Identification of natural regulatory T cell epitopes reveals convergence on a dominant autoantigen

SUMMARY

Regulatory T (Treg) cells expressing the transcription factor Foxp3 are critical for the prevention of autoimmunity and the suppression of anti-tumor immunity. The major self antigens recognized by Treg cells remain undefined, representing a substantial barrier to the understanding of immune regulation. Here, we have identified natural Treg cell ligands in mice. We found that two recurrent Treg cell clones, one prevalent in prostate tumors and the second associated with prostatic autoimmune lesions, recognized distinct non-overlapping MHC class-II-restricted peptides derived from the same prostate-specific protein. Notably, this protein is frequently targeted by autoantibodies in experimental models of prostatic autoimmunity. Based on these findings, we propose a model in which Treg cell responses at peripheral sites converge on those self proteins that are most susceptible to autoimmune attack, and we suggest that this link may be exploited as a generalizable strategy to identify the Treg cell antigens relevant to human autoimmunity.

eTOC

The endogenous antigens recognized by thymus-derived Treg cells have remained largely undefined. Leonard et al. identify natural Treg cell ligands in mice, demonstrating that two recurrent Treg cell clones recognize distinct non-overlapping peptides derived from a single prostate-specific protein.

INTRODUCTION

The immune system generates a diverse repertoire of conventional CD4⁺ T cell clones capable of responding to foreign antigens, while restricting immune responses directed at self antigens. Each CD4⁺ T cell expresses a unique T cell receptor (TCR) capable of recognizing major histocompatibility complex class II molecules (MHC-II) complexed with short peptides, which are generated from intact proteins by antigen processing. In the process of negative selection, many CD4⁺ T cells exhibiting strong reactivity to self peptide-MHC-II (pMHC-II) are eliminated from the conventional T (Tconv) cell repertoire by clonal deletion or differentiation into innate-like T cell lineages (Klein et al., 2014). In contrast, some CD4⁺ thymocytes exhibiting overt reactivity to self pMHC-II ligands differentiate into regulatory T (Treg) cells expressing the transcription factor Foxp3 (Hsieh et al., 2012), which function in the periphery to maintain immune homeostasis and suppress autoreactive Tconv cells that evade negative selection. Thus, reactivity to self antigen is crucial for the establishment of two major tolerance mechanisms - negative selection and Treg cell development, which function in concert to restrict immune responses to self tissues. Beyond the role of self pMHC-II recognition in directing Treg cell development in the thymus, the continued recognition of self antigen outside the thymus is critical for orchestrating Treg cell differentiation, homeostasis, and suppressor activity. Given that self pMHC-II recognition is central to many facets of Treg cell biology, it is essential to identify the endogenous peptides that trigger Treg cell development in the thymus and are engaged by Treg cells to coordinate immune suppression in the periphery. However, due to technical challenges associated with identifying MHC-II-restricted self peptides, the natural antigens recognized by thymus-derived Treg (tTreg) cells have remained undefined. Without this knowledge, it has not been possible to gain a complete understanding of why Treg cell-mediated suppression is subverted in autoimmune and inflammatory diseases, and how Treg cells are co-opted by developing cancers to suppress anti-tumor immunity.

The paradigm that self pMHC-II recognition via the TCR drives both the thymic development and peripheral function of Treg cells is supported by a large body of evidence. Early studies in mice utilizing TCR repertoire profiling revealed that the TCRs expressed by peripheral Treg cells are largely distinct from those expressed by Tconv cells (Hsieh et al., 2004; Hsieh et al., 2006; Lathrop et al., 2008; Wong et al., 2007), demonstrating that the formation of the Treg cell repertoire is an antigen-driven, TCR-dependent process. Consistent with this, developmental studies show that Treg cell-derived TCRs facilitate thymic differentiation into the Treg cell lineage, whereas Tconv-expressed TCRs are inefficient at directing this process (Bautista et al., 2009; Leung et al., 2009). Similarly, transgenic expression of a model antigen containing a pMHC-II-binding peptide in the thymus promotes the development or survival of antigen-specific Treg cells (Hsieh et al., 2012), indicating that TCR-dependent agonist signals promote thymic (t)Treg cell development. More recent studies demonstrate that the thymic development of some Treg cell specificities is dependent on the expression of Autoimmune regulator (Aire) (Malchow et al., 2013; Perry et al., 2014), a transcription factor that drives the promiscuous expression of tissue-specific antigens in the thymus (Anderson et al., 2002; Derbinski et al., 2005). In the periphery, a substantial proportion of Treg cells proliferate (Fisson et al., 2003; Smigiel et al., 2014) or perceive strong TCR signals (Moran et al., 2011) at steady state, suggesting that many peripheral Treg cells actively recognize agonist pMHC-II ligands in the absence of inflammation. Likewise, T cells transduced with Treg-derived TCRs can proliferate following transfer into lymphopenic mice, suggesting that some Treg cell TCRs confer reactivity to self pMHC-II complexes (Hsieh et al., 2004). Finally, recent work demonstrates that conditional ablation of the TCR on Foxp3⁺ cells results in systemic autoimmunity (Levine et al., 2014; Vahl et al., 2014), demonstrating an important role for TCR-dependent signals in Treg cell differentiation and function in the periphery.

The gap in knowledge regarding the identity of endogenous antigens recognized by Treg cells has restricted progress in several fundamental areas of Treg cell biology. First, it has not been possible to analyze the phenotype, frequency, anatomical distribution, and repertoire complexity of Treg cells reactive to natural self antigens that direct Treg cell differentiation in the thymus. Second, it is unknown whether tTreg TCRs recognize endogenous ligands with unique biochemical or structural characteristics, such as binding affinity, docking mode, or conformational changes, that might provide insight into the mechanisms of Treg cell-mediated suppression. Third, given the vast array of potential self peptides in the body, it is unclear whether the protection of tissues from autoimmune attack is dependent on the presence of Treg cells reactive to particular self pMHC-II complexes. Fourth, it has been challenging to clearly elucidate why some self-reactive T cells are purged by clonal deletion whereas others are directed into the Treg cell lineage, and whether the affinity of TCR-pMHC-II binding is a primary determinant of these alternate cell fates.

Here we report the identification of the peptide antigens recognized by two recurrent prostate-specific Treg cell clones, one prevalent in mouse prostate tumors and the second associated with autoimmune lesions of the prostate. These endogenous Treg cell clones recognize distinct non-overlapping MHC-II-restricted peptides from a prostate-specific protein that is frequently targeted by autoantibodies in multiple mouse models of autoimmunity. The link between the antigen specificity of Treg cells and that of autoantibodies that emerge in settings of immune dysregulation has implications for our understanding of Treg cell-mediated tolerance, and reveals a potential strategy for identifying additional Treg cell antigens in mice and humans.

RESULTS

Identification of a Self pMHC-II Antigen Recognized by MJ23 Treg Cells

To reveal the nature of self peptides recognized by tTreg cells, we aimed to identify a pMHC-II ligand recognized by a naturally occurring Treg cell clone named “MJ23”, which we identified previously based on its predominance in mouse prostate tumor lesions(Malchow et al., 2013). The MJ23 TCR confers reactivity to a prostate-associated, MHC-II-restricted antigen of unknown identity, and mediates Treg cell development in the thymus via an Aire-dependent process(Malchow et al., 2013). To screen for antigenic activity, we used an in vitro culture system to monitor the proliferation of TCR transgenic T cells expressing the MJ23 TCR (MJ23tg T cells). Using this system, we found that splenic dendritic cells (DCs) cultured with protein extracts from the dorsolateral or anterior (but not ventral) lobes of the prostate induced proliferation of MJ23tg T cells (Figure 1A), demonstrating that the relevant peptide can be isolated and detected using T cell-based assays.

Figure 1.

