Crossreactive αβ T Cell Receptors Are the Predominant Targets of Thymocyte Negative Selection

Abstract

The precise impact of thymic positive and negative selection on the T cell receptor (TCR) repertoire remains controversial. Here, we used unbiased, high-throughput cloning and retroviral expression of individual pre-selection TCRs to provide a direct assessment of these processes at the clonal level in vivo. We found that 15% of random TCRs induced signaling and directed positive (7.5%) or negative (7.5%) selection, depending on strength of signal, whereas the remaining 85% failed to induce signaling or selection. Most negatively selected TCRs exhibited promiscuous crossreactivity toward multiple other major histocompatibility complex (MHC) haplotypes. In contrast, TCRs that were positively selected or non-selected were minimally crossreactive. Negative selection of crossreactive TCRs led to clonal deletion but also recycling into intestinal CD4(-)CD8β(-) intraepithelial lymphocytes (iIELs). Thus, broadly crossreactive TCRs arise at low frequency in the pre-selection repertoire but constitute the primary drivers of thymic negative selection and iIEL lineage differentiation.

Figure 1. Expression of pre-selection TCRs

A) Pre-selection DP thymocytes were sorted from MHC-deficient or Trac^−/− mice as shown, and TCRα chains were cloned and inserted into a conditional retroviral vector, as diagrammed. TCRα RV^cond contains a loxP-flanked GFP, the TCRα chain of interest, and an IRES-Thy1.1. Only after Cd4-Cre-mediated excision of the upstream GFP sequence can the TCRα chain be translated. Bone marrow from TCRVβ transgenic Cd4-Cre mice was transduced with TCRα RV^cond, and mixed with control CD45.1-congenic WT bone marrow transduced with Empty RV^cond, before injection into lethally irradiated, CD45.1-congenic recipient. B) Representative flow cytometry of the thymus and spleen of mixed RV chimeras, gated as indicated, demonstrating the different types of selection outcomes observed. All plots are pre-gated on transduced (Thy1.1⁺) CD45.2 cells. The TCR^hi gate was determined using antibodies specific for the relevant Vβ and Vα expressed in a given chimera. No Sel, no selection; PS, positive selection; NS, negative selection. Additional information about each TCR can be found in Table S1 and additional controls in Figures S1 and S2. Data shown are representative of more than 100 independent chimeras as detailed in subsequent figures.

Figure 2. Frequencies of positive selection, negative selection, and No Selection outcomes for pre-selection and post-selection (DP CD69⁺) TCRs

A) Summary of outcomes. B) Flow cytometry of thymus and spleen from a TCR identified as inducing medullary rather than cortical negative selection. C–D) Summary of all 92 pre-selection and post-selection (DP CD69⁺) TCRs analyzed in RV chimeras. CD5 and PD-1 expression are presented relative to control amounts in CD69⁻ DPs. Cell numbers were normalized based on the frequency of transduction in splenic non-T cells. Open and filled circles denote chimeras constructed from Trac^+/− and Trac^−/− bone marrow, respectively. E) The numbers of thymocytes were normalized as in C–D. F) TCRα amino acid sequences from TCRs scored as no selection, positive selection, or negative selection were compared with a reference pool of TCRα sequences derived from Vβ-only transgenic sple-nocytes, to determine the frequency of overlap. p values were calculated by Student’s t test. Data pooled from 20 independent experiments.

Figure 3. Identification of MHC crossreactive TCRs

A) Unsignaled CD69⁻ PD-1⁻ DP thymocytes from no selection TCR Vβ8.2Vα8-75 or negative selection TCR Vβ8.2Vα2-20 were cultured with a mixture of mouse thymocytes and splenocytes or with SV-40 transformed fibroblasts derived from indicated strains of various MHC haplotypes prior to staining for CD69 and PD-1. Black boxes highlight MHC-reactive TCRs. B) Flow cytometry of Vβ8.2Vα2-20 expressing cells in a B6 (H2^b) or B10.D2 (H2^d) backgrounds. Similar in vivo results were obtained with 3 out of 3 additional TCRs rective against the H2^b or H2^d backgrounds in vitro.

Figure 4. Most negative selection TCRs are MHC crossreactive

TCRs are listed and classified according to their negative selection (A), positive selection (B), or no selection (C) phenotype, determined as in Fig. 3A, with filled box denoting reactivity against corresponding MHC expressed by a mixture of spleen and thymic cells, or by a panel of SV40-transformed fibroblasts as in Fig. 3. MHC reactivity was defined by induction of greater than 10% increase in CD69⁺PD-1⁺ cells over MHC-deficient stimulator cells. Most TCRs were from the pre-selection pool, but some TCRs obtained from iIELs or post-selection (CD69⁺) pools (as listed in Table S1) were added to increase the size of the positive selection and negative selection TCR collections. D) Summary of MHC crossreactivity for TCRs of each group. p values were calculated by Mann-Whitney test. Data pooled from 15 independent experiments.

Figure 5. Crossreactive TCRs are frequently reactive against both MHC class I and II

A) Unsignaled CD69⁻ PD-1⁻ DP thymocytes purified from the indicated retrogenic mice were cultured with stimulator cells derived from either WT or various MHC-deficient mouse strains in the B6 background to determine the relative increase in CD69⁺ PD-1⁺ cells. Representative TCRs are shown. B–C) In vivo validation by examining thymocyte development in MHC I-deficient, MHC II-deficient, or MHC-deficient backgrounds for Vβ8.2Vα2-107 (B) and for Vβ7Vα2-U4 (C). In total, we confirmed the expected MHC reactivity in vivo for 4 out of 4 different TCRs. D) Summary of MHC class I and class II reactivities for 19 negative selection TCRs. Data are pooled from 8 independent experiments.

Figure 6. Cortical negative selection requires store-operated calcium entry

A) Flow cytometry analysis of thymus from Vav-icre Stim1^fl/flStim2^fl/fl (dKO) and littermate control mice gated as indicated. B) Flow cytometry analysis of mixed bone marrow chimeras composed of a 1:1 mixture of WT (CD45.1) and Vav-iCre Stim1^fl/flStim2^fl/fl (CD45.2) cells. C) Summary plots with data points pooled from experiments using WT and dKO mice (black circles), mixed bone marrow chimeras (gray filled circles), and non-competitive bone marrow chimeras (open circles). D) Left: Flow cytometry analysis of sorted DP CD69⁻ thymocytes from mice of the indicated genotypes stimulated for 18 hours with plate-bound anti-TCRβ + DCs. Right: Summary plots. p values were calculated by Student’s t test. Data pooled from 4 experiments (C) or 2 experiments (D).