Incomplete TCRβ allelic exclusion accelerates spontaneous autoimmune arthritis in K/BxN TCR transgenic mice

Summary

Allelic exclusion of antigen receptor loci is a fundamental mechanism of immunological self tolerance. Incomplete allelic exclusion leads to dual T cell receptor (TCR) expression and can allow developing autoreactive αβ T lymphocytes to escape clonal deletion. Because allelic exclusion at the TCRβ locus is more stringent than at the TCRα locus, dual TCRβ expression has not been considered a likely contributor to autoimmunity. We show here that incomplete TCRβ allelic exclusion permits developing thymocytes bearing the autoreactive, transgene-encoded KRN TCR to be positively selected more efficiently, thereby accelerating the onset of spontaneous autoimmune arthritis. Our findings highlight dual TCRβ expression as a mechanism that can enhance the maturation of autoreactive pathogenic T cells and lead to more rapid development of autoimmune disease.

Introduction

A diverse repertoire of lymphocytes provides broad immune defense against pathogens. This diversity arises via somatic genetic recombination of antigen receptor loci in developing T and B cells (1). These recombination events also result in lymphocytes that express autoreactive antigen receptors. Multiple modes of immunological tolerance act to delete or restrain these potentially autoreactive cells (2, 3). Autoimmune diseases occur because immunological tolerance fails.

Via allelic exclusion most lymphocytes express only one antigen receptor specificity (4). For most αβ T cells, this means that the T cell receptors on the cell surface are encoded by one TCRα (designated Tcra) allele and one TCRβ (Tcrb) allele. Dual receptor T cells that express two productively rearranged TCRα or TCRβ chains due to incomplete allelic exclusion do exist, however, in mice and humans (5–9), and this population is known to contain autoreactive clones (10). In mice, an estimated 10% of αβ T cells express dual TCRα chains; allelic exclusion at the TCRβ locus is more stringent, with only 1–3% of αβ T cells expressing dual TCRβ chains (6). The discovery of dual receptor T cells raised concern that such cells posed a special challenge to immunological tolerance (8). If one TCR recognized foreign peptide antigen presented by self MHC and the other recognized self peptide:MHC complexes, stimulation of that T cell with foreign antigen through one TCR might allow that cell to mediate autoimmunity through the other TCR. Recent evidence in a mouse model of multiple sclerosis showed that this scenario can occur (11).

Expression of two different TCR specificities can also impair immunological tolerance by allowing a potentially autoreactive T cell to escape clonal deletion in the thymus, a phenomenon observed in several TCR transgenic mouse models. The level of cell surface expression of the autoreactive TCR is relatively lower in the presence of an additional TCR in the same cell, compromising negative selection (12, 13). Dual TCRα chain expression can also rescue positive selection of thymocytes otherwise destined to die by neglect (14, 15). Although expression of dual TCRs can allow potentially autoreactive T cells to escape clonal deletion, these dual TCR-expressing escapees have not been shown to provoke autoimmune disease (16–20), except in a situation in which the T cells were engineered to express two different transgene-encoded TCRs (21). However, because TCRβ allelic exclusion is so stringent, prior studies of dual TCR expression in autoimmunity have focused on incomplete TCRα allelic exclusion (13, 16–18, 20). The possibility that dual TCRβ expression might contribute to the development of autoimmunity has not been rigorously investigated.

In the “K/BxN” TCR transgenic mouse model of autoantibody-mediated arthritis, CD4⁺ T cells expressing the “KRN” transgene-encoded TCRα and TCRβ chains recognize a peptide derived from the ubiquitously-expressed enzyme glucose-6-phosphate isomerase (GPI) presented by the MHC class II allele I-A^g7 derived from the NOD mouse strain. Activated CD4⁺ KRN T cells provide help to GPI-specific B cells, resulting in the production of arthritis-inducing anti-GPI autoantibodies (22–24). An unresolved conundrum is how the CD4⁺ KRN T cells reactive to a ubiquitously-expressed self antigen escape clonal deletion in the thymus. Such escape begins around three weeks of age, followed shortly thereafter by the development of arthritis (23).