MJ23 T cells recognize an antigen derived from the prostatic protein Tcaf3. CD4⁺ T cells were isolated from MJ23tg Rag1^−/− CD45^.1/.1 or OT-IItg Rag1^−/− CD45^.1/.1 mice, labeled with CellTrace-Violet (CTV), and used as a probe for antigen. (a) In vitro stimulation of MJ23tg T cells by prostatic secretory extracts. 1 × 10⁴ MJ23tg T cells were cultured with 5 × 10⁴ CD11c⁺ cells from B6 spleen, plus secretory extracts prepared from the anterior, dorsolateral, or ventral prostate lobes of tumor-bearing TRAMP males, with or without anti-MHC-II blocking antibody. Dilution of CTV was assessed by flow cytometry on day 5. (b) In vitro stimulation of MJ23tg T cells by Tcaf3 protein. As in (a), MJ23tg or OT-IItg T cells were stimulated in vitro with 2 µg/mL recombinant Tcaf3 protein or 1 µM Ova_323–339 peptide, and assayed on day 5. (c) In vivo stimulation of MJ23tg T cells by Tcaf3 protein. 1 × 10⁵ MJ23tg T cells were transferred i.v. into congenically disparate B6 female hosts. 2 hours after transfer, recipients were immunized with 5 µg Tcaf3 protein, 5 µg Tgm4 protein, or PBS alone. CD4⁺ T cells from the spleen (left panel) and pooled skin-draining lymph nodes (right panel) were analyzed for CTV dilution on day 5. (d) In vitro stimulation of MJ23tg T cells by Tcaf3 peptide. As in (a), MJ23tg or OT-IItg T cells were stimulated in vitro with 5 nM Tcaf3_646–658 peptide, with or without anti-MHC-II Ab or isotype control. Dilution of CTV was analyzed on day 3. (e) Tcaf3_646–658 peptide truncation analysis. As in (a), MJ23tg T cells were stimulated in vitro with 33 nM Tcaf3_646–658 peptide variants, comprising serial truncations from the N- and/or C-termini. Dilution of CTV was analyzed on day 3. The core nonamer epitope predicted computationally is denoted by red shading. Percent of cells proliferated is shown as the mean ± SEM of three replicates. (f) Tcaf3_646–658 peptide is required for the thymic development of MJ23tg Treg cells. Bulk thymocytes from MJ23tg Rag1^−/− CD45^.1/.1 females were transferred into 4-6-week-old Tcaf3^+/+, Tcaf3^+/tm1, or Tcaf3^tm1/tm1 hosts, both male and female, and analyzed at day 7 for expression of CD25 and Foxp3. Left panels, representative flow cytometric analysis of Foxp3 and CD25 expression on MJ23tg and polyclonal thymocytes for recipients of the indicated genotype. Right panel, summary plot of data for MJ23tg thymocyte transfer into Tcaf3^tm1/tm1 or Tcaf3⁺ (Tcaf3^+/+ and Tcaf3^+/tm1) recipients. Significance testing was performed using the Student’s t-test. ** indicates p < 0.01. (g) Tcaf3-specific autoantibodies can be detected in the serum of Aire^−/−males. Recombinant Tcaf3 protein was resolved by SDS-PAGE, and subjected to Western blotting using serum from Aire-deficient (Aire^−/−) or Aire⁺ (Aire^+/+ or Aire^+/−) littermates of the indicated ages. Data are representative of multiple independent experiments: (a) N = 4, (b) N = 3, (c) N = 3, (d) N = 5, (e) N = 3, (f) N = 2, (g) N = 3. See also Figure S1.

We reasoned that the MJ23 peptide antigen would fit three criteria: 1) the protein containing the peptide would be exclusively or preferentially expressed in the prostate; 2) the gene encoding this protein would be a transcriptional target of Aire in medullary thymic epithelial cells(Anderson et al., 2002; Derbinski et al., 2005; Sansom et al., 2014); 3) this gene would not be located on the Y chromosome, because MJ23tg Treg cells undergo thymic development in both male and female mice (Malchow et al., 2013).

Using data from a transcriptional profiling study (Sansom et al., 2014), we identified 20 candidate genes fitting these criteria (Table S1), and expressed the corresponding proteins recombinantly in insect cells. We cultured DCs with individual proteins, and assayed for their capacity to induce proliferation of MJ23tg T cells (Figure 1B and S1A–B). One of these candidates, encoded by TRPM8 channel-associated factor 3 (Tcaf3, also known as Eapa2 or Fam115e), induced robust stimulation of MJ23tg T cells in vitro (Figure 1B). Stimulation was abolished by anti-MHC-II blocking antibody, and Tcaf3 did not induce stimulation of transgenic T cells of an irrelevant specificity(Barnden et al., 1998) (Figure 1B). Immunization with recombinant Tcaf3 protein, but not with another Aire-dependent prostatic protein (transglutaminase 4, Tgm4; Table S1) , induced robust proliferation of MJ23tg T cells in vivo (Figure 1C), demonstrating that the Tcaf3-derived peptide can be efficiently processed and presented in vivo. To identify the peptide epitope recognized by MJ23tg T cells, we used in silico approaches(Vita et al., 2015; Zhu et al., 2003) to generate a list of Tcaf3-derived peptides predicted to bind I-A^b, the only functional MHC class II gene expressed in these mice. Of these candidate peptides, we identified one peptide of sequence THSKAPWGELATD, spanning residues 646–658 of Tcaf3, that robustly stimulated MJ23tg T cells in vitro (Figure 1D). Subsequent analysis of peptide truncations revealed the minimal core epitope to be the nonamer SKAPWGELA (residues 648–656) (Figure 1E), the same core sequence that was predicted computationally to bind to I-A^b. Taken together, our results identify Tcaf3_646–658 as the I-A^b-restricted peptide recognized by prostate-specific MJ23 Treg cells.

To verify that Tcaf3_646–658 is the antigenic peptide recognized by prostate-specific MJ23 Treg cells in vivo, we generated mice with targeted deletion of the region encoding the Tcaf3_646–658 13-mer peptide (hereafter referred to as Tcaf3^tm1 mice). To test the requirement for this epitope in driving MJ23 Treg development in vivo, we isolated thymocytes from MJ23tg Rag1^−/− females, which harbor no Foxp3⁺ cells due to intraclonal competition (Malchow et al., 2013), and transferred these cells into new hosts by intrathymic injection. Donor MJ23tg thymocytes readily differentiated into Foxp3⁺ cells in Tcaf3^+/+ hosts, but failed to do so in Tcaf3^tm1/tm1 mice (Figure 1F). Consistent with this, prostatic extracts isolated from Tcaf3^tm1/tm1 mice failed to stimulate MJ23tg T cells in vitro (Figure S1C), demonstrating a specific requirement for Tcaf3_646–658 peptide. These data provide loss-of-function evidence that the Tcaf3_646–658 peptide is a natural epitope recognized by Treg populations in vivo, and demonstrates that the thymic development of this tissue-specific Treg cell clone requires thymic expression of a single self peptide.

Mouse Tcaf3 is a 102-kDa protein of unknown function that is exclusively expressed in the dorsolateral and anterior lobes of the mouse prostate(Fujimoto et al., 2006), consistent with our MJ23tg T cell stimulation data (Figure 1A). Of note, previous studies have demonstrated that Tcaf3 is targeted by autoantibodies in multiple experimental mouse models characterized by prostatic autoimmunity, including neonatal thymectomy(Setiady et al., 2006), castration(Meng et al., 2011), and Aire deficiency(Setiady et al., 2006). Concordant with these published results, we recurrently detected serum antibodies reactive to recombinant Tcaf3 protein in Aire-deficient males greater than 10 weeks of age, but not in age-matched, Aire-sufficient controls (Figure 1G). Thus, cumulative evidence demonstrates that Tcaf3 is highly susceptible to recognition by autoantibodies in settings of immune dysregulation.

Identification of Endogenous Tcaf3-specific T Cells Using pMHC-II Tetramers

Having identified the peptide ligand recognized by the MJ23 Treg cell clone, we aimed to directly identify, enumerate, and characterize endogenous, polyclonal Tcaf3_646–658-specific T cells using pMHC-II tetramers. We generated fluorescently labeled I-A^b tetramers bearing a variant of the antigenic Tcaf3_646–658 peptide (hereafter referred to as Tcaf3/I-A^b tetramers)(Malhotra et al., 2016). We used a peptide variant in which serine 648, predicted to lie at an MHC-binding anchor position, was changed to tyrosine, a preferred anchor residue at this position(Liu et al., 2002; Zhu et al., 2003). This alteration enabled production of stable Tcaf3/I-A^b tetramers, and slightly enhanced the potency of peptide stimulation of MJ23tg T cells in vitro (Figure 2A). Tcaf3/I-A^b tetramers bearing this variant peptide stained MJ23tg T cells, but did not bind non-specifically to polyclonal cells (Figure 2B), demonstrating specificity of binding.

Figure 2.

