“Promiscuous” thymic expression of peripheral autoantigens can contribute to immunological tolerance in some cases. However, in this study we show that thymic mRNA expression alone cannot predict a contribution to thymic tolerance. Autoimmune gastritis is caused by CD4+ T cells directed to the α (H/Kα) and β (H/Kβ) subunits of the gastric membrane protein the H+/K+ ATPase. H/Kα mRNA is expressed in the thymus, but H/Kβ expression is barely detectable. In this study, we demonstrate that thymic H/Kα in wild-type mice or mice that overexpressed H/Kα did not result in negative selection of pathogenic anti-H/Kα T cells. However, negative selection of anti-H/Kα T cells did occur if H/Kβ was artificially overexpressed in the thymus. Given that H/Kα cannot be exported from the endoplasmic reticulum and is rapidly degraded in the absence of H/Kβ, we conclude that H/Kα epitopes are unable to access MHC class II loading compartments in cells of the normal thymus. This work, taken together with our previous studies, highlights that thymic autoantigen expression does not necessarily result in the induction of tolerance.

Generating tolerance to autoantigens is likely to be a multistep process. There is abundant evidence of T cells being either deleted or rendered inactive in both the thymus and periphery. There is also compelling evidence that regulatory T cells play a role controlling self-reactive T cells that escape deletion or inactivation. The argument for a role for the thymus in tolerance to Ags of peripheral (nonthymic) tissues was strengthened when it became apparent that several Ags, previously considered to be expressed only in peripheral tissues (peripheral Ags), also appeared to be present in the thymus (reviewed in Ref.1). More recently, systematic analysis of gene expression in the thymus has shown that whereas all non-T cell populations in the thymus, including dendritic cells, macrophages, and epithelial cells, express peripheral Ags, the gene expression by medullary epithelial cells is particularly promiscuous (1, 2, 3). Medullary epithelial cells probably express 1200–1300 genes encoding peripheral Ags (2). A number of experiments, often involving thymic transplantation and the production of bone marrow chimeras, have demonstrated that promiscuous thymic expression can contribute to tolerance and the prevention of autoimmune disease (4, 5, 6). Deletion and anergy are among the mechanisms involved (5, 7, 8), and generation of regulatory T cells remains a distinct although unsubstantiated possibility (9).

The product of the Aire gene induces thymic expression of a large number of promiscuously expressed genes in thymic medullary epithelia (10, 11). AIRE is mutated in humans with a polyglandular autoimmune syndrome (12), APS1, and aire-deficient mice display a similar autoimmune condition (10). Using a transgenic system, it has been demonstrated that aire-mediated expression is sufficient to induce thymic tolerance (7).

The gastric H+/K+ ATPase is the major autoantigen targeted in autoimmune gastritis (13, 14). We and others (13, 15, 16, 17, 18, 19) have demonstrated that a CD4+ T cell response to this Ag is necessary and sufficient to induce the disease in mice. T cell responses to the H+/K+ ATPase have also been demonstrated in humans with autoimmune gastritis and pernicious anemia (20). The H+/K+ ATPase is comprised of two subunits, α (H+/K+ ATPase α-subunit encoded by Atp4a; H/Kα) and β (H+/K+ ATPase β-subunit encoded by Atp4b; H/Kβ), and immune responses to both subunits have been detected in individuals with autoimmune gastritis. Expression of the H/Kα gene in humans was readily detected in the thymus in medullary epithelia and dendritic cells (DCs),4 but H/Kβ mRNA was absent from that organ (2). In mice, H/Kα mRNA is readily detected in epithelia and APCs, but H/Kβ expression in the thymus is, at best, barely detectable and appears to be found only in thymic epithelia (1). Microarray analyses of thymic RNA revealed that H/Kα expression levels in aire-deficient mice differed only marginally from wild-type (WT) mice, and H/Kβ mRNA was at the limit of detection (10).

In this paper, we have investigated the role of thymic expression of H/Kα in the induction of tolerance of gastritogenic T cells. Using mice expressing a TCR directed to a peptide in H/Kα and H/Kα-deficient mice (H/Kα−/−), we found that thymic H/Kα did not induce thymic deletion of H/Kα-specific T cells. Transgenic overexpression of H/Kβ in the thymus demonstrated that the failure of H/Kα expressed in the thymus to induce tolerance was due to low levels of H/Kβ.

The A23 TCR transgenic (19), H/Kα−/− (21), H/Kβ−/− (22), IE-H/Kα (23), and IE-H/Kβ (13) mice have been previously described. All strains had been backcrossed at least 10 times to BALB/cCrSlc. A23 mice were intercrossed with H/Kα−/−, IE-H/Kα, and IE-H/Kβ mice in some experiments. All of the experimental animals were maintained in a conventional animal facility at the Departments of Biochemistry or Microbiology, The University of Melbourne. In all experiments, mice between 7 and 13 wk were used. Experiments were approved by The University of Melbourne Animal Experimentation Ethics Committee.

