Physical contact between thymocytes and the thymic stroma is essential for T cell maturation and shapes the T cell repertoire in the periphery. Stromal elements that control these processes still remain elusive. We used a mouse strain with mutant NF-κB-inducing kinase (NIK) to examine the mechanisms underlying the breakdown of self-tolerance. This NIK-mutant strain manifests autoimmunity and disorganized thymic structure with abnormal expression of Rel proteins in the stroma. Production of immunoregulatory T cells that control autoreactive T cells was impaired in NIK-mutant mice. The autoimmune disease seen in NIK-mutant mice was reproduced in athymic nude mice by grafting embryonic thymus from NIK-mutant mice, and this was rescued by supply of exogenous immunoregulatory T cells. Impaired production of immunoregulatory T cells by thymic stroma without normal NIK was associated with altered expression of peripheral tissue-restricted Ags, suggesting an essential role of NIK in the thymic microenvironment in the establishment of central tolerance.
Autoimmune disease is a pathological condition in which the immune system turns on itself and causes serious damage to the organism’s tissues by as yet unknown mechanisms (1). A unifying concept for the mechanisms underlying the development of autoimmune disease has been one of the major challenges of immunological studies. Breakdown of self-tolerance is considered to be the key event for the disease process, and understanding the pathogenesis of this process is crucial for developing a therapeutic approach to these diseases. Because the thymus is the primary lymphoid organ for the establishment of self-tolerance, it is important to investigate how this process is controlled by the organization of the thymic microenvironment.
NF-κB-inducing kinase (NIK)3 is structurally related to mitogen-activated protein kinase kinase kinase (2) and has been shown to phosphorylate both IκB kinase (IKK)-α and IKK-β, which sequentially activate the downstream IκB proteins necessary for NF-κB activation (3, 4). The alymphoplasia (aly) strain of mice carries a natural mutation of the NIK gene (5, 6) in which a G855R substitution in the C terminus of the protein results in inability to bind to IKK-α (7). NIKaly/aly mice have provided a unique model for the abnormal development of lymphoid organs; NIKaly/aly mice lack all lymph nodes and Peyer’s patches, and spleen architecture such as development of germinal centers and follicular dendritic cell clusters is disturbed (5, 6, 8). We and others have demonstrated that this is due to defective NF-κB activation through the lymphotoxin (LT)-β receptor (LTβR) (6, 7, 9), a receptor essential for the development of lymphoid organs (10). Thymic structure is also disorganized in NIKaly/aly mice, which is not observed in mice deficient for LT-α or LT-β (10); the medulla in NIKaly/aly mice is smaller than that in NIKaly/+ mice, and the boundary of the cortex and medulla is unclear (5, 6, 11). In addition to this abnormal lymphoid organogenesis, NIKaly/aly mice also serve as a model of autoimmune disease, but of unknown etiology (12); histopathological analysis of NIKaly/aly mice has revealed chronic inflammatory changes in several organs, including the liver, pancreas, lung, salivary gland, and lacrimal gland (Refs. 5 and 12 and K. Izumi, Y. Bando, and M. Matsumoto, unpublished observation). Furthermore, these inflammatory changes in exocrine organs could be transferred into recombination-activating gene 2-deficient mice by a T cell-enriched fraction of spleen cells from NIKaly/aly mice (12). We reasoned that the autoimmune disease phenotype seen in NIKaly/aly mice might be associated with the altered thymic microenvironment.
One important role of the thymic stroma in establishing self-tolerance is the elimination of pathogenic autoreactive T cells by negative selection (13, 14). For this purpose, thymic epithelial cells need to express a set of self-Ags encompassing all the self-Ags expressed by parenchymal organs. Supporting this hypothesis, analysis of gene expression in thymic stroma has demonstrated that epithelial cells of the medulla are a specialized cell type in which promiscuous expression of a broad range of peripheral tissue-restricted genes is an autonomous property (15). Consistent with this notion, an autoimmune regulator (AIRE) in thymic epithelial cells, a putative transcription factor, has been demonstrated to regulate this process (16), and deficiency of AIRE both in humans and in animal models results in the development of organ-specific autoimmune disease (16, 17, 18).
In addition to negative selection, self-tolerance is maintained by another mechanism involving immunoregulatory T cells. CD4+CD25+ T cells are immunoregulatory T cells that prevent CD4+ T cell-mediated organ-specific autoimmune diseases (19, 20, 21, 22). The significance of this cell type in the maintenance of self-tolerance has been demonstrated by the fact that elimination of CD4+CD25+ T cells leads to the development of organ-specific autoimmune diseases in otherwise normal mice (23). Although it has been demonstrated that immunoregulatory T cells are substantially reduced in mice deficient in B7, CD28 (24), or CD40 (25), and that Foxp3, a transcription factor genetically defective in an autoimmune disease termed immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome, is a key regulator for the development of immunoregulatory T cells (26, 27, 28), the thymic stromal elements that control the production of immunoregulatory T cells remain elusive.