Analysis of Tcaf3-specific MJ23tg T cells using pMHC-II tetramers bearing a Tcaf3_646–658 variant. (a) Potency of wild-type and variant Tcaf3 peptides. MJ23tg T cells were stimulated in vitro as in Figure 1 with wild-type Tcaf3_646–658 peptide (Tcaf3-WT) or Tcaf3_646–658 harboring a serine to tyrosine mutation at residue 648 (Tcaf3-648Y). Left, representative flow cytometry plots of in vitro cultures with 4 nM peptide, analyzed at day 3. Right, dose response curves fit to a cooperative model. Points denote the mean ± SEM of three replicates. (b) Staining of MJ23tg T cells by Tcaf3/I-A^b tetramers. MJ23tg Rag1^−/− CD45^.1/.1 T cells (red) were spiked into polyclonal CD45^.2/.2 B6 splenocytes (blue) and co-stained with PE- and APC-labeled tetramers of I-A^b bearing the Tcaf3-648Y peptide (Tcaf3/I-A^b tetramers) at the indicated concentrations. (c–e). Expansion of endogenous Tcaf3-specific Treg cells by immunization with Tcaf3 peptide plus CFA. B6 (Aire⁺) or Aire^−/− male mice were immunized with CFA only, Tcaf3_646–658 peptide in CFA, or 2W1S peptide in CFA, as indicated. 14 days post-challenge, lymphocytes from the pooled spleen and lymph nodes were co-stained with PE- and APC-labeled Tcaf3/I-A^b tetramers or 2W1S/I-A^b tetramers, as indicated, and tetramer-binding cells were magnetically enriched. (c) Representative flow cytometric analysis of Foxp3 expression by magnetically enriched CD4⁺ T cells. Plots in the left column depict tetramer-PE vs. tetramer-APC staining, with double-tetramer^neg and double-tetramer⁺ gates shown. The middle and right columns present histograms of Foxp3 expression by cells within the double-tetramer^neg and double-tetramer⁺ gates, respectively. (d–e) Summary plots of the tetramer analysis in (c), depicting the total number of double-tetramer⁺ cells that express Foxp3 (d), and the percentage of double-tetramer⁺ T cells (e). Data are representative of multiple independent experiments: (a) N = 5, (b) N = 5, (c–e) N = 2. The mean ± SEM is indicated. Significance testing was performed using the nonparametric Mann-Whitney test. * indicates p < 0.05.

Using this reagent, we first sought to characterize endogenous, polyclonal Tcaf3_646–658-specific T cells in healthy mice. To do this, we employed an established approach in which rare antigen-specific T cells were expanded in vivo by immunization with peptide plus Complete Freund’s Adjuvant (CFA), and expanded T cells were analyzed 14 days later following magnetic enrichment of Tcaf3/I-A^b-tetramer⁺ T cells from the pooled spleen and lymph nodes(Legoux and Moon, 2012; Malhotra et al., 2016; Moon et al., 2007). T cells were co-stained with both PE- and APC-labeled tetramers simultaneously to reduce false-positive staining(Stetson et al., 2002; Tubo et al., 2013). As a control, we immunized mice with the I-A^b-restricted foreign peptide 2W1S and used 2W1S/I-A^b tetramers to characterize antigen-specific T cells(Moon et al., 2007). Challenge of male mice with CFA plus Tcaf3_646–658 peptide induced expansion of Tcaf3/I-A^b-tetramer⁺ T cells, yielding an average of ~955 (± 240 SEM) Tcaf3/I-A^b-tetramer⁺ cells (Figure 2C, 2D). Approximately 64% (± 4.7 SEM) of Tcaf3/I-A^b-tetramer⁺ T cells in immunized mice expressed Foxp3, a critical Treg cell transcriptional regulator and marker of the Treg cell lineage (Figure 2C, 2E), demonstrating that expanded Tcaf3_648–656-specific T cells are predominantly found in the Treg cell subset. In contrast, challenge with CFA plus 2W1S peptide yielded ~4,150 (± 1,500 SEM) 2W1S/I-A^b-tetramer⁺ cells (Figure 2C, 2D), of which only a minor fraction (~11.7% ± 3.2 SEM) expressed Foxp3 (Figure 2C, 2E), consistent with previous studies (Malhotra et al., 2016). Additionally, analysis of Aire^−/− mice immunized with CFA plus Tcaf3_646–658 peptide revealed that only a minor fraction of Tcaf3/I-A^b-tetramer⁺ cells expressed Foxp3 following expansion (Figure 2D, 2E), consistent with previous work demonstrating a role for Aire in directing MJ23 T cells into the Treg cell lineage (Malchow et al., 2016; Malchow et al., 2013). Thus, our data demonstrate that polyclonal Tcaf3_646–658-specific T cells can be expanded in vivo by immunization with peptide plus CFA, and expanded Tcaf3/I-A^b-tetramer⁺ T cells are skewed to the Treg cell lineage in wild-type B6 mice.

The Tcaf3-specific MJ23 TCR was originally identified based on recurrent expression of this TCR by Foxp3⁺ Treg cells infiltrating the prostate tumors of TRAMP mice(Greenberg et al., 1995; Malchow et al., 2013). Given that the MJ23 TCR represents a single Tcaf3_646–658-specific TCR clone, it was not previously possible to quantify the total contribution of polyclonal Tcaf3_646–658-specific Treg cells to the T cell infiltrate of TRAMP prostate tumor lesions. To address this, we performed tetramer analysis of T cells isolated from 6–7-month-old TRAMP males bearing late-stage prostate tumors. Tcaf3/I-A^b tetramer⁺ T cells were readily detected in all TRAMP prostate tumors examined, comprising, on average, ~2.1% (± 0.63 SEM) of all tumor-infiltrating Treg cells (Figure 3A, 3B). Consistent with the above results and previous studies of the MJ23 clone(Malchow et al., 2016; Malchow et al., 2013), prostate tumor-infiltrating Tcaf3/I-A^b tetramer⁺ T cells were strongly skewed to the Foxp3⁺ Treg cell subset (Figure 3A, 3B).

Figure 3.

Detection of polyclonal Tcaf3/I-A^b-specific T cells in prostate tumors and prostatic autoimmune lesions. Tcaf3/I-A^b tetramer analysis of T cells from the prostates of tumor-bearing TRAMP males (a–b), Aire^+/+males (c–d) and Aire^−/− males (e–f). (a) Representative flow cytometric analysis of CD4⁺ T cells isolated from the dorsolateral lobe of prostate tumors from 6-7-month old TRAMP^+/+ males, stained with Tcaf3/I-A^b or negative control 2W1S/I-A^b tetramers. Numbers indicate percentage of cells co-stained with PE- and APC-labeled tetramers (oval gates). (b) Summary plots of tetramer analysis in (a) for N = 12 mice. The mean ± SEM is shown. (c) Representative flow cytometric analysis of tetramer staining, as in (a), of CD4⁺ cells isolated from the prostates of 7-8-month-old Aire^+/+ males. (d) Summary plots of analysis in (c). (e) Representative flow cytometric analysis of tetramer staining of CD4⁺ cells isolated from the prostates of 7-8-month-old Aire^−/− males. (f) Summary plots of analysis in (e). (g) Representative analysis of Tcaf3/I-A^b tetramer-PE vs. Foxp3 staining of CD4⁺ T cells from the prostates of TRAMP vs. Aire^−/− mice. (h) Summary plot of the proportion of Tcaf3/I-A^b double tetramer⁺ cells (oval gates in (a) and (e)) that express Foxp3. Data are representative of multiple independent experiments: (a–b) N = 2, (c–d) N = 2, (e–f) N = 2. Tetramer analyses were performed with 12.5 nM (Aire^−/− and Aire^+/+) or 20 nM (TRAMP) of each tetramer. Significance testing was performed using the nonparametric Mann-Whitney test. ** indicates p < 0.001; *** indicates p < 0.0001; n.s., not significant.

In previous studies, we demonstrated that Aire is critical for the thymic development of monoclonal MJ23 Treg cells(Malchow et al., 2013), and showed that in the absence of Aire, MJ23 T cells emerge in the Foxp3^neg Tconv cell subset and recurrently infiltrate autoimmune prostatic lesions of Aire^−/− males(Malchow et al., 2016). We therefore predicted that polyclonal Tcaf3_646–658-reactive T cells would exhibit a similar pattern. Consistent with this hypothesis, analysis of Tcaf3/I-A^b tetramer⁺ cells from the pooled spleen and lymph nodes of Aire^−/− males immunized with CFA alone or CFA plus Tcaf3_646–658 peptide revealed that only a minor fraction of Tcaf3_646–658-specific cells expressed Foxp3 (Figure 2D, 2E). Likewise, direct analysis of prostate-infiltrating T cells demonstrated that Tcaf3/I-A^b tetramer⁺ cells were nearly undetectable in the prostates of 7–8-month-old tumor-free Aire^+/+ mice, which harbor very few T cells due to a lack of inflammation (Figure 3C, 3D), but were readily detected in the prostates of age-matched Aire^−/− males (Figure 3E, 3F). Notably, Tcaf3_646–658-specific T cells from the prostates of Aire^−/− mice were strongly skewed to the Foxp3^neg Tconv cell subset (Figure 3G, 3H), contrasting with polyclonal Tcaf3-specific T cells infiltrating TRAMP prostate tumors (Figure 3G, 3H). Thus, our cumulative data demonstrate that Aire plays a critical role in directing polyclonal Tcaf3_646–658-specific T cells into the Treg cell lineage.

