T-bet Controls Pathogenicity of CTLs in the Heart by Separable Effects on Migration and Effector Activity

T-bet-deficient T cells are markedly less pathogenic than control T cells in a model of CD8⁺ T cell-mediated myocarditis. A, Survival was determined after adoptive transfer of 250,000 (n = 22/group) and 500,000 (n = 12/group) T-bet-deficient (T-bet ^−/−) or control (wild-type (WT)) OT-I effector T cells into cMy-mOVA mice (p = 0.001 and 0.01, for 250,000 and 500,000 T cell doses, respectively). B, Serum levels of cardiac TnI were determined in samples taken between days 4–5 after adoptive transfer of 250,000 and 500,000 T-bet ^−/− or WT OT-I effector T cells. The data represent values from individual mice in one experiment for each cell dose (p = 0.037 and < 0.0001 for 250,000 and 500,000 T cell doses, respectively). Similar results were obtained in five independent experiments. C, The histological score of myocarditis was determined in hearts removed from animals sacrificed at day 5 after adoptive transfer of 250,000 and 500,000 T-bet ^−/− or WT OT-I effector T cells. The data represent scores from individual mice (p = 0.0004 and 0.0014 for 250,000 and 500,000 T cell doses, respectively). Similar results were obtained in three separate experiments. D, Serum samples were collected from cMy-mOVA mice at days 4–5 after adoptive transfer of 500,000 T-bet-deficient (T-bet ^−/−) or control (WT) OT-I effector T cells. The concentrations of the indicated cytokines were determined by FACS-based bead assay. Data points are from individual mice in one representative experiment of three performed. Horizontal bars, Means. Values of p between WT and T-bet ^−/− groups are: 0.011 (TNF-α); 0.0004 (IFN-γ); 0.001 (MCP-1).

Adoptively transferred T-bet-deficient T cells are not detectable in hearts of cMy-mOVA mice. Photomicrographs are shown of H&E-stained heart sections from cMy-mOVA mice sacrificed 5 days after adoptive transfer of 250,000 control OT-I cells (A and C) or T-bet-deficient OT-I cells (B and D). Frozen sections from the same hearts were also stained for CD90.2, which is expressed on the transferred T cells but not on host cells (E and F). The data shown are typical of stains performed on ∼30 hearts from each experimental group.

Effector phenotype of wild-type (WT) and T-bet-deficient CD8⁺ T cells. A, Effector T cells were removed from primary cultures at day 5, and were restimulated in microwells 1 μg/ml SIINFEKL peptide plus APCs, as described in Materials and Methods. Culture supernatants were removed 48 h later, and analyzed for the indicated cytokines by FACS-based cytokine bead assays. The data represent the mean ± SEM of determinations made in four experiments. Values of p indicating significant differences between WT and T-bet ^−/− groups were found for the following cytokines: IL-2 (p = 0.009); IL-10 (p = 0.028); IFN (p = 0.008); TNF (p = 0.045). B, Control (WT) and T-bet^−/− OT-I effector cells were assayed for cytotoxic activity against SIINFEKL-loaded EL4 target cells by ⁵¹Cr release assay. The data shown are representative of three independent cytotoxicity assays. C, Quantitative (real-time) RT-PCR analysis of FasL, perforin, and granzyme B were performed on RNA samples from WT or T-bet^−/− OT-I effector cells. Data represent means ± SEM or RNA samples run in duplicate from four to five independent cell preparations.

T-bet-deficient CD8⁺ T cell effectors have an impaired migratory phenotype characterized by lymph node retention and decreased expression of receptors for chemokines expressed in inflamed heart. A, The cardiac lymph node was removed from CD90.1-homozygous cMy-mOVA mice 5 days after adoptive transfer of 1 × 10⁶ CD90.2 homozygous control (wild-type (WT)) or T-bet-deficient (T-bet^−/−) effector OT-I cells, and lymph node cell suspensions were stained for CD90.2 and CD8. The numbers in the upper right quadrants indicate the percent of total gated cells positive for both CD90.2 and CD8. B, After 6 days of primary activation with peptide and APCs, WT or T-bet^−/− OT-I effector cells were restimulated for 48 additional hours with anti-CD3ε. RNA was then isolated, and quantitative (real-time) RT-PCR analysis of chemokine receptor gene expression was performed. Data represent means ± SEM of RNA samples from four independent experiments. C, WT and T-bet^−/− OT-I effector cells were stained with purified anti-CXCR3 Ab and analyzed by flow cytometry. Isotype control-stained T-bet^−/− cells are shown; histograms of isotype control-stained WT OT-I cells were superimposable. D, Quantitative (real-time) RT-PCR analysis of chemokine gene expression was performed on RNA samples of heart tissue of cMy-mOVA mice sacrificed 5 days posttransfer of 500,000 WT or T-bet^−/− OT-I effector cells. Data represent means ± SEM of RNA samples from five mice per group in two independent experiments.

CXCR3-deficient effector CD8⁺ T cells are less pathogenic than control effector CD8⁺ T cells. A, CXCR3 ^−/− and control effector T cells were removed from primary cultures at day 5. Chemotaxis assays were performed with the indicated concentrations of CXCL11 (ITAC). Data represent the mean of two samples ± SD in a typical experiment of three performed. B, cMy-mOVA mice were sacrificed at day 4 postadoptive transfer of 250,000 control (wild-type (WT)) or CXCR3-deficient (CXCR3^−/−) OT-I effector cells. Representative H&E-stained sections of heart from the two experimental groups are shown. C, Histopathological scores for myocarditis were determined for each animal described in B. The mean scores for the two groups (horizontal bars) were significantly different (p = 0.0003). Serum TnI (D) and serum cytokine concentrations (E) were determined in samples taken at time of sacrifice of the animals described in B. Mean values of each group (horizontal bars) were significantly different for: troponin, p = 0.021; TNF-α, p = 0.018; IFN-γ, p = 0.032; and MCP-1, p = 0.014.

Retroviral-mediated expression of CXCR3 in T-bet-deficient CD8⁺ T cells reconstitutes cardiac migration, but not cardiomyocyte toxicity. A, Control or T-bet^−/− OT-I cells were infected with control or CXCR3-expressing retrovirus during primary activation cultures, as described in Materials and Methods, with ∼30% transduction efficiency, as determined by expression of virally encoded GFP. An aliquot of T cells from each group was stained for human CXCR3; the histograms show CXCR3 staining for GFP⁺ cells. B, cMy-mOVA mice were injected with 500,000 virally transduced cells, and sacrificed at day 5 posttransfer. Representative H&E-stained sections of heart from mice receiving control or CXCR3 virus-infected T-bet^−/− OT-1 cells are shown. C, Histopathological scores for myocarditis were determined for each animal. The mean scores (horizontal bars) for the recipients of control virus- vs CXCR3 virus-infected T-bet^−/− OT-I cells were significantly different; p = 0.0023. D, Serum TnI concentrations were determined in samples taken at time of sacrifice. Mean values (horizontal bars) were significantly different between the recipients of control virus- vs CXCR3-virus-infected T-bet^−/− OT-I cells (p = 0.0321), and dramatically decreased in comparison to recipients of control OT-I cells (Fig. 2B). E, Serum cytokine concentrations were determined in the samples described in D; mean values were significantly different between the two groups only for IL-6 (p = 0.0351).