Given that NIKaly/aly mice have disorganized thymic structure together with organ-specific autoimmune disease, we hypothesized that NIK in the thymic stroma is essential for the establishment of self-tolerance. To examine this possibility, we have grafted embryonic thymus from NIK-mutant mice onto athymic nude mice. In this system, mature T cells derived from NIK-sufficient recipient bone marrow are produced de novo through interaction with thymic stroma derived from NIK-mutant mice. We found that the autoimmune disease phenotype was transferable into the recipient animals by this treatment. We also found that NIKaly/aly mice have impaired production of immunoregulatory T cells, and that this was associated with abnormal expression of Rel proteins in the thymic stroma together with reduced gene expression for peripheral tissue-restricted Ags. Furthermore, we show that exogenous supply of immunoregulatory T cells rescues the autoimmune disease phenotype in this mouse model, supporting the potential benefits of this particular cell type for the treatment of autoimmune diseases. Thus, our results implicate thymic epithelial cells as a therapeutic target for the manipulation of processes for the establishment of self-tolerance and for the control of autoimmunity.
Materials and Methods
NIKaly/+ mice, NIKaly/aly mice (5), and BALB/cA Jcl-ν mice (BALB/cnu/nu mice) were purchased from CLEA Japan (Osaka, Japan). LT-α−/− mice are the generous gift from Dr. D. D. Chaplin (University of Alabama, Birmingham, AL). The mice were maintained under pathogen-free conditions, and were handled in accordance with the Guidelines for Animal Experimentation of Tokushima University, School of Medicine (Tokushima, Japan). The experiments were initiated at 8–12 wk of age.
Immunohistochemical analysis of the thymus was performed as previously described (8, 9). Briefly, frozen tissue sections were fixed in cold acetone, and stained by first incubating with peanut agglutinin (PNA)-biotin and Ulex europaeus agglutinin 1 (UEA-1)-biotin (Vector Laboratories, Burlingame, CA). After being washed, the sections were further incubated with streptavidin conjugated with alkaline phosphatase (AP) (Zymed Laboratories, San Francisco, CA). Color development for bound AP was with an AP reaction kit (Vector Laboratories). For the confocal microscopic analysis, Alexa 488 (Molecular Probes, Eugene, OR) -conjugated anti-mouse CD4 (clone GK1.5), Alexa 546-conjugated CD8 (clone 53-6.7) and ER-TR5 mAbs were used. Alexa 633-conjugated anti-rat IgG (Molecular Probes) was used for the detection of ER-TR5 binding. For the detection of antinuclear Ab (ANA), serum from thymic chimeras was incubated with HepG2 cells grown on slide glasses. FITC-conjugated anti-mouse IgG Ab (Southern Biotechnology Associates, Birmingham, AL) was used for the detection.
Thymic lobes were isolated from embryos at 14.5 days postcoitus, and were cultured for 4 days on top of Nuclepore filters (Whatman, Clifton, NJ) placed on RPMI 1640 medium (Invitrogen, San Diego, CA) supplemented with 10% heat-inactivated FBS (Invitrogen), 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μM 2-ME, hereafter referred to as R10, containing 1.35 mM 2′-deoxyguanosine (2-DG) (Sigma-Aldrich, St. Louis, MO). Five pieces of thymic lobes were grafted under the renal capsule of BALB/cnu/nu mice. After 6–8 wk, reconstitution of peripheral T cells was determined by flow cytometric analysis (BD Biosciences, Mountain View, CA) with anti-CD4 (clone GK1.5; BD PharMingen, San Diego, CA) and anti-CD8 (clone 53-6.7; BD PharMingen) mAbs, and then thymic chimeras were used for the analyses. In some cases, mice were injected with CD4+ T cells or CD4+CD25− T cells (2 × 107 cells per mouse) isolated from BALB/cnu/+ mice just after thymus graft.
Bone marrow (BM) transfer
BM transfer was performed as described previously (8). In brief, BM cells were suspended in R10 medium containing anti-CD90 (Thy1.2) mAb (clone 5a-8; Cedarlane Laboratories, Ontario, Canada) plus low toxicity rabbit C (Cedarlane Laboratories). After incubation at 37°C for 45 min, the cells were washed twice and adjusted to 3 × 107 viable cells/ml in R10. Each recipient mouse was lethally irradiated (10 Gy) and treated with 0.5 ml of donor BM cells i.v. on the same day. The recipient mice were used in the analyses 6–10 wk after BM transfer.
