CD8 T cell tolerance results from eviction of immature autoreactive cells from the thymus

Abstract

CD8 T cell tolerance is thought to result from clonal deletion of autoreactive thymocytes before they differentiate into mature CD8 T cells in the thymus. However, we report that, in mice, CD8 T cell tolerance instead results from premature thymic eviction of immature autoreactive CD8 thymocytes into the periphery, where they differentiate into self-tolerant mature CD8 T cells. Premature thymic eviction is triggered by T cell receptor (TCR)-driven down-regulation of the transcriptional repressor Gfi1, which induces expression of sphingosine-1-phosphate receptor-1 (S1P1) on negatively selected immature CD8 thymocytes. Thus, premature thymic eviction is the basis for CD8 T cell tolerance and is the mechanism responsible for the appearance in the periphery of mature CD8 T cells bearing autoreactive TCRs that are absent from the thymus.

Fig. 1.. Characterization of developing thymocytes and T cells that preferentially survive strong TCR signaling during negative selection in the thymus.

(A) Thymocyte profiles from mice containing either MHC-II–selected or MHC-I–selected T cells. Shown are TCR profiles of total thymocytes (left), PD-1 profiles on TCR^hi (CD1d–PBS57⁻ non-iNKT) thymocytes (middle), and cleaved caspase-3 expression in PD-1⁺TCR^hi (non-iNKT) thymocytes (right). (B) Percent PD-1⁺ cells among TCRβ^hi thymocytes (left) and quantification of their surface PD-1 expression as mean fluorescence intensity (MFI) (right). (C) TCR profiles of male versus female thymocytes from Marilyn and HY^cd4 transgenic mice. (D) TCR^hi thymocyte numbers in Marilyn and HY^cd4 transgenic mice. (E) CD4 versus CD8α expression on TCR^hi thymocytes (top) and TCR⁺ lymph node (LN) cells (bottom) in male and female HY^cd4 mice. (F) Thymocyte and LN T cell numbers in HY^cd4 transgenic mice. (G) Expression of cleaved caspase-3 versus Bcl-2 (left) or CCR7 (right) in HY-specific thymocytes from HY^cd4 male mice. (H) TCR^hi thymocyte numbers in WT and B7-deficient [B7 double-knockout (B7^DKO)] HY^cd4 male mice. (I) CD4 versus CD8α expression on TCR^hi thymocytes on WT and B7-deficient HY^cd4 male mice. [(A) to (C), (E), (G), and (I)] Representative data are from three independent experiments. [(D), (F), and (H)] Data are from two to four independent experiments. P values were determined using unpaired two-tailed Student’s t tests (means ± SEMs).

Fig. 2.. Clonal eviction of autoreactive CD8 thymocytes signaled to undergo negative selection.

(A) CD4 versus CD8α expression on TCR^hi thymocytes from HY^cd4 male and female mice (left) and expression of various proteins in the indicated thymocyte subsets (single-color histograms). (B) Expression of Runx3d mRNA (top) and RUNX3 protein (bottom) in thymocyte subsets from HY^cd4 male and female mice. (C) To identify cells exiting the thymus, thymocytes were harvested 5 min after in vivo i.v. injection of phycoerythrin (PE)–conjugated anti-CD5 mAb. Thymocytes that bound the injected anti-CD5 mAb were referred to as CD5(i.v.)⁺ cells. (D) CD4 versus CD8 profiles of HY-specific T3.70⁺ peripheral T cells in HY^cd4 male and female mice. Only negatively selecting male mice contained peripheral CD4/8^dull HY-specific T cells. (E) PD-1 and CD24 expression on peripheral CD4/8^dull and CD4⁻8⁺ HY-specific T cells in HY^cd4 male mice. (F) CD4/8^dull and CD8⁺ HY-specific T3.70⁺ T cells were sorted from the spleen and LN of HY^cd4 male mice and were adoptively transferred into separate Rag2^−/− hosts. CD4 versus CD8α expression on each transferred cell population was overlayed before and 3 days after transfer into host mice. (G) CD4/8^dull and CD8⁺ HY-specific T3.70⁺T cells were sorted from the spleen and LN of HY^cd4 male mice and were placed into separate in vitro cultures containing IL-7 for 4 days. Overlayed CD4 versus CD8α expression on sorted T cells before and after IL-7 culture is shown. (H) CD4 versus CD8α expression on HY-specific T3.70^hi thymocytes and splenocytes in male and female embryonic day 18.5 (E18.5) HY^cd4 mice. (I) Comparison of HY-specific T3.70⁺ T cells in HY^cd4 male versus female neonatal mice at indicated ages. P0, postnatal day 0. (J) Frequencies of CD4/8^dull and CD8⁺ cells among T3.70⁺TCRβ⁺ splenocytes in HY^cd4M neonatal mice. [(A) and (C) to (G)] Data are representative of three or four independent experiments. (B) Data are from three independent experiments with technical triplicates. (H) Representative data are from three or four HY^cd4 embryos of each gender. [(I) and (J)] n = 3 to 8 from each gender at each time point. Data are expressed as means ± SEMs.

Fig. 3.. Strong TCR signaling of immature thymocytes induces S1P1 expression and clonal eviction by down-regulating Gfi1.