Preparation of cell suspensions, staining with Abs, and flow cytometric analysis was as described (24).

Enriched CD4+ cells were prepared from the lymph nodes of A23 mice. Single-cell suspensions were incubated with Abs directed to CD8, B220, and F4/80, and the Ab-bound cells were removed using anti-rat IgG-coated magnetic beads (Dynal). The resulting population was stained with CFSE as described (25) and injected i.v. into WT or H/Kβ−/− mice. Recipient mice were killed 2 days later, and lymph node cells were stained with PerCP-conjugated anti-CD4, biotinylated anti-Vα2, and streptavidin-APC, and analyzed by flow cytometry.

WT mice were thymectomized at 4–5 wk of age by aspiration of both thymic lobes through a small incision in the sternum. Thymus grafting was performed 2–4 wk later by inserting two thymic lobes from a neonatal thymus under the left kidney capsule. Mice were killed 12 wk later, and the stomachs were removed for histological examination. The absence of thymic remnants was confirmed by autopsy, and cells were recovered from the graft to confirm graft function.

Thymic expression of Ags has the capability to influence thymic-positive or -negative selection of developing T cells. To determine whether thymic expression of H/Kα influences thymic selection of H/Kα-specific T cells, we used the A23 TCR transgenic mouse (19). The Vα2-Vβ2 TCR expressed by these mice was from a T cell clone isolated from mice with autoimmune gastritis. All A23 mice display signs of autoimmune gastritis as early as 10 days after birth, and as few as 103 T cells transferred from A23 mice to athymic recipients was sufficient to induce autoimmune gastritis (19), hence these T cells are highly pathogenic and can respond to endogenous levels of gastric H/Kα. In this study we produced mice that expressed the A23 TCR but were deficient in H/Kα (A23.H/Kα−/− mice). Thymocyte populations of A23, A23.H/Kα+/−, and A23.H/Kα−/− were compared by flow cytometry, and representative flow cytometric profiles are shown in Fig. 1. H/Kα+/− mice carry one defective H/Kα allele and have similar levels of H/Kα protein to WT mice and are phenotypically normal. Profiles of CD4, CD8, and the Vα2 were not significantly different among A23, A23.H/Kα+/−, and A23.H/Kα−/− mice. Very similar levels of cell surface TCR and CD4 were found on thymic T cells in mice of these three genotypes (data not shown). Table I shows the census for the thymocyte populations we examined. There were no significant differences in the number of cells in any thymocyte population among the mice of A23, A23.H/Kα+/−, and A23.H/Kα−/− genotypes including the most mature CD3high and Vα2high cells. Collectively, this data demonstrates that selection of T cells expressing this highly pathogenic anti-H/Kα TCR were not influenced by thymic expression of H/Kα.

FIGURE 1.

Flow cytometric analysis of thymocytes. The thymus was removed from 7- to 8-wk-old mice and analyzed for CD4, CD8, CD3, and Vα2 expression by flow cytometry. Live cells were gated according to the forward and side scatter profiles. Representative scatter plots show the gating for single CD4-positive, single CD8-positive, CD4CD8 double-positive, CD4CD8 double-negative, CD3high, and Vα2high cells for each of the genotypes indicated.

FIGURE 1.

Flow cytometric analysis of thymocytes. The thymus was removed from 7- to 8-wk-old mice and analyzed for CD4, CD8, CD3, and Vα2 expression by flow cytometry. Live cells were gated according to the forward and side scatter profiles. Representative scatter plots show the gating for single CD4-positive, single CD8-positive, CD4CD8 double-positive, CD4CD8 double-negative, CD3high, and Vα2high cells for each of the genotypes indicated.

Close modal
Table I.

Census of thymocytes in A23.H/Kα−/− micea

A23A23.H/Kα+/−A23.H/Kα−/−
n = 4n = 6n = 8
Total 114 ± 13 95.0 ± 42.7 114 ± 74 
CD4CD8 3.2 ± 0.87 2.9 ± 0.5 3.8 ± 1.7 
CD4+CD8+ 90.7 ± 10.4 75.3 ± 36.0 87.5 ± 60.5 
CD8+ 2.5 ± 0.4 3.4 ± 1.9 3.8 ± 1.9 
CD4+ 13.9 ± 2.4 10.3 ± 5.2 13.0 ± 5.4 
CD4+CD3high 11.3 ± 1.6 10.1 ± 4.5 12.7 ± 6.0 
CD4+Vα2high 13.4 ± 2.0 10.6 ± 4.3 13.0 ± 6.2 
A23A23.H/Kα+/−A23.H/Kα−/−
n = 4n = 6n = 8
Total 114 ± 13 95.0 ± 42.7 114 ± 74 
CD4CD8 3.2 ± 0.87 2.9 ± 0.5 3.8 ± 1.7 
CD4+CD8+ 90.7 ± 10.4 75.3 ± 36.0 87.5 ± 60.5 
CD8+ 2.5 ± 0.4 3.4 ± 1.9 3.8 ± 1.9 
CD4+ 13.9 ± 2.4 10.3 ± 5.2 13.0 ± 5.4 
CD4+CD3high 11.3 ± 1.6 10.1 ± 4.5 12.7 ± 6.0 
CD4+Vα2high 13.4 ± 2.0 10.6 ± 4.3 13.0 ± 6.2 
a

