Using a previously described human keratin 14 (K14) promoter, we created mice expressing a peptide Ag (OVAp) in epithelial cells of the skin, tongue, esophagus, and thymus. Double transgenic mice that also express a TCR specific for this Ag (OT-I) showed evidence for Ag-driven receptor editing in the thymus. Surprisingly, such mice exhibited a severe autoimmune disease. In this work we describe the features of this disease and demonstrate that it is dependent on CD8 T cells. Consistent with the Ag expression pattern dictated by the human K14 promoter, an inflammatory infiltrate was observed in skin and esophagus and around bile ducts of the liver. We also observed a high level of TNF-α in the serum. Given that Ag expression in the thymus induced development of T cells with dual TCR reactivity, and that dual-reactive cells have been suggested to have autoimmune potential, we tested whether they were a causal factor in the disease observed here. We found that OT-I/K14-OVAp animals on a recombinase-activating gene-deficient background still suffered from disease. In addition, OT-I animals expressing OVA broadly in all tissues under a different promoter did not experience disease, despite having a similar number of dual-specific T cells. Thus, in this model it would appear that dual-reactive T cells do not underlie autoimmune pathology. Finally, we extended these observations to a second transgenic system involving 2C TCR-transgenic animals expressing the SIY peptide Ag with the hK14 promoter. We discuss the potential relationship between autoimmunity and self-Ags that are expressed in stratified epithelium.

Tolerance reflects the ability of the immune system to distinguish between healthy and infected tissue. For T cells, an important mechanism to establish tolerance occurs during development in the thymus. In this study, developing T cells that express a self-Ag-reactive TCR are induced to die via apoptosis, thereby eliminating dangerous T cell clones before they complete development (1). At least two mechanisms exist to ensure that a wide variety of self-Ags are presented to T cells in the thymus. First, bone marrow-derived APCs can cross-present exogenous Ags that are not synthesized within thymic APC (2). Second, medullary epithelial cells are apparently specialized for promiscuous expression of otherwise tissue-specific gene products (3). Peptides from these are presented to developing T cells and induce immune tolerance. Nonetheless, in both humans and mice, autoreactive T cells escape deletion in the thymus and circulate through the body. Their autoaggressive potential after maturation in the thymus is further controlled by peripheral tolerance mechanisms (4).

There are numerous ways in which T cells with autoreactive receptors could escape clonal deletion in the thymus. T cells can modulate the level of surface molecules, such as the TCR, coreceptor, and CD5, to reduce their sensitivity and avoid deletion (5, 6, 7). T cells can also avoid deletion in the thymus by secondary TCR gene rearrangement, which occurs during allelic inclusion (8) and receptor editing (9). When the secondary gene rearrangement event occurs on the same allele as the primary, the primary TCR chain is replaced. When it occurs on the alternate allele it results in the synthesis of two TCR chains. In this later case, the second chain can compete with the first during pairing and assembly and thereby alter the T cell reactivity. T cells with two productive rearrangements of the same TCR chain are referred to as dual-reactive, even though one specificity may predominate.

Although receptor editing is viewed as a tolerance mechanism, dual-reactive T cells have been shown to have autoimmune potential. In one model, dual-reactive CD4+ T cells were observed in mice that expressed a TCR transgene specific for an influenza hemagglutinin along with the hemagglutinin Ag (10). The Ag was expressed in hemopoietic cells with the Ig κ promoter and in the pancreas with the rat insulin promoter. Although the majority of hemagglutinin-specific T cells were deleted in the thymus, CD4+ T cells that expressed the transgenic TCR were detected in the peripheral lymphoid organs and the pancreas. While these T cells were anergic when tested in vitro, 25% of the mice developed mild diabetes. Adoptive transfer experiments demonstrated that it was the CD4+ T cells that expressed low levels of the transgenic α-chain along with an endogenously rearranged α-chain that induced diabetes. In another system, constitutive expression of two transgenic TCRs allowed T cells to escape deletion in the thymus (11). These cells were tolerant in the periphery but could be activated in vitro through either TCR. These experiments suggest that dual-reactive T cells may play a dominant role in induction of spontaneous autoimmune disease (12).

We previously reported a TCR-transgenic strain where expression of self-Ag in the thymus resulted in receptor editing. These mice expressed the OT-I TCR, specific for a peptide epitope from OVA in the context of the Kb MHC molecule. When the OVA peptide Ag was expressed in the thymus under the human keratin 14 (K14)3 promoter, only a modest reduction in thymic CD4+CD8+ double positive cells was observed. TCRα gene rearrangement was highly elevated in these precursors, and the predominant T cell population that matured expressed both the transgenic receptor and an endogenous TCRα chain. Surprisingly, OT-I/K14-OVAp animals exhibited a lethal disease between 2 and 6 wk of age. In this report we describe the features of this disease and provide evidence that it is a CD8-mediated autoimmune disease with significant manifestation in the skin. Because of the large dual-reactive T cell population in these mice, and because such cells have previously been reported to have autoimmune potential, we tested the hypothesis that dual-reactive T cells play a dominant role in spontaneous disease induction in this model.

