Autoimmune susceptibility imposed by public TCRβ chains

Although the TCR repertoire is highly diverse, a small fraction of TCR chains, referred to as public, preferentially form and are shared by most individuals. Prior studies indicated that public TCRβ may be preferentially deployed in autoimmunity. We hypothesized that if these TCRβ modulate the likelihood of a TCRαβ heterodimer productively engaging autoantigen, because they are widely present in the population and often high frequency within individual repertoires, they could also broadly influence repertoire responsiveness to specific autoantigens. We assess this here using a series of public and private TCRβ derived from autoimmune encephalomyelitis-associated TCR. Transgenic expression of public, but not private, disease-associated TCRβ paired with endogenously rearranged TCRα endowed unprimed T cells with autoantigen reactivity. Further, two of six public, but none of five private TCRβ provoked spontaneous early-onset autoimmunity in mice. Our findings indicate that single TCRβ are sufficient to confer on TCRαβ chains reactivity toward disease-associated autoantigens in the context of diverse TCRα. They further suggest that public TCR can skew autoimmune susceptibility, and that subsets of public TCR sequences may serve as disease- specific biomarkers or therapeutic targets.

Scientific RepoRts | 6:37543 | DOI: 10.1038/srep37543 repertoire is preferentially deployed relative to the non-shared, or private, pre-immune repertoire 17 . Here we assess the contribution of individual public and private TCRβ sequences to the autoimmune response during MOG 35-55 -induced EAE. We describe mice that transgenically express 15 public or private disease-associated TCRβ , each of which pairs with endogenously rearranged TCRα . Public but not private TCRβ selectively imposed autoimmune risk, fostering autoantigen reactivity and even the development of spontaneous fulminant autoimmunity. Our findings demonstrate that single TCR chains can broadly influence repertoire reactivity and support the hypothesis that recognition biases imposed by public TCR contribute to autoimmune responses.

Results
Spontaneous autoimmunity mediated by a public TCRβ. To understand the composition and dynamics of autoimmune effector and regulatory repertoires, we previously performed saturation sequencing of splenic and CNS T cells from 12 mice with MOG 35-55 -induced EAE and 5 healthy controls, analyzing > 18 × 10 6 CD4 + Foxp3 -(Tconv) and Foxp3 + (Treg) TRBV13-2 + TCRβ 6,17-21 . TRBV13-2 is the dominant TCRβ in MOGspecific T cells 22 . Our results indicated the presence of a diverse, public TCR repertoire within the autoimmune response, and that T cells bearing public TCR were preferentially deployed relative to private TCR from the pre-immune repertoire 17 . This suggested a role for public sequences in predisposing the repertoire toward autoreactivity. To better define the impact of public TCRβ , we generated retroviral transgenic (retrogenic) mice on a TCRβ −/− background that enforced the expression of public and private TCRβ sequences identified through these sequencing analyses and from MOG 35-55 specific T cell hybridomas (Table 1). Mice retrogenic for the ovalbumin-specific OTIIβ chain were generated as an autoantigen non-specific control.
TCRβ 1 was found in splenic and CNS Treg and Tconv of all mice with EAE that were studied (Table 1) 17 . Impressively, TCRβ 1 retrogenic mice uniformly developed spontaneous EAE at 4 weeks, corresponding to very early T cell engraftment (Fig. 1A). Indeed, numbers of T cells infiltrating the CNS at this early time were similar to numbers in the spleen (Fig. 1B). Mortality was > 50% (Fig. 1A). CD4 + Foxp3 − , CD4 + Foxp3 + , and CD8 + T cells engrafted, and the CD69 activation marker was elevated in splenic and CNS T cells from diseased TCRβ 1 mice relative to OTIIβ retrogenic mice ( Supplementary Fig. S1). TCRβ 1 + T cells proliferated vigorously in response to MOG 35-55 (Fig. 1C). Splenic and CNS cells from TCRβ 1 retrogenic mice also demonstrated Th1 and Th17 subset differentiation, which is associated with pathogenicity in EAE ( Fig. 1D and Supplementary Fig. S1). Histologic analyses of the CNS of TCRβ 1 retrogenic mice showed a mixed infiltrate of lymphocytes, macrophages, and granulocytes, gliosis and perivascular cuffing in the septum, meninges, optic nerve, and white tracts of the lumbar  Table 1. TCRβ retrogenic mice. TRBV13-2 + TCRβ chains that were shared in the indicated number of total CD4 + , CD4 + Foxp3 − , and CD4 + Foxp3 + populations in the CNS' and spleens of mice with EAE were transduced into TCRβ −/− HPCs to generate retrogenic mice. Sequences β 1-β 6 were identified in the CNS' and spleens of multiple mice (CNS shared, public). Sequences β 7-β 10 were identified in the spleens of multiple mice, but in CNS tissue of only a single mouse (CNS non-shared, public). Sequence β 11 was identified in the spleen and CNS of a single mouse, and sequences β 12-β 15 were isolated from TRBV13-2 + MOG 35-55 -specific T cell hybridomas, and were not observed in any of the mice evaluated for the repertoire analyses (private). For TCRβ chains identified in a single mouse, the percent of total TRBV13-2 + TCR sequences in the CNS bearing the indicated sequence is listed in parentheses. OTIIβ comprises the TRBV13-2 + TCRβ chain from the OTII ovalbumin 323-229-specific TCR, and was assessed as a negative control.