The thymus was removed from 7- to 8-wk-old mice, and the number of CD4CD8 double-negative (CD4CD8), CD4CD8 double-positive (CD4+CD8+), CD4 single-positive (CD4+), CD8 single-positive (CD8+), CD4+CD3high, and CD4+Vα2high cells was determined by flow cytometry. Examples of the gating used to define these populations are shown in Fig 1. n = number of mice analyzed. The average number of cells and SD × 10−6 for each cell subset is shown. No significant differences were found in any thymocyte population across genotypes.

The inability to delete A23 T cells in the thymus may have been due to the level of expression of H/Kα being too low. Previously, we constructed a mouse that expresses H/Kα under the control of the I-Eα promoter (IE-H/Kα transgene), and these mice overexpressed H/Kα in the thymus (23). In this study, we analyzed mice expressing both the A23 TCR and the IE-H/Kα transgene (A23.IE-H/Kα; Fig. 1 and Table II). We did not observe clonal deletion or down-regulation of TCR or CD4 (data not shown) on the A23 T cells despite elevated levels of H/Kα expression.

Table II.

Census of thymocytes in A23.IE-H/Kβ micea

WTIEβA23A23.IE-H/KαA23.IE-H/Kβ
n = 7n = 7n = 10n = 3n = 9
Total 106 ± 39 96 ± 37 167 ± 51 196 ± 80 127 ± 36 
CD4CD8 2.3 ± 0.76 3.4 ± 2.0 5.2 ± 1.4 4.6 ± 1.6 4.1 ± 1.4 
CD4+CD8+ 84.5 ± 32 86.7 ± 17 146 ± 48 165 ± 72 115 ± 33 
CD8+ 3.3 ± 1.0 3.5 ± 1.3 3.8 ± 1.4 3.4 ± 1.2 1.9 ± 0.5b 
CD4+ 10.2 ± 3.6 8.2 ± 3.0 13.8 ± 6.1 15.0 ± 3.4 3.5 ± 1.6c 
CD4+CD3high 9.2 ± 2.7 8.6 ± 3.3 14.1 ± 5.6 14.9 ± 3.3 3.4 ± 1.6c 
CD4+Vα2high 0.83 ± 0.37 0.90 ± 0.50 15.3 ± 5.7 16.9 ± 3.4 4.1 ± 2.0c 
WTIEβA23A23.IE-H/KαA23.IE-H/Kβ
n = 7n = 7n = 10n = 3n = 9
Total 106 ± 39 96 ± 37 167 ± 51 196 ± 80 127 ± 36 
CD4CD8 2.3 ± 0.76 3.4 ± 2.0 5.2 ± 1.4 4.6 ± 1.6 4.1 ± 1.4 
CD4+CD8+ 84.5 ± 32 86.7 ± 17 146 ± 48 165 ± 72 115 ± 33 
CD8+ 3.3 ± 1.0 3.5 ± 1.3 3.8 ± 1.4 3.4 ± 1.2 1.9 ± 0.5b 
CD4+ 10.2 ± 3.6 8.2 ± 3.0 13.8 ± 6.1 15.0 ± 3.4 3.5 ± 1.6c 
CD4+CD3high 9.2 ± 2.7 8.6 ± 3.3 14.1 ± 5.6 14.9 ± 3.3 3.4 ± 1.6c 
CD4+Vα2high 0.83 ± 0.37 0.90 ± 0.50 15.3 ± 5.7 16.9 ± 3.4 4.1 ± 2.0c 
a

The thymus was removed from 7- to 8-wk-old mice, and the number of CD4CD8 double-negative (CD4CD8), CD4CD8 double-positive (CD4+CD8+), CD4 single-positive (CD4+), CD8 single-positive (CD8+), CD4+CD3high, and CD4+Vα2high cells was determined by flow cytometry. Examples of the gating used to define these populations are shown in Fig 1. n = number of mice analyzed. The average number of cells and SD × 10−6 for each cell subset is shown.

b

, p = 0.0015, Mann-Whitney.

c

, p < 0.0001, Mann-Whitney.