C57BL/6 (B6) and RAG1null.B6 mice were obtained from The Jackson Laboratory (Bar Harbor, ME). OT-I mice express a transgenic receptor specific for the OVA257–264 peptide (OVAp) in the context of the H-2Kb (13). 2C mice (14) express an alloreactive receptor that also has reactivity to a synthetic peptide (SIYRYYGL), in the context of H-2Kb (15). K14-OVAp and K14-SIYp transgenic constructs were generated using a multistep PCR procedure as previously described (16). Act-mOVA mice were obtained from B. Ehst and M. Jenkins (University of Minnesota, Minneapolis, MN). These mice express a transmembrane form of OVA under control of the β-actin promoter and CMV immediate-early enhancer. OVA could be detected on the surface of all white blood cells by flow cytometry with an OVA-specific Ab (Sigma-Aldrich, St. Louis, MO) and in all tissues by immunohistochemistry (data not shown). All four major thymic APCs (cortical and medullary epithelial cells, dendritic cells, and macrophages) from Act-mOVA mice were able to stimulate OT-I T cells (data not shown). All mice were treated in accordance with federal guidelines approved by the University of Minnesota Institutional Animal Care and Use Committee.

Thymus, lymph nodes, and spleen were harvested from the indicated mice between 2 and 8 wk of age and stained with Abs to CD4 (L3T4, RMA-5), CD8 (53-6.7), CD69 (H1.2F3), Thy 1.1 (HIS51), and TCRβ (H57-597), all obtained from BD PharMingen (San Diego, CA). Vβ5 (MR9.1)- and Vα2 (B20.1.1)-specific Abs were used to detect the OT-I transgene, although it should be noted that these Abs also recognize endogenous Vβ5 and Vα2 TCR chains. A clonotype-specific Ab (1B2) was used to detect the 2C receptor along with an Ab to Vβ8 (F23.1). Data were collected using a FACSCalibur (BD Biosciences, San Jose, CA) and analyzed with FlowJo software (TreeStar, San Carlos, CA).

Tissues from OT-I/K14-OVAp or OT-I control animals were arranged in aluminum foil cups, snap-frozen in liquid nitrogen, and stored at −80°C. Cryosections were cut 5 mm thick, mounted onto glass slides, fixed for 5 min in acetone, and stained with H&E.

Serum was tested in duplicate with TNF-α or IL-1β ELISA kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s protocol. A standard curve was generated with known amounts of cytokine provided in the kit, and the values in the serum were extrapolated from this.

Mice were injected i.p. with depleting Abs to CD8α (2.43), CD8β (H35-17), or rat Ig beginning at day 1 after birth and continuing every other day until 6 wk of age. An initial dose of 50 μg was increased to 100 μg after 2 wk of age and to 150 μg after 4 wk. A neutralizing Ab to TNF-α (TN3-19.12; a gift of K. Sheehan, Washington University, St. Louis, MO) was injected i.p. every fourth day from birth to wk 1 (50 μg), from wk 1 to wk 3 (100 μg), and from wk 3 to wk 6 (150 μg). All mice were weighed and monitored for signs of disease every other day. Mice remaining at the end of wk 6 were sacrificed. Serum was taken for analysis as described above. Lymph node and spleen were analyzed for T cell depletion by flow cytometry. Depletion >95% was observed in all animals.

We studied TCR-transgenic mice that expressed the antigenic peptide ligand in the thymus, skin, and esophagus (9). The transgenic TCR was the OT-I TCR that recognizes a peptide from chicken OVA (OVAp) in the context of the MHC class I molecule, Kb. The antigenic peptide, OVAp, was also expressed as a transgene under the control of a human keratinocyte-specific promoter, K14. This promoter directs expression in epithelial cells of the thymus as well as the skin, esophagus, and tongue (17). We previously reported that in double transgenic (OT-I/K14-OVAp) mice the presence of the antigenic peptide in the thymus did not delete the OT-I transgenic T cells. On the contrary, extensive endogenous TCRα gene rearrangement occurred (9). This receptor editing was required for efficient development of mature T cells. Thus, a significant number of CD4 and CD8 T cells were present in periphery of OT-I/K14-OVAp mice (Table I). These cells expressed the transgenic TCRβ chain (Vβ5) at a high level and the transgenic α-chain (Vα2) at a reduced level (9), implying the usage of an endogenous TCRα chains. Interestingly, the level of Vα2 on CD8 T cells from the spleen or lymph nodes of OT/K14-OVAp mice was consistently higher than the level of Vα2 on mature CD8 thymocytes in the same animals (Fig. 1 A). In addition, the cell surface phenotype of CD8 T cells from OT-I/K14-OVAp animals differed from control OT-I animals, being CD44high, CD62Llow, and CD69+ (9). Together these features suggest that peripheral CD8 T cells in OT-I/K14-OVAp mice were activated.

Table I.

T lymphocyte cellularity in various strains of micea

StrainCD8CD4Total CD4 + CD8
B6 6.7 ± 1.1 9.7 ± 1.8 16.4 
OT-I 14.6 ± 4.9 1.95 ± .77 16.6 
OT-I/K14-OVAp 1.28 ± .38 4.19 ± 1.9 5.5 
OT-I/Act-mOVA 4.1 ± 4.7 4.3 ± 5.7 8.4 
OT-I/RAGnull 7.30 ± 5.57 0.12 ± .08 7.4 
OT-I/K14-OVAp RAGnull 0.33 ± .27 0.06 ± .05 0.4 
2C 2.8 ± 2.1 0.8 ± 0.5 3.2 
2C/K14-SIY 2.1 ± 1.1 4.4 ± 5.0 6.5 
StrainCD8CD4Total CD4 + CD8
B6 6.7 ± 1.1 9.7 ± 1.8 16.4 
OT-I 14.6 ± 4.9 1.95 ± .77 16.6 
OT-I/K14-OVAp 1.28 ± .38 4.19 ± 1.9 5.5 
OT-I/Act-mOVA 4.1 ± 4.7 4.3 ± 5.7 8.4 
OT-I/RAGnull 7.30 ± 5.57 0.12 ± .08 7.4 
OT-I/K14-OVAp RAGnull 0.33 ± .27 0.06 ± .05 0.4 
2C 2.8 ± 2.1 0.8 ± 0.5 3.2 
2C/K14-SIY 2.1 ± 1.1 4.4 ± 5.0 6.5 
a