spinal cord, consistent with optico-spinal encephalomyelitis ( Fig. 1E and F, Supplementary Fig. S1). Notably, disease in TCRβ 1 retrogenic mice was markedly accelerated, increased in incidence, and more severe than our prior results with retrogenic mice expressing five different disease-associated private MOG-specific TCRα β heterodimers 23 .
TCRβ1 imposes MOG specificity on TCRαβ heterodimers. We hypothesized that TCRβ 1 supports MOG 35-55 recognition by TCR with diverse TCRα . To establish pairing requirements, we first co-expressed TCRβ 1 in CD4 + TCRα β − 4G4 hybridomas together with 7 TCRα chains that were isolated from non-TCRβ 1 TCR. All α -TCRβ 1 combinations were expressed, and two of the seven hybrid TCR responded to MOG 35-55, indicating that TCRβ 1 can drive MOG 35-55 responsiveness ( Fig. 2A and B). TCRα form through recombination of the endogenous locus in developing thymocytes in TCRβ 1 mice, and would thus be anticipated to be highly diverse. To assess the diversity of TCRα associated with TCRβ 1 during the autoimmune response, we isolated TCRα cDNA from CNS-infiltrating T cells from 3 TCRβ 1 mice by 5′ RACE. These were heterogeneous and did not overlap between mice (Supplemental Table S1), indicating that TCRβ 1 is associated with diverse TCRα .
To quantify the functional responsiveness of TCRβ 1 + TCR compared with private TCR, five TCRα derived from CNS-infiltrating TCRβ 1 + T cells were cloned together with TCRβ 1 into polycistronic retroviral constructs.  In addition, six TCRα β from private MOG 35-55 -specific T cell hybriodomas were similarly cloned. These included previously described clones 1MOG9 and 5MOG113 23 . All TCR constructs were transduced into 4G4 CD4 + TCRα β − T cell hybridomas. Cells were sorted for co-expressed GFP and similar TCR levels. TCR avidity and MOG-sensitivity was functionally determined by stimulation with titrations of MOG 35-55, using IL-2 production as a readout. Four of the five TCRβ 1-derived TCRs demonstrated a > 2-3 log10 increased sensitivity for antigen and a dramatically increased maximal IL-2 response relative to any of the private TCR (Fig. 2C). This indicates that TCRβ 1 imposes on TCRα β an unusually high degree of responsiveness to MOG 35-55 autoantigen.
To determine whether non-transgenic T cells impede spontaneous EAE mediated by TCRβ 1, we generated chimeric retrogenic mice. We mixed wild type (WT) CD45.1 + CD45.2 − and smaller numbers of congenic TCRβ 1-transduced CD45.1 − CD45.2 + hematopoietic progenitor cells (HPCs). Approximately 40% of mice were protected from spontaneous disease. When EAE developed, symptoms were milder and mortality diminished, consistent with a protective role for the co-engrafted WT cells (Figs 1A and 3A). The ratio of TCRβ 1 and WT T cells was measured in the peripheral blood with early engraftment (d28). TCRβ 1 + cells were most often a minority, and significantly less frequent in mice that did not develop EAE compared with those that did (Fig. 3B). We anticipated that TCRβ 1 would impose MOG-recognition on unprimed T cells in healthy animals. To test this, we analyzed MOG 35-55 responsiveness in unprimed disease-free chimeric mice. T cells were labeled with cell trace violet, and stimulated either with MOG 35-55 or α CD3/α CD28. TCRβ 1 + CD45.2 + but not WT CD45.1 + T cells from unprimed disease-free mice proliferated strongly to MOG 35-55 ( Fig. 3C and D). An estimated 15.6 ± 7.8% of the initial population of CD45.2 + T cells responded to MOG 35-55 compared to 49.2 ± 12.3% to control α CD3/CD28. Alternative analyses measuring 3 H-thymidine incorporation in sorted and stimulated CD45.1 + and CD45.2 + T cells yielded similar results (Fig. 3E).

