.e., HEL specific) T cells in the thymus (20). Because Aire-deficient mTEC showed a reduction in transcription of a group of genes encoding peripheral Ags analyzed by the gene-chip technique (19), it has been hypothesized that pathogenic autoreactive T cells could not be eliminated efficiently due to the reduced expression of corresponding target Ags in the Aire-deficient thymus (20). However, as this transgenic study did not demonstrate the effect of Aire loss on the thymic expression of HEL, there is still a lack of experimental evidence to connect the postulated roles of Aire in the transcriptional regulation of tissue-specific Ag expression with efficient elimination of autoreactive T cells. Thus, beyond transcriptional control of self-Ags in the thymus, other mechanisms of AIRE-dependent tolerance remain to be investigated. Furthermore, the effect of Aire deficiency on the production and/or function of Tregs has not yet been fully documented (19, 20, 21). Finally, the factors contributing to the complexity of the APECED phenotype (i.e., involvement of various target organs among patients) are unknown. Although intrafamilial variation in the clinical pictures suggests that factors other than the specific AIRE mutations might be involved in the disease process (22), this hypothesis cannot be easily proven in human subjects. To approach these issues, we have generated Aire-deficient mice by gene targeting. Identification of a target Ag associated with the tissue destruction caused by Aire deficiency together with strain-dependent target-organ specificity of the autoimmune disease has suggested unique properties of AIRE in the establishment and maintenance of self-tolerance.
Materials and Methods
Aire-deficient mice were generated by gene targeting. Briefly, the targeting vector was constructed by replacing the genomic Aire locus starting from exon 5 to exon 12 with the neomycin resistance gene (neor). The targeting vector was introduced into TT2 embryonic stem cells (H-2b/k) (23), and the homologous recombinant clones were first identified by PCR and confirmed by Southern blot analysis. After the targeted cells had been injected into ICR 8 cell embryos (CLEA Japan), the resulting chimeric male mice were mated with C57BL/6 females to establish the germline transmission. C57BL/6 mice, BALB/c mice, and BALB/cA Jcl-ν mice were purchased from CLEA Japan. The mice were maintained under pathogen-free conditions and handled in accordance with the Guidelines for Animal Experimentation of Tokushima University School of Medicine. The experiments were initiated when the mice were 8–12 wk of age.
Formalin-fixed tissue sections were subjected to H&E staining, and two pathologists independently evaluated the histology without being informed of the condition of each individual mouse. Histological changes were scored as 0 (no change), 1 (mild lymphoid cell infiltration), or 2 (marked lymphoid cell infiltration).
Measurement of tear secretion
Measurement of tear secretion was performed as previously described (24, 25). Briefly, anesthetized mice were injected i.p. with 100 μl of pilocarpine hydrochloride (1 mg/ml) to stimulate tear production. Secreted tears were absorbed every 5 min with a cotton thread treated with a pH indicator phenol red (ZONE-QUICK; Menicon), and the length of the red portion of the thread was measured each time. Total length of the red portion of the thread during the first 20 min after pilocarpine injection was normalized by body weight.
ELISA and Western blot analysis
Various forms of recombinant α-fodrin were expressed with pGEX-4Ts plasmids (26). Western blot analysis and ELISA for the detection of auto-Abs against various forms of recombinant α-fodrin were performed with anti-mouse IgG Ab (Vector Laboratories), as described previously (25, 27, 28, 29, 30, 31). For the ELISA, absorbance values greater than the mean ± 3 SD in wild-type sera were considered positive. Western blot analysis of α-fodrin expression from the proteins extracted from the thymus and lacrimal glands was performed with mouse anti-α-fodrin mAb (Affiniti) and rabbit anti-AFN-A polyclonal Ab (25, 27, 28, 29, 30, 31).
Autoreactive responses against α-fodrin
For in vitro stimulation with α-fodrin, total splenocytes were stimulated with 10 μg/ml recombinant α-fodrin. For the last 8 h of the 32-h culture period, the cells were pulsed with [3H]thymidine, and 3H incorporation was determined as described previously (25).