We and others have previously shown that during biosynthesis, the H/Kβ aids in membrane insertion of H/Kα (26) and that transport of the H/Kα from the endoplasmic reticulum (ER) requires pairing with H/Kβ (22, 27). Because H/Kβ levels in the thymus are very low we reasoned that perhaps H/Kα was unstable, unable to leave the ER, and therefore failed to reach the post-Golgi compartments in which MHC class II molecules are loaded with peptide, thus resulting in low presentation of H/Kα epitopes to developing A23 T cells. To determine whether a lack of H/Kβ reduces H/Kα presentation, we injected CFSE-labeled A23 CD4+ T cells into H/Kβ−/− mice and examined their proliferation in the lymph node draining the stomach, the paragastric lymph node, 2 days later. H/Kβ−/− mice have a similar number of H+/K+ ATPase-expressing parietal cells in the gastric mucosa as WT mice, although levels of H/Kα protein are significantly lower than WT levels due to rapid degradation (22, 28). Far fewer A23 T cells had divided in the H/Kβ−/− nodes compared with WT, demonstrating that presentation of H/Kα is compromised in the absence of H/Kβ (Fig. 2).

FIGURE 2.

Proliferation of A23 T cells is reduced in H/Kβ−/− mice. CFSE-labeled A23 CD4+ T cells were injected i.v. into WT and H/Kβ−/− mice. At 48 h, the paragastric lymph node cells were stained for CD4 and Vα2 and analyzed by flow cytometry. Plots show CFSE staining on CD4+Vα2+ lymphocytes, and are representative of seven mice from three independent experiments. Percentage figures refer to the proportion of CFSE-labeled cells in the gate indicated, which represent the undivided A23 T cells.

FIGURE 2.

Proliferation of A23 T cells is reduced in H/Kβ−/− mice. CFSE-labeled A23 CD4+ T cells were injected i.v. into WT and H/Kβ−/− mice. At 48 h, the paragastric lymph node cells were stained for CD4 and Vα2 and analyzed by flow cytometry. Plots show CFSE staining on CD4+Vα2+ lymphocytes, and are representative of seven mice from three independent experiments. Percentage figures refer to the proportion of CFSE-labeled cells in the gate indicated, which represent the undivided A23 T cells.

Close modal

We next determined whether low levels of expression of thymic H/Kβ would influence selection of A23 T cells. To do this, we produced mice that express the A23 TCR and also overexpress H/Kβ in the thymus under the control of the I-Eα promoter (A23.IE-H/Kβ mice). We have previously shown that the I-Eα promoter drives thymic overexpression of H/Kβ (13). Representative flow cytometric profiles are shown in Fig. 1, and a census of thymic T cell populations is shown in Table II. The numbers of thymic T cell populations in WT and IE-H/Kβ mice were not significantly different, indicating that expression of the IE-H/Kβ transgene does not nonspecifically influence thymocyte development. The number of CD4 single-positive, CD3high, and Vα2high cells were reduced by ∼4-fold in A23.IE-H/Kβ mice compared with A23 mice (Fig. 1 and Table II). The number of CD4CD8 double-negative and CD4CD8 double-positive cells was similar in both genotypes (Table II). This data demonstrates that most T cells expressing the anti-H/Kα TCR are deleted at the CD4 single-positive stage in mice that overexpress H/Kβ.

We also examined the levels of CD4, CD8, and TCR on CD4 single-positive thymocytes and a representative analysis is shown in Fig. 3. The levels of CD4, Vα2, and Vβ2 on the residual CD4 single-positive cells in the thymus of the A23.IE-H/Kβ mice were consistently lower than on the corresponding cells in A23 mice. In mice in which a clear mature CD4+ population could be discerned, the mean fluorescence of CD4 and TCR staining was on average 60 and 74%, respectively, of the levels found in A23 mice.

FIGURE 3.

Down-regulation of CD4 and TCR on thymocytes from A23.IE-H/Kβ mice. The thymus was removed from 7- to 8-wk-old mice of the genotypes indicated and analyzed for CD4, CD8, Vα2, and Vβ2 expression by flow cytometry. Live cells were gated according to the forward and side scatter profiles. Representative scatter plots of CD4 and CD8 levels are shown. The histograms show the levels of CD4, Vα2, and Vβ2 (as indicated) on CD4 single-positive thymocytes from A23 (shaded) or A23.IE-H/Kβ mice (clear). The mean fluorescence intensities for each population are indicated below the histograms.

FIGURE 3.

Down-regulation of CD4 and TCR on thymocytes from A23.IE-H/Kβ mice. The thymus was removed from 7- to 8-wk-old mice of the genotypes indicated and analyzed for CD4, CD8, Vα2, and Vβ2 expression by flow cytometry. Live cells were gated according to the forward and side scatter profiles. Representative scatter plots of CD4 and CD8 levels are shown. The histograms show the levels of CD4, Vα2, and Vβ2 (as indicated) on CD4 single-positive thymocytes from A23 (shaded) or A23.IE-H/Kβ mice (clear). The mean fluorescence intensities for each population are indicated below the histograms.