The average number of CD4 or CD8 T cells (×106) from spleens of various transgenic mice is indicated, along with the SD. The data are from 8–16 adult animals per group for B6 or OT-I strains. The data are from six 2-wk-old 2C or 2C/K14-SIY mice.

FIGURE 1.

T cell phenotype in lymph nodes of various transgenic strains. A, Lymph node cells from various adult transgenic mice were stained for CD4, CD8, Vα2, and CD69. Histograms show the level of Vα2 or CD69 on the CD8-positive cells. For comparison, the level of Vα2 on CD8 single positives in the thymus of the double transgenic is also included. B, Lymph nodes from OT-I RAGnull and the OT-I/K14-OVAp RAGnull mice were stained as above. These cells were also stained for Thy 1 due to a high number of cells in the OT-I/K14-OVAp RAGnull lymph nodes that were not T cells. C, Spleens from 2C or 2C/K14-SIY mice at 2 wk of age were harvested and stained as in A.

FIGURE 1.

T cell phenotype in lymph nodes of various transgenic strains. A, Lymph node cells from various adult transgenic mice were stained for CD4, CD8, Vα2, and CD69. Histograms show the level of Vα2 or CD69 on the CD8-positive cells. For comparison, the level of Vα2 on CD8 single positives in the thymus of the double transgenic is also included. B, Lymph nodes from OT-I RAGnull and the OT-I/K14-OVAp RAGnull mice were stained as above. These cells were also stained for Thy 1 due to a high number of cells in the OT-I/K14-OVAp RAGnull lymph nodes that were not T cells. C, Spleens from 2C or 2C/K14-SIY mice at 2 wk of age were harvested and stained as in A.

Close modal

OT-I/K14-OVAp double transgenic animals experienced a lethal disease. The outward signs included hair loss, skin lesions, weight loss, and a hunched appearance (Fig. 2). Disease was not observed in OT-I or K14-OVAp single transgenic littermates (Fig. 3,A and data not shown). Onset occurred at an average age of 2–4 wk and was incompletely penetrant, with 20% of animals showing no signs of disease (Fig. 3,A). Animals with outward signs of the disease showed, on average, higher levels of TNF-α in the serum (Fig. 4). In addition, the level of TNF-α in the serum was elevated in a few of the double transgenic mice that were not acutely ill. However, there were no major differences in cellular composition or T cell surface phenotype of lymph node and spleen cells from acutely sick double transgenics and age-matched littermates that showed no signs of disease (data not shown).

FIGURE 2.

Disease in OT-I/K14-OVAp mice. Photograph was taken at 4 wk of age. The OT-I/K14-OVAp mouse is on the left next to an OT-I control littermate.

FIGURE 2.

Disease in OT-I/K14-OVAp mice. Photograph was taken at 4 wk of age. The OT-I/K14-OVAp mouse is on the left next to an OT-I control littermate.

Close modal
FIGURE 3.

Percentage of survival of double transgenic mice. Mice were monitored every other day for signs of disease until 12 wk of age. The age at which the mice died or were euthanized due to severe disease was recorded and plotted as the percentage of mice that were alive at each time point. The number of mice in each group is indicated in parentheses.

FIGURE 3.

Percentage of survival of double transgenic mice. Mice were monitored every other day for signs of disease until 12 wk of age. The age at which the mice died or were euthanized due to severe disease was recorded and plotted as the percentage of mice that were alive at each time point. The number of mice in each group is indicated in parentheses.

Close modal
FIGURE 4.

High levels of TNF in the serum of the double transgenic mice. Mice were euthanized between 1 and 12 wk of age. A, Serum TNF-α was measured using an ELISA. The SDs were 8, 43, 151, and 53 for OT-I, OT-I/K14-OVAp, OT-I/K14-OVAp acutely ill, and OT-I/K14-OVAp RAGnull acutely ill, respectively. The p values obtained from a two-tailed t test, assuming unequal variance in comparison to OT-I, were 0.26, 0.01, and 0.03 for OT-I/K14-OVAp, OT-I/K14-OVAp acutely ill, and OT-I/K14-OVAp RAGnull acutely ill, respectively. B, Serum from nine acutely ill animals was tested for both IL-1β and TNF-α. Lines though the axis and abscissa indicate the assay sensitivity threshold.

FIGURE 4.

High levels of TNF in the serum of the double transgenic mice. Mice were euthanized between 1 and 12 wk of age. A, Serum TNF-α was measured using an ELISA. The SDs were 8, 43, 151, and 53 for OT-I, OT-I/K14-OVAp, OT-I/K14-OVAp acutely ill, and OT-I/K14-OVAp RAGnull acutely ill, respectively. The p values obtained from a two-tailed t test, assuming unequal variance in comparison to OT-I, were 0.26, 0.01, and 0.03 for OT-I/K14-OVAp, OT-I/K14-OVAp acutely ill, and OT-I/K14-OVAp RAGnull acutely ill, respectively. B, Serum from nine acutely ill animals was tested for both IL-1β and TNF-α. Lines though the axis and abscissa indicate the assay sensitivity threshold.