Proteins extracted from embryonic thymic lobes prepared as described above were analyzed using an ECL Western blotting detection system (Amersham, Piscataway, NJ). The Abs used were rabbit antipeptide Abs directed against p50 (cat. no. sc-114), p52 (cat. no. sc-298), RelA (cat. no. sc-109), RelB (cat. no. sc-226), and c-Rel (cat. no. sc-71), all purchased from Santa Cruz Biotechnology (Santa Cruz, CA) (8).
Formalin-fixed tissue sections were subjected to H&E staining, and two pathologists independently evaluated the histology without being informed of the detailed condition of the individual mouse. Histological changes were scored as 0 (no change), 1 (mild lymphoid cell infiltration), or 2 (marked lymphoid cell infiltration).
Isolation and functional analysis of immunoregulatory T cells
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 a FACS (ALTRA; Beckman Coulter, Fullerton, CA) as previously described (23). 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 NIKaly/+ 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 the culture was measured. For inoculation into thymus-grafted BALB/cnu/nu mice, CD4+ T cells were isolated from spleen and lymph node of BALB/cnu/+ mice by using MACS CD4 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany), as described previously (8). CD4+CD25− cells were prepared by the depletion of CD25+ cells with anti-CD25 Ab plus low toxicity rabbit C.
Real-time PCR for the quantification of Foxp3 was conducted with cDNA prepared from RNA extracted from CD4+ splenocytes and whole thymus. The primers and the probe used were as previously described (26). Real-time PCR were performed in a final volume of 20 μl with 400 nM of the forward and reverse primers and 200 nM of the probe by use of a QuantiTect Probe PCR kit (Qiagen, Valencia, CA). Reactions were run on an ABI/PRIZM 7700 sequence detection system (Applied Biosystems, Foster City, CA) in triplicate. Cycling conditions were a single denaturing step at 95°C for 15 min followed by 45 cycles of 94°C for 15 s and 60°C for 1 min.
RNA was extracted from whole thymus with TRIzol (Invitrogen), and treated with DNase to eliminate any contaminating DNA. After phenol/chloroform extraction and ethanol precipitation, 5 μg of total RNA were subjected to oligo(dT)-primed reverse transcription with a cDNA Cycle kit (Invitrogen). The primer pairs used for PCR were as previously described (16). PCR was conducted in a final volume of 30 μl with 1.5 U of ExTaq DNA polymerase (Takara Biomedicals, Otsu, Japan) and 250 nM of each primer. Cycling conditions were a single denaturing step at 94°C for 3 min followed by either 35 cycles (for tissue-specific Ags) or 25 cycles (for β-actin) of 94°C for 45 s, 50–54°C for 45 s, and 72°C for 1 min, followed by a final extension step of 72°C for 3 min. For AIRE, a single denaturing step at 95°C for 3 min was followed by 35–40 cycles of 95°C for 30 s and 65°C for 5 min.
NIK in the thymic stroma is required for self-tolerance
Thymic structure is disorganized in NIKaly/aly mice; the boundary of the cortex and medulla revealed with H&E staining is unclear, and the medulla contained fewer epithelial cell components compared with that from NIKaly/+ mice (Fig. 1,Aa). We then used immunohistochemistry to investigate thymic organization in NIKaly/aly mice. The medulla in NIKaly/aly mice was smaller than that in NIKaly/+ mice, and ER-TR5-positive medullary epithelial cells were sparse (Fig. 1,Ab). Because thymocytes in the cortex bind with PNA (29), the medulla can be identified as a PNA-negative area. We found that medullary epithelial cells that bind with the lectin UEA-1 (30) were absent from NIKaly/aly mice (Fig. 1 Ac). Although signals through LTβR control the production of UEA-1+ cells (31), membrane-bound LT is dispensable for this action, because they are present in the thymus of LT-α−/− mice (F. Kajiura and M. Matsumoto, unpublished observation).