Public but not private TCRβ confer myelin specificity and provoke spontaneous autoimmunity.
TCRβ 1 is to our knowledge the first example of a single TCR chain endowing a heterogeneous population of T cells with overt spontaneous autoreactivity in mice not otherwise susceptible to spontaneous autoimmunity. To more comprehensively define the impact of public TCRβ , we generated 14 additional TCRβ retrogenic mice (Table 1). Like TCRβ 1, TCRβ 2 − 6 were identified in ≥ 9 of 12 CNS' and all spleens of mice with EAE that were analyzed (group 1; CNS-shared, public) 17 . TCRβ 7-10 were seen in a single CNS at high frequency and shared in splenocytes to varying extents (group 2; CNS non-shared, public). TCRβ 11-15 were wholly private (group 3; private). As previously reported, a large fraction of, though not all, CNS-infiltrating T cells in MOG-EAE recognize the MOG 35-55 epitope 17,24 . To minimize the possibility that TCR selected for analysis were derived from non-specific bystander T cells, group 2 and 3 TCRβ were derived from high frequency CNS-infiltrating clones (β 7-11) or from TCRα β sequences isolated from private T cell clones demonstrated to recognize MOG 35-55 autoantigen (β 12-15). For each TCRβ , retrogenic mice were monitored for clinical disease for ≥ 120 days or until the development of disease, at which time all major organs were histologically assessed. T cells from additional disease-free mice were assayed for MOG 35-55 -specific responsiveness.
Of the additional group 1 TCRβ , none developed spontaneous EAE ( Supplementary Fig. S2), though 2 of the 5 mice showed autoimmune features. Unprimed T cells from TCRβ 4 mice proliferated strongly in response to MOG 35-55 as measured both by 3 H-thymidine incorporation and membrane-associated dye dilution assays ( Fig. 4A-C). Therefore, like TCRβ 1, TCRβ 4 endows a large proportion of TCRα β with specificity for the MOG 35-55 autoantigen. TCRβ 3T cells did not respond to MOG 35-55 (Supplementary Fig. S3). However, with early engraftment these mice developed spontaneous alopecia and esophagitis (Fig. 4D-F). This was associated with prominent T cell infiltrates in these locations indicating that this CNS-associated public TCRβ can provoke alternative types of spontaneous autoimmunity.
There was no histologic or clinical evidence of disease in mice expressing any of the 4 group 2 TCRβ that were identified in a single CNS but public in the spleen (Supplementary Fig. S2). However, T cells from one of these, TCRβ 7, proliferated weakly to MOG 35-55 . This was detectable by 3 H-thymidine incorporation but not the less sensitive dye dilution assay (Fig. 4G). Mice expressing the 5 private group 3 TCRβ did not show evidence of spontaneous myelin reactivity or clinical or histologic disease ( Supplementary Figs S2 and S3). In total, three public TCRβ , two in group 1 (TCRβ 1, TCRβ 4) and one in group 2 (TCRβ 7), endowed unprimed T cells with MOG 35-55 responsiveness in combination with endogenous TCRα . Two group 1 public TCRβ chains provoked spontaneous autoreactivity (TCRβ 1, TCRβ 3). No autoimmune phenotype was observed with the enforced expression of private (group 3) TCRβ .