A Second Prostate-Associated Treg Cell Clone Recognizes a Distinct Tcaf3-Derived Peptide

Having identified Tcaf3_646–658 as the self peptide recognized by MJ23 Treg cells, we reasoned that a similar approach could be used to identify antigenic peptides recognized by additional Aire-dependent, prostate-specific Treg cell clones. In previous work, we identified multiple T cell clones that are strongly skewed to the Treg cell compartment in Aire^+/+ mice, but are misdirected to become pathogenic, prostate-infiltrating Tconv cells in Aire^−/− mice (Malchow et al., 2016). The most abundant of these Aire-dependent, prostate-associated clones, named “SP33” in the current study, expressed a TCRα chain of complementarity determining region 3 (CDR3)α sequence TRAV9D–ALSMSVNYQLI paired with the same fixed transgenic TCRβ chain as the MJ23 clone ((Malchow et al., 2016) and Methods). Of note, compared to the MJ23 TCRα (TRAV14-LYYNQGKLI), the SP33 TCRα chain utilizes a different V region segment and exhibits little sequence similarity within the CDR3α. To study the peptide specificity of the SP33 TCR, we generated SP33 TCR “retrogenic” (SP33rg) mice (McDonald et al., 2015; Turner et al., 2010) in which bone marrow progenitors retrovirally transduced with an SP33 TCR expression construct were engrafted into host mice. Using resulting SP33rg T cells as a probe for pMHC-II ligand, we found that DCs cultured with protein extracts from the dorsolateral or anterior prostatic lobes stimulated CD4⁺ SP33rg T cells in vitro (Figure 4A), mirroring the results of MJ23 T cell stimulation assays (Figure 1A). Screening of the candidate prostate-specific proteins described above revealed that Tcaf3, the same protein containing the peptide ligand recognized by MJ23 T cells, robustly stimulated SP33rg T cells in vitro (Figure 4B) and in vivo (Figure 4C).

Figure 4.

SP33 T cells recognize a distinct epitope derived from the prostatic protein Tcaf3. CD4⁺ T cells were sorted from SP33rg or OT-IItg Rag1^−/− CD45^.1/.1 mice, labeled with CellTrace-Violet (CTV), and used as a probe for antigen. (a) In vitro stimulation of SP33rg T cells by prostatic secretory extracts. 1 × 10⁴ SP33rg T cells were cultured with 5 × 10⁴ CD11c⁺ cells from B6.SJL spleen, plus secretory extracts prepared from the anterior, dorsolateral, or ventral prostate lobes of tumor-bearing TRAMP males with or without anti-MHC-II blocking antibody. Dilution of CTV was assessed by flow cytometry on day 5. (b) In vitro stimulation of SP33rg T cells by Tcaf3 protein. As in (a), SP33rg or OT-IItg T cells were stimulated in vitro with 2 µg/mL recombinant Tcaf3 protein or 1 µM Ova_323–339 peptide, and assayed on day 3. (c) In vivo stimulation of SP33rg T cells by Tcaf3 protein. 6.6 × 10⁴ MJ23tg T cells were transferred i.v. into congenically disparate B6.SJL female hosts. 2 hours after transfer, recipients were immunized with 5 µg Tcaf3 protein, 5 µg Tgm4 protein, or PBS alone. CD4⁺ T cells from the spleen (left panel) and pooled skin-draining lymph nodes (right panel) were analyzed for CTV dilution on day 5. (d) In vitro stimulation of SP33rg T cells by Tcaf3_88–107 peptide. As in (a), SP33rg or OT-IItg T cells were stimulated in vitro with 5 nM Tcaf3_88–107 peptide, with or without anti-MHC-II Ab or isotype control, and 5 nM Tcaf3_646–658 peptide. Dilution of CTV was analyzed on day 3. (e) Tcaf3_88–107 peptide truncation analysis. As in (a), SP33rg T cells were stimulated in vitro with 33 nM Tcaf3_88–107 peptide variants, comprising truncations from the N- and/or C-termini. Dilution of CTV was analyzed on day 3. The core nonamer epitope predicted computationally is denoted by red shading. Percent of cells proliferated is shown as the mean ± SEM of three replicates. Data are representative of multiple independent experiments: (a) N = 3, (b) N = 3, (c) N = 3, (d) N = 3, (e), N = 3. Significance testing was performed using the nonparametric Mann-Whitney test. * indicates p < 0.05. See also Figure S1.

Given that the TCRα chains of the MJ23 and SP33 TCRs exhibit little sequence similarity, we predicted that the MJ23 and SP33 TCRs may recognize distinct Tcaf3-derived peptides. Consistent with this, screening of the panel of synthetic Tcaf3 peptides demonstrated that SP33rg T cells were robustly stimulated by a peptide of sequence CPGAPIAVHSSLASLVNILG (Tcaf3_88–107) (Figure 4D and 4E), and were not stimulated by the MJ23 agonist peptide THSKAPWGELATD (Tcaf3_646–658) (Figure 4D). Taken together, our data demonstrate that two recurrent Treg cell clones, the first identified in mouse prostate tumors (MJ23) and the second identified in prostatic autoimmune lesions in Aire^−/− mice (SP33), recognize two distinct peptides derived from a single prostate-specific protein, Tcaf3.

DISCUSSION

Due to the technical challenges associated with identifying MHC-II-restricted self peptides, the natural antigens recognized by recurrent thymus-derived Treg cell populations had thus far remained elusive. In this study, we identified endogenous Treg cell ligands in mice, demonstrating that two recurrent Treg cell clones, the first identified in mouse prostate tumors (MJ23) and the second associated with prostatic autoimmune lesions (SP33), recognize distinct non-overlapping peptides derived from a single prostate-specific protein, Tcaf3. By focusing our efforts on Aire-dependent Treg specificities reactive to tissue-specific antigens, we generated a tractable list of candidates using available transcriptional profiling data, which enabled screening for antigenic activity using standard immunological assays. As discussed below, the findings and experimental approach described here have implications for basic Treg cell biology and for the discovery of additional Treg cell ligands in mice and humans.

Previous reports demonstrate that a minor fraction of Foxp3⁺ Treg cells can be identified within antigen-specific T cell populations reactive to both self and foreign pMHC-II (Malhotra et al., 2016; Moon et al., 2011; Su et al., 2016). However, these populations do not reflect naturally occurring antigen specificities that drive the robust selection of Treg cells in the thymus. For example, immunization with self peptide plus adjuvant is commonly used to induce tissue-specific autoimmunity in animal models such as experimental autoimmune uveitis and experimental autoimmune encephalomyelitis (Chen et al., 2015; Rangachari and Kuchroo, 2013). In such models, immunization induces the emergence of both Tconv and Treg cells specific for the antigenic self peptide (Kieback et al., 2016; Korn et al., 2007; Silver et al., 2015). However, in these settings, antigen-specific Treg cells are outnumbered by pathogenic Tconv cells of the same specificity and fail to prevent autoimmune pathology (Kieback et al., 2016; Korn et al., 2007; Silver et al., 2015). Furthermore, it is unclear whether Foxp3⁺ Treg cells specific for these self peptides are present in the endogenous repertoire in the absence of peptide immunization. These findings suggest that these self peptides do not efficiently direct T cells into the Treg cell lineage. In contrast to these studies, our previous work using TCR repertoire profiling has demonstrated that at steady state, in the absence of peptide immunization, the Tcaf3-specific T cell clones MJ23 and SP33 are strongly skewed to the Foxp3⁺ Treg cell lineage (Malchow et al., 2016; Malchow et al., 2013). The importance of directing these clonotypes into the Treg cell lineage is revealed when this process is dysregulated in Aire^−/− mice, in which the MJ23 and SP33 clones are misdirected into the Tconv subset and infiltrate autoimmune lesions of the prostate (Malchow et al., 2013). Our current data using pMHC-II tetramers corroborate and extend these findings, demonstrating that endogenous polyclonal Tcaf3_648–656/I-A^b-specific T cells isolated from prostate tumors are strongly biased to the Treg cell subset, and that Tcaf3_648–656/I-A^b-specific T cells expanded by immunization with peptide plus CFA are skewed to the Foxp3⁺ subset. Overall, our cumulative data demonstrate that Tcaf3-derived peptides efficiently direct T cells into the Treg cell lineage, and serve as the antigenic targets of naturally occurring prostate-associated Treg cell populations.

Because little is known about the identity of the self antigens that direct Treg cell development in the thymus, it has not been possible to determine whether the differentiation of a given Treg cell clone relies on the specific recognition of a single agonist ligand, or can be conferred by TCR cross-reactivity to an array of self ligands. By generating Tcaf3^tm1 mice harboring a targeted deletion of the Tcaf3_646–658 13-mer peptide, we demonstrated that MJ23 Treg cell development was abolished in Tcaf3^tm1/tm1 mice. Thus, the thymic development of this tissue-specific Treg cell clone requires thymic expression of a single self peptide, and cannot be mediated by reactivity to other self ligands presented in the thymus.