Close modal

Histology of tissue sections revealed an extensive infiltrate of mononuclear cells directly beneath the epidermis and also around the hair follicles (Fig. 5). This infiltrate consisted of CD4 T cells, CD8 T cells, and macrophages, but not B cells (data not shown). Hyperkeratosis and sloughing were observed in some areas of the skin. In the esophagus (Fig. 5) and tongue (data not shown), inflammation was again localized to the epithelial layer. Double transgenic animals showed occasional incidence of diarrhea and intestinal necrosis. However, this was not a consistent feature of this disease and in most animals the epithelium of the stomach, small intestine, and colon appeared intact, with no significant mononuclear cell infiltrate. Transgenes driven by the hK14 promoter were shown to be expressed in a manner similar to endogenous K14, specifically in basal cells of stratified epithelium in skin, esophagus, and tongue (17, 18, 19, 20, 21). Therefore, the pattern of infiltration reproducibly observed in skin and esophagus is consistent with the predicted Ag expression pattern in these animals. We also observed an inflammatory infiltrate in the liver (Fig. 5). The infiltrate was not observed in the parenchyma of the liver or around hepatic venules, but surrounded the bile duct. Although transgene expression was not detected in total liver RNA by PCR (9), it was detected in RNA from bile duct (data not shown), perhaps explaining the unique pattern of infiltration in this organ. Nonetheless, bile duct was not obliterated and function was not impaired, at least as reflected by the normal levels of bilirubin and γ-glutamyl transpeptidase present in the serum (data not shown). In skin, esophagus, and liver, the composition of the infiltrate was similar, being comprised of both CD4 and CD8 T cells and macrophages but not B cells (data not shown). Other organs that were analyzed and appeared normal included heart, lung, pancreas, and kidney (data not shown).

FIGURE 5.

Histopathology in OT-I/K14-OVAp mice. Tissues were taken from healthy OT-I control (lower panels) or diseased OT-I/K14-OVAp mice. Shown are representative H&E-stained tissue sections from liver (left panels), esophagus (middle panels), and skin (right panels). In liver, note inflammation around the bile duct (B) but not hepatic venules (V). In esophagus, note the extensive infiltrate beneath the basal cell layer (bracket). “L” indicates the lumen of the esophagus. In skin, note the hyperkeratosis (∗), sloughing, and follicular infiltrate (arrow).

FIGURE 5.

Histopathology in OT-I/K14-OVAp mice. Tissues were taken from healthy OT-I control (lower panels) or diseased OT-I/K14-OVAp mice. Shown are representative H&E-stained tissue sections from liver (left panels), esophagus (middle panels), and skin (right panels). In liver, note inflammation around the bile duct (B) but not hepatic venules (V). In esophagus, note the extensive infiltrate beneath the basal cell layer (bracket). “L” indicates the lumen of the esophagus. In skin, note the hyperkeratosis (∗), sloughing, and follicular infiltrate (arrow).

Close modal

The activated phenotype of the CD8 T cells and the fact that there was a mononuclear cell infiltrate found in areas of Ag expression suggested that the disease in the OT-I/K14-OVAp mice might be due to the autoaggressive CD8 T cells. To test this, CD8 T cells were depleted with administration of Abs to CD8α or CD8β from birth to 6 wk of age. As a control, littermates were injected with rat Ig, and the mice were monitored for signs of disease. All mice injected with CD8 Abs survived, whereas only 3 of the 10 mice injected with the control Ab survived (Table II). The percentage of survival in mice injected with the control Ab was similar to untreated OT-I/K14-OVAp mice. Thus, CD8 T cells are required for disease induction.

Table II.

Disease was dependent on CD8αβ T cellsa

AbNumber of Mice That Died Before 6 wk
CD8α 0 /6 
CD8β 0 /5 
TNF-α 5 /6 
Rat Ig 7 /10 
AbNumber of Mice That Died Before 6 wk
CD8α 0 /6 
CD8β 0 /5 
TNF-α 5 /6 
Rat Ig 7 /10 
a

Neonatal animals were given 50 μg of Ab i.p. every other day (for CD8 depletion) or every fourth day (for TNF neutralization) until 6 wk of age.

It had previously been shown that T cells expressing CD8αα homodimers could mature extrathymically in TCR-transgenic mice (22, 23). FACS analysis of the spleen and lymph nodes from OT-I/K14-OVAp mice indicated that there was only a very small increase in CD8αα cells, but that these cells also had an activated phenotype (data not shown). To test whether it was these cells or conventional CD8αβ heterodimer expressing T cells of thymic origin that induced disease, we treated mice with a CD8β Ab. Similar to the CD8α depletion, all of the mice injected with the CD8β Ab survived, indicating that the CD8αα T cells alone did not cause disease. Mice depleted with either CD8α or CD8β Ab did not display elevated levels of TNF in the serum (data not shown). Because of the correlation between elevated levels of TNF-α and disease, we also tested whether TNF-α was required for disease. Six OT-I/K14-OVAp mice were injected with a neutralizing Ab to TNF-α (24), beginning at day 1 after birth and continuing every 4 days until 6 wk of age. Surprisingly, five of the six mice suffered from acute disease and died by 6 wk of age (Table II). This indicated that, while TNF-α levels were high in acutely sick mice, this cytokine was not the sole mediator of disease. In this regard, we noted that IL-1β was elevated in seven of nine serum samples from acutely ill mice (Fig. 4 B); therefore, it is possible that IL-1β could mediate inflammation in TNF-α neutralized mice.