Close modal

Our experiments with A23.H/Kα−/− mice demonstrated that thymic selection of cells expressing the A23 TCR were not influenced by the presence or absence of H/Kα in the thymus. This experiment only examined a single TCR specificity. To determine whether the absence of H/Kα in the thymus results in the accumulation of pathogenic T cells capable of causing autoimmune gastritis in mice with a polyclonal repertoire, we performed thymus-grafting experiments similar to those described by Avichezer et al. (6). In that work, mice grafted with thymi from interphotoreceptor retinoid-binding protein-deficient mice were highly susceptible to autoimmune uveitis due to the lack of thymic deletion of interphotoreceptor retinoid-binding protein-specific T cells. Thymi from H/Kα−/− or WT mice were grafted under kidney capsule of 6- to 7-wk-old WT mice that had been thymectomized 2 wk earlier. At 12 wk after grafting, stomachs were examined for gastric inflammation, and sera were assayed for anti-gastric autoantibodies, both hallmarks of autoimmune gastritis. All five mice that received thymi from H/Kα−/− mice and three that received thymi from WT mice were free from significant autoimmune pathology. These data suggest that thymic expression of H/Kα was not required for the prevention of autoimmune gastritis in mice with a normal T cell repertoire.

Although there is great interest in the implications of promiscuous thymic expression of self-Ags, to date few studies have formally assessed the outcome of thymic expression of specific Ags on the induction of T cell tolerance. In the cases of Ags responsible for driving experimental autoimmune encephalomyelitis (4) and uveitis (6) and in model Ag systems (5, 7), thymic expression did appear to be capable of inducing tolerance and, in some cases, preventing autoimmunity. In contrast to this data, we did not find that thymic expression of the H/Kα autoantigen was able to induce tolerance.

We and others have shown that mRNA encoding H/Kα is expressed widely in the thymus including in DCs and medullary epithelia that have been demonstrated to mediate negative selection. To determine whether thymic H/Kα was able to mediate deletion of T cells, we crossed A23 mice expressing a highly pathogenic anti-H/Kα TCR with mice deficient in H/Kα. The numbers of the thymocyte populations and the levels of TCR and CD4 in mature thymocytes in A23.H/Kα−/− mice were very similar to A23 mice that express normal levels of H/Kα, indicating that thymic selection events were comparable between the two strains. Because the A23 T cells are clearly able to respond to H/Kα Ag in the periphery, as evidenced by their ability to potently induce gastritis in transgenic or athymic mice, and from the experiments described in Fig. 2, we consider that the lack of central deletion is indicative of an insufficient presentation of H/Kα peptides in the thymus.

H/Kβ is expressed at very low levels in the thymus compared with H/Kα (23). Because it is known that in the absence of H/Kβ, H/Kα is trapped in the ER and rapidly degraded (22, 26, 27) we reasoned that the low level of H/Kβ in the thymus may compromise presentation of H/Kα epitopes. To determine whether the absence of H/Kβ indeed had this effect, we examined proliferation of A23 T cells in the lymph node, draining the stomach of H/Kβ−/− mice as a measure of presentation of H/Kα peptide. We found that 2 days after transfer, A23 T cells had undergone up to three divisions in the paragastric lymph node of WT mice, whereas very little cell division had occurred in H/Kβ−/− mice. The situation in the paragastric lymph node may not be directly comparable to the situation in the thymus because the cells responsible for presentation in the lymph node may not be those that synthesize H/Kα; however, this data demonstrates that the absence of H/Kβ is detrimental to Ag presentation.

To demonstrate that limiting levels of H/Kβ were responsible for the poor presentation of H/Kα in the thymus, we examined selection of A23 T cells in A23.IE-H/Kβ mice in which H/Kβ was overexpressed in MHC class II-expressing cells in the thymus. We found in these mice that A23 T cells were efficiently deleted at the single-positive stage of development. The number of mature T cells in the thymi of A23.IE-H/Kb mice was ∼4-fold lower than in A23 mice. Furthermore, the levels of TCR and CD4 on mature thymocytes were lower in A23.IE-H/Kβ mice than A23 mice. This data suggests that the minority of anti-H/Kα T cells that escape deletion in A23.IE-H/Kβ mice do so because they have a lower level of TCR and coreceptor, or that these molecules are down-regulated. As mentioned above, we did not see differences in TCR or CD4 levels in A23 mice relative to A23.H/Kα−/− mice, which further supports the finding that the anti-H/Kα T cells in the A23 mouse are not subject to negative selection.

We consider the most likely explanation for these observations is that H/Kα in thymic epithelia and DCs of WT mice is rapidly degraded in the ER, as has been observed in other cell types (22, 26), and thus does not traffic to the post-Golgi compartments in which peptides are generated and loaded into the MHC class II binding cleft (29). Low levels of presentation of H/Kα-derived peptide would thus result. It is also possible that APC derive H/Kα peptides from neighboring cells in the thymus. Stabilizing H/Kα by overexpression of H/Kβ would be expected to facilitate this pathway as well.