The data presented here, combined with evidence from previous studies (Malchow et al., 2016; Malchow et al., 2013), support a model in which the Aire-dependent presentation of Tcaf3-derived peptides directs Tcaf3-specific thymocytes into the Treg cell lineage. In the absence of Aire, Tcaf3-specific T cells, including the MJ23 and SP33 clonotypes, become skewed to the Tconv subset and infiltrate the prostate. Informally, we refer to these cells as “T-rogue cells” – Treg cell-biased clonotypes that “go rogue” in a setting of Aire deficiency (Malchow et al., 2016). Of note, Aire^−/− mice also develop autoantibodies against Tcaf3, likely aided by the provision of T cell help by T-rogue cells. Consistent with this notion, a recent study demonstrated that Aire-deficient human subjects harbor highly mutated, high-affinity autoantibodies (Meyer et al., 2016), indicative of T helper cell-dependent affinity maturation. Thus, our data reveal a link between the pMHC-II antigens recognized by Aire-dependent, tissue-specific Treg cells and the protein antigens recognized by autoantibodies in settings of Aire deficiency. Based on this conceptual link, we hypothesize that the specificities of autoantibodies that arise in human subjects with loss-of-function AIRE mutations (Landegren et al., 2016; Meyer et al., 2016) may reveal the specificities of recurrent Aire-dependent Treg cell populations in humans. One such candidate is the prostatic protein Tgm4, which is recurrently targeted by autoantibodies in AIRE-deficient men and Aire^−/− male mice(Landegren et al., 2015; Meyer et al., 2016). Broadly speaking, we suggest that this approach may serve as a general strategy to identify the constellation of tissue-specific self peptides that are targeted by recurrent Treg cell populations relevant to human autoimmune diseases and cancer.

The concept of immunodominance, in which the immune response is focused on a limited set of antigenic determinants, has been demonstrated extensively for effector T cell responses directed at foreign antigens (Sercarz et al., 1993; Yewdell, 2006). Our finding that two recurrent prostate-associated Treg cell clones recognize distinct peptides derived from a single self protein raises the possibility that the concept of immunodominance may extend to Treg cell antigen recognition at regional sites, and that Treg cell populations at a given site may be largely focused on those autoantigens that are most susceptible to autoimmune attack. Early studies of TCR repertoire complexity in the secondary lymphoid organs have demonstrated that the diversity of the bulk Treg cell repertoire in naive mice is comparable to that of the Tconv cell repertoire (Hsieh et al., 2004; Pacholczyk et al., 2006; Wong et al., 2007), suggesting that the peripheral Treg cell repertoire is not focused on a limited array of antigens. However, a subsequent survey revealed that Treg cell-expressed TCRs are asymmetrically distributed in lymph nodes throughout the body, suggestive of antigen-driven enrichment of distinct Treg cell clones in regional lymph nodes (Lathrop et al., 2008). More recently, Treg cell repertoire analysis in mouse prostate tumors (Malchow et al., 2013) and non-lymphoid organs such as muscle (Burzyn et al., 2013) and visceral adipose tissue (Feuerer et al., 2009; Kolodin et al., 2015) have revealed the enrichment of oligoclonal Treg cell populations at these sites, some of which were recurrent. Integrating this evidence with our current findings, we envision a model of “regional Treg cell immunodominance”, in which select Treg cell specificities are drawn from the diverse peripheral Treg cell pool and enriched and/or expanded in response to inflammatory reactions at different non-lymphoid sites.

In addition to Tcaf3, there are other studies in the literature describing autoimmune responses converging on a single tissue-specific autoantigen. For example, it has been shown that Aire deficiency in mice on the B6 background leads to the development of effector T cell and autoantibody responses directed at unique structures of the eye (Anderson et al., 2002; Jiang et al., 2005). Intriguingly, the development of uvea-specific autoimmunity in Aire^−/− mice is dependent on the peripheral expression of a single target antigen, retinol binding protein 3 interstitial (Rbp3, also known as interphotoreceptor retinoid binding protein [IRBP] (DeVoss et al., 2006)). Moreover, in an Aire^+/+ setting, deficiency of Rbp3 specifically in the thymus is sufficient to induce uveitis (DeVoss et al., 2006). Although it remains unclear whether T-rogue cells specific for Rbp3 are implicated in this system, the results support the idea that provocation of autoimmune responses to the eye in mice may be uniquely focused on a limited number of autoantigens such as Rbp3.

Finally, the identification of endogenous self peptides recognized by Aire-dependent tTreg populations will enable the interrogation of fundamental questions in Treg cell biology which have thus far been experimentally intractable. First, the role of cognate antigen in driving the thymic development and peripheral homeostasis of Tcaf3-specific Treg cells, and the role of TCR-pMHC affinity in coordinating these processes can be addressed in future studies. Second, the establishment of Tcaf3/I-A^b tetramers will permit the enumeration and phenotypic analysis of endogenous Tcaf3-specific T cells at different stages of ontogeny and in various settings of health and disease. Third, our findings will enable studies to determine the structural and biochemical basis of antigen recognition by the MJ23 and SP33 TCRs, to determine whether tTreg TCRs recognize self antigen with unique binding characteristics. Lastly, the Tcaf3 system may provide a model for the development of additional approaches for the treatment of autoimmune or inflammatory disorders based on the selective in vivo expansion of organ-specific Treg cells, and the recruitment of such cells to inflamed sites.

STAR METHODS

Contact for Reagent and Resource Sharing

Further information and requests for reagents should be directed to and will be fulfilled by the Lead Contact, Peter A. Savage (psavage@bsd.uchicago.edu).

Experimental Model and Subject Details

Mice

The following mice were purchased from the Jackson Laboratory, and bred and maintained at the University of Chicago: C57BL/6J (B6) mice, CD45^.1/.1 B6.SJL-Ptprc^a Pepc^b/BoyJ mice, TRAMP C57BL/6-Tg(TRAMP)8247Ng/J mice, Rag1^−/− B6.129S7-Rag1^tm1Mom/J mice, Aire^−/− B6.129S2-Aire^tm1.1Doi/J mice, TCRα^−/−B6.129S2-Tcra^tm1Mom/J mice, CD4-Cre B6.Cg-Tg(Cd4-cre)1Cwi/BfluJ mice, and OT-II transgenic C57BL/6-Tg(TcraTcrb)425Cbn/Crl mice. MJ23tg Rag1^−/− CD45^.1/.1 and “TCRβtg” mice expressing a fixed TCRβ chain of sequence TRBV26-ASSLGSSYEQY were generated as described previously (Malchow et al., 2013). All mice were generated on a pure B6 background or were fully backcrossed to the B6 background. All mice were bred and maintained under specific pathogen free conditions in accordance with the animal care and use regulations of the University of Chicago. Mice were housed in sterile and ventilated micro-isolation cages, up to five mice per cage, and fed irradiated standard pellet chow and reverse osmosis water ad libitum in a 12-hour light/dark cycle, with room temperature at 22+/− 1°C. All cages contained sterile ¼-inch corncob bedding and a nestlet for environmental enrichment. Mice for experiments were age-matched, were littermates when possible, and were assigned to experimental groups based on genotype.

Generation of Tcaf3^tm1 mice

Tcaf3^tm1 mice harboring genetic deletion of the region encoding the Tcaf3_646–658 peptide were generated on the B6 background by Cyagen Biosciences using CRISPR/Cas9-based targeting and homology-directed repair. In founder lines, the lack of mutations in off-target genes was verified by PCR amplification and sequencing of select candidate genes. Intercross of heterozygous Tcaf3^+/tm1 mice yielded healthy offspring at Mendelian ratios. The targeted region of the Tcaf3 gene is presented below, with the deleted segment shown in bold, and the genotyping primer sites underlined: CCACTTAACTTCATCCCAGACATCCATCTCCTGGCTCTATGGGGGCCTTCTCTACATCATGGTCCCCAACAAATATAATCAGGATAATGTGTCTGTCACCATCCGTGGGGCTGTATCTGCTCCATACTTCAGGCTGGGTAAGACGACCCAGGAGGAATGGAAGAATCTTATCACACACAGCAAAGCTCCGTGGGGAGAACTAGCCACAGACAATATCATCCTGACAATTCCAACGGTAAACCTCAAGGAGCTTCAGGACCCCTATCCACTGCTCCAACTCTG.

Cell lines

High Five insect cells (Trichoplusia ni, female, ovarian) were used for production of recombinant prostatic proteins (Figures 1B, 1C, 1G, 4B, 4C, S1A); cells were grown in Insect-XPRESS Protein-Free Insect Cell Medium supplemented with additional L-glutamine (2 mM) and gentamicin sulfate (50 µg/mL), in suspension culture shaking at 120 rpm and 27°C. Drosophila S2 cells were used for recombinant production of I-A^b; cells were transfected according to the Drosophila Expression System manual (Thermo Fisher) in Schneider’s Drosophila medium supplemented with 10% FBS, 1X Pen/Strep (100 U/mL Penicillin, 0.1 mg/mL Streptomycin), and 20 µg/mL Gentamicin, and maintained in stationary cultures at 27°C. S2 transfectants were selected with 25 µg/mL Blasticidin, and stable lines were expanded for expression in Express Five SFM supplemented with 25 µg/mL Blasticidin, 1X Pen/Strep (100 U/mL Penicillin, 0.1 mg/mL Streptomycin), and 20 µg/mL Gentamicin, in suspension culture shaking at 120 rpm and 27°C.