Because it was previously shown that T cells with two TCRs could induce autoimmune disease, and because the majority of CD8 T cells in these animals express two receptors, we wanted to test the role of dual-reactive T cells in disease induction in this model. To do this, we crossed the OT-I/K14-OVAp mice to the recombinase-activating gene (RAG)1null background. In the absence of RAG1 protein, rearrangement of endogenous TCRs cannot occur, thereby eliminating dual-reactive T cells. As shown previously, this dramatically limited the development of CD8 T cells in the thymus (9). Nevertheless, some T cells were present in the periphery of the OT-I/K14-OVAp RAGnull mice (Table I), and these cells expressed the OT-I TCR and CD8, albeit at levels slightly lower than controls (Fig. 1,B). Fig. 3 shows that these mice also suffered from acute disease, indicating that dual reactivity is not required. However, the disease in the OT-I/K14-OVAp RAGnull mice was delayed in the time of onset compared with the double transgenics on a wild-type background (Fig. 3,A). The symptoms of the disease were similar to wild-type OT-I/K14-OVAp mice including hair loss, skin lesions, and weight loss. There were also elevated levels of TNF-α in the serum of the OT-I/K14-OVAp RAGnull mice, but the levels on average were not as high as the OT-I/K14-OVAp mice on a wild-type background (Fig. 4). These data show that T cell dual reactivity is not required for cells to escape deletion in the thymus or to induce autoimmune disease in this model.

Not only are dual-reactive T cells dispensable for disease, but data from a second Ag transgenic we studied suggests that dual receptor-expressing T cells are not sufficient to induce disease either. In this case we obtained transgenic mice where an actin promoter was used to drive OVA expression at a high level in all tissues (Act-mOVA). Interestingly, a thymic phenotype consistent with receptor editing was observed in this mouse, as in OT-I/K14-OVAp mice (D. Mayerova and K. A. Hogquist, manuscript in preparation). The thymus and spleen showed a large number of CD4 and CD8 T cells with low-level expression of the transgenic α-chain (Fig. 1,A and Table I). The transgenic β-chain was not decreased, indicating endogenous TCRα chain usage. Nonetheless, these mice were completely healthy (Fig. 3,B), again suggesting that autoimmune disease in OT-I/K14-OVAp is not a reflection of receptor editing. In Fig. 1 A, note that CD8 T cells from both thymus and lymph node of OT-I/Act-mOVA mice expressed a very low level of the transgenic α-chain and had not up-regulated CD69. However, peripheral CD8 T cells from OT-I/K14-OVAp always expressed a higher level of Vα2 than CD8 thymocytes from the same mice, although this level was not as high as CD8 T cells expressing only the transgenic α- and β-chains. This up-regulation on peripheral cells might be a result of T cell activation to the K14-OVAp Ag. Indeed, the CD69+ CD8 T cells seen in OT-I/K14-OVAp are predominantly among this higher Vα2-expressing subset (data not shown). These results suggest that disease may be associated with Ags expressed under control of the hK14 promoter rather than with receptor editing as a tolerance mechanism.

To test whether Ag expression by K14 results in disease in other TCR-transgenic models, we studied the 2C TCR-transgenic strain. The 2C TCR recognizes the peptide mimetope SIYRYGGL when presented by H-2Kb (15). We created transgenic mice expressing this peptide under the control of the human K14 promoter. In the thymus of 2C/K14-SIY mice there was a profound reduction in the number of CD4+CD8+ precursor cells and no significant receptor editing at the TCRα locus, in contrast to that observed in OT-I mice (D. Mayerova and K. A. Hogquist, manuscript in preparation). The small number of CD8 T cells in the thymus, as well as those in the spleen, expressed both the transgenic α- and β-chains of the 2C receptor (Fig. 1). Despite the small number of CD8 T cells maturing in the thymus, 2-wk-old mice displayed a significant number of CD8 T cells in the spleen (Table I). These cells were apparently able to initiate disease, because these mice suffered from a similar lethal disease to the OT-I/K14-OVAp strain (Fig. 3 B). Altogether, these data suggest that the K14 promoter results in an Ag expression pattern or level that is difficult to achieve complete immunological tolerance to. The breakdown of tolerance in this case does not correlate with the extent of clonal deletion or receptor editing in the thymus and may reflect a breakdown in an as-yet-unidentified peripheral tolerance process.

We report in this work the observation of spontaneous autoimmune disease in OT-I TCR-transgenic animals when the Ag is expressed under control of the human K14 promoter. This is a lethal disease, although the precise cause of death was not determined. An inflammatory infiltrate was observed surrounding the bile ducts of the liver, presumably related to transgene expression in the bile duct. Nonetheless, liver and bile duct function were apparently normal. The high levels of TNF-α observed in the serum of these mice might account for the lethality, because aspects of this disease are similar to those seen in mice expressing TNF-α as a transgene (20). Nonetheless, serum levels of TNF-α varied, and a TNF-α neutralization treatment did not prevent lethal disease. IL-1α and IL-1β were also elevated in double transgenics (Fig. 4), and in the absence of TNF-α these cytokines might contribute to disease. Thus, disease initiation does not require TNF-α, and the end stages of the disease process could involve multiple factors including TNF-α. CD8 T cells were required for the disease, because treatment with Ab to either CD8α or CD8β prevented lethality and no disease characteristics were observed in treated animals.