Despite substantial thymic deletion of T cells in A23.IE-H/Kβ mice we still noted that all mice of this genotype developed autoimmune gastritis, and the proportion of T cells that expressed the activation markers CD44 and CD69 in peripheral lymph nodes was similar to that observed in A23 mice (data not shown). We think that autoimmune gastritis in A23.IE-H/Kβ mice, despite substantial negative selection, is the result of the high pathogenicity of A23 T cells. Even though only a small number of these T cells escape the thymus, they are still sufficient to cause autoimmunity (19). Whether this would occur in a situation where only a small proportion of T cells expressed such a pathogenic TCR, as would be the case in normal mice, is not known.

At first glance, our data presented in this study indicating that, in WT mice, thymic H/Kα does not cause negative selection conflicts with previous studies of thymically expressed Ags where tolerance was observed. However, the H/Kα may represent a special case where presentation to the immune system does not occur because transport to MHC class II-loading compartments and stability of the H/Kα protein is compromised. Although the situation with the H+/K+ ATPase may be somewhat unusual, most components of multiprotein complexes must be assembled in the ER before export. Hence, it is very likely that many of the ∼1000 proteins that are promiscuously expressed in the thymus will have similarly complex biosynthetic programs.

We have recently demonstrated that thymic expression of the dominant epitope of H/Kβ induced increased positive selection of low-affinity gastritogenic T cells and exacerbated autoimmune gastritis (24). These data taken with the results reported in this study demonstrate that autoantigen expression in the thymus may have disparate outcomes that are dependent on the level and site of expression and the efficiency of processing a particular Ag. Depending on Ag levels and affinity of TCRs, the presence of an autoantigen in cortical epithelia cells may promote positive selection of pathogenic T cells, thus increasing the predisposition to autoimmune disease, whereas expression in medullary epithelia or APC may be more likely to promote deletion (30). Thymic expression may also lead to the generation of regulatory T cells capable of protecting from disease (9, 30). However, if the antigenic peptide is not efficiently delivered to MHC molecules, as is probably the case for H/Kα, then despite high levels of gene expression there may be no affect at all on thymic selection and autoimmunity. Hence, it will be necessary to directly test the functional effect of thymic Ag rather than drawing conclusions from expression data alone.

This work also addresses the question of the relative importance of T cell responses to the individual H+/K+ ATPase subunits in autoimmune gastritis. Previously, we found that overexpression of H/Kβ in the thymus resulted in the abrogation of autoimmune gastritis following neonatal thymectomy (13), which we suggested was due to tolerance of anti-H/Kβ T cells. We also discovered that H/Kα was normally expressed in the thymus (23) and that overexpression of H/Kα had no effect on the onset of autoimmune gastritis (23). Based on these earlier observations, we hypothesized and that anti-H/Kβ T cells were of prime importance in the induction of autoimmune gastritis. However, it is clear that anti-H/Kα T cells are prevalent in mice with autoimmune gastritis and can cause autoimmune disease (18, 19, 31). Our work in this study reconciles these findings. It demonstrates that the expression of H/Kα alone was not able to tolerize anti-H/Kα T cells and that thymic overexpression of H/Kβ may induce negative selection of T cells directed to both H+/K+ ATPase subunits. This scenario would explain why thymic expression of H/Kβ prevented autoimmune gastritis even though anti-H/Kα T cells are at least partly responsible for the disease.

Overall, our data in this study suggest that thymic expression of the H/Kα gene plays no role in the induction of immunological tolerance to the H+/K+ ATPase. A corollary of this conclusion is that events in the periphery of normal mice are responsible for preventing anti-H/Kα T cells from causing autoimmune disease. It is likely that CD4+CD25+ regulatory T cells play a role in tolerance to the H+/K+ ATPase (32). However, our experiments with H/Kβ-deficient mice suggest that T cell-mediated regulation is insufficient to prevent a pathogenic autoimmune response to this protein (33). The availability of mice with TCRs direct to the H+/K+ ATPase subunits (16, 19, 34) and mice deficient in H/Kα and H/Kβ will be of great use in investigating these additional mechanisms of peripheral tolerance to gastric autoantigens.

We thank Max Walker and Troy Taylor for excellent animal husbandry. We are very grateful to Prof. Gary Shull (University of Cincinnati College of Medicine, Cincinnati, OH) for providing the H/Kα−/− mice.

The authors have no financial conflict of interest.

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.

1

This work was supported by the National Health and Medical Research Council of Australia.

4

Abbreviations used in this paper: DC, dendritic cell; WT, wild type; ER, endoplasmic reticulum.