Thymic stroma preparation
Thymic stroma was prepared as described previously with slight modification (32). Briefly, thymic lobes were isolated from three mice for each group and cut into small pieces. The fragments were gently rotated in RPMI 1640 medium (Invitrogen) supplemented with 10% heat-inactivated FCS (Invitrogen), 20 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μM 2-ME, hereafter referred to as R10, at 4°C for 30 min, and dispersed further with pipetting to remove the majority of thymocytes. The resulting thymic fragments were digested with 0.15 mg/ml collagenase IV (Sigma-Aldrich) and 10 U/ml DNase I (Roche Molecular Biochemicals) in RPMI 1640 at 37°C for 15 min. The supernatants that contained dissociated TECs were saved, whereas the remaining thymic fragments were further digested with collagenase IV and DNase I. This step was repeated twice, and the remaining thymic fragments were digested with collagenase IV, DNase I, and 0.1 mg/ml dispase I (Roche Applied Science) at 37°C for 30 min. The supernatants from this digest were combined with the supernatants from the collagenase digests, and the mixture was centrifuged for 5 min at 450 × g. The cells were suspended in PBS containing 5 mM EDTA and 0.5% FCS and kept on ice for 10 min. CD45− thymic stromal cells were then purified by depleting CD45+ cells with MACS CD45 microbeads (Miltenyi Biotec) according to the manufacturer’s instructions. The resulting preparations contained ∼60% Ep-CAM+ cells and <10% thymocytes (i.e., CD4/CD8 single-positive and CD4/CD8 double-positive cells), as determined by flow cytometric analysis.
RNA was extracted from thymic stromal cells with High Pure RNA isolation kit (Roche Applied Science) and made into cDNA with cDNA Cycle kit (Invitrogen) according to the manufacturer’s instructions. The following primer pairs for the α-fodrin gene were used: 5′-GCTTCAAGGAGCTCTCTACC-3′ and 5′-GCAGTTTGATTCCTTTCTCC-3′ (encompassing α-fodrin exons 1–3; accession no. XM_355324), 5′-CCAGCAGCAACAATTTAATC-3′ and 5′-AGCAGATTCTGGACTCCAAT-3′ (encompassing the α2-spectrin exons 2–4; accession no. XM_207079), and 5′-GTGCAGAAATCAGCTGAGAA-3′ and 5′-GCTTGTGTTTCTTCCTCAGA-3′ (encompassing the α2-spectrin exons 24–27). PCR was conducted in a final volume of 20 μl with 1.5 U of ExTaq DNA polymerase (Takara Biomedicals) and 250 nM each primer. Cycling conditions comprised a single denaturing step at 94°C for 10 min followed by 35 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 1.5 min, followed by a final extension step of 72°C for 10 min. For β-actin, a single denaturing step at 94°C for 3 min was followed by 25 cycles of 94°C for 45 s, 50°C for 45 s, and 72°C for 1 min, followed by a final extension step of 72°C for 3 min (33).
Real-time PCR for quantification of α-fodrin, Foxn1, and tissue-specific Ag genes was conducted with thymic stroma cDNA prepared as described above. The primers and the probes are as follows. α2-spectrin primers: 5′-GACAGCCAGTGATGAGTCATACAAG-3′ and 5′-CACGGATTCGGTCAGCATT-3′; α2-spectrin probe: 5′-FAM-ACCCCACCAACATCCAGAGCAAGC-3′; Foxn1 primers: 5′-GACATGCACCTCAGCACTCTCTA-3′ and 5′-CTGATGTTGGGCATAGCTCAAG-3′; Foxn1 probe: 5′-FAM-CCCGGCTCAAAGCCATTGGCTC-3′; insulin primers: 5′-AGACCATCAGCAAGCAGGTC-3′ and 5′-CTGGTGCAGCACTGATCCAC-3′; insulin probe: 5′-FAM-CCCGGCAGAAGCGTGGCATT-3′; salivary protein 1 primers: 5′-ACTCCTTGTGTTGCTTGGTGTTT-3′ and 5′-TCGACTGAATCAGAGGAATCAACT-3′; salivary protein 1 probe: 5′-FAM-TTCACCAGCAGAATCAGCAGTTCCAGAA; C-reactive protein primers: 5′-TACTCTGGTGCCTTCTGATCATGA-3′ and 5′-GGCTTCTTTGACTCTGCTTCCA-3′; C-reactive protein probe: 5′-FAM-CAGCTTCTCTCGGACTTTTGGTCATGA-3′; fatty acid binding protein primers: 5′-CGTGTAGACAATGGAAAGGAGCT-3′ and 5′-AAGAATCGCTTGGCCTCAACT-3′; fatty acid binding protein probe: 5′-FAM-TCATTACCAGAAACCTCTCGGACAGCA-3′; glutamic acid decarboxylase 67 primers: 5′-TCCTCCAAGAACCTGCTTTCC-3′ and 5′-GCTCCTCCCCGTTCTTAGCT-3′; glutamic acid decarboxylase 67 probe: 5′-FAM-CCGACTTCTCCAACCTGTTTGCTCAAGA-3′. Foxp3 expression was examined with cDNAs prepared from splenocytes (CD4+CD25+ or CD4+CD25−) and total thymus. The primers, the probes, and the reactions used for Foxp3 and Hprt were those described previously (33, 34).