T cells with dual TCR reactivity were previously shown to have autoimmune potential (10, 11). Because the OT-I/K14-OVAp animals showed a high number of dual-reactive CD8 T cells, we considered whether this was a factor in disease induction. OT-I/K14-OVAp animals on a RAGnull background still experienced disease, although the average age of onset was delayed by ∼4 wk. This experiment formally proved that dual-reactive T cells are not required for disease. However, the delayed onset of disease in RAGnull animals might suggest that dual-reactive T cells can accelerate or exacerbate disease. However, such a conclusion would be premature, because RAGnull animals also have a B cell deficiency, severe lymphopenia, and a greatly reduced total number of CD8 T cells. Such factors could also contribute to the age of disease onset. The study of TCRαnull double transgenics and development of an adoptive transfer system to mimic disease are being used to approach this issue further.

The lack of correlation between the number of dual-reactive T cells and disease occurrence in two additional strains we examined also suggested that dual reactivity was not a key factor in this model of autoimmune disease. OT-I/Act-mOVA animals had a slightly higher number of dual receptor-expressing CD4 and CD8 T cells in the spleen compared with OT-I/K14-OVAp animals (Fig. 1 and Table I). However, this strain showed no signs of disease. The fact that OT-I/K14 and OT-I/mOVA strains had a similar thymic tolerance phenotype but differed absolutely in disease occurrence suggests that the immune dysregulation in OT-I/K14 may be related to how or where the Ag is presented in the peripheral tissues of K14 mice. Consistent with this, we also observed disease in 2C animals expressing Ag under control of the human K14 promoter. Thymic tolerance (via clonal deletion) was very profound and these animals matured few CD8 T cells in thymus, similar to 2C animals expressing Ld (14). Thus again, the disease incidence did not correlate with thymic tolerance, at least in terms of whether clonal deletion or receptor editing was more predominant. In other TCR-transgenic models of spontaneous autoimmune disease 1) thymic tolerance did not occur (25), 2) deletion occurred but was partial (26), 3) deletion occurred and was profound (10), or 4) allelic inclusion allowed selective maturation of cells with dual reactivity (10). Thus, a breakdown of clonal deletion does not appear to correlate with spontaneous disease in other transgenic models.

Recent studies suggest that the thymus also directs the selection of regulatory CD25+CD4+ T cells (27). A potential deficiency of thymus-derived regulatory CD4 T cells in the above models has not been addressed. In our model, we might consider the possibility that Ags expressed under the K14 promoter do not efficiently select regulatory CD25+CD4+ T cells in the thymus. Such cells have recently been shown to suppress CD8 T cell responses (28). In light of this possibility, it is interesting that class II molecules expressed under the control of the K14 promoter did allow selection of CD25+ regulatory CD4 T cells (29). However, in our case, where the transgene encodes only the 8-aa class I binding peptide, this would not be sufficient to select class II-restricted, OVA-specific regulatory CD4 T cells. Nonetheless, it is unlikely that this would be the sole factor contributing to disease in this situation. If the absence of Ag-specific regulatory CD4 T cells resulted in disease, then disease should be a universal feature of all transgenic animals expressing class I binding peptide minigenes. Two other articles describing peptide transgenic mice did not report health problems, even when crossed to TCR transgenics specific for the Ag (30, 31).

Based on the data presented in this work, it is possible that use of the K14 promoter to express self-Ags is associated with the breakdown of peripheral tolerance. We note with interest that the localization of hK14-driven transgenes correlates rather tightly with that of the specialized subset of dendritic cells termed Langerhans cells (LC). These cells reside specifically in stratified epithelial tissue, including the epidermis, hair follicle, and esophagus (32). Within the epidermis, LC form a semicontinuous network directly underneath the basal cell layer of keratinocytes. LC become activated and migrate to the lymph node during skin irritation, injury, or infection. Of APCs, LC represent a likely candidate APC for the presentation of OVAp to lymph node T cells in OT-I/K14-OVAp mice, although this remains to be tested directly. Thus, it is interesting to consider whether skin injury or infection might be a cofactor for disease induction or escalation in this murine autoimmune disease model. Perhaps in the context of other predisposing factors, peripheral tolerance more frequently fails to control the T cell response to self-Ags that are repeatedly or chronically presented by APC from the skin but not from other tissues. Skin injury has been shown to lead to pathologic lesions in individuals predisposed to the human autoimmune disease psoriasis—a phenomenon referred to as Koebner’s effect (33).

A predisposing factor in this model is clearly the high Ag-specific T cell clone frequency. We emphasize that both K14-OVAp and K14-SIYp single transgenic animals appear healthy and viable; disease was observed only after crossing K14 transgenic animals to a TCR-transgenic specific for the Ag. This is consistent with CTL tolerance to a viral Ag driven by the K14 promoter, observed in a non-TCR-transgenic system (34). Thus, an exaggerated frequency or continued onslaught of T cells may be required to shift the tolerance/immunity balance in this model, as in all TCR-transgenic models of spontaneous disease (10, 35, 36, 37, 38, 39). In humans, expansion of T cell clones occurs during the immune response (40, 41) and under conditions of lymphopenia (42, 43). However, it is not clear whether this is a predisposing factor in human autoimmune disease (44). Because of this, murine TCR-transgenic models may not represent ideal systems to study the basis of immune tolerance loss. Nonetheless, TCR transgene-based models of autoimmune disease can potentially be useful to define pathogenic processes and autoimmune spreading mechanisms, and to test treatment protocols. Future experiments will be aimed at determining the potential roles of skin injury in disease induction and escalation in this model.

