A narrow T cell receptor repertoire instructs thymic differentiation of MHC class Ib-restricted CD8+ regulatory T cells

Abstract

Although most CD8+ T cells are equipped to kill infected or transformed cells, a subset may regulate immune responses and preserve self-tolerance. Here, we describe a CD8 lineage that is instructed to differentiate into CD8 T regulatory cells (Tregs) by a surprisingly restricted set of T cell receptors (TCRs) that recognize MHC-E (mouse Qa-1) and several dominant self-peptides. Recognition and elimination of pathogenic target cells that express these Qa-1-self-peptide complexes selectively inhibits pathogenic antibody responses without generalized immune suppression. Immunization with synthetic agonist peptides that mobilize CD8 Tregs in vivo efficiently inhibit antigraft antibody responses and markedly prolong heart and kidney organ graft survival. Definition of TCR-dependent differentiation and target recognition by this lineage of CD8 Tregs may open the way to new therapeutic approaches to inhibit pathogenic antibody responses.

Keywords: Immunology; T cell receptor.

Figure 1. Identification of Qa-1–FL9–specific TCR.

(A) WT B6 mice were immunized with Kb^–/–Db^–/– DC loaded with FL9 peptide on days 0, 8, and 15. At day 22, Qa-1–FL9–specific CD8 T cells were detected by tets (Qa-1–FL9–PE and Qa-1–FL9–APC) within CD44⁺CD122⁺Ly49⁺ CD8 T cells. (B) TCR repertoire of Qa-1–FL9 Tet⁺ CD8 T cells. Single Qa-1–FL9–PE⁺ Qa-1–FL9–APC⁺ cells were sorted and subjected to sequencing for TCRα and TCRβ. Thirty-nine of TCRα and TCRβ pairs were analyzed based on their TCR V gene segments. Relative usage of TCRα and TCRβ V genes by these Tet⁺ single cells is depicted by donut charts. (C) TCR repertoire of Qa-1–Hsp60 Tet⁺ CD8 T cells. Single Qa-1–Hsp60–PE⁺ Qa-1–Hsp60–APC⁺ cells were sorted and sequenced for TCRα and TCRβ. Relative usage of TCRα and TCRβ V genes by these Tet⁺ single cells is depicted by donut chart. (D) Frequency and phenotype of Vα3.2⁺Vβ5⁺ cells within Ly49⁺ CD8 cells in the spleens and LNs of WT B6, Qa-1.D227K–KI and Qa-1–KO mice at 8 weeks of age (n = 6/group). (E) Qa-1–dependent differentiation of FL9 T cells: tet-mediated detection of TCR in 58C hybridoma transduced with FL9.2 and FL9.8 TCR (upper panel). Responsiveness of FL9.2-TCR– and FL9.8-TCR–expressing hybridoma upon stimulation with increasing dose of peptides measured by CD69 expression (lower panel). (F) Measurement of Qa-1–FL9 binding affinity of FL9.2 and FL9.8 TCR. FL9.2 TCR⁺ and FL9.8 TCR⁺ hybridoma were labelled with Qa-1–FL9–PE tets and incubated in the presence of anti-Qa-1 antibodies for the indicated time. Percentages of PE⁺ cells were measured at different time points as a measurement of tet dissociation level. (G) Tg TCR⁺ cells in TCR⁺ thymocytes and percent of active-Caspase–3⁺PD1⁺ cells in DP (CD4⁺CD8⁺) thymocytes in OT-I → WT B6, FL9.2 Tg → WT B6 BM chimera 8 weeks after BM reconstitution. (H) Ki67 and CD44 expression by OT-I and FL9.2 TCR Tg CD8⁺ T cells was measured as an indication of Ag encounter in the spleen and liver of OT-I → WT B6 and FL9.2 Tg → WT B6 BM chimera 8 weeks after BM reconstitution. OT-1 is also used as a control in Supplemental Figure 6B. *P < 0.05, **P < 0.01, ***P < 0.001, according to Wilcoxon-Mann-Whitney rank sum test.

Figure 2. Qa-1–dependent differentiation of FL9 T cells.