Bacteria

E. coli DH5α were cultured at 37°C shaking at 250 rpm. E. coli W3110 strain 33D3 were used for production of recombinant prostatic proteins (Figure S1B); 33D3 bacteria were cultured at 37°C shaking at 250 rpm, and then switched to 30°C for protein expression.

Method Details

Preparation of prostatic extracts

The procedure for the preparation of prostatic secretory extracts was adapted from Fujimoto et al. (Fujimoto et al., 2006). Prostates of tumor-bearing TRAMP male mice of 6 months of age, or age-matched Tcaf3^+/+ or Tcaf3^tm1/tm1 males of 6–8 weeks of age, were dissected to separate the anterior, ventral, and dorsolateral lobes. The dissected lobes were incubated separately in 1–2 mL PBS for 5 minutes at room temperature to extract secreted proteins, and then spun for 5 minutes at 10,000 × g at 4°C. The supernatant was transferred t o a fresh tube and spun again for 5 minutes at 13,200 × g at 4°C. The supernatant from this second spin wa s retained and total protein content was quantified by BCA protein assay (Pierce). These secretory extracts were flash frozen in liquid nitrogen and store at −80°C until use.

Production of recombinant prostatic proteins

Tcaf3, Tgm4 and Svs2 proteins used in Figures 1C, 4C and S1A were produced recombinantly in High Five insect cells (Thermo Fisher) with an N-terminal gp67 secretion signal and 8xHis tag. Tagged proteins were purified from culture supernatant by nickel affinity chromatography, and their His tags were removed by cleavage with 3C protease. Protease and uncleaved proteins were removed by a second pass over nickel resin. Pate4, Sbpl, Wfdc3, Msmb, and Svs4 proteins used in Figure S1B were produced recombinantly in 33D3 E. coli by periplasmic secretion as previously described (Maynard et al., 2005). Briefly, proteins were expressed with an N-terminal PELB secretion signal and C-terminal 6xHis tag. Periplasmic proteins were isolated by osmotic shock in a solution of sucrose, EDTA and lysozyme. Following neutralization with MgCl₂, recombinant proteins were purified by nickel affinity chromatography. All recombinant proteins were exchanged into PBS using a Zeba desalting spin column (Thermo Fisher) and sterilized by 0.22 µm filtration.

Tcaf3 (NCBI: NM_203396.1)

The Tcaf3 protein sequence (underlined) was fused at its N-terminus to the gp67 secretion signal sequence, an 8xHis tag, a 3C protease cleavage site, and intervening linker sequences. A slash (“/”) indicates where N-terminal tags were cleaved from the mature protein. MLLVNQSHQGFNKEHTSKMVSAIVLYVLLAAAAHSAFAADLHHHHHHHHGSGGLEVLFQ/GPERGSHMAMATTPDAAFETLMNGVTSWDLPKEPIPSELLLTGESAFPVMVNDKGQVLIAASSYGQGRLVVVSHESYLLHDGLVPFLLNVVKWLCPCPGAPIAVHSSLASLVNILGDSGINALVQPEPGEALGVYCIDAYNDALTEKLIQFLKNGGGLLIGGQALNWAAHHGHDKVLSIFPGNQVTSVAGVYFTDISANRDWFKVSKEIPNLRLYVQCEDELEDDQQQLLKGMSEIYIEAGVIPSQLLVHGQRAFPLGVDNSLNCFLAAARYGRGRVVLGGNESLILNQTMLPFVLNALHWLMGNQTGRIGLASDMKVLKSMLPNSSFQWSESELLTSDLSVFCCCSLANIDSEEVEEFVAEGGGLLIGAEAWSWGRRNPYSSCMTQYPDNIVLKRFGLGITSHVAQRGSFPFPNPEGTNYHFRRALSQFESVIYSRGSSLHESWLNKLSQDCFYMFQMTHQRISIYDSVKKHALKMIQSKDFPSVTEQYPIARGSSQAFLLSLAYELFKSGVDRSQLLPPPALLPPTESPITIKISTDNDNSWVSTGLYLPEGQVAQVLLPSEATHAKLKVLIGCHRDNISQARTYFRPPVMTYVYHLTSSQTSISWLYGGLLYIMVPNKYNQDNVSVTIRGAVSAPYFRLGKTTQEEWKNLITHSKAPWGELATDNIILTIPTVNLKELQDPYPLLQLWDKMVRAVAKLAARPFPFQRAERVVLDKQISFGFLHSGYPIMGLISIVEGIISEFKIRSHGIWGVIHELGHNHQKSGWTFPPHTTEALCNLWTIYVHETVLNIPREQAHPSLNPELRRQRIKYHLNKGAPLSNWIMWTALETYLQLQEGFGWEPFIQVFADYRTLSGLPQNNEDKMNLWVKKFSEAVHKNLAPFFEAWGWPVKYAVAKSLASLPEWQENPMKRYTAEGTEGRE

Svs2 (NCBI: NM_017390.4

The Svs2 protein sequence (underlined) was fused at its N-terminus to the gp67 secretion signal sequence, an 8xHis tag, a 3C protease cleavage site, and intervening linker sequences. A slash (“/”) indicates where N-terminal tags were cleaved from the mature protein. MLLVNQSHQGFNKEHTSKMVSAIVLYVLLAAAAHSAFAADLHHHHHHHHGSGGLEVLFQ/GPEGGSQYGATKGHFQSSSSEGFMLGQKGRLSFGIKGGSDEAAEESLFMQSQRRVYGQGGGDMTQTRVSQEHTSVKGAALCRNGQVSQLKSQESQIKSYGQVKSSGQLKSGGSAFGQVKSSVSQIKSYGQLKSGGQLKSGGPAFGQVKSQESQIKSYGQLKSSGQLKSGGSAFGQVKSSVSQIKSYGQLKSGGSQVKSYGQTKSYGEEGQLNSFSQLKSQGAQLKSYGQQKSQQQSSFSQVKSQSSQLKSYGQQKSLKGFSQQTQHKGFAMDEGMSQVRKQFSDDDLSVQQKSTQQMKTEEDLSQFGQQRQYGQERSQSYKGYLEQYRKKVQEQQRKNFNPGNYFTKGGADLYQAQLKG

Tgm4 (NCBI: NM_177911.4)

The Tgm4 protein sequence (underlined) was fused at its N-terminus to the gp67 secretion signal sequence, an 8xHis tag, a 3C protease cleavage site, and intervening linker sequences. A slash (“/”) indicates where N-terminal tags were cleaved from the mature protein. MLLVNQSHQGFNKEHTSKMVSAIVLYVLLAAAAHSAFAADLHHHHHHHHGSGGLEVLFQ/GPERGSHMAANVLIIYAVNVERKLNAAAHHTSEYQTKKLVLRRGQIFTLKVILNRPLQPQDELKVTFTSGQRDPPYMVELDPVTSYRSKGWQVKIAKQSGVEVILNVISAADAVVGRYKMRVNEYKAGVFYLLFNPWCSDDSVFMASEEERAEYILNDTGYMYMGFAKQIKEKPWTFGQFEKHILSCCFNLLFQLENNEMQNPVLVSRAICTMMCAANGGVLMGNWTGDYADGTAPYVWTSSVPILQQHYVTRMPVRYGQCWVFSGILTTALRAVGIPARSVTNFESAHDTEKNLTVDIYLDESGKTIPHLTKDSVWNFHVWTDAWMKRQDLPHGYDGWQVLDSTPQEISDGGFRTGPSPLTAIRQGLIQMKYDTTFVFTEVNGDKFIWLVKQNQEREKNILIAVETASIGKKISTKMVGENRREDITLQYKFPEGSPEERKVMAKASGKPSDDKLNSRTLNNSLQISVLQNSLELGAPIYLTITLKRKTATPQNVNISCSLNLQTYTGNKKTNLGVIQKTVQIHGQESRVFLTMDASYYIYKLGMVDDEMVIGGFIIAEIVDSGERVATDTTLCFLYSAFSVEMPSTGKVKQPLVITSKFTNTLPIPLTNIKFSVESLGLANMKSWEQETVPPGKTITFQMECTPVKAGPQKFIVKFISRQVKEVHAEKVVLISK

Pate4 (NCBI: NM_020264.4)