The initial report characterizing K/BxN mice noted that many CD4⁺ T cells expressed endogenously-encoded TCRα and TCRβ chains in addition to the ones encoded by the KRN transgenes. These dual receptor T cells from K/BxN (henceforth KRN.H-2^b/g7) mice were less responsive to peptide:MHC when compared to KRN T cells isolated from KRN.H-2^b mice lacking the MHCII restriction element I-A^g7, raising the possibility that the dual TCR expression in the KRN.H-2^b/g7 mice resulted in decreased TCR signal strength and allowed the autoreactive T cells to escape clonal deletion (23). We explored this possibility in greater depth and found unexpectedly that expression of dual TCRβ (or dual TCRα) chainsdue to incomplete allelic exclusion allowed autoreactive T cells to be positively selected more efficiently, thereby accelerating the onset of spontaneous autoimmune arthritis.

Results

Incomplete TCRβ allelic exclusion is common among KRN T cells

We first sought to estimate the frequency of KRN T cells that express endogenously-encoded TCRβ chains in addition to the KRN transgene-encoded one. Using a panel of monoclonal antibodies that recognizes sixteen mouse Vβ regions, we found that cells expressing both the KRN transgene-encoded, Vβ6-containing TCRβ chain plus an endogenously-encoded TCRβ chain were common among CD4 single positive (CD4 SP) thymocytes (~25–30%) and CD4⁺ splenic T cells (~40%) (Figure 1). This Vβ antibody panel comprises approximately two-thirds of all the mouse Vβ regions, and thus these values likely underestimate the true frequency of dual TCRβ expressing cells. We detected low levels of dual TCRβ expression in mice lacking the KRN transgenes, consistent with prior reports (5, 6). The lack of a monoclonal antibody specific for Vα4 (the KRN-encoded TCRα chain) precluded similar analysis of dual TCRα expression, though KRN T cells have been shown to express endogenously-encoded TCRα chains (23). Intriguingly, we observed dual TCRβ expression in KRN.H-2^b mice in which the KRN T cells are positively selected but remain naïve, as well as in KRN.H-2^b/g7 mice in which the developing KRN T cells encounter GPI peptide:I-A^g7 complexes in the thymus yet escape deletion to provoke arthritis. In both KRN.H-2^b and KRN.H-2^b/g7 mice, the most frequently-used endogenous Vβ chains among those detectable with the available reagents included Vβ14, Vβ8.1/8.2, and Vβ11 (data not shown). Thus, expression of endogenously-encoded TCRβ chains in addition to the transgene-encoded one is frequent among KRN T cells.

Figure 1. Dual TCRβ expression is common in KRN TCR transgenic mice.

Flow cytometric analysis of CD4 SP thymocytes and CD4⁺ splenic T cells of the indicated mice is shown. The KRN transgene-encoded Vβ6 chain is on the x-axis. Staining with a panel of Vβ antibodies is represented on the y-axis (Vβ2, 3, 4, 5.1/5.2, 7, 8.1/8.2, 8.3, 9, 10b, 11, 12, 13, 14, 17a). Numbers indicate percentages; cells expressing dual TCRβ chains are represented in the top right quadrant of each plot (gray font). The results are representative of data from two independent experiments.

Incomplete TCRβ allelic exclusion allows autoreactive T cells to undergo thymic maturation efficiently

We hypothesized that this incomplete allelic exclusion might allow autoreactive CD4⁺ KRN T cells to escape clonal deletion. To test this hypothesis, we bred KRN.H-2^b/g7 mice that lacked the capacity to express endogenously-encoded TCRα chains, TCRβ chains, or both (25). We found that KRN.H-2^b/g7 mice unable to express both endogenous TCRα and TCRβ chains (henceforth TCRα/β-deficient) did not develop arthritis during the first 8 weeks of life, whereas expression of either endogenous TCRα or endogenous TCRβ chains was sufficient to provoke the early onset of arthritis typical in this model (Figure 2). The disease phenotype did not differ among mice carrying the functional endogenous Tcra and/or Tcrb alleles in homozygosity versus heterozygosity; i.e., there was no apparent gene-dosage effect (data not shown). The TCRα/β-deficient KRN.H-2^b/g7 mice had a significant reduction in the frequency and absolute number of CD4 SP thymocytes and CD4⁺ splenic T cells relative to mice able to express endogenous TCRα and/or TCRβ chains, as well as a reduction in the level of arthritogenic anti-GPI IgG autoantibodies (Figure 3 and Supplemental Figure 1 (S1)).

Figure 2. Both endogenous TCRα and TCRβ chains can facilitate spontaneous autoimmune arthritis.