Thymus grafting was performed as previously performed (33). Briefly, thymic lobes were isolated from embryos at 14.5 days postcoitus, and then cultured for 4 days on Nucleopore filters (Whatman) placed on R10 containing 1.35 mM 2′-deoxyguanosine (Sigma-Aldrich). Five pieces of thymic lobes were grafted under the renal capsule of BALB/c nude mice. After 6–8 wk, reconstitution of peripheral T cells was determined by flow cytometric analysis with anti-CD4 (clone GK1.5; BD Pharmingen) and anti-CD8 (clone 53-6.7; BD Pharmingen) mAbs, and then the thymic chimeras were used for analysis.
Immunohistochemical analysis of the thymus was performed as described previously (35, 36). For the detection of auto-Abs, mouse serum was incubated with various organs obtained from Rag2-deficient mice. FITC-conjugated anti-mouse IgG Ab (Southern Biotechnology Associates) was used for the detection (33).
Isolation and functional analysis of Tregs
Spleen cell suspensions were stained with FITC-conjugated anti-CD25 (clone 7D4) and PE-conjugated anti-CD4 (clone H129.19) (BD Pharmingen), and sorted by FACS (ALTRA; Beckman Coulter) as described previously (37). The purity of the CD25− and CD25+CD4+ populations was >90 and 95%, respectively. Spleen cells sorted as described above were cocultured with RBC-lysed and irradiated (15 Gy) spleen cells (5 × 104) from wild-type mice as APC for 3 days in 96-well round-bottom plates in R10. Anti-CD3 mAb (clone 145-2C11) (Cedarlane Laboratories) at a final concentration of 10 μg/ml was added to the culture for stimulation, and 3H incorporation during the last 6 h of culture was measured.
Development of Sjögren’s syndrome (SS)-like pathologic changes in exocrine organs from Aire-deficient mice
To investigate the roles of AIRE in the establishment and maintenance of self-tolerance in vivo, we generated Aire-null mutant mice. To this end, we deleted a large proportion of the known functional domains of Aire including SAND, PHD1, and PHD2 (6) (Fig. 1,A). The correct targeted event was confirmed by Southern blot analysis and genomic PCR of material from the gene-targeted mice (Fig. 1, B and C). Offspring homozygous for Aire deficiency were born in the numbers expected from the heterozygous crossing, and homozygous Aire-deficient mice were grossly normal. Although both male and female homozygous Aire-deficient mice are fertile when crossed with wild-type mice, homozygous crossing produced offspring only occasionally (F. Kajiura and M. Matsumoto, unpublished observation). Total spleen cell numbers and total thymocyte numbers were indistinguishable between control and Aire-deficient mice. Flow cytometric analysis showed similar expression of B220, CD3, CD4, and CD8 in the spleen and thymus of control and Aire-deficient mice. Proliferative responses and Ig production from the B cells after various stimuli, and proliferative responses and IL-2 production from the T cells stimulated with anti-CD3 mAb, were also unchanged by the Aire deficiency (S. Sun and M. Matsumoto, unpublished observation).
To assess the impact of Aire deficiency on the breakdown of self-tolerance, we inspected various organs (i.e., salivary glands, lacrimal glands, thyroid, heart, lung, liver, stomach, pancreas, kidney, small intestine, testis, and ovary) from Aire-deficient mice of original mixed background (i.e., H-2b/k × H-2b). The most marked changes were evident in the lacrimal glands (Fig. 2, A and B); all the Aire-deficient mice showed infiltration of many lymphoid cells in the lacrimal glands, whereas no such changes were observed in the control mice. We also observed infiltration of many lymphoid cells in the parotid glands (8 of 8 Aire-deficient mice) and submandibular glands (10 of 16 Aire-deficient mice) (Fig. 2 A). Consistent with these SS-like pathologic changes in exocrine organs from Aire-deficient mice, secretion of tears per unit of mouse body weight was decreased in the affected mice (0.89 ± 0.33 mm/20 min/body weight (g) from control mice (n = 5) vs 0.46 ± 0.08 mm/20 min/body weight (g) from Aire-deficient mice (n = 4); p < 0.05). In 1 of 10 Aire-deficient mice, lymphoid cell infiltration in either the stomach or pancreas was also observed. There were no obvious pathologic changes in other organs from Aire-deficient mice during follow-up to the age of 8 mo.