(A) S1pr1 mRNA (left) and S1P1 protein (right) expression in HY^cd4 male and female thymocytes. (B) CD4 versus CD8α profiles of HY-specific T3.70⁺ splenocytes in HY^cd4 male mice after five daily injections of either FTY720 or vehicle. (C) CD4 versus CD8α and Qa-2 profiles of HY-specific T3.70⁺ TCR^hi thymocytes from S1P1-deficient [S1P1 conditional knockout (S1P1^cKO)] and WT HY^cd4 male mice. (D) Numbers of mature (CD24^lo) CD8SP thymocytes in S1P1^cKO and WT HY^cd4 male mice. (E) Ifng mRNA in thymocyte subsets from S1P1^cKO and WT HY^cd4 male and female mice. (F) Frequency of CD317⁺ preselection (CD69⁻DP) thymocytes in S1P1^cKO and WT HY^cd4 male mice. (G) Number of total thymocytes in S1P1^cKO and WT HY^cd4 male mice. (H) Nr4a1, Foxo1, and S1pr1 mRNA expression in immature (CD69⁻DP) and mature (CD4SP) B6 thymocytes after overnight stimulation with immobilized plate-bound anti-TCRβ. (I) Gfi1 mRNA expression in HY^cd4 male and female thymocytes. (J) Gfi1 mRNA expression in immature (CD69⁻DP) B6 thymocytes after overnight stimulation with immobilized plate-bound anti-TCRβ. (K) Donor HY^cd4M bone marrow cells (containing CD4-cre) were infected in vitro with FLEX/Gfi1 or FLEX/Thy1.1 lentiviruses, and the lentivirus-infected bone marrow cells were then injected into lethally irradiated B6 CD45.1 host mice. Donor-origin HY^cd4M thymocytes were analyzed 4 to 6 weeks later, and those constitutively expressing Gfi1 or Thy1.1 lentiviral proteins were identified by surface expression of hCD2 reporter protein. [(A), (E), and (H) to (K)] Representative experiment from two to four independent experiments with technical triplicates. [(B) to (D)] Representative experiment from three or four independent experiments. [(F) and (G)] Data are from two to three independent experiments. P values were determined using unpaired two-tailed Student’s t tests. Data are expressed as means ± SEMs.

Fig. 4.. Peripherally generated CD8 T cells are self-tolerant.

(A) In vitro proliferative responses of CD8 T cells from HY^cd4 male and female mice to 3 days of stimulation with varying doses of HY peptide presented by B6 splenocytes, as measured by the dilution of cell trace violet (CTV). (B) In vivo response as measured by CD107a expression of CD8 T cells from draining and nondraining popliteal LNs in HY^cd4 male and female mice to footpad injection with HY peptide and LPS 16 hours earlier. (C) In vivo growth of female (MC38) or male (B16-F10) tumor cells subcutaneously injected into HY^cd4 mice. (D) Surface TCR and CD8 expression and quantification of HY tetramer binding to CD8 LN T cells from HY^cd4 male and female mice. (E) Ly49 versus CD122 profiles of peripheral CD8 T cells from HY^cd4 male and female mice. (F) Relative body weights of male Rag2^−/− mice after transfer of naïve CD4⁺CD25⁻CD45RB^hi B6 male T cells alone or together with male HY^cd4 CD8 T cells. (G) Ly49 versus CD122 profiles of CD8SP thymocytes or CD8 spleen T cells from HY^cd4F mice, S1P1^cKO.HY^cd4M mice, and HY^cd4M mice. (H) Ly49 versus CD8α expression on CD8⁺T3.70⁺ T cells from HY^cd4M mice analyzed 7 days after adoptive transfer into male and female B6 hosts. [(A), (B), and (E)] Representative data are from three independent experiments. (D) Data are pooled from three independent experiments. [(C) and (F)] Representative experiment from two independent experiments (n = 5 mice from each group). P values were determined using unpaired two-tailed Student’s t tests except for (C) and (F), where two-way analysis of variance (ANOVA) was used instead (means ± SEMs).

Fig. 5.. Assessment of clonal eviction in normal polyclonal mice.

(A) Analysis of MHC-I–selected thymocytes from MHC-II⁻ polyclonal mice. PD-1 expression on TCR^hi (non-NKT) thymocytes (left), and CD4 versus CD8α profiles on PD-1⁺ and PD-1⁻ TCR^hi thymocytes (right). (B) CD4 versus CD8α profiles on PD-1⁺TCRβ^hi thymocytes from MHC-II⁻ male and female mice. (C) Gfi1 and S1pr1 mRNA expression in sorted thymocyte subsets from MHC-II⁻ mice. (D) Analysis of RTEs in the spleens of RAG2-GFP.MHC-II⁻ mice. PD-1 expression (left) and CD4 versus CD8α profiles on PD-1⁺ and PD-1⁻ RTEs (right). (E) Quantification of RAG-GFP fluorescence in spleen RTEs from RAG2-GFP.MHC-II⁻ mice. (F) Expression of Cd4, Cd8a, and Runx3d mRNA in sorted spleen and LN T cell subsets in Rag-GFP.MHC-II⁻ mice. (G) RTE CD4/8^dullPD-1⁺ and CD8⁺PD-1⁺ were sorted from the spleens and LNs of RAG-GFP.MHC-II⁻ mice and placed into separate IL-7 in vitro cultures for 4 days. (H) Ly49 versus CD8α profile of total CD8 T cells from MHC-II⁻ spleens (left) and comparison of TCRβ and CD8α expression on Ly49⁻ and Ly49⁺ CD8 T cell subsets (bar graphs). (I) Comparison between B6 male and B6 female mice of HY–IA^b tetramer binding to CD4 T cells (left), HY–H-2D^b tetramer binding to total and Ly49⁺ CD8 T cells (middle), and Flu–H-2D^b tetramer binding to Ly49⁺CD8 T cells (right). [(A) to (H)] Representative experiment from two to three independent experiments. (I) Data are from two independent experiments. P values were determined using unpaired two-tailed Student’s t tests. Data are expressed as means ± SEMs.