Aire-dependent thymic development of tumor-associated regulatory T cells

Abstract

Despite considerable interest in the modulation of tumor-associated Foxp3(+) regulatory T cells (T(regs)) for therapeutic benefit, little is known about the developmental origins of these cells and the nature of the antigens that they recognize. We identified an endogenous population of antigen-specific T(regs) (termed MJ23 T(regs)) found recurrently enriched in the tumors of mice with oncogene-driven prostate cancer. MJ23 T(regs) were not reactive to a tumor-specific antigen but instead recognized a prostate-associated antigen that was present in tumor-free mice. MJ23 T(regs) underwent autoimmune regulator (Aire)-dependent thymic development in both male and female mice. Thus, Aire-mediated expression of peripheral tissue antigens drives the thymic development of a subset of organ-specific T(regs), which are likely coopted by tumors developing within the associated organ.

Fig. 1. CD4⁺Foxp3⁺ Tregs expressing a canonical TCR are recurrently enriched in TRAMP prostate tumors

Analysis of Vα2 (TRAV14)⁺ TCRα chains in tumor-bearing TRAMP^+/− Foxp3^gfp TCRβtg mice. CD4⁺Foxp3^neg and CD4⁺Foxp3⁺ T cells were FACS-purified from different anatomical sites of ~27-week-old male mice, and cDNA was subjected to molecular analysis. (A) CDR3 length distribution analysis of Vα2⁺ TCRα chains of T cell subsets isolated from prostate tumors. The mouse number is indicated. The red boxes denote TCRα transcripts encoding a CDR3 of 9 amino acids (AA) in length. (B) Vα2⁺ TCRα transcripts from the indicated populations were PCR amplified and subjected to deep sequencing (see Methods (24)). The percentage of all Vα2⁺ TCRα transcripts encoding the canonical Vα2-LYYNQGKLI chain is plotted for each sample. The mean ± SEM is indicated. The asterisk indicates p < 0.05 (t-test). (C) Plots of the percentage of Vα2⁺ transcripts encoding a CDR3 of 9 AAs (based on peak area) for CD4⁺Foxp3⁺ T cells isolated from different anatomical sites of the mice depicted in panel (A). PR, prostate; pLN, periaortic lymph nodes; SP, spleen; bLN, brachial lymph nodes; TH, thymus. Samples from each mouse are color-coded. The plot does not depict data from mice 1 and 2, and includes data from two additional mice (numbers 17 and 18). Samples with values above the indicated threshold (dashed line) are considered “overrepresented”. This threshold is defined as the mean percent peak area plus three standard deviations for Vα2⁺ TCRα transcripts encoding a 9 AA CDR3 from the spleen of female TCRβtg mice. (D) Heat map of the Morisita-Horn (MH) similarity index for prostatic T cell subsets from mice 1–16. Samples are oriented in ascending numerical order, from top to bottom and left to right (not shown). A motif table of predicted amino acid sequences is presented in Table S1. CD4⁺Foxp3^neg and CD4⁺Foxp3⁺ subsets from the same prostate tumor intersect at the red diagonal lines. Data are pooled from N = 3 independent FACS-sorting experiments.

Fig. 2. MJ23 T cells recognize a prostate-associated self antigen

(A–C) Spontaneous T cell autoreactivity in the prostates of tumor-free male MJ23tg Rag1^−/− mice. (A) Representative flow cytometric analyses of T cells isolated from the indicated organs of 6-week-old male or female mice. SP, spleen; pLN, periaortic lymph nodes; bLN, brachial lymph nodes; mLN, mesenteric lymph nodes; PR, prostate; Panc, pancreas. (B) Summary plot of the absolute number of CD4⁺Foxp3⁺ T cells from the indicated organs of 6-week-old male or female mice. (C) Hematoxylin and eosin staining of dorsolateral prostatic lobes from 18-week-old MJ23tg Rag1^−/− and B6 mice. Scale bar = 100 μm. (D) CFSE-labeled CD4⁺Foxp3^neg MJ23tg T cells or OT-IItg cells were cultured with FACS-purified CD45⁺CD11c⁺F4/80^neg cells isolated from TRAMP prostate tumors in the presence of recombinant mouse IL-2. In addition, MHC-II antibody or isotype control antibody was added to the culture. Dilution of CFSE was assessed by flow cytometry on day 4. The mean ± SEM is indicated. Asterisks indicate p < 0.05, ns = not significant. For (A–B), data are pooled from N = 2 independent experiments. For (B), t-tests were used to compare data from male and female mice at each site. Data in (D) are representative of N = 7 independent experiments.