We thank Kelly McCarthy, Melinda Berthold, and John Hermanson for excellent technical assistance and members of the Hogquist and Jameson labs for their input.

1

This work was funded by National Institutes of Health Grants PO1 AI 35296 and RO1 AI 39560 and the University of Minnesota Graduate School.

3

Abbreviations used in this paper: K14, human keratin 14; RAG, recombinase-activating gene; LC, Langerhans cell.

1
Amsen, D., A. M. Kruisbeek.
1998
. Thymocyte selection: not by TCR alone.
Immunol. Rev.
165
:
209
2
Huseby, E. S., B. Sather, P. G. Huseby, J. Goverman.
2001
. Age-dependent T cell tolerance and autoimmunity to myelin basic protein.
Immunity
14
:
471
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
4
Stockinger, B..
1999
. T lymphocyte tolerance: from thymic deletion to peripheral control mechanisms.
Adv. Immunol.
71
:
229
5
Schonrich, G., U. Kalinke, F. Momburg, M. Malissen, A. M. Schmitt Verhulst, B. Malissen, G. J. Hammerling, B. Arnold.
1991
. Down-regulation of T cell receptors on self-reactive T cells as a novel mechanism for extrathymic tolerance induction.
Cell
65
:
293
6
Azzam, H. S., A. Grinberg, K. Lui, H. Shen, E. W. Shores, P. E. Love.
1998
. CD5 expression is developmentally regulated by T cell receptor (TCR) signals and TCR avidity.
J. Exp. Med.
188
:
2301
7
Tarakhovsky, A., S. B. Kanner, J. Hombach, J. A. Ledbetter, W. Muller, N. Killeen, K. Rajewsky.
1995
. A role for CD5 in TCR-mediated signal transduction and thymocyte selection.
Science
269
:
535
8
Borgulya, P., H. Kishi, Y. Uematsu, H. von Boehmer.
1992
. Exclusion and inclusion of α and β T cell receptor alleles.
Cell
69
:
529
9
McGargill, M. A., J. M. Derbinski, K. A. Hogquist.
2000
. Receptor editing in developing T cells.
Nat. Immunol.
1
:
336
10
Sarukhan, A., C. Garcia, A. Lanoue, H. von Boehmer.
1998
. Allelic inclusion of T cell receptor α genes poses an autoimmune hazard due to low-level expression of autospecific receptors.
Immunity
8
:
563
11
Zal, T., S. Weiss, A. Mellor, B. Stockinger.
1996
. Expression of a second receptor rescues self-specific T cells from thymic deletion and allows activation of autoreactive effector function.
Proc. Natl. Acad. Sci. USA
93
:
9102
12
Stockinger, B., T. Zal.
1997
. T-cells with two T-cell receptors: are they important in autoimmunity?.
Biochem. Soc. Trans.
25
:
617
13
Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, F. R. Carbone.
1994
. T cell receptor antagonist peptides induce positive selection.
Cell
76
:
17
14
Sha, W. C., C. A. Nelson, R. D. Newberry, D. M. Kranz, J. H. Russell, D. Y. Loh.
1988
. Positive and negative selection of an antigen receptor on T cells in transgenic mice.
Nature
336
:
73
15
Udaka, K., K. H. Wiesmuller, S. Kienle, G. Jung, P. Walden.
1996
. Self-MHC-restricted peptides recognized by an alloreactive T lymphocyte clone.
J. Immunol.
157
:
670
16
Stefanski, H. E., D. Mayerova, S. C. Jameson, K. A. Hogquist.
2001
. A low-affinity TCR ligand restores positive selection of CD8+ T cells in vivo.
J. Immunol.
166
:
6602
17
Vassar, R., M. Rosenberg, S. Ross, A. Tyner, E. Fuchs.
1989
. Tissue-specific and differentiation-specific expression of a human K14 keratin gene in transgenic mice.
Proc. Natl. Acad. Sci. USA
86
:
1563
18
Vasioukhin, V., L. Degenstein, B. Wise, E. Fuchs.
1999
. The magical touch: genome targeting in epidermal stem cells induced by tamoxifen application to mouse skin.
Proc. Natl. Acad. Sci. USA
96
:
8551
19
Vassar, R., M. E. Hutton, E. Fuchs.
1992
. Transgenic overexpression of transforming growth factor α bypasses the need for c-Ha-ras mutations in mouse skin tumorigenesis.
Mol. Cell. Biol.
12
:
4643
20
Cheng, J., K. Turksen, Q. C. Yu, H. Schreiber, M. Teng, E. Fuchs.
1992
. Cachexia and graft-vs.-host-disease-type skin changes in keratin promoter-driven TNF α transgenic mice.
Genes Dev.
6
:
1444
21
Turksen, K., T. Kupper, L. Degenstein, I. Williams, E. Fuchs.
1992
. Interleukin 6: insights to its function in skin by overexpression in transgenic mice.
Proc. Natl. Acad. Sci. USA
89
:
5068
22
Rocha, B., H. von Boehmer, D. Guy-Grand.
1992
. Selection of intraepithelial lymphocytes with CD8 α/α co-receptors by self-antigen in the murine gut.