(A) Arthritis development expressed as ankle thickening over time in mice of the indicated genotypes (all KRN.H-2^b/g7). The “+” denotes that the endogenous Tcra and Tcrb alleles are present either in heterozygosity or homozygosity; “−“ denotes homozygosity of the knockout allele. Each line represents one mouse (n = 5–6 mice per group). (B) Representative ankle sections from 8-week-old mice of the indicated genotypes. H&E staining; scale bar in left panel represents 500 microns and all images are at the same magnification.

Figure 3. Endogenous TCRα/β expression allows autoreactive CD4⁺ T cells to escape clonal deletion.

Flow cytometry of thymocytes (top) and CD3⁺ splenocytes (bottom) obtained from mice of the indicated genotypes (all KRN.H-2^b/g7) and stained with anti-CD4 and anti-CD8 monoclonal antibodies. Numbers indicate the percentage of cells in each of the boxed gates. The plots are representative of 6 experiments with a total of 5–6 mice/group. The mice were 8–10 weeks of age.

Because the development of arthritis in this model depends on autoantibody production, we sought to verify that the lack of arthritis development in the TCRα/β-deficient KRN.H-2^b/g7 mice was not due to an unanticipated B cell defect. Adoptive transfer of CD4⁺ splenic T cells from KRN.H-2^b mice to the TCRα/β-deficient KRN.H-2^b/g7 mice provoked the rapid onset of arthritis (Supplemental Table I), demonstrating that the failure of the TCRα/β-deficient KRN.H-2^b/g7 mice to develop arthritis was due to their lack of CD4⁺ KRN T cells. Taken together, these findings show that incomplete TCRβ (or TCRα) allelic exclusion facilitates the thymic maturation of CD4⁺ T cells expressing the autoreactive KRN TCR, allowing them to provoke autoantibody-mediated arthritis.

Dual TCR expression is readily detected among CD4 SP thymocytes

To determine how dual TCR expression enhances thymic maturation of autoreactive KRN T cells, we performed further analyses to determine at what stage of T cell differentiation dual TCR expression was first detectable. Expression of endogenously-encoded TCRα and TCRβ chains was readily detected among CD4⁺ splenic T cells (Supplemental Figure 2) and CD4 SP thymocytes (Supplemental Figure 3). Among splenocytes, the frequency of cells expressing dual TCRα chains was higher in mice lacking endogenous TCRβ chains relative to the endogenous TCRα/β-sufficient mice (34.7% versus 18.3%); similarly the frequency of cells expressing dual TCRβ chains was higher in mice lacking endogenous TCRα chains relative to the endogenous TCRα/β-sufficient mice (41.7% versus 32.4%), further demonstrating that allelic inclusion can occur readily in the KRN system. Cells co-expressing an endogenously-encoded TCRα chain plus an endogenously-encoded TCRβ chain in addition to the KRN-encoded TCRβ chain were detectable, but not common. Expression of dual TCRβ chains was associated with decreased expression of the KRN transgene-encoded Vβ6 chain in the presence of I-A^g7 (Figures 1, S2) (23). We did not detect significant endogenous TCRα or TCRβ expression at the earlier CD4⁺ CD8⁺ double-positive (DP) and CD4⁻ CD8⁻ double-negative (DN) stages of thymic development (Figure S4 and data not shown). The simplest interpretation of these findings is that dual TCR expression is rare through the DP stage, but that dual TCR-expressing DP thymocytes have a selective advantage at the DP to CD4 SP transition, resulting in their increased frequency among the CD4 SP population.