Fig. 3. Thymic development of MJ23 Tregs

The MJ23 TCR facilitates thymic Treg development at low clonal frequencies. T cell-depleted bone marrow cells from MJ23tg Rag1^−/− CD45.1⁺ female donor mice were engrafted, along with polyclonal “filler” cells from B6 females (CD45^.2/.2), into sublethally irradiated CD45^.2/.2 recipient mice. This approach resulted in seeding of MJ23tg precursors at a low frequency (<1%). 6 weeks post-engraftment, the fate of MJ23tg cells was analyzed. (A) Representative flow cytometric analyses of CD45.1⁺ MJ23tg T cells from different anatomical sites of male or female mice of the indicated ages and strain. Abbreviations are the same as in Fig. 2. For the thymus, the left column (CD4 vs. CD8) depicts undepleted samples, the right column (Foxp3 vs. CD4) depicts CD8-depleted samples. For the periphery, the percentage of all CD4⁺Foxp3⁺ T cells that are CD45.1⁺ (MJ23tg⁺) is indicated. (B) Summary plots of the “efficiency” of MJ23tg Treg development in the thymus, in which the percentage of CD4⁺CD8^neg CD45.1⁺ MJ23tg cells that express Foxp3 is plotted vs. the frequency of CD45.1⁺ MJ23tg thymocytes (as a percentage of all CD4⁺CD8^neg cells) for cells isolated from host mice of the indicated strain, gender, and age. Dashed lines indicate best-fit semi-log curves. (C) Summary plots of the percentage of CD45.1⁺ MJ23tg T cells amongst all CD4⁺Foxp3⁺ cells isolated from various organs of the indicated hosts. The mean ± SEM is shown. Asterisks indicate p < 0.05. ANOVA was used to compare the secondary lymphoid sites (spleen and lymph nodes) within chimeric hosts of a given type. Data are pooled from at least N = 3 independent experiments.

Fig. 4. Aire-dependent thymic development of antigen-specific Tregs

(A–C). The thymic development of MJ23tg Tregs is Aire-dependent. T cell-depleted bone marrow cells from MJ23tg Rag1^−/− CD45.1⁺ female donor mice were engrafted, along with “filler” cells from Aire^−/− females (CD45^.2/.2), into sublethally irradiated Aire^−/− or Aire^+/+ recipient mice (CD45^.2/.2), both male and female. 6 weeks post-engraftment, the fate of MJ23tg cells was analyzed. (A) Representative flow cytometric analyses of CD45.2⁺ polyclonal T cells (left) and CD45.1⁺ MJ23tg T cells from different anatomical sites of 12–16-week-old male or female mice of the indicated Aire genotype. Abbreviations are the same as in Fig. 2. The percentage of CD4⁺ cells that are Foxp3⁺ is shown. For the thymus, the left column (CD4 vs. CD8) depicts undepleted samples, the right column (Foxp3 vs. CD4) depicts CD8-depleted samples. (B) Summary plots of the “efficiency” of MJ23tg Treg development, in which the percentage of CD45.1⁺ MJ23tg cells that express Foxp3 is plotted vs. the frequency of CD45.1⁺ MJ23tg thymocytes (as a percentage of all CD4⁺CD8^neg cells) for cells isolated from host mice of the indicated sex and genotype. Dashed lines indicate best-fit semi-log curves. (C) Summary plots of the percentage of CD45.2⁺ polyclonal CD4⁺CD8^neg thymocytes that express Foxp3 in chimeric mice of the indicated sex and genotype. The mean ± SEM is shown. Asterisks indicate p < 0.05 for the comparison of Aire^−/− vs. Aire^+/+ mice (t-test). Data in (A–C) are pooled from N = 3 independent experiments. (D) Aire-dependent thymic development of “RT83tg” Tregs. RT83tg bone marrow chimeras were generated in Aire^−/− or Aire^+/+ hosts, as described above for MJ23tg T cells. Representative flow cytometric analyses of CD45.1⁺ RT83tg T cells from the thymus of 12–16-week-old male or female mice of the indicated Aire genotype. The percentage of CD4⁺ cells that are Foxp3⁺ is shown. For the thymus, the left column (CD4 vs. CD8) depicts undepleted samples, the right column (Foxp3 vs. CD4) depicts CD8-depleted samples. Data in (D) are pooled from N = 2 independent experiments.