Proc. Natl. Acad. Sci. USA
89
:
5336
23
Poussier, P., H. S. Teh, M. Julius.
1993
. Thymus-independent positive and negative selection of T cells expressing a major histocompatibility complex class I restricted transgenic T cell receptor α/β in the intestinal epithelium.
J. Exp. Med.
178
:
1947
24
Sheehan, K. C., N. H. Ruddle, R. D. Schreiber.
1989
. Generation and characterization of hamster monoclonal antibodies that neutralize murine tumor necrosis factors.
J. Immunol.
142
:
3884
25
Goverman, J., A. Woods, L. Larson, L. P. Weiner, L. Hood, D. M. Zaller.
1993
. Transgenic mice that express a myelin basic protein-specific T cell receptor develop spontaneous autoimmunity.
Cell
72
:
551
26
Kouskoff, V., A. S. Korganow, V. Duchatelle, C. Degott, C. Benoist, D. Mathis.
1996
. Organ-specific disease provoked by systemic autoimmunity.
Cell
87
:
811
27
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
28
Piccirillo, C. A., E. M. Shevach.
2001
. Cutting edge: control of CD8+ T cell activation by CD4+CD25+ immunoregulatory cells.
J. Immunol.
167
:
1137
29
Bensinger, S. J., A. Bandeira, M. S. Jordan, A. J. Caton, T. M. Laufer.
2001
. Major histocompatibility complex class II-positive cortical epithelium mediates the selection of CD4+25+ immunoregulatory T cells.
J. Exp. Med.
194
:
427
30
Goldrath, A. W., M. J. Bevan.
1999
. Low-affinity ligands for the TCR drive proliferation of mature CD8+ T cells in lymphopenic hosts.
Immunity
11
:
183
31
Levelt, C. N., E. Mizoguchi, X. Huang, R. Zacks, A. K. Bhan, S. Tonegawa.
1998
. Inhibition of intrathymic T cell development by expression of a transgenic antagonist peptide.
Proc. Natl. Acad. Sci. USA
95
:
14349
32
Jakob, T., M. C. Udey.
1999
. Epidermal Langerhans cells: from neurons to nature’s adjuvants.
Adv. Dermatol.
14
:
209
33
Melski, J. W., J. D. Bernhard, R. S. Stern.
1983
. The Koebner (isomorphic) response in psoriasis: associations with early age at onset and multiple previous therapies.
Arch. Dermatol.
119
:
655
34
Frazer, I. H., G. J. Fernando, N. Fowler, G. R. Leggatt, P. F. Lambert, A. Liem, K. Malcolm, R. W. Tindle.
1998
. Split tolerance to a viral antigen expressed in thymic epithelium and keratinocytes.
Eur. J. Immunol.
28
:
2791
35
Katz, J. D., B. Wang, K. Haskins, C. Benoist, D. Mathis.
1993
. Following a diabetogenic T cell from genesis through pathogenesis.
Cell
74
:
1089
36
Morgan, D. J., R. Liblau, B. Scott, S. Fleck, H. O. McDevitt, N. Sarvetnick, D. Lo, L. A. Sherman.
1996
. CD8+ T cell-mediated spontaneous diabetes in neonatal mice.
J. Immunol.
157
:
978
37
Verdaguer, J., D. Schmidt, A. Amrani, B. Anderson, N. Averill, P. Santamaria.
1997
. Spontaneous autoimmune diabetes in monoclonal T cell nonobese diabetic mice.
J. Exp. Med.
186
:
1663
38
Geiger, T., L. Gooding, R. Flavell.
1992
. T-cell responsiveness to an oncogenic peripheral protein and spontaneous autoimmunity in transgenic mice.
Proc. Natl. Acad. Sci. USA
89
:
2985
39
Goverman, J..
1999
. Tolerance and autoimmunity in TCR transgenic mice specific for myelin basic protein.
Immunol. Rev.
169
:
147
40
Altman, J. D., P. A. Moss, P. J. Goulder, D. H. Barouch, M. G. McHeyzer-Williams, J. I. Bell, A. J. McMichael, M. M. Davis.
1996
. Phenotypic analysis of antigen-specific T lymphocytes.
Science
274
:
94
41
Callan, M. F., L. Tan, N. Annels, G. S. Ogg, J. D. Wilson, C. A. O’Callaghan, N. Steven, A. J. McMichael, A. B. Rickinson.
1998
. Direct visualization of antigen-specific CD8+ T cells during the primary immune response to Epstein-Barr virus in vivo.
J. Exp. Med.
187
:
1395
42
Masuko, K., S. Kato, M. Hagihara, F. Tsuchida, Y. Takemoto, K. Izawa, T. Kato, S. Yamamori, Y. Mizushima, K. Nishioka, et al
1996
. Stable clonal expansion of T cells induced by bone marrow transplantation.
Blood
87
:
789
43
Gaschet, J., C. Denis, N. Milpied, M. M. Hallet, F. Romagne, A. Necker, R. Vivien, J. David-Ameline, F. Davodeau, M. Bonneville, et al
1995
. Alterations of T cell repertoire after bone marrow transplantation: characterization of over-represented subsets.
Bone Marrow Transplant.
16
:
427
44
Goronzy, J. J., C. M. Weyand.
2001
. Thymic function and peripheral T-cell homeostasis in rheumatoid arthritis.
Trends Immunol.
22
:
251