Given that NIK plays an essential role in the thymic organogenesis through the NF-κB activation pathway, we investigated the expression of Rel proteins in NIKaly/aly thymic stroma by Western blotting. Thymic lobes were isolated from NIKaly/+ and NIKaly/aly embryos at 14.5 days postcoitus and cultured for 4 days in the presence of 2-DG to eliminate thymocytes. Such thymic lobes did not contain any live thymocytes as determined by flow cytometric analysis (Fig. 1,B). In contrast, thymic lobes cultured in the absence of 2-DG supported maturation of thymocytes under the same conditions. This was also confirmed by the absence or presence of lck expression from thymic lobes cultured with or without 2-DG, respectively (Fig. 1,C, top). Although the expression of p100, a precursor form of p52, was more abundant in the thymic stroma from NIKaly/aly mice compared with that from NIKaly/+ mice, p52 expression was significantly reduced (Fig. 1 C); the ratio between p52 and p100 in NIKaly/+ thymic stroma was roughly equal, whereas it was extremely low in NIKaly/aly thymic stroma. Thus, NIK-dependent generation of p52 from p100 in thymic stroma could constitute a second NF-κB signaling pathway, as originally observed for hemopoietic cells (8). RelB expression in the thymic stroma was also reduced in NIKaly/aly mice compared with that from NIKaly/+ mice. These results suggest that disturbed thymic organogenesis in NIKaly/aly mice is due to the abnormal regulation of the NF-κB activation pathway in the thymic stroma in the absence of normal NIK. Expression of other members of Rel proteins was comparable in NIKaly/+ and NIKaly/aly mice.
To investigate the impact of the altered thymic microenvironment on the development of the autoimmune disease phenotype in NIKaly/aly mice, we generated thymic chimeras. 2-DG-treated embryonic thymic lobes were prepared as described above, and then grafted under the renal capsule of BALB/cnu/nu mice. Grafting both NIKaly/+ and NIKaly/aly embryonic thymus induced T cell maturation of BALB/cnu/nu mice in the periphery to a similar extent: CD4+ T cells plus CD8+ T cells were 7.05 ± 2.45% in BALB/cnu/nu mice grafted with NIKaly/+ thymus (n = 13) compared with 7.35 ± 2.96% in BALB/cnu/nu mice grafted with NIKaly/aly thymus (n = 12). It is important to note that the de novo produced mature T cells in both cases originate from NIK-sufficient BALB/cnu/nu mouse bone marrow. It is also noteworthy that BALB/cnu/nu mice grafted with either NIKaly/+ or NIKaly/aly thymus are in the context of normal lymph node development. Histological evaluation of the grafted thymus revealed that the NIKaly/+ thymus contained UEA-1+ cells, as observed in untreated adult NIKaly/+ thymus, whereas NIKaly/aly embryonic thymus grafted onto nude mice did not acquire UEA-1+ cells (Fig. 1,Ad), suggesting that T cells with normal NIK (derived from BALB/cnu/nu mice) cannot restore UEA-1+ medullary epithelial cells in the NIKaly/aly thymus when the T cell-thymic stromal cell interaction was initiated from the embryonic stage. Thus, production of UEA-1+ medullary epithelial cells seems to require normal NIK in the thymic stromal element. Remarkably, histological examination of NIKaly/aly thymus-grafted mice revealed many lymphoid cell infiltrations in the liver (mainly in the portal area) (Fig. 2, Aa and C) and pancreas (interlobular periductal and perivascular areas near islets) (Fig. 2, Ab and C). In contrast, we saw few such changes in NIKaly/+ thymus-grafted mice. Consistent with these histological findings, serum harvested from NIKaly/aly thymus-grafted mice, but not from NIKaly/+ thymus-grafted mice, demonstrated elevated levels of transaminases (glutamic oxaloacetic transaminase and glutamic pyruvic transaminase) from the damaged liver (Fig. 2,B). Autoimmune-disease phenotypes in NIKaly/aly thymus-grafted mice were slightly milder at the later time point (i.e., 3 mo after thymus grafting), suggesting that the autoimmune-disease phenotype develops predominantly during the period of the proliferative burst of homeostatic expansion in this experimental system (F. Kajiura, K. Izumi, Y. Bando, and M. Matsumoto, unpublished observation). Mild gastritis and sialoadenitis were observed similarly in both NIKaly/+ thymus- and NIKaly/aly thymus-grafted mice; similar changes have been observed even in isogenic thymus-grafted nude mice of the BALB/c background (S. Sakaguchi, unpublished observation). When the serum from those thymic chimeras was tested for ANA using HepG2 as Ag, NIKaly/aly thymus-grafted mice showed a high incidence of the IgG class of autoantibody as detected with immunofluorescence (Fig. 2 Ac); of eight NIKaly/aly thymus-grafted mice tested, one mouse showed strong ANA activity and four showed moderate activity, whereas ANA activity was observed in only one of five NIKaly/+ thymus-grafted mice, and this activity was only weak.