Dual TCRβ expression facilitates positive selection

The reduced number and frequency of CD4 SP thymocytes in the TCRα/β-deficient KRN.H-2^b/g7 mice could be due to less efficient positive selection, more efficient negative selection, or both (12, 13, 15, 26). Our results above suggested that dual TCR expression was readily detected among the positively-selected CD4 SP population. We therefore sought to determine if dual TCR expression facilitated positive selection of KRN T cells. It is known that CD4⁺ KRN T cells can develop in C57BL/6 mice (I-A^b), but remain naïve in the absence of I-A^g7, the MHCII molecule that presents GPI-derived peptides to provoke arthritis. However, we had observed that dual TCRβ expression was common even in KRN.H-2^b mice (Figure 1). To determine if endogenous TCR expression enhanced positive selection of CD4⁺ KRN T cells on I-A^b, we bred TCRα/βdeficient KRN.H-2^b mice. The key difference between these mice and those used earlier in this report (Figures 2 and 3) is their absence of I-A^g7. The absence of TCRα, TCRβ, or both did not significantly alter the numbers of total, DP, or DN thymocytes (Figure S5), demonstrating that the KRN TCR can mediate β-selection and allow progression to the DP stage efficiently. Importantly, in TCRα/β-deficient KRN.H-2^b mice we again observed a selective deficiency of CD4⁺ T cells relative to littermates capable of expressing endogenous TCRα or TCRβ chains, demonstrating that the KRN receptor itself is very weakly selected on I-A^b (Figure 4). The frequency and number of CD4 SP thymocytes did not differ significantly between TCRα/β-deficient KRN.H-2b mice and TCRα/β-deficient KRN.H-2b/g7 mice (Figures S1A and S6A), suggesting that the primary effect of dual TCR expression in this system is to enhance positive selection of thymocytes expressing the KRN TCR rather than to allow them to escape negative selection more efficiently. Although the frequency of CD4 SP thymocytes and CD4⁺ splenic T cells increased in older TCRα/β-deficient KRN.H-2^b mice, they remained lower than those in mice capable of expressing endogenous TCRα or TCRβ chains (Figures 4 and S6). We observed a similar defect in positive selection in Rag1-deficient KRN.H-2^b mice (Supplemental Figure 7).

Figure 4. Endogenous TCRα/β expression promotes positive selection of KRN CD4⁺ T cells.

Flow cytometry of thymocytes (top) and CD3⁺ splenocytes (bottom) from mice of the indicated genotypes (all KRN.H-2^b) and stained with anti-CD4 and anti-CD8 monoclonal antibodies. Numbers indicate the percentage of cells in each of the boxed gates. Mice were 8–16 weeks of age; within this range, age-related differences were not apparent except in the TCRα/TCRβ-deficient mice, as shown. Gates were set based on isotype control staining within individual experiments (e.g. the panels in the 4^th column are from a separate experiment from the others). The plots are representative of 3 experiments with a total of 4–6 mice/group.

Taken together, our findings are consistent with a model in which expression of dual TCRs allows autoreactive CD4⁺ KRN T cells to be positively selected more efficiently. This effect increases the likelihood that an autoreactive clone will escape the thymus to provoke autoimmune disease.

Dual TCR expression engenders promiscuity of KRN T cells

The KRN TCR was originally derived from an H-2^k mouse, suggesting that it can be positively selected on I-A^k or I-E^k. CD4⁺ KRN T cells can also be positively selected on I-A^b, although as we show, this was enhanced by endogenous TCR expression (Figures 4 and S7). Similarly, positive selection of CD4⁺ T cells bearing the KRN TCR on H-2^k was more efficient in Rag1-sufficient mice relative to Rag1-deficient mice (Laura Mandik-Nayak, personal communication), analogous to what has been described for KB, another transgenic TCR (26). Most studies of arthritis in the KRN transgenic system utilize mice that express the arthritis-provoking H-2^g7 haplotype in heterozygosity with either H-2^b or H-2^k (23, 27, 28). Based on our finding that dual TCR expression facilitates positive selection of CD4⁺ KRN T cells, we investigated the possibility that H-2^b or H-2^k was necessary for positive selection of CD4⁺ KRN T cells. We therefore bred KRN.H-2^g7 mice. The development of arthritis in KRN.H-2^g7 mice was indistinguishable from that in KRN.H-2^b/g7 mice (Supplemental Figure 8), demonstrating that I-A^b was not necessary for arthritogenesis. Thus, mature CD4⁺ KRN T cells can develop in H-2^g7, H-2^b, and H-2^k mice. Because the H-2^g7 haplotype is mixed (comprising the MHC class I molecules K^d and D^b and the MHC class II molecule I-A^g7) it remains formally possible but unlikely that the CD4⁺ KRN T cells are positively selected on the MHC class I molecule D^b. Our data collectively suggest that this apparent promiscuity of CD4⁺ KRN T cells is due not to degeneracy of the transgenically-encoded KRN TCR per se, but rather to their frequent co-expression of endogenously-encoded TCR chains.