Autoreactive responses against α-fodrin in Aire-deficient mice
We have previously reported that NFS/sld mutant mice thymectomized 3 days after birth (3d-Tx) exhibit SS-like phenotypes with autoreactivity against α-fodrin, a ubiquitously expressed actin-binding protein (27, 38). Because of the similarity of SS-like phenotypes between Aire-deficient mice and the 3d-Tx-SS model, we investigated whether Aire-deficient mice exhibit autoreactivity against α-fodrin. We first tested the production of auto-Ab against various forms of recombinant α-fodrin in sera from Aire-deficient mice using Western blot analysis (Fig. 3, A and B). Sera from 3d-Tx mice showed reactivity predominantly against the JS-1 fragment (27). Four of five Aire-deficient mice showed reactivity against 2.7A, and two mice showed reactivity against 3′DA (Fig. 3,B). Sera from control mice showed no such reactivities. Production of auto-Ab against α-fodrin in Aire-deficient mice was also evaluated by ELISA using additional forms of recombinant α-fodrin (31) and larger numbers of mice. Ten of 11 Aire-deficient mice showed significantly higher reactivities against at least one form of recombinant α-fodrin fragment compared with those from 11 control mice (Fig. 3 C). Interestingly, each Aire-deficient mouse showed reactivity against different forms of α-fodrin.
We also confirmed the development of autoimmunity against α-fodrin using splenocytes from Aire-deficient mice (25). Such splenocytes cultured with recombinant α-fodrin showed significant proliferative responses; four Aire-deficient mice tested showed a response to 2.7A, but not to JS-1, whereas no such reactivities were observed from age-matched control mice (Fig. 4).
Unrepressed expression of corresponding target Ag in Aire-deficient thymus
The mechanism controlling the thymic microenvironment necessary for the establishment of self-tolerance in an Aire-dependent manner is of considerable interest. It has been suggested that “promiscuous” expression of a broad range of peripheral tissue-specific genes by TECs is essential for establishing self-tolerance (18), and Aire has been implicated in the control of this promiscuous gene expression through a transcriptional mechanism (19). Supporting this notion, real-time PCR has revealed that expression of insulin and salivary protein 1 was significantly reduced in the Aire-deficient thymic stroma (Fig. 5,A). Because Aire-deficient mice developed autoimmunity against the defined target Ag, α-fodrin, we examined whether the expression of α-fodrin mRNA in the thymic stroma is changed in Aire-deficient mice. Using real-time PCR together with semiquantitative RT-PCR with three sets of primers encompassing the entire coding region of α-fodrin, we detected unrepressed α-fodrin expression from Aire-deficient thymic stroma when compared with that from control thymic stroma (Fig. 5, A and B); this was observed under the condition where the expression of Foxn1, which encodes a transcription factor involved in thymus development (39), was indistinguishable between the samples (Fig. 5 A). Thus, our results suggest that Aire regulates self-tolerance beyond the transcriptional control of self-protein expression in the thymus, at least against this ubiquitously expressed protein.
To test whether autoreactivity against α-fodrin is associated with the development of inflammatory lesions in exocrine organs from Aire-deficient mice, we performed Western blot analysis using proteins extracted from the lacrimal glands. Both lacrimal glands and thymus from younger Aire-deficient mice (i.e., 3 mo) contained larger quantities of intact form α-fodrin (240 kDa) than the cleaved form (150 kDa), as observed for proteins from the control mice (Fig. 6,A); this was demonstrated with two different kinds of Abs recognizing the C-terminal half (anti-α-fodrin mAb) and N-terminal half (anti-AFN-A polyclonal Ab) of α-fodrin. However, lacrimal glands from some aged Aire-deficient mice (i.e., 8 mo) contained a reduced amount of the intact form (Fig. 6 B), although no detectable changes in α-fodrin expression in the thymus were observed in either form or quantity. This result suggests that autoreactivity against α-fodrin is associated with the pathogenetic process responsible for destruction of the lacrimal glands in this SS-like model, as observed in 3d-Tx-SS model (27, 38).