Consistent with the significance of the thymic stromal element for the development of the autoimmune-disease phenotype seen in NIKaly/aly mice as well as in NIKaly/aly thymus-grafted mice, NIKaly/aly mice reconstituted with NIKaly/+ BM displayed a spectrum of inflammatory lesions similar to that in NIKaly/aly mice, as exemplified by the pancreas and the salivary gland (Fig. 2 D). In contrast, NIKaly/+ mice reconstituted with NIKaly/aly BM showed few such inflammatory changes.
Impaired production of immunoregulatory T cells in NIK-mutant mice
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. CD4+CD25+ T cells are immunoregulatory T cells that prevent CD4+ T cell-mediated organ-specific autoimmune diseases (19, 20, 21, 22). We reasoned that the development of the autoimmune-disease phenotype in NIKaly/aly mice might be associated with the impaired production of immunoregulatory T cells in the altered thymic microenvironment. Spleen from NIKaly/aly mice contained reduced percentages as well as total numbers of CD4+CD25+ T cells compared with spleen from NIKaly/+ mice (Fig. 3,A, top panel). Reduced numbers of CD4+CD25+ T cells in NIKaly/aly mice were also observed in the thymus (Fig. 3 A, bottom panel).
We and others have recently demonstrated that Foxp3, a gene responsible for the multiorgan autoimmune disease called immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome, is a functional marker for immunoregulatory T cells (26, 27, 28). To confirm the reduction of immunoregulatory T cells in NIKaly/aly mice, we performed a quantitative RT-PCR for Foxp3 using CD4+ T cells isolated from the spleen. Hypoxanthine phosphoribosyltransferase (HPRT) expression level served as an internal control for the assay. The Foxp3 expression level in CD4+ T cells from NIKaly/aly mice was almost half of that in cells from NIKaly/+ mice (Table I), with a good correlation with the percentages as well as with the numbers of CD4+CD25+ T cells determined by flow cytometric analysis. Although less dramatic, Foxp3 expression from the whole thymus was also reduced in NIKaly/aly mice, again consistent with the results obtained by flow cytometric analysis.
|.||.||.||Foxp3 .||HPRT .||Foxp3/HPRT .|
|Expt. 1b||CD4+ T cellc||aly/+||20.04||10.71||1.87|
|Expt. 2f||CD4+ T cell||aly/+||72.67||95.18||0.76|
|Expt. 3f||CD4+ T cell||aly/+||65.50||51.59||1.27|
|.||.||.||Foxp3 .||HPRT .||Foxp3/HPRT .|
|Expt. 1b||CD4+ T cellc||aly/+||20.04||10.71||1.87|
|Expt. 2f||CD4+ T cell||aly/+||72.67||95.18||0.76|
|Expt. 3f||CD4+ T cell||aly/+||65.50||51.59||1.27|
Quantitative RT-PCR for Foxp3 was performed with HPRT expression level as an internal control for the assay.
cDNAs prepared from CD4+CD25+ T cells from BALB/c mouse spleen were used as the standard for real-time PCR.
For each experiment, CD4+ T cells were freshly isolated from the spleen with CD4 microbeads. The purity of CD4+ T cells was >90% for both NIKaly/+ mice and NIKaly/aly mice.
Relative abundance of Foxp3 was calculated from the ratio between Foxp3/HPRT values from NIKaly/+ mice and those from NIKaly/aly mice, and is shown in parentheses.
cDNAs were prepared from the whole thymus.
cDNAs prepared from CD4+ T cells from NIKaly/+ mouse spleen were used as the standard.
Although reduced in number, CD4+CD25+ T cells isolated from NIKaly/aly mice dose-dependently suppressed [3H]thymidine uptake by native T cells cocultured in vitro with an efficiency nearly identical to that of CD4+CD25+ cells from NIKaly/+ mice (Fig. 3 B); addition of IL-2 abolished the suppressive function of immunoregulatory T cells in both cases. Thus, although NIK plays an important role in the production of immunoregulatory T cells, NIK is not required for the suppressive function of this cell type. Together with the results from thymic chimeras, these results suggest that NIK plays an important role in the establishment of self-tolerance, in part through the generation of the thymic microenvironment that controls the production of immunoregulatory T cells.