Discussion
We have previously shown that public TCRβ are preferentially incorporated into the CNS-infiltrating repertoire during MOG 35-55 -induced EAE 17 . By assessing 15 distinct TCRβ in vivo, we further define the differential impact of public and private receptor chains implicated in the autoimmune response. Three public, but no private TCRβ constructs incorporating six private MOG 35-55 -reactive TCRα β (1MOG213, 1MOG203, 1MOG202, 5MOG113, 1MOG9, and 1MOG23) were similarly generated. Constructs were transduced into CD4 + 4G4 TCRα β − hybridoma cells and sorted for GFP-positivity and similar levels of TCR. Cell lines were stimulated with the indicated concentration of MOG 35-55 or anti-CD3, and IL-2 production was measured at 24 h by ELISA. Private (non-TCRβ 1) TCR are shown in the left panel and CNS-infiltrating TCRβ 1 + TCR are shown in the right panel. Private TCR 1MOG9 is included in both panels to illustrate enhanced sensitivity of TCRβ 1 + TCR relative to those from private MOG-reactive TCR. Data are representative of at least 3 independent experiments performed in duplicate. were able to confer overt MOG 35-55 -reactivity to unprimed T cells expressing endogenously rearranged TCRα . Enforced expression of two of six CNS-shared TCRβ provoked spontaneous autoimmunity in a mouse strain that does not otherwise develop spontaneous disease. This implies that public TCRβ can distort repertoire responses and foster reactivity to specific autoantigens.
One hypothesis for the preferential incorporation of public sequences into the autoimmune repertoire is that these sequences predispose TCRα β toward self reactivity. Other repertoire studies have also identified public sequences among autoreactive T cells 11,25,26 . That public TCR may generically confer responsiveness to self-antigens is also suggested by our finding that transgenic expression of the public, EAE-associated TCRβ 3 chain led to the development of spontaneous alopecia areata and not EAE. Therefore, a single TCRβ may promote reactivity to disease-associated autoantigens from different tissues. In this regard, it is noteworthy that a previously isolated TRBV13-2 + TCRβ from a MOG 35-55 -specific hybridoma, 1MOG244.2, possesses two TCRα chains. Transgenic expression of one TCRα β led to MOG 35-55 reactive T cells. The second TCRα β also provoked spontaneous alopecia areata, potentially suggesting a broader association between CNS and skin reactivities 27 . It is also possible that the preferential deployment of public TCR during EAE reflects a generic increase in TCR responsiveness to antigen. Indeed, though speculative, it is possible that TCR co-evolved with MHC such that increased recognition fitness is present in the high frequency public sequences that are most likely to form. Either model is supported by our finding that TCRs utilizing the public beta chain, TCRβ 1, exhibit markedly enhanced sensitivity and maximal response when compared with control private TCR. Thus this public β chain may promote MOG 35-55 recognition by endowing TCR with a particularly high functional avidity for antigen.
We found that 2 of the 6 group 1 TCRβ (CNS-shared and public), and altogether 3 public TCRβ broadly imposed MOG-specificity on TCRα β . MOG-responsiveness was particularly prominent in mice expressing TCRβ 1, where nearly one-third the number of CD4 + T cells from disease-free animals responding to α CD3 proliferated in response to MOG 35-55. Unlike antibody-antigen interactions, which may rely on a single Ig chain, the TCR-MHC interface extensively involves both the TCRα and β surfaces. Implicitly, TCRβ 1 dominates interactions defining specificity during MOG 35-55 -IA b recognition, and this is accompanied by more generic interactions with TCRα that are simply non-disruptive and provide requisite supplemental association energy for effective T cell stimulation. It cannot be excluded that TCRβ 1 and other public TCRβ chains bind autoantigens in non-conventional manners that minimize reliance on the TCRα , and structural studies will be necessary to better resolve the physical nature of the reactivity imposed by these sequences 28 . In summary, we show that individual TCRβ sequences foster myelin antigen recognition in unprimed T cells. In a limited in vivo sampling of 15 transgenic TCRβ chains, this property was selectively observed