1
Kyewski, B., J. Derbinski, J. Gotter, L. Klein.
2002
. Promiscuous gene expression and central T-cell tolerance: more than meets the eye.
Trends Immunol.
23
:
364
.-371.
2
Gotter, J., B. Brors, M. Hergenhahn, B. Kyewski.
2004
. Medullary epithelial cells of the human thymus express a highly diverse selection of tissue-specific genes colocalized in chromosomal clusters.
J. Exp. Med.
199
:
155
.-166.
3
Derbinski, J., A. Schulte, B. Kyewski, L. Klein.
2001
. Promiscuous gene expression in medullary thymic epithelial cells mirrors the peripheral self.
Nat. Immunol.
2
:
1032
.-1039.
4
Klein, L., M. Klugmann, K. A. Nave, V. K. Tuohy, B. Kyewski.
2000
. Shaping of the autoreactive T-cell repertoire by a splice variant of self protein expressed in thymic epithelial cells.
Nat. Med.
6
:
56
.-61.
5
Klein, L., B. Roettinger, B. Kyewski.
2001
. Sampling of complementing self-antigen pools by thymic stromal cells maximizes the scope of central T cell tolerance.
Eur. J. Immunol.
31
:
2476
.-2486.
6
Avichezer, D., R. S. Grajewski, C.-C. Chan, M. J. Mattapallil, P. B. Silver, J. A. Raber, G. I. Liou, B. Wiggert, G. M. Lewis, L. A. Donoso, R. R. Caspi.
2003
. An immunologically privileged retinal antigen elicits tolerance: major role for central selection mechanisms.
J. Exp. Med.
198
:
1665
.-1676.
7
Liston, A., S. Lesage, J. Wilson, L. Peltonen, C. C. Goodnow.
2003
. Aire regulates negative selection of organ-specific T cells.
Nat Immunol.
4
:
350
.-354.
8
Hoffmann, M. W., J. Allison, J. F. Miller.
1992
. Tolerance induction by thymic medullary epithelium.
Proc. Natl. Acad. Sci. USA
89
:
2526
.-2530.
9
Jordan, M. S., A. Boesteanu, A. J. Reed, A. L. Petrone, A. E. Holenbeck, M. A. Lerman, A. Naji, A. J. Caton.
2001
. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide.
Nat. Immunol.
2
:
301
.-306.
10
Anderson, M. S., E. S. Venanzi, L. Klein, Z. Chen, S. P. Berzins, S. J. Turley, H. von Boehmer, R. Bronson, A. Dierich, C. Benoist, D. Mathis.
2002
. Projection of an immunological self shadow within the thymus by the aire protein.
Science
298
:
1395
.-1401.
11
Johnnidis, J. B., E. S. Venanzi, D. J. Taxman, J. P. Ting, C. O. Benoist, D. J. Mathis.
2005
. Chromosomal clustering of genes controlled by the aire transcription factor.
Proc. Natl. Acad. Sci. USA
102
:
7233
.-7238.
12
Peterson, P., K. Nagamine, H. Scott, M. Heino, J. Kudoh, N. Shimizu, S. E. Antonarakis, K. J. Krohn.
1998
. APECED: a monogenic autoimmune disease providing new clues to self-tolerance.
Immunol. Today
19
:
384
.-386.
13
Alderuccio, F., B. H. Toh, S. S. Tan, P. A. Gleeson, I. R. van Driel.
1993
. An autoimmune disease with multiple molecular targets abrogated by the transgenic expression of a single autoantigen in the thymus.
J. Exp. Med.
178
:
419
.-426.
14
van Driel, I. R., A. G. Baxter, K. L. Laurie, T. D. Zwar, N. L. La Gruta, L. M. Judd, K. L. Scarff, P. A. Silveira, P. A. Gleeson.
2002
. Immunopathogenesis, loss of T cell tolerance and genetics of autoimmune gastritis.
Autoimmun. Rev.
1
:
290
.-297.
15
De Silva, H. D., P. A. Gleeson, B. H. Toh, I. R. van Driel, F. R. Carbone.
1999
. Identification of a gastritogenic epitope of the H/K ATPase β-subunit.
Immunology
96
:
145
.-151.
16
Alderuccio, F., V. Cataldo, I. R. van Driel, P. A. Gleeson, B. H. Toh.
2000
. Tolerance and autoimmunity to a gastritogenic peptide in TCR transgenic mice.
Int. Immunol.
12
:
343
.-352.
17
Suri-Payer, E., P. Kehn, A. Cheever, E. Shevach.
1996
. Pathogenesis of post-thymectomy autoimmune gastritis: identification of anti-H/K adenosine triphosphatase-reactive T cells.
J. Immunol.
157
:
1799
.-1805.
18
Suri-Payer, E., A. Z. Amar, R. McHugh, K. Natarajan, D. H. Margulies, E. M. Shevach.
1999
. Post-thymectomy autoimmune gastritis: fine specificity and pathogenicity of anti-H/K ATPase-reactive T cells.