(A) Frequency of Tg TCR⁺ cells in the TCR⁺ thymic cells in OT-I– or FL9.2 TCR Tg mice (Va2⁺Vβ5⁺ for OT-I and Vα3.2⁺Vβ5⁺ for FL9 T cells) and levels of Helios expression. (B) Frequency of Tg TCR⁺ cells in the TCR⁺ splenic cells in OT-I or FL9 TCR Tg mice and levels of Helios and Ly49 expression. (C) Frequency of FL9 TCR Tg T cells in the TCR⁺ thymic cells in WT or Qa-1^–/–.FL9 TCR Tg mice. Frequency of FL9 Tg T cells (Vα3.2⁺Vβ5⁺) in the total thymocytes is shown in graph (right). (D) Frequency of Vα3.2⁺Vβ5⁺ T cells in TCRβ⁺ spleen cells from FL9.2 TCR Tg mice on Qa-1–WT and –KO backgrounds (at 8 weeks old). Representative FACS plots for the detection of Vα3.2⁺Vβ5⁺ cells in spleen are shown (left panel). (E) Expression of CD44 and NKG2D by FL9.2 T cells in spleens of WT.FL9.2 TCR Tg and Qa-1^–/–.FL9.2 TCR Tg mice. Representative FACS plots for NKG2D⁺CD44⁺ cells in the spleens of FL9.2 TCR Tg mice are shown on the left. (F) Expression of Ki67 by FL9.2 T cells in the LNs of WT.FL9.2 TCR Tg and Qa-1^–/–.FL9.2 TCR Tg mice. (G) CFSE-labeled FL9.2 T cells developed in Qa-1–WT or Qa-1–KO mice were transferred into irradiated (800 rads) Qa-1–WT, Qa-1–KO and D227K-KI adoptive hosts. Seven days after transfer, Qa-1–WT or Qa-1–KO FL9.2 T cells were recovered from spleens of adoptive hosts. Numbers of FL9.2 T cells in the spleens of adoptive hosts are shown. (H) Percentage of Qa-1 WT FL9.2 T cells that undergo more than 3 divisions in Qa-1–WT, Qa-1–KO and D227K-KI hosts. Mean ± SEM is indicated. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05.

Figure 3. FL9 Tg CD8 T cells recognize and suppress activated CD4 T cells.

(A) in vitro, conA-stimulated CD4 cells from WT B6, Qa-1.D227K–KI, KbDb-KO and ERAAP-KO mice were cocultured with FL9.2 T cells isolated from FL9.2 TCR Tg mice. WT CD4 cells loaded with the FL9 peptide were used as a positive control. After 20 hours, CD69 expression on FL9 Tg T cells were measured as a readout of TCR stimulation. (B) In vivo, WT or D227K mice were immunized with OT-II peptides in CFA. After 7 days, CD4 (CD4⁺CD25^–) cells were isolated from immunized mice and transferred into WT B6 hosts with or without FL9 TCR Tg T cells followed by immunization with OT-II/CFA. Detection of I-A^b/Ova_323-339-specific CD4 T cells in the spleen of hosts by I-A^b/Ova_323-339 tets (upper). Percent and numbers of I-A^b/Ova_323-339 tet⁺ (upper) and I-A^b/Ova_323-339 tet^– activated (lower) CD4 cells recovered from adoptive hosts (middle and right). (C) Qa-1 expression by I-A^b/Ova_323-339 tet⁺ and tet^– CD4 cells. (D) WT B6 and D227K mice were immunized with Ova/CFA and injected with isotype or anti-Vα3.2 antibodies on day 0, before boosting on day 8 with Ova/IFA along with antibody injection. Frequency of I-Ab/Ova_323-339 tet⁺ CD4 cells in blood were assessed on day 15. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05.

Figure 4. Identification of superagonists for FL9 T cells.

(A) A library composed of 96 FL9 peptide variants (crude peptides) was generated by amino acid mutagenesis at the Qa-1 anchoring positions (p2, p3, p6, p7, and p9). FL9 TCR⁺ hybridomas were incubated with EL4 cells (Qa-1⁺) loaded with each FL9 peptide variant for 12 hours and CD69 expression and TCR downregulation were measured as an indication of TCR stimulation. (B) Activation of FL9 TCR⁺ 58C hybridoma after stimulation with FL9 peptide variants. CD69 expression by FL9 TCR⁺ 58C hybridoma after stimulation with each FL9 peptide variant (left). Downregulation of TCR is shown as ΔTCR MFI based on calculation 100–(Testing TCR MFI/Control TCR MFI) × 100 (%) (middle). Expression of Vα3.2 and Vβ5 on EL4 cells (trogocytosis) was measured (right). (C) Activation of FL9.2 T cells after stimulation with FL9 variants selected from library screen above. Dose-dependent activation of FL9 T cells was measured by culturing FL9.2 T cells with EL4 (Qa-1⁺) at various concentrations of indicated peptides (0, 1, 3, and 10 μg/mL). (D) CD45.1⁺ B6 hosts were adoptively transferred with FL9.2 T cells and immunized i.p. with PBS, FL9, or FL9-68 in CFA on day 0. After 6 days, proliferation of FL9.2 T cells (CD45.2⁺Vα3.2⁺Vβ5⁺) was measured by CFSE dilution (left). (E) CD45.1⁺ B6 mice that were vaccinated with FL9-68 in CFA or CFA alone on day 0 were immunized with Ova_323-339 peptide in CFA on day 6. The frequency of I-A^b/Ova_323-339 Tet⁺ CD4 cells in activated (CD44⁺) CD4 cells was analyzed on day 14. (F) Comparison of high-affinity antibody and auto-antibody responses in WT and Qa-1.D227K mice. WT B6 and Qa-1.D227K.KI mice were immunized with NP₁₉-KLH/CFA and boosted with NP₁₉-KLH/IFA on day 10. High affinity anti-NP responses were measured on day 15. Levels of anti-dsDNA antibody were measured on day 21. Mean ± SEM is indicated. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05.