The KRN TCR is sufficient to drive arthritis

When we allowed the TCRα/β-deficient KRN.H-2^b/g7 mice to age beyond 8 weeks, arthritis did eventually develop (Figures 5A, B). This was associated with the emergence of CD4⁺ splenic T cells, as well as the production of anti-GPI autoantibodies (Figure S1). Impaired thymic maturation as still apparent; however, a clear population of mature CD4⁺ splenic T cells was now detectable (13.6%), albeit at lower frequency than in the mice capable of expressing endogenous TCRα or TCRβ chains (~45%) (Figures 3 and 5C). As expected, the mature CD4⁺ splenic T cells expressed only the KRN transgene-encoded Vβ6-containing chain and not endogenously-encoded TCRα or TCRβ chains (Figure 5C). The relative paucity of CD4 SP thymocytes also suggested that re-entry of mature CD4⁺ T cells into the thymus is not prominent in this model (29). We postulate that given sufficient time and a high frequency of potentially autoreactive immature thymocytes, autoreactive CD4⁺ KRN T cells rarely can exit the thymus, expand, and provide productive help to B cells. We are currently investigating whether the peripheral CD4⁺ T cells in aged TCRα/β-deficient KRN.H-2^b/g7 mice arise from single or multiple thymic escapees.

Figure 5. KRN.H-2^b/g7 mice lacking endogenous TCRα and TCRβ chains develop late-onset arthritis as autoreactive T cells escape clonal deletion.

(A) Arthritis development expressed as ankle thickening over age in Tcra⁻ Tcrb⁻ KRN.H-2^b/g7 mice. Each line represents one mouse. (B) Representative ankle section from an arthritic Tcra⁻ Tcrb⁻ KRN.H-2^b/g7 mouse. H&E staining; scale bar represents 500 microns. (C) Flow cytometry of thymocytes and CD3⁺ splenocytes from an arthritic 25-week-old Tcra⁻ Tcrb⁻ KRN.H-2^b/g7 mouse, stained with anti-CD4, anti-Vβ6 (the KRN clonotype), and a mixture of antibodies recognizing 3 endogenous Vα chains and 3 endogenous Vβ chains. Numbers indicate the percentage of cells in each box. The plots on the right depict the cells in the red box. Gating was determined based on control mice with cell populations present in each of the gates.

The finding of late-onset arthritis in the TCRα/β-deficient KRN.H-2^b/g7 mice suggested that the KRN TCR alone was sufficient to provoke arthritis. We utilized adoptive transfer experiments to address this point more thoroughly. Transfer of splenocytes from TCRα/β-deficient KRN.H-2^b mice provoked arthritis in T cell-deficient H-2^b/g7 hosts, verifying that the KRN TCR alone was sufficient to provoke disease in the right context. In contrast, splenocytes from young, non-arthritic TCRα/β-deficient KRN.H-2^b/g7 mice did not transfer disease, consistent with inefficient maturation of CD4⁺ KRN T cells among these donors (Supplemental Table II). Transfer of splenocytes from older, arthritic TCRα/β-deficient KRN.H-2^b/g7 mice in which some autoreactive T cells have escaped the thymus provoked arthritis in half of the recipients (Supplemental Table II). The most likely explanation for the failure of some of the recipients in this latter group to develop arthritis is that the absolute number of mature autoreactive CD4⁺ T cells in the donor spleen was quite low – enough to drive anti-GPI autoantibody production in the donor, but at a survival or proliferative disadvantage when transferred to lymphopenic hosts. These observations, in concert with our primary finding that unmanipulated TCRα/β-deficient KRN.H-2^b/g7 mice do eventually develop arthritis, are consistent with the conclusion reached in a prior report that expression of endogenously-encoded TCRα and TCRβ chains is not required for CD4⁺ KRN T cells to provoke autoimmune arthritis (30).

We considered the possibility that KRN T cells lacking endogenous TCRs might be functionally impaired, thereby contributing to the delay in arthritis onset observed in the TCRα/β-deficient KRN.H-2^b/g7 mice. We therefore adoptively transferred purified, congenically-marked CD4+ T cells from KRN.H-2^b mice and Rag1-deficient KRN.H-2^b mice into lymphopenic H-2^b/g7 hosts. Five days after transfer, we enumerated the donor-derived cells and found that the KRN T cells lacking endogenous TCR expression had proliferated slightly less than those able to express endogenous TCRs (Supplemental Figure 9). Similarly, it was recently reported that Rag1-deficient CD4⁺ KRN T cells induce arthritis after transfer to lymphopenic H-2^b/g7 hosts, albeit with slightly slower kinetics than wildtype KRN T cells (31), consistent with our findings in Supplemental Table II. However, these minor delays in proliferation and arthritis induction alone are unlikely to account for the several-week delay in arthritis onset that we observed in the KRN.H-2^b/g7 mice lacking endogenous TCR expression.