Loss of Aire in the thymic stroma is responsible for the breakdown of self-tolerance
Despite the predominant Aire expression in TECs, thymic structure was not apparently affected by the absence of Aire. Results of H&E staining as well as immunohistochemistry with the lectin Ulex europaeus agglutinin 1 (40) and ER-TR5 mAb (41), both recognizing a subset of mTEC, were indistinguishable between control and Aire-deficient mice (F. Kajiura, T. Ueno, Y. Takahama, and M. Matsumoto, unpublished observation). Organization of dendritic cells in the thymus identified with the mAb CD11c was also unaffected by Aire deficiency. Thus, Aire may not affect thymic organogenesis. Alternatively, relatively low frequencies of Aire-expressing cells among mTECs may account for the apparently normal thymic structure in Aire-deficient mice.
To investigate the impact of Aire deficiency in the thymic microenvironment, we generated thymic chimeras. Thymic lobes were isolated from control and Aire-deficient embryos of mixed background (H-2b/k × H-2b) and cultured for 4 days in the presence of 2′-deoxyguanosine to eliminate thymocytes. Such thymic lobes did not contain any live thymocytes, as determined by flow cytometric analysis and Western blot analysis with anti-lck Ab (33). The lobes were then grafted under the renal capsule of BALB/c nude mice (H-2d). Grafting of both control and Aire-deficient embryonic thymus induced T cell maturation in BALB/c nude mice at the periphery to a similar extent: CD4+ T cells plus CD8+ T cells were 12.5 ± 2.2% in nude mice grafted with control thymus (n = 6), compared with 12.3 ± 1.6% in nude mice grafted with Aire-deficient thymus (n = 7). It is important to note that the mature T cells produced de novo in both cases originated from Aire-sufficient nude mouse bone marrow (BM). Remarkably, histological examination of Aire-deficient thymus-grafted mice revealed infiltration of many lymphoid cells in the liver (mainly in the portal area) and pancreas (interlobular periductal and perivascular areas near islets) (Fig. 7, A and B). In contrast, we observed few such changes in control thymus-grafted mice.
To confirm that T cells developing in a thymic microenvironment without Aire are autoreactive per se, we injected splenocytes obtained from BALB/c nude mice grafted with Aire-deficient thymus into another group of BALB/c nude mice. We observed similar lymphoid cell infiltration in the liver of the recipient mice, whereas injection of splenocytes obtained from nude mice grafted with control thymus induced no such changes in the recipient mice (Fig. 7 B). These results clearly indicate the significance of Aire as a thymic stromal element required for the establishment of self-tolerance.
Impaired regulation of autoreactivity in the absence of Aire
There is accumulating evidence that T cell-mediated dominant control of autoreactive T cells represents an important mechanism for the maintenance of immunologic self-tolerance (16, 17). We investigated whether loss of Aire in the thymus has a major impact on the production and/or function of Tregs. Spleen and thymus from adult Aire-deficient mice contained similar percentages as well as total numbers of CD4+CD25+ T cells compared with those from control mice (Fig. 8,A). Real-time PCR for quantification of Foxp3 mRNA (34, 42, 43) did not show any reduction of Tregs in the spleen of Aire-deficient mice (Fig. 8 B). Expression of Foxp3 in the whole thymus was also comparable between control mice and Aire-deficient mice (Foxp3/Hprt from wild-type mice = 1.8 vs Foxp3/Hprt from Aire-deficient mice = 2.4).
Recently, it has been demonstrated that functional alterations of Tregs could contribute to the development of autoimmune disease. A significant decrease in the effector function of CD4+CD25+ T cells from peripheral blood of patients with multiple sclerosis has been reported (44). It is of particular interest that the suppressor function of CD4+CD25+ T cells has been demonstrated to be defective in patients with autoimmune polyglandular syndrome type II, which is phenotypically closely related to APECED (also called autoimmune polyglandular syndrome type I) but whose pathogenesis is currently unknown (45). It is therefore important to test the function of Tregs from Aire-deficient mice. CD4+CD25+ T cells isolated from Aire-deficient mice dose-dependently suppressed [3H]thymidine uptake by naive T cells cocultured in vitro with an efficiency nearly identical to that of CD4+CD25+ cells from control mice (Fig. 8,Ca). This was also the case when responder cells (CD4+CD25− cells) isolated from Aire-deficient mice were used for the assay (Fig. 8 Cb). Thus, Aire does not have a major impact on the production and/or function of Tregs, at least as assessed in those assays.