Eur. J. Immunol.
29
:
669
.-677.
19
McHugh, R. S., E. M. Shevach, D. H. Margulies, K. Natarajan.
2001
. A T cell receptor transgenic model of severe, spontaneous organ-specific autoimmunity.
Eur. J. Immunol.
31
:
2094
.-2103.
20
Bergman, M. P., A. Amedei, M. M. D’Elios, A. Azzurri, M. Benagiano, C. Tamburini, R. van der Zee, C. M. Vandenbroucke-Grauls, B. J. Appelmelk, G. Del Prete.
2003
. Characterization of H+, K+-ATPase T cell epitopes in human autoimmune gastritis.
Eur. J. Immunol.
33
:
539
.-545.
21
Spicer, Z., M. L. Miller, A. Andringa, T. M. Riddle, J. J. Duffy, T. Doetschman, G. E. Shull.
2000
. Stomachs of mice lacking the gastric H, K-ATPase α-subunit have achlorhydria, abnormal parietal cells, and ciliated metaplasia.
J. Biol. Chem.
275
:
21555
.-21565.
22
Scarff, K. L., L. M. Judd, B. H. Toh, P. A. Gleeson, I. R. van Driel.
1999
. Gastric H+, K+-adenosine triphosphatase β subunit is required for normal function, development, and membrane structure of mouse parietal cells.
Gastroenterology
117
:
605
.-618.
23
Alderuccio, F., P. A. Gleeson, S. P. Berzins, M. Martin, I. R. van Driel, B. H. Toh.
1997
. Expression of the gastric H/K-ATPase α-subunit in the thymus may explain the dominant role of the β-subunit in the pathogenesis of autoimmune gastritis.
Autoimmunity
25
:
167
.-175.
24
Laurie, K. L., N. L. La Gruta, N. Koch, I. R. van Driel, P. A. Gleeson.
2004
. Thymic expression of a gastritogenic epitope results in positive selection of self-reactive pathogenic T cells.
J. Immunol.
172
:
5994
.-6002.
25
Marzo, A. L., B. F. Kinnear, R. A. Lake, J. J. Frelinger, E. J. Collins, B. W. Robinson, B. Scott.
2000
. Tumor-specific CD4+ T cells have a major “post-licensing” role in CTL mediated anti-tumor immunity.
J. Immunol.
165
:
6047
.-6055.
26
Beggah, A. T., P. Beguin, K. Bamberg, G. Sachs, K. Geering.
1999
. β-Sub-unitassembly is essential for the correct packing and the stable membrane insertion of the H, K-ATPase α-subunit.
J. Biol. Chem.
274
:
8217
.-8223.
27
Gottardi, C. J., M. J. Caplan.
1993
. Molecular requirements for the cell-surface expression of multisubunit ion-transporting ATPases: identification of protein domains that participate in Na, K-ATPase and H, K-ATPase subunit assembly. [Published erratum appears in 1993 J. Biol. Chem. 268: 25260.].
J. Biol. Chem.
268
:
14342
.-14347.
28
Franic, T. V., L. M. Judd, D. Robinson, S. P. Barrett, K. L. Scarff, P. A. Gleeson, L. C. Samuelson, I. R. van Driel.
2001
. Regulation of gastric epithelial cell development revealed in H+/K+-ATPase β-subunit- and gastrin-deficient mice.
Am. J. Physiol. Gastrointest. Liver Physiol.
281
:
G1502
.-G1511.
29
Trombetta, E. S., I. Mellman.
2005
. Cell biology of antigen processing in vitro and in vivo.
Annu. Rev. Immunol.
23
:
975
.-1028.
30
Starr, T. K., S. C. Jameson, K. A. Hogquist.
2003
. Positive and negative selection of T cells.
Annu. Rev. Immunol.
21
:
139
.-176.
31
Nishio, A., M. Hosono, Y. Watanabe, M. Sakai, M. Okuma, T. Masuda.
1994
. A conserved epitope on H+, K+-adenosine triphosphatase of parietal cells discerned by a murine gastritogenic T-cell clone.
Gastroenterology
107
:
1408
.-1414.
32
Suri-Payer, E., A. Z. Amar, A. M. Thornton, E. M. Shevach.
1998
. CD4+CD25+ T cells inhibit both the induction and effector function of autoreactive T cells and represent a unique lineage of immunoregulatory cells.
J. Immunol.
160
:
1212
.-1218.
33
Laurie, K. L., I. R. van Driel, T. D. Zwar, S. P. Barrett, P. A. Gleeson.
2002
. Endogenous H/K ATPase β-subunit promotes T cell tolerance to the immunodominant gastritogenic determinant.
J. Immunol.
169
:
2361
.-2367.
34
Candon, S., R. S. McHugh, G. Foucras, K. Natarajan, E. M. Shevach, D. H. Margulies.
2004
. Spontaneous organ-specific Th2-mediated autoimmunity in TCR transgenic mice.
J. Immunol.
172
:
2917
.-2924.