In sum, our findings demonstrate that the critical contribution of the endogenously-encoded TCRs in this system is to enhance positive selection of CD4 SP thymocytes bearing the autoreactive KRN TCR, thereby leading to earlier onset of autoimmune arthritis.

Endogenous TCR expression is required for valvular carditis in K/BxN mice

We have recently reported that KRN.H-2^b/g7 mice develop spontaneous mitral valve inflammation (endocarditis) in addition to their well-described arthritis (32). The development of endocarditis depended on the sustained presence of CD4⁺ cells (33). We found that TCRα/β-deficient KRN.H-2^b/g7 mice were protected from developing endocarditis. Intriguingly, we observed scant mitral valve inflammation even among the aged TCRα/β-deficient KRN.H-2^b/g7 mice, despite that they had active arthritis for 6–17 weeks. Furthermore, as with arthritis, endocarditis developed normally in KRN.H-2^g7 mice; I-A^b was not necessary (Figure 6). The absence of endocarditis in TCRα/β-deficient KRN.H-2^b/g7 mice raises the possibility that an antigen other than GPI peptide:I-A^g7 is responsible for the cardiac pathology, although other explanations are possible.

Figure 6. Endogenous TCRα/β expression is required for autoimmune endocarditis in KRN.H-2^b/g7 mice.

(A) Representative mitral valves of mice of the indicated genotypes. H&E, scale bar represents 100 microns. (B) The maximum mitral thickness of the mitral valve is depicted. Circles represent individual mice, bars represent the mean; P-values were calculated using Student’s t-test.

Discussion

Our findings demonstrate that dual TCRβ expression can facilitate positive selection of thymocytes bearing an autoreactive TCR. The binding of autoreactive TCRs to self peptide:MHC is suboptimal relative to the binding of TCR to foreign peptide:MHC. This suboptimal binding has generally been felt to increase the risk that T cells bearing autoreactive TCRs might escape negative selection (34). However, suboptimal self peptide:MHC binding by an autoreactive TCR could also impair positive selection; in that setting, expression of a second TCR specificity may rescue positive selection and increase the likelihood that autoimmunity will develop. Several prior studies have reported that dual TCRα expression can facilitate positive selection in TCR transgenic systems (4, 14, 35, 36); here we report that dual TCRβ expression can lead to similar effects on positive selection and in this case accelerate the development of autoimmunity.

Although the high frequency of dual TCRβ expression we observed clearly arises due to the expression of the transgenically-encoded TCR, incomplete TCRβ allelic exclusion does occur in normal mice and humans. Prior studies of a role for dual TCR T cells in non-TCR-transgenic mouse autoimmune disease models concluded that dual TCR expression was not necessary for immunological tolerance to be breached, based on the finding that TCRα-hemizygous mice whose T cells are unable to express dual TCRα chains remained susceptible to disease (16–18). Dual TCR expression due to incomplete TCRβ allelic exclusion was still possible in those mice, however, and our findings demonstrate that dual TCRβ expression can indeed promote the development of autoimmunity in certain situations. The possibility that dual TCR expression can contribute to autoimmunity in non-TCR-transgenic models and in humans deserves to be revisited more rigorously, focusing on a potential contribution of incomplete TCRβ allelic exclusion.

The primary mechanism by which dual TCRβ (or dual TCRα) expression drives autoimmunity in our model is by allowing CD4⁺ KRN T cells to be positively selected more efficiently. Dual TCR expression is not, however, required for the development of frank autoimmunity in this model. It is thus not necessary in this model to invoke a scenario in which a dual TCR T cell is activated by one TCR and then provokes autoimmunity through its second TCR. However, this finding does not exclude the possibility that foreign antigens could stimulate mature CD4⁺ KRN T cells, through either the KRN transgene- and/or endogenously-encoded TCRs (37).

Prior studies in which dual TCR expression has been shown to drive autoimmune pathology have depended on experimental manipulation including adoptive transfer to a lymphopenic host (13), expression of two different transgene-encoded TCRs (21), or infection (11). In contrast, the autoreactive CD4⁺ KRN T cells that enter the periphery in KRN.H-2^b/g7 mice also escape peripheral immunological tolerance mechanisms, resulting in spontaneous systemic autoimmune disease. Although the T cell repertoire in these TCR transgenic mice is highly biased toward autoreactivity, our findings serve as a proof of principle that spontaneous autoimmune diseases can arise more efficiently in the presence of dual TCR-expressing T cells with an increased likelihood of surviving thymic maturation. Determining whether dual TCRβ expression can contribute to spontaneous autoimmunity in mice with a polyclonal T cell repertoire is the next logical extension of this work.