Figure 5. Superagonist peptide vaccination inhibits AMR in heart transplantation.

(top panel) Schematic of experimental design. B6 mice were vaccinated with FL9-68 peptide in IFA or PBS in IFA on the indicated days, followed by sensitization with BALB/c skin on day 10. Mice were further vaccinated with FL9-68 in IFA or PBS/IFA on days 10, 13, and 16. On day 27, BALB/c hearts were heterotopically transplanted to the abdominal cavity of B6 recipients. 250 μg of CTLA-4 Ig was administered i.v. after transplantation and recipients were analyzed on day 34. (A) Frequency of Qa-1-FL9-Tet⁺ CD8 (CD44⁺CD122⁺Ly49⁺) T cells in dLN of IFA- or FL9-68/IFA-vaccinated B6 hosts is shown. (B) Qa-1 expression by total, naive CD4 and Tfh cells in the draining lymph node (dLN) of skin-sensitized B6 hosts after BALB/c heart graft. (C) Numbers of Tfh, GC B and plasma cells in dLNs of naive B6 mice or PBS/IFA- or FL9-68/IFA-vaccinated B6 recipients. (D) Donor-specific antibodies (IgG1) in naive mice or PBS/IFA-, OT-I/IFA-, FL9/IFA- or FL9-68/IFA–vaccinated B6 recipients of skin grafts were measured in the serum collected on day 26, the day before heart transplantation. BALB/c donor splenocytes were incubated with serially diluted serum followed by detection with fluorescence-labeled anti-mouse IgG1 antibodies. Statistical analysis was performed with 2-way Anova involving mixed-effect analysis. (E) C4d deposition (blue) in heart allografts (upper panel). Tissue sections from heart grafts of B6 mice that were vaccinated with PBS/IFA alone or FL9-68/IFA were stained with anti-C4d antibodies. H&E stain of heart allograft showing graft infiltrating lymphocytes (lower panel). (F) Heart graft survival in mice vaccinated with FL9-68/IFA, FL9-68/IFA with α-Ly49F Ab, an irrelevant peptide, Ova/IFA, and PBS/IFA alone as control. Median Survival (days): FL9-68/IFA, 25; FL9-68/IFA with α-Ly49F Ab, 10; Ova/IFA, 12; and IFA alone, 11. Mean ± SEM is indicated. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05.

Figure 6. FL9-68 ameliorates AMR and prolongs graft survival in kidneys.

(A) Schematic of experimental design. (B) Frequency of FL9-Qa–1–specific CD8 (CD44⁺CD122⁺Ly49⁺) T cells in mice with or without FL9-68 immunization (day 20 after transplant). (C) Frequency of Tfh cells (PD-1⁺CXCR5⁺CD4⁺), activated GC B cells (FAS⁺GL-7⁺B220⁺), and plasma cells (B220^–CD138⁺) in the graft recipients with or without FL9-68 immunization. (D) Donor-specific antibodies (IgG1) in PBS/ Adj- or FL9-68/ Adj-vaccinated B6 recipients of kidney grafts. (E) Proliferation of activated CD4⁺ T cells from PBS/Adj- or FL9-68/Adj-immunized recipients, when cocultured with irradiated donor (BALB/c) splenocytes. Mixed lymphocyte reaction, comparison of the proliferative response of memory (CD44⁺) CD4 T cells isolated from the indicated group of kidney allograft recipients at day 20 upon exvivo stimulation with irradiated donor (BALB/c) splenocytes, as judged by CellTrace Violet on day 3 of coculture. (F) Gross anatomy of kidney allograft at day 20. (G) IHC for C4d deposition (blue). (H) Survival of kidney allograft measured by survival of recipients with absence of native kidney function. Median Survival (days): FL9-68/Adj, 40; PBS/Adj, 20.5. Mean + SEM is indicated. ***P < 0.001, **P < 0.01, *P < 0.05.