The Pate4 protein sequence (underlined) was fused at its N-terminus to the PELB secretion signal sequence and a short linker, and at its C-terminus to a short linker and a 6xHis tag. A slash (“/”) indicates the predicted boundary between the secretion signal and the mature protein. MKSLLPTAAAGLLLLAAQ/PAMALICNSCEKSRDSRCTMSQSRCVAKPGESCSTVSHFVGTKHVYSKQMCSPQCKEKQLNTGKKLIYIMCCEKNLCNSFAASGADHHHHHH

Sbpl (NCBI: NM_001077421.1)

The Sbpl protein sequence (underlined) was fused at its N-terminus to the PELB secretion signal sequence and a short linker, and at its C-terminus to a short linker and a 6xHis tag. A slash (“/”) indicates the predicted boundary between the secretion signal and the mature protein. MKSLLPTAAAGLLLLAAQ/PAMAQNVLGNAAGKYFYVQGEDQGQLKGMRIFLSVFKFIKGFQLQFGNNWTDVYGSRSDNFIDFLLEDGEHVIKVEGSAVICLTSLTFTTNKGRVATFGVRRGRYFSDTGGSDKHLVTVNGMHAPGLCVTGMGFKWEDNAKDLGSPEPVKEPKDSSDSSNKKEDEGRGKDDDDNDEDEDDNDEDENNYGNDDDDDDNDDQKDESAASGADHHHHHH

Wfdc3 (NCBI: NM_027961.1)

The Wfdc3 protein sequence (underlined) was fused at its N-terminus to the PELB secretion signal sequence and a short linker, and at its C-terminus to a short linker and a 6xHis tag. A slash (“/”) indicates the predicted boundary between the secretion signal and the mature protein. MKSLLPTAAAGLLLLAAQ/PAMAGEHALRGECPADPLPCQELCTGDESCPQGHKCCSTGCGHACRGDIEGGRDGQCPRILVGLCIVQCMMDENCQSGERCCKSGCGRFCIPGLQPLQQLKDSNLTDGFNSKLEAQAPAASGADHHHHHH

Msmb (NCBI: NM_020597.3)

The Msmb protein sequence (underlined) was fused at its N-terminus to the PELB secretion signal sequence and a short linker, and at its C-terminus to a short linker and a 6xHis tag. A slash (“/”) indicates the predicted boundary between the secretion signal and the mature protein. MKSLLPTAAAGLLLLAAQ/PAMAVCSIENREIFPNQMSDDCMDADGNKHFLNTPWKKNCTWCSCDKTSITCCTNATRPLSYDKDNCDVQFHPENCTYSVVDRKNPGKTCRVDSWTMAASGADHHHHHH

Svs4 (NCBI: NM_009300.3)

The Svs4 protein sequence (underlined) was fused at its N-terminus to the PELB secretion signal sequence and a short linker, and at its C-terminus to a short linker and a 6xHis tag. A slash (“/”) indicates the predicted boundary between the secretion signal and the mature protein. MKSLLPTAAAGLLLLAAQ/PAMAKKTKEKFLQSEETVRESFSMGSRGHMSRSSEPEVFVRPQDSIGDEASEEMSSSSSSRRRSKIISSSSDGSNMEGESSYSKRKKSRFSQDALEAASGADHHHHHH

In vitro T cell stimulation

CD4⁺ T cells were isolated from MJ23tg Rag1^−/− CD45^.1/.1 female or OT-IItg Rag1^−/− CD45^.1/.1 donor mice and purified by MACS (Miltenyi Biotech) enrichment. CD4⁺ T cells were CellTrace-Violet (ThermoFisher) labeled per manufacturer instructions with slight modification. Briefly, cells were pelleted, resuspended in CellTrace-Violet (CTV) at 1:1000 dilution and incubated for 20 minutes at 37°C. The reaction was quenched by the addition of 13 mL of complete culture media. To isolate splenic dendritic cells, splenocytes were isolated from C57BL/6 CD45^.2/.2 mice and enriched for CD11c⁺ cells by MACS-based (Miltenyi Biotech) positive selection. 1 × 10⁴ CTV-labeled T cells were co-cultured with 5 × 10⁴ CD11c⁺ splenocytes, 100 U/mL recombinant mouse interleukin-2 (IL-2), and prostatic extract, proteins, or peptides as indicated. Additionally, anti-MHC-II antibody clone M5 (eBioscience) or IgG2b, κ isotype control antibody (BD Pharmingen) was added to indicated cultures at a final concentration of 10 µg/mL. Cell cultures were set up in 384-well ultra low attachment, round-bottom plates (Corning). Dilution of CTV was assessed by flow cytometry on day 3 or day 5 as indicated.

In vivo T cell stimulation

CD4⁺ T cells were isolated from MJ23tg Rag1^−/− CD45^.1/.1 female or OT-IItg Rag1^−/− CD45^.1/.1 donor mice and purified by MACS (Miltenyi) enrichment. CD4⁺ T cells were CellTrace-Violet (ThermoFisher) labeled as described above. Cells were resuspended in incomplete media and retro-orbitally injected into B6 CD45^.2.2 female hosts. 2 hour later mice were immunized i.v. with 5 µg of Tcaf3 or control protein in PBS, or with PBS alone. CD4⁺ T cells from the spleen or pooled inguinal, axillary, and brachial lymph nodes were MACS-enriched, and dilution of CTV was assessed by flow cytometry on day 5.

I-A^b-binding peptide prediction

Tcaf3-derived peptides most likely to bind I-A^b were predicted by two computational methods: the Immune Epitope Database (IEDB, (Vita et al., 2015)) and an in-house peptide scoring script based on the amino acid position probabilities reported by Zhu et al. (Zhu et al., 2003). Top-scoring peptides from either method were synthesized and assayed as described below. Of note, the validated antigenic peptide (Tcaf3_646–658) was predicted by Zhu et al., but not by the IEDB.

Peptides

The top 36 candidate peptides were synthesized in Sigma’s PepScreen format, and the crude, unpurified peptides were screened for in vitro stimulation of MJ23tg T cells as described above. Truncation mutants were also synthesized in the PepScreen format and assayed as crude peptides. All other peptide experiments were performed with peptides synthesized and purified to >98% purity (GenScript).

Tcaf3 Western blot

Serum was isolated from Aire-sufficient (Aire^+/+ or Aire^+/−) or Aire-deficient (Aire^−/−) males at varying ages. 3 µg Tcaf3 protein was loaded onto a 1-well 4–20% SDS-PAGE gel (Bio-Rad; ~38 ng protein per mm lane width) and transferred to nitrocellulose membrane. The membrane was blocked for 1 h at room temperature with 3% w/v BSA in TBSt (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% v/v Tween 20), and then assembled into a Mini-PROTEAN II Multiscreen Apparatus (Bio-Rad). Sera were diluted 1:400 in TBSt + 3% w/v BSA + 0.1% w/v sodium azide, loaded into separate channels of the Multiscreen Apparatus, and incubated overnight at 4°C. Channels were washed with TBSt in the Apparatus, then the membrane was removed from the Apparatus, washed again with TBSt, and blotted for 1 hour at room temperature with bovine anti-mouse-IgG HRP conjugate (Santa Cruz Biotechnology, sc-2371) diluted 1:10,000 in TBSt + 5% w/v nonfat dried milk. The membrane was washed with TBSt, incubated with SuperSignal West Pico chemiluminescent substrate (Thermo Fisher) and imaged on a ChemiDoc imager (Bio-Rad).

I-A^b tetramer production

Tcaf3/I-A^b tetramers bearing the Tcaf3_646–658(648Y) peptide (THYKAPWGELATD) and 2W1S/I-A^b tetramers bearing the 2W1S peptide (EAWGALANWAVDSA) were produced using methods similar to those described previously (Moon et al., 2007). I-A^b was expressed in Drosophila S2 cells, using separate plasmids to encode the alpha and beta chains, as described previously (Moon et al., 2007). Constructs were co-transfected into Drosophila S2 cells together with a plasmid encoding the BirA biotin ligase. Protein expression was induced with the addition of 0.8 mM CuSO₄, in the presence of 2 µg/mL biotin (Sigma-Aldrich). Biotinylated I-A^b protein was purified from culture supernatant by nickel affinity chromatography with His Bind Ni-IDA resin (EMD Millipore) and by avidin affinity chromatography with Pierce Monomeric Avidin UltraLink Resin (Thermo Fisher). Tetramers were formed by mixing biotinylated I-A^b with streptavidin-APC (Prozyme PJ27S) or streptavidin-PE (Prozyme PJRS34) at a slight molar excess of I-A^b to biotin binding sites. Saturation of the streptavidin conjugate was verified by non-reducing SDS-PAGE without boiling samples.