We were surprised by the finding that the aged TCRα/β-deficient KRN.H-2^b/g7 mice developed arthritis but not endocarditis. It is possible that their protection from endocarditis was due simply to their lower number of mature CD4⁺ KRN T cells or to the presence of anti-GPI IgG at slightly lower concentrations or for a shorter time. However, whether the arthritogenic antigen GPI-peptide:I-A^g7 is also the key autoantigen driving endocarditis is not known. We are intrigued by the possibility that endocarditis depends on expression of a TCR comprising a KRN transgene-encoded chain paired with an endogenously-encoded one, a possibility that would strongly suggest that an antigen other than GPI peptide:I-A^g7 drives pathology in this target tissue. In such a scenario, the expression of dual TCRs could create a situation analogous to epitope spreading in which activation of a dual TCR CD4⁺ KRN T cell via GPI peptide:I-A^g7 interacting with the KRN transgene-encoded TCR results in cross-reactivity mediated by recognition of a cardiac antigen by a second TCR containing an endogenously-encoded TCR chain.

In sum, our findings suggest that the potential contribution of dual TCR T cells to the pathogenesis of autoimmune diseases needs to be explored in greater depth, based on the novel observation that incomplete TCRβ allelic exclusion can enhance positive selection of autoreactive T cells and accelerate autoimmunity.

Materials and Methods

Mice

KRN T cell receptor transgenic mice on the C57BL/6 background (23) were a gift from Drs. Diane Mathis and Christophe Benoist (Harvard Medical School, Boston, MA) and the Institut de Génétique et de Biologie Moléculaire et Cellulaire (Strasbourg, France); C57BL/6 mice congenic for H-2^g7 (B6.g7) (32) were also a gift from Drs. Mathis and Benoist. Tcra^−/− (B6.129S2-Tcra^tm1Mom/J), Tcrb^−/− (B6.129P2-Tcrb^tm1Mom/J) (25), and Rag1^−/− mice (B6.129S7-Rag1^tm1Mom/J) (38) on the C57BL/6 background were purchased from The Jackson Laboratory (Bar Harbor, ME). The genotypes of mice were determined using both PCR and flow cytometry.

Arthritis scoring, autoantibody titers, and histology

Assessment of ankle thickening, measurement of anti-GPI antibody titers, histological preparation of tissue specimens were performed as described (32). Microscopy was performed at room temperature using an Olympus BX51 microscope equipped with Olympus UPLAN objectives, an Olympus DP71 12.5 megapixel digital camera, and Olympus DP-BSW software (version 3.02).

Cell transfer experiments

Splenocyte transfer experiments were performed as previously described (32). Separation of CD4⁺ T cells from whole splenocytes was performed according to manufacturer’s instructions (Miltenyi Biotec).

Flow cytometry

Monoclonal antibodies used for flow cytometry included those that recognize B220 (clone RA3-6B2), CD3 (clone 17A2), CD8 (53-6.7), and CD90.1 (HIS51) (eBioscience) and CD4 (RM4-5), CD45.1 (A20), CD45.2 (104), Vα2 (B20.1), Vα3.2 (RR3-16), Vα8.3 (KT50), Vβ2 (B20.6), Vβ3 (KJ25), Vβ4 (KT4), Vβ5.1/5.2 (MR9-4), Vβ6 (RR4-7), Vβ7 (TR310), Vβ8.1/8.2 (MR5-2), Vβ8.3 (1B3.3), Vβ9 (MR10-2), Vβ10b (B21.5), Vβ11 (RR3-15), Vβ12 (MR11-1), Vβ13 (MR12-3), Vβ14 (14-2), and Vβ17a (KJ23) (BD Pharmingen). Fixable viability dye eFluor 450 was used per the manufacturer’s instructions (eBioscience). Cells were analyzed using a BD LSRII flow cytometer (BD Biosciences). Data were analyzed with FlowJo software (Treestar). An example of the gating scheme used throughout the study is shown in Supplemental Figure 10.

Abbreviation used

GPI

glucose-6-phosphate isomerase