in public TCR, providing a potential explanation for the preferential incorporation of public receptors into the autoimmune response. Flow cytometry. Cells were stained for 20 min at 4 °C in PBS containing 0.1% sodium azide and 2% (vol/vol) fetal bovine serum (FBS). Monoclonal antibodies specific for CD4 (clone RM4-5), CD8 (clone 53-6.7), TCRβ (clone H57-597), CD69 (clone H1-2F3), CD45.1 (clone A20) and CD45.2 (clone 104) were purchased from BD Biosciences. Intracellular staining of Foxp3 (clone FJK-16s) was performed using the Foxp3 Staining Buffer Set (eBioscience). For cytokine staining, cells were cultured for 4 h at 37 °C with Cell Stimulation Cocktail (eBioscience) in the presence of 10 μ g/mL monensin (eBioscience), followed by fixation, permeabilization, and staining for IL-17A (clone eBio17B7, eBioscience) and IFN-γ (clone XMG1.2, BD Biosciences). Flow cytometric analysis was performed on an LSRFortessa (BD Biosciences) and analyzed with FlowJo software (Tree Star).  -TC GA GT TG GC TA CC CC CT  CT CA GA CA TC AG TG TA CT TC TG TG CC AG CG GT GA GA CT GG GG GA AA CT AT GC TG AG CA GT  TCTTCGGACCAGGGACACGACTCACCGTCCTAGAA-3′ ; 5′ -GATCTTCTAGGACGGTGAGTCG  TGTCCCTGGTCCGAAGAACTGCTCAGCATAGTTTCCCCCAGTCTCACCGCTGGCACAGAAGTACAC  TGATGTCTGAGAGGGGGTAGCCAAC-3′ ). This was subcloned as a XhoI/BglII fragment into the previously cloned V-C region of the 1MOG244.2 TRBV13-2 TCRβ to synthetically recreate TCRβ 1 23 . Other TCRβ constructs were similarly constructed. The OTII TCRβ was PCR amplified (5′ -GCCGAATTCGCCACCATGTCT AACACTGCCTTC-3′ ; 5′ -GTCACATTTCTCAGATCTTCTAG-3′ ) and then subcloned into the EcoRI/BglII sites of MSCV-TCRβ 1-GFP to replace the TCRβ 1V and J domains. For polycistronic TCRα β constructs, TCRα counts (black line), cell count data fit to proliferation model (red line), and individual proliferative generations (blue lines). (D) Summary data from T cells from individual mice stimulated as in (C). The magnitude of each division peak was divided by 2 n , where n = division peak number, to estimate numbers of parental cells whose progeny populated an individual peak. (E) CD4 + TCR + CD45.2 + (TCRβ 1 + ) and CD4 + TCR + CD45.1 + (WT) T cells were sorted from 8 wk chimeric mice without current or historical signs of EAE. The cells were stimulated as indicated and proliferation measured on day 3 by 3 H-thymidine incorporation. ***p ≤ 0.001.
Clinical evaluation. Cohorts of retrogenic mice were clinically monitored for ≥ 120 days. Mice were submitted for histopathologic examination either during peak disease or after 120 days if healthy. Full necropsy, including of CNS tissues, was performed on at least three mice for each cohort. Paraffin-embedded tissue samples were stained with hematoxylin and eosin (H&E) and, where appropriate, CD3. The severity of spontaneous EAE was scored by using the predetermined qualitative and semi-quantitative criteria: 0, lesions absent, 1; minimal to mild inconspicuous lesions; 2, conspicuous lesions; 3, prominent multifocal lesions; 4, marked coalescing lesions.
Mouse T-Activator CD3/CD28 Dynabeads (Invitrogen) were added where indicated at a 1:1 bead-to-cell ratio. Alternatively, cells were labeled with 5 μ M CellTrace Violet (Invitrogen) prior to stimulation according to the manufacturer's instruction. Cells were stained with surface markers and 7-AAD (BD Biosciences) and T cell proliferation was measured by dye dilution. Proliferation analysis was performed with Flowjo software. Cytokine analysis. Culture supernatants from primary T cells were collected at 48 h and analyzed for IL-2, IL-4, IL-10, IFN-γ , and IL-1α using the Milliplex MAP mouse cytokine/chemokine immunoassay kit (Millipore) on a Luminex (Bio-Rad) instrument. For hybridomas, supernatant was assessed at 24 h for IL-2 only. For intracellular cytokine staining, cells were cultured with Cell Stimulation Cocktail and 10 μ g/ml monensin (eBioscience), for 4 h at 37 °C, followed by fixation, permeabilization, and intracellular staining for IL-17A and IFN-γ .
Statistics. Means, SDs, and Kaplan Meier curves were calculated in Excel or PRISM. Plots demonstrate mean ± 1SD. Two-tailed student t-tests were applied to compare any two groups and ANOVA for three or more groups. For multiple comparisons, significance is shown only for indicated groups. A p ≤ 0.05 was considered statistically significant.