Rescue of self-tolerance by immunoregulatory T cells
We investigated whether the autoimmune-disease phenotype observed in BALB/cnu/nu mice grafted with embryonic thymus from NIKaly/aly mice can be rescued by the supply of sufficient numbers of immunoregulatory T cells. When we injected total CD4+ T cells (2 × 107) (containing ∼10% of CD25+ cells by FACS analysis; data not shown) isolated from BALB/cnu/+ mouse spleen and lymph nodes into NIKaly/aly thymus-grafted BALB/cnu/nu mice, we found no development of inflammatory lesions (Fig. 2,C). Titration studies using purified CD4+CD25+ T cells demonstrated that >2 × 105 cells per mouse were required to prevent the disease in this system (F. Kajiura, S. Sun, K. Izumi, Y. Bando, and M. Matsumoto, unpublished observation). Although CD25 is a marker for CD4+ T cells that have immunoregulatory properties, it has been reported that T cells with this function are also found in CD25− subpopulations in the periphery (32). Consistent with this notion, injection of CD4+ T cells depleted of CD25+ cells also rescued the chronic inflammatory changes in NIKaly/aly thymus-grafted BALB/cnu/nu mice to some extent, although the preventive effect was weaker compared with CD4+ T cells containing CD25+ cells (Fig. 2 C). These results further support the idea that NIK in thymic stroma maintains self-tolerance by producing immunoregulatory T cells.
We also grafted NIKaly/+ and NIKaly/aly embryonic thymus simultaneously onto BALB/cnu/nu mice. Inflammatory changes of the liver and pancreas in these animals were milder compared with those from BALB/cnu/nu mice singly grafted with NIKaly/aly embryonic thymus, but were definitely still present, especially in the liver, when compared with BALB/cnu/nu mice singly grafted with NIKaly/+ embryonic thymus (Fig. 2 C). This result suggests that impaired production of immunoregulatory T cells (which should have been corrected by the grafted NIKaly/+ thymus) is not the only defect caused by thymic stroma without normal NIK. There must be an additional defect(s) determined by the NIKaly/aly thymus that cannot be completely rescued by the NIKaly/+ thymus (see Discussion).
NIK mutation is associated with altered gene expression in the thymus
The mechanism that controls the thymic microenvironment in a NIK-dependent fashion is of considerable interest. We wanted to determine the gene(s) relevant to the role of the thymic stroma in the development of the autoimmune-disease phenotype in NIKaly/aly mice. One candidate gene is AIRE, mutation of which is responsible for the autoimmune-polyendocrinopathy-candidiasis ectodermal dystrophy syndrome, which shows monogenic autosomal recessive inheritance (18). Because AIRE is strongly expressed in medullary epithelial cells in the thymus (15, 16, 18), we suspected that expression of AIRE in the thymus might be changed in the absence of normal NIK. Therefore, we prepared RNA from whole thymus and performed RT-PCR for AIRE. AIRE expression in NIKaly/aly thymus was dramatically reduced compared with that from NIKaly/+ thymus (Fig. 4 A, left, top). Real-time PCR demonstrated that AIRE expression level from NIKaly/aly thymus was 1–5% of that from NIKaly/+ thymus (N. Kuroda, H. Han, and M. Matsumoto, unpublished observation).
“Aberrant” or “promiscuous” expression of a broad range of peripheral tissue-restricted genes by thymic epithelial cells has been implicated in the essential process of establishing self-tolerance (15), and AIRE in thymic epithelial cells has been demonstrated to control this promiscuous gene expression (16). We used semiquantitative RT-PCR to investigate the expression of peripheral tissue-restricted genes in RNAs extracted from the whole thymus. The NIKaly/+ thymus showed easily detectable expression of tissue-specific Ags, including salivary protein 1, fatty acid binding protein, glutamate decarboxylase 67, and C-reactive protein (Fig. 4,A, left). In contrast, those messages were hardly, if at all, detected in NIKaly/aly thymus. Thus, tissue-specific Ag expression was dramatically reduced in NIKaly/aly thymus. As expected from normal thymic structure in LT-α−/− mice, LT-α−/− thymus demonstrated indistinguishable levels of AIRE and tissue-specific Ag expression from those of control thymus (Fig. 4 A, right).