To gain further insight into how Aire contributes to the establishment of self-tolerance, we grafted control (Aire sufficient) and Aire-deficient embryonic thymus simultaneously into BALB/c nude mice. Inflammatory changes in the liver and pancreas of these animals were still present (Fig. 7 B), supporting the hypothesis that impaired dominant control of autoreactive T cells by Tregs may not be the major defect caused by a thymic stroma lacking Aire; if impaired production of Tregs were the major defect caused by a thymic stroma lacking Aire, we assume that the defect should have been corrected by the grafted Aire-sufficient thymus. Therefore, it is reasonable to speculate that overproduction of autoreactive T cells plays an important role in the disease process triggered by Aire deficiency.
Strain-dependent target-organ specificity of the autoimmune disease caused by Aire deficiency
Although APECED is a monogenic disorder, it has been postulated that there may be additional factor(s) that determine the clinical features of the disease, such as the spectrum of affected organs (5, 6, 22). To test this hypothesis, we backcrossed our original strain of Aire-deficient mice to either the C57BL/6 (H-2b) or BALB/c (H-2d) strain for six generations. Both backcrossed strains showed autoimmune phenotypes similar to those from an original strain of Aire-deficient mice of mixed background (i.e., infiltration of many lymphoid cells in the salivary glands) (Fig. 9,B, top). However, Aire-deficient BALB/c mice additionally demonstrated lymphoid cell infiltration in the gastric mucosa (Fig. 9, A and B, bottom), a feature that has been observed only rarely in the original Aire-deficient mice of mixed background (1 of 10) or Aire-deficient C57BL/6 mice (Fig. 9,B, bottom). Consistent with these histological findings, serum harvested from Aire-deficient BALB/c mice (4 of 4) demonstrated strong auto-Abs against gastric mucosa (Fig. 9 C), whereas this activity was observed in only one of four Aire-deficient C57BL/6 mice, and it was weak. Thus, the genetic background of the mice clearly influences the target-organ specificity of the disease caused by Aire deficiency.
Using gene-targeted mice, we have investigated the mechanisms controlling the establishment and maintenance of self-tolerance by Aire. Both the numbers and suppressive function of CD4+CD25+ Tregs were not changed in Aire-deficient mice, when assessed in the adult mice. Using thymic chimeras, we also investigated possible defects in the production of any cell types (including CD4+CD25+ Tregs) that are involved in the prevention of T cell-mediated organ-specific autoimmune diseases in the absence of Aire. When Aire-deficient and Aire-sufficient thymus were grafted simultaneously into nude mice, the development of inflammatory lesions was not completely inhibited. These results suggest that impaired production of Tregs may not be the major mechanism responsible for the breakdown of self-tolerance in Aire-deficient mice, and it is reasonable to speculate that the Aire-deficient thymus allows production of more pathogenic autoreactive T cells than could be controlled by the Tregs. However, it is important to emphasize that other aspects of Tregs, such as their repertoire formation, still remain unsolved; we cannot rule out the possibility that Aire may affect the Ag specificity of the Treg repertoire, because most of the analysis of the Tregs in the present study was quantitative rather than qualitative.
We have demonstrated that anti-α-fodrin autoimmunity developed in Aire-deficient mice despite the fact that the transcription of corresponding Ag (i.e., α-fodrin) in the thymic stroma was not down-regulated. Based on this finding, we suggest that Aire may regulate the processing and/or presentation of self-Ags by TECs, possibly through a coordinated action with BM-derived cells (see below), so that the maturing T cells can recognize the corresponding self-Ags in a form capable of efficiently triggering autoreactive T cells. It would be important to know whether our proposed model of Aire function in the establishment of self-tolerance is confined to ubiquitous self-Ags, such as α-fodrin, or applicable to tissue-specific Ags as well. In this regard, it is critical to investigate first whether autoimmunity develops bona fide against transcriptionally repressed tissue-specific Ags in the thymus in Aire-deficient mice. Definitively, identification of the substrate(s) for E3 ubiquitin ligase activity by AIRE should help to clarify the actual mechanisms of AIRE-dependent tolerance (10).