I-A^b alpha chain

The extracellular domain of the I-A^b alpha chain (underlined) was fused at its N-terminus to a secretion signal sequence (boundary denoted by “/”), and at its C-terminus to an acidic leucine zipper, and a recognition sequence for the BirA biotin ligase. MPCSRALILGVLALTTMLSLCGG/EDDIEADHVGTYGISVYQSPGDIGQYTFEFDGDELFYVDLDKKETVWMLPEFGQLASFDPQGGLQNIAVVKHNLGVLTKRSNSTPATNEAPQATVFPKSPVLLGQPNTLICFVDNIFPPVINITWLRNSKSVADGVYETSFFVNRDYSFHKLSYLTFIPSDDDIYDCKVEHWGLEEPVLKHWEPEIPAPMSELTETGGGGSTTAPSAQLEKELQALEKENAQLEWELQALEKELAQGGSGGSGLNDIFEAQKIEWHE

I-A^b beta chain with Tcaf3_646–658(S1Y) peptide

The extracellular domain of the I-A^b beta chain (underlined) was fused at its N-terminus to a secretion signal sequence (boundary denoted by “/”), the Tcaf3_646–658(S1Y) peptide (in bold), and a linker sequence, and at its C-terminus to a basic leucine zipper and a 6xHis tag. MALQIPSLLLSAAVVVLMVLSSPGTEG/THYKAPWGELATDGGGGTSGGGSGGSERHFVYQFMGECYFTNGTQRIRYVTRYIYNREEYVRYDSDVGEHRAVTELGRPDAEYWNSQPEILERTRAELDTVCRHNYEGPETHTSLRRLEQPNVVISLSRTEALNHHNTLVCSVTDFYPAKIKVRWFRNGQEETVGVSSTQLIRNGDWTFQVLVMLEMTPRRGEVYTCHVEHPSLKSPITVEWRAQSESAWSKGGGGSTTAPSAQLKKKLQALKKKNAQLKWKLQALKKKLAQHHHHHH

I-A^b beta chain with 2W1S peptide

The extracellular domain of the I-A^b beta chain (underlined) was fused at its N-terminus to a secretion signal sequence (boundary denoted by “/”), the 2W1S peptide (in bold) and a linker sequence, and at its C-terminus to a basic leucine zipper and a 6xHis tag. MALQIPSLLLSAAVVVLMVLSSPGTEG/GDSEAWGALANWAVDSAGGGGSLVPRGSGGGGSERHFVYQFMGECYFTNGTQRIRYVTRYIYNREEYVRYDSDVGEHRAVTELGRPDAEYWNSQPEILERTRAELDTVCRHNYEGPETHTSLRRLEQPNVVISLSRTEALNHHNTLVCSVTDFYPAKIKVRWFRNGQEETVGVSSTQLIRNGDWTFQVLVMLEMTPRRGEVYTCHVEHPSLKSPITVEWRAQSESAWSKGGGGSTTAPSAQLKKKLQALKKKNAQLKWKLQALKKKLAQHHHHHH

I-A^b tetramer staining

Tetramer staining was adapted from Tungatt et al. (Tungatt et al., 2015). Cells were treated with dasatinib (AdooQ Bioscience) at a final concentration of 50 nM for 30 minutes at 37°C. PE- and APC-labeled tetramers were added directly to dasatinib-treated cells (without washing) at a final concentration of 12.5–100 nM for 1 hour at room temperature. Cells were washed and incubated with unconjugated mouse anti-PE antibody (clone PE001, Biolegend) and mouse anti-APC antibody (clone APC003, Biolegend) at a concentration of 10 µg/mL for 20 minutes on ice. Cells were washed and stained for flow cytometric analysis as described below. Staining of cells from Aire^+/+ and Aire^−/− prostates was performed with 12.5 nM of each tetramer. Staining of cells from TRAMP prostate tumors was performed with 20 nM of each tetramer.

I-A^b tetramer-based enrichment

Tetramer enrichment was adapted from Legoux et al. (Legoux and Moon, 2012). Cells were isolated from the spleen, inguinal, axillary, brachial, cervical, mesenteric, and periaortic lymph nodes. Cells were co-stained with APC- and PE-labeled tetramer as described above at a final tetramer concentration of 20–200 nM depending the experiment. Following staining, cells were incubated with anti-PE microbeads (Miltenyi Biotech) and mouse anti-APC antibody (clone APC003, Biolegend) and incubated for 20 min on ice. Cells were enriched over a magnetic column and bound cells were eluted by pushing 5 mL of buffer though the column with a plunger. The resulting bound fraction was stained and analyzed as described above. The total number of tetramer-positive events was calculated as described by Legoux et al. (Legoux and Moon, 2012).

Antibodies and flow cytometry

All antibodies used were purchased from Biolegend, eBioscience, or BD Biosciences. Cells were stained with conjugated antibodies specific for the following proteins (with clone name in parentheses): CD4 (GK1.5), CD8α (53–6.7) or CD8β (Ly-3), CD3 (17A2), CD45.1 (A20), CD45.2 (104), CD69 (H1.2F3), Foxp3 (FJK-16s), B220 (RA3–6B2), CD11b (M1/70), CD11c (N418), and F4/80 (BM8). Cells were stained for 20 minutes on ice in staining buffer (phosphate-buffered-saline with 2% FCS, 0.1% NaN₃, 5% normal rat serum, 5% normal mouse serum, 5% normal rabbit serum, with all sera from Jackson Immunoresearch), and 10 µg/mL 2.4G2 antibody). Intracellular staining for Foxp3 was preformed using fixation and permeabilization buffers from eBioscience. Flow cytometry was performed on an LSR Fortessa (BD Biosciences) and data was analyzed using FlowJo software (Tree Star).

CFA immunization

Mice were give subcutaneous injection of 100 µg of peptide in 100 µL of CFA emulsion (Sigma-Aldrich) or 100 µL of CFA emulsion alone. Mice were analyzed 14 days later.

Intrathymic injection of thymocytes

4 × 10⁶ bulk thymocytes from MJ23tg Rag1^−/− CD45^.1/.1 females were injected intrathymically into 4–6-week-old Tcaf3-decient (Tcaf3^tm1/tm1) or Tcaf3-sufficient (Tcaf3^tm1/+, Tcaf3^+/+) mice, and analyzed 7 days post-transfer. Donor thymocytes were Foxp3^neg due to intracloncal competition in MJ23tg Rag1^−/− CD45^.1/.1 females.

Retrovirus production, infection, and generation of SP33rg mice

SP33rg mice were generated as described previously (McDonald et al., 2014). Briefly, the SP33 TCRα was cloned into a retroviral construct modified from Turner et al. (McDonald et al., 2014; Turner et al., 2010). Plat-E cells, also previously described (Morita et al., 2000), were used to generate retrovirus. TCRα^−/− CD4-Cre TCRβtg mice on a C57BL/6 background were injected with 5-fluorouracil (APP Pharmaceuticals) 3 days prior to bone marrow harvest. Bone marrow cells were cultured for 2 days in X-Vivo 10 (Lonza) containing 15% FCS, 1% penicillin/streptomycin, mouse SCF, mouse IL-3 and mouse IL-6 (Biolegend). Cells were infected with retrovirus by spinfection in the presence of 6 µg/mL polybrene (EMD Millipore) and cultured for an additional 24 hours. 1 × 10⁶ spinfected cells were then mixed with 5 × 10⁶ freshly harvested bone marrow “filler” cells from Rag1^−/− mice and injected into irradiated (800 rad) CD45^.1/.1 B6.SJL recipient mice to generate TCR “retrogenic” (rg) mice. SP33rg cells were isolated from retrogenic mice 6–8 weeks after bone marrow reconstitution. CD4⁺ T cells were FACS-purified from SP33rg mice following staining with the following antibodies: anti-CD8β (Ly-3), anti-CD45.1 (A20), anti-CD45.2 (104), and anti-Thy1.1 (OX-7). Isolated cells were CellTrace-Violet labeled and used for in vitro and in vivo experiments as described above.

TCR Gene Usage and CDR3 Sequences

The MJ23 TCR contains the alpha chain TRAV14-LYYNQGKLI, utilizing TRAJ23, and the beta chain TRBV26-ASSLGSSYEQY, utilizing TRBJ2-7. The SP33 TCR contains the alpha chain TRAV9D–ALSMSVNYQLI, utilizing TRAJ33, and the beta chain TRBV26-ASSLGSSYEQY, utilizing TRBJ2–7.

Quantification and Statistical Analysis

Data were analyzed using Prism software (GraphPad). Significance testing was performed using the nonparametric Mann-Whitney test or the Student’s t-test (two-tailed). No statistical methods were used to predetermine sample size.

Highlights.

Two Treg cell clones recognize distinct peptides from the prostatic protein Tcaf3

Tcaf3 is recurrently targeted by autoantibodies in settings of immune dysregulation

Thymic development of the MJ23 Treg cell clone is dependent on a single Tcaf3 peptide

Peptide-MHC tetramers identify endogenous Tcaf3-specific Treg cells in prostate tumors