Although AIRE has been proposed to regulate promiscuous gene expression of many peripheral tissue-restricted Ags in the thymus (16), it is not clear whether the promiscuity in gene expression is confined to AIRE-expressing cells. It is possible that many tissue-specific Ags are also expressed in medullary epithelial cells which do not express AIRE, and NIK may affect the development of such cell types as well, resulting in the dramatic reduction of tissue-specific Ag expression. To investigate this, we first adjusted the amount of cDNA to normalize the AIRE expression level between NIKaly/+ thymus and NIKaly/aly thymus (Fig. 4,B). In this experimental setting, we detected much stronger housekeeping gene (β-actin) expression from NIKaly/aly thymus cDNA compared with NIKaly/+ thymus cDNA (Fig. 4,B). Given that AIRE is predominantly expressed by the thymic medulla (16, 18) (from both UEA-1+ and UEA-1− cells) (15), this result supports that AIRE-expressing medullary epithelial cells are fewer in NIKaly/aly thymus compared with NIKaly/+ thymus. We then tested the expression of peripheral tissue-restricted Ags by these AIRE-normalized cDNA samples, which should represent RNAs derived from roughly equal amounts of AIRE-expressing cells from both NIKaly/+ thymus and NIKaly/aly thymus. Expression of peripheral tissue-restricted Ags was still weaker in NIKaly/aly thymus compared with NIKaly/+ thymus (Fig. 4 B); by real-time PCR, the salivary protein 1 expression level from NIKaly/aly thymus was less than one third of that from NIKaly/+ thymus when the AIRE expression level was normalized (N. Kuroda, H. Han, and M. Matsumoto, unpublished observation). These results suggest that reduced AIRE expression in NIKaly/aly thymus alone does not fully account for the reduced expression of peripheral tissue-restricted Ags. Rather, our results suggest the existence of AIRE-negative thymic epithelial cells which equally play important roles in the expression of peripheral tissue-restricted Ags; development of such cell types is defective in NIKaly/aly mice, as observed for the development of AIRE-expressing cells.
We have demonstrated that NIK plays an essential role in the organization of the thymic microenvironment that is required for the establishment of self-tolerance. The breakdown of self-tolerance in the absence of normal NIK in thymic stroma involved the processes for the production of immunoregulatory T cells, and the defect was rescued by the exogenous supply of immunoregulatory T cells. NIK was originally identified as a kinase required for NF-κB activation induced by a wide variety of ligand binding (2). It is now clear, however, that the requirement for NIK for NF-κB activation is strictly signal-dependent; NF-κB activation induced by TNF-α takes place without NIK, whereas NIK is essential for NF-κB activation downstream of the LTβR (7, 33). The NIK-related signaling pathway(s) involved in the establishment of self-tolerance in a thymic-stroma dependent fashion are presently unknown. Because LT-α−/− mice possess normal thymic architecture with normal expression of AIRE and peripheral tissue-Ags, and grafting of the embryonic thymus from LT-α−/− mice (which had been developed in the absence of LT in its early ontogeny) onto BALB/cnu/nu mice did not induce any inflammatory changes in the liver and pancreas of recipient mice (F. Kajiura, K. Izumi, Y. Bando, and M. Matsumoto, unpublished observation), it is reasonable to speculate that membrane-bound LT is not solely responsible for this action. Given that signals through LTβR control thymic organogenesis (31), LIGHT and/or other unidentified LTβR ligand(s) might be responsible for the phenotypes described in the present study. However, it remains also possible that NIK is acting downstream of other receptor(s) beyond LTβR in this process. Of note, the pathway(s) regulate the processes involved in the generation of p52 from a precursor p100 (Fig. 1 C), as observed for signals through LTβR (34, 35, 36, 37).
We have established that NIK is required for the production of immunoregulatory T cells; this was demonstrated both by the reduction of CD4+CD25+ T cell numbers and by the reduced expression of Foxp3, a newly identified functional marker for immunoregulatory T cells (26, 27, 28). Supporting this, an exogenous supply of sufficient numbers of immunoregulatory T cells was able to rescue the autoimmune disease phenotype not only in NIKaly/aly thymus-grafted BALB/cnu/nu mice (Fig. 2 C), but also in NIKaly/aly mice themselves (M. Minami, M. Nakazawa, and C. Tamura, personal communication). It will be important to study the nature of the TCR ligands and the mechanisms involved in the NIK-dependent production of immunoregulatory T cells. It has been proposed that immunoregulatory T cells may arise from relatively high avidity interactions with self-peptide-MHC complexes just below the threshold for negative selection in the thymus (38, 39). Accordingly, thymic stroma which have been developed in the absence of normal NIK may not be able to present TCR ligands (most likely containing self-peptides) efficiently enough, resulting in insufficient avidity for the production of immunoregulatory T cells.
Given that alteration in the pattern of gene promiscuity in the thymic stroma affects the production of immunoregulatory T cells, the process for negative selection might also be affected in NIKaly/aly mice. We have tested this possibility by the use of a TCR transgenic model specific for male H-Y Ag (14). Unexpectedly, elimination of autoreactive T cells clonotypic for H-Y TCR in the male was not obviously changed in NIKaly/aly mice crossed with H-Y TCR-expressing transgenic mice (F. Kajiura and M. Matsumoto, unpublished observation). However, the possibility remains that this ubiquitous H-Y Ag-specific TCR experimental model may not be relevant for assessment of the negative selection-inducing activity of thymic stromal cells for peripheral Ags. In fact, the following observation suggests the probable involvement of NIK in the thymic stroma in the negative selection process. When NIKaly/+ thymus and NIKaly/aly thymus were grafted simultaneously onto BALB/cnu/nu mice, the development of inflammatory lesions was not completely inhibited (Fig. 2 C). This result suggests that impaired production of immunoregulatory T cells is not the only defect, but is accompanied by the existence of a dominant process for the breakdown of self-tolerance determined by the thymic stroma without normal NIK. In this scenario, it is possible that the grafted NIKaly/aly thymus allows production of more pathogenic autoreactive T cells in the recipient mice than can be controlled by the immunoregulatory T cells that are produced by the grafted NIKaly/+ thymus.