We have demonstrated that α-fodrin is one of the target Ags involved in the autoimmune-disease process caused by Aire deficiency. Because transfer of sera from affected mice did not result in the development of sialoadenitis or disruption of α-fodrin in the recipient mice (N. Ishimaru, R. Arakaki, and Y. Hayashi, unpublished observation), the disease process in Aire-deficient mice is most likely elicited by a cell-mediated immunity, as observed in the 3d-Tx-SS model (29, 30). Consistent with this hypothesis, splenocytes from Aire-deficient mice demonstrated proliferative responses in vitro when cultured with recombinant α-fodrin (Fig. 4).
Reduction of the intact form of α-fodrin in the affected lacrimal glands of some aged Aire-deficient mice (Fig. 6 B) suggests that elicitation of autoreactivity against α-fodrin could be the primary pathogenetic process that leads to tissue destruction (27). In fact, adoptive transfer of α-fodrin-reactive T cells into ovariectomized B6 and SCID mice resulted in the development of autoimmune exocrinopathy quite similar to SS (30). However, based on the fact that α-fodrin is a ubiquitous protein and that the tissue destruction is confined to exocrine organs, it is reasonable to speculate that other undetermined tissue-specific target Ag(s) in exocrine organs might be additionally involved in the tissue destruction. Identification of precise target Ags involved in the disease process in Aire-deficient mice should help unravel the molecular mechanisms by which loss of Aire contributes to disease development.
We have demonstrated Aire-dependent disease development using allogeneic thymic chimeras; autoimmune disease commences in BALB/c nude recipients (H-2d) of Aire-deficient, but not of wild-type, thymic transplants from mice of original mixed background (H-2b/k × H-2b) (Fig. 7). The roles of TECs vs BM-derived cells in T cell repertoire selection in allogeneic thymic chimeras have been an issue of long-standing interest and debate. Given that nude mice reconstituted with an MHC-incompatible thymus generate effector T cells that are specific for the host and not for the thymic MHC (46), a novel mechanism may be responsible for the Aire-dependent negative selection; Aire expressed on TECs acts on BM-derived cells “in trans” as an important factor in organizing the “negative selection niche” in the thymus (47). This scenario is in good accordance with our results demonstrating the impaired tolerance to a ubiquitously expressed auto-Ag (i.e., α-fodrin) in Aire-deficient mice, because tolerance to ubiquitous self-proteins is mediated mainly by BM-derived cells in the thymus (48). Further study is required to test this intriguing hypothesis.
There is increasing evidence for the genetic complexity that underlies monogenic diseases (49, 50). In fact, the spectrum of the APECED phenotype is broad; the number of symptoms as well as the onset of each manifestation varies among affected patients. In our backcrossed mice, gastritis was observed predominantly in the BALB/c strain. In light of the fact that the individual HLA class II alleles modify the APECED phenotype (22), it is possible to speculate that MHC could be a candidate for the factor that determines this target-organ specificity. However, a genetic study with congenic strains has demonstrated that BALB/c (H-2d), BALB.B (H-2b), and BALB.K (H-2k) were all susceptible to experimentally induced gastritis, whereas B10.D2 (H-2d) were resistant, suggesting the predominant role of non-MHC gene(s) in determining susceptibility to autoimmune gastritis (51). Thus, MHC genes as well as non-MHC genes may together contribute to the complex phenotypes of APECED.
In conclusion, integration of detailed phenotypic analyses of Aire-deficient mice with current perspectives of thymus biology promises to illuminate many aspects of the molecular mechanisms responsible for the establishment and maintenance of self-tolerance. With the production of inbred strains of Aire-deficient mice, it may also be feasible to assess the impact of environmental factors that could influence the clinical features of APECED.
We thank Drs. W. van Ewijk and M. Itoi for the gift of ER-TR5 mAb. We thank Drs. T. Yamada, T. Yabuki, M. Kasai, and K. Iwabuchi for valuable suggestions. We also thank K. Awahayashi and F. Saito for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by Special Coordination Funds of the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government (MEXT), and by a grant-in-aid for Scientific Research from the MEXT.
Abbreviations used in this paper: AIRE, autoimmune regulator; APECED, autoimmune-polyendocrinopathy-candidiasis ectodermal dystrophy; TEC, thymic epithelial cell; mTEC, medullary TEC; PHD, plant homeodomain; HEL, hen egg lysozyme; 3d-Tx mice, mice thymectomized 3 days after birth; SS, Sjögren’s syndrome; Treg, immunoregulatory T cell; BM, bone marrow.