Although the exact mechanism by which NIK regulates the thymic microenvironment that is required for the establishment of central tolerance is unknown, the disorganized thymic structure, together with reduced AIRE expression in mice with a mutation disrupting the RelB gene merits attention (40, 41). Because of the phenotypic similarities between NIKaly/aly and RelB−/− mice (42) (multi-inflammatory lesions together with the absence of UEA-1+ medullary epithelial cells in the thymus), we speculate that NIK regulates the thymic microenvironment through the activation of the NF-κB complex containing RelB. Based on the findings that production of p52 is impaired in NIKaly/aly thymic stroma (Fig. 1 C) and that NIK has been demonstrated necessary for the production of NK T cells through the action of RelB (43, 44), we further speculate that the NIK-related signaling pathway(s) activate the NF-κB complex in thymic stroma mainly consisting of p52/RelB heterodimers to generate the thymic microenvironment that is necessary for the establishment of self-tolerance. The molecular link between the NF-κB activation pathway and the regulatory mechanisms of the thymic microenvironment as well as of autoimmunity requires further study.
We have demonstrated that AIRE expression is significantly reduced in NIKaly/aly thymus. Although we suggest that a developmental effect of mutated NIK on thymic epithelial cells expressing AIRE (i.e., fewer medullary epithelial cells) seems to be the likely explanation, we cannot exclude the possibility that reduced AIRE expression in NIKaly/aly thymus is due to down-regulation of AIRE by the mutated NIK through a transcriptional mechanism. The latter possibility needs to be tested in future work.
Promiscuous gene expression of many peripheral tissue-restricted Ags in the thymic epithelial cells play essential roles in the establishment of self-tolerance, and AIRE has been proposed to be essential in this process (15, 16). We have demonstrated that promiscuous gene expression is dramatically reduced in the NIKaly/aly thymus, and we speculate that developmental effect of NIK on thymic epithelial cell components (including both AIRE-expressing cells and AIRE-negative cells) is responsible for the reduced gene expression of many peripheral tissue-restricted Ags in NIKaly/aly thymus. In contrast, AIRE seems to regulate expression of peripheral tissue-restricted Ags predominantly through a transcriptional mechanism, because the thymic structure is apparently unaffected by the absence of AIRE (16, 17); this possibility might be consistent with the demonstration that AIRE can interact with CREB-binding protein/p300 (45). Thus, there must be a group of genes that together control promiscuous gene expression in the thymus through their unique actions, as exemplified by NIK and AIRE. Therefore, it is critical to determine whether the reduced promiscuous gene expression in the thymus and the development of autoimmune diseases are directly linked in future study.
Finally, although both impaired production of immunoregulatory T cells and possibly an impaired negative selection process contribute to the development of autoimmunity in NIKaly/aly mice, the exogenous supply of sufficient numbers of immunoregulatory T cells was able to rescue the autoimmune disease phenotype, supporting the therapeutic benefits of immunoregulatory T cells. With the advent of thymic organogenesis using thymic precursor cells (46, 47), it may be feasible to manipulate the thymic microenvironment, thereby controlling the processes for the establishment of self-tolerance.
We thank Dr. W. Van Ewijk for mAb ER-TR5, and Drs. K. Iwabuchi, P. Peterson, S. Hori, E. Nishimura, N. Ishimaru, M. Minami, M. Nakazawa, and C. Tamura for valuable suggestions. We also thank K. Awahayashi and F. Saito for technical assistance.
This work was supported in part by Special Coordination Funds of the Ministry of Education, Culture, Sports, Science, and Technology, by a grant-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology, and by the Novartis Foundation.
Abbreviations used in this paper: NIK, NF-κB-inducing kinase; IKK, IκB kinase; aly, alymphoplasia; LT, lymphotoxin; AIRE, autoimmune regulator; PNA, peanut agglutinin; AP, alkaline phosphatase; ANA, antinuclear Ab; BM, bone marrow; UEA-1, Ulex europaeus agglutinin 1; 2-DG, 2′-deoxyguanosine; HPRT, hypoxanthine phosphoribosyltransferase.