Plasticity and lineage commitment of individual TH1 cells are determined by stable T-bet expression quantities

Abstract

T helper 1 (T_H1) cell identity is defined by the expression of the lineage-specifying transcription factor T-bet. Here, we examine the influence of T-bet expression heterogeneity on subset plasticity by leveraging cell sorting of distinct in vivo-differentiated T_H1 cells based on their quantitative expression of T-bet and interferon-γ. Heterogeneous T-bet expression states were regulated by virus-induced type I interferons and were stably maintained even after secondary viral infection. Exposed to alternative differentiation signals, the sorted subpopulations exhibited graded levels of plasticity, particularly toward the T_H2 lineage: T-bet quantities were inversely correlated with the ability to express the T_H2 lineage-specifying transcription factor GATA-3 and T_H2 cytokines. Reprogramed T_H1 cells acquired graded mixed T_H1 + T_H2 phenotypes with a hybrid epigenetic landscape. Continuous presence of T-bet in differentiated T_H1 cells was essential to ensure T_H1 cell stability. Thus, innate cytokine signals regulate T_H1 cell plasticity via an individual cell-intrinsic rheostat to enable T cell subset adaptation to subsequent challenges.

Fig. 1.. In vivo–primed T_H1 cells maintain plasticity toward the T_H2 program.

(A to F) Naive LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into C57BL/6 mice, followed by LCMV infection. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (A) Experimental scheme illustrating the adoptive cell transfer procedure and post-infection (p.i.) analysis time points. (B) Ten days after infection, expression of the indicated lineage-specifying transcription factors and signature cytokines were determined in splenic CD4⁺ Thy1.1⁺ donor cells by FACS. Geometric mean indices + SEM of T-bet (T_H1), GATA-3 (T_H2), RORγt (T_H17), BCL6 (T_FH), or FoxP3 (T_reg) protein expression in CD4⁺ Thy1.1⁺ donor cells versus endogenous naive CD62L^hi CD44^lo CD4⁺ Thy1.2⁺ T cells are shown. Dot plot shows IFN-γ and IL-4 expression after PMA/ionomycin stimulation of splenic CD4⁺ Thy1.1⁺ donor cells. (C) IL-4–induced STAT6 phosphorylation and total STAT6 protein levels in naive and effector splenic CD4⁺ Thy1.1⁺ donor cells. Geometric mean indices + SEM of pSTAT6 staining of IL-4–exposed versus untreated cells (left bars) and of total STAT6 staining versus isotype control staining (right bars) within CD4⁺ Thy1.1⁺ donor cells are depicted. Statistics: Mann-Whitney test. (D) Cytokine production of CD4⁺ Thy1.1⁺ donor cells treated as indicated was measured after stimulation for 24 hours with PMA/ionomycin using cytometric bead arrays. Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test. [(E) and (F)] GATA-3 and T-bet protein expression in reactivated CD4⁺ Thy1.1⁺ donor cells were determined by FACS. (F) Dot plot overlays illustrate GATA-3 and T-bet coexpression. GATA-3 relative protein expression, normalized to classic T_H2 cells, is depicted (left bar graph). Relative increase in GATA-3 protein expression is plotted compared to levels of GATA-3 in naive T cells or T_H1 cells (right bar graph). [(A) to (F)] Data are pooled from three independent experiments (n = 9).

Fig. 2.. IL-4 drives in vivo–primed T_H1 cells to acquire an additional T_H2-like epigenetic imprint.

(A to C) Naive LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into C57BL/6 mice, which were subsequently infected with LCMV. Effector T_H1 Thy1.1⁺ donor T cells were isolated on day 10 of infection and reactivated for two rounds of 4 days under neutral (anti–IL-4, anti–IL-12, and anti–IFN-γ) or T_H2 (IL-4, anti–IL-12, and anti–IFN-γ) conditions. (A) Experimental scheme. [(B) and (C)] H3K27ac, H3K4me3, H3K4me1, H3K27me3, and H3K9me3 modifications of selected T_H2 and T_H1 genes are presented. (B) Z scores of all 87 T_H2 or 93 T_H1 genes of each respective histone modification of the different cell types are depicted. (C) Histone modifications in individual characteristic T_H2 or T_H1 signature genes in the cell types indicated at the top of the panels are plotted. The color coding depicts the z score value. A z score equal to 0 means that the score for a particular gene is the same as the mean for all conditions, while a z score of +1 indicates that the corresponding value is one SD above the mean. Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. [(B) and (C)] Data represent one experiment with two independent biological replicates pooled from four to eight independent mice.

Fig. 3.. Type I IFNs restrain maximal GATA-3 induction in T_H1 cells.

(A to F) Naive WT, Ifnar1^−/−, Ifngr1^−/−, and Tbx21^−/− LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into C57BL/6 mice or Il12b^−/− mice. Recipient mice were infected with LCMV. On day 10 after LCMV infection, CD4⁺ Thy1.1⁺ donor T cells were isolated, characterized, and reactivated for two rounds of 4 days under T_H2 conditions. [(A) to (E)] Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; **P < 0.01; ***P < 0.001; ****P < 0.0001. (A) Experimental scheme. (B) T-bet protein expression was determined in splenic CD4⁺ Thy1.1⁺ donor cells by FACS on day 10 after infection. Geometric mean indices + SEM of T-bet staining within CD4⁺ Thy1.1⁺ donor T cells are depicted. [(C) and (D)] CD4⁺ Thy1.1⁺ donor T cells were analyzed by FACS for expression of the indicated surface molecules on day 10 after infection. [(E) and (F)] Geometric mean indices + SEM of GATA-3 within CD4⁺ Thy1.1⁺ donor T cells isolated directly after infection (day 10 after infection) or after reactivation for two rounds of 4 days under T_H2 conditions. [(A) to (F)] Data are pooled from three independent experiments (n = 6 to 13).

Fig. 4.. T-bet acts in a dose-dependent manner to stabilize the T_H1 program and repress the T_H2 program.

(A to E) Naive Tbx21^+/+, Tbx21^+/−, and Tbx21^−/− LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into C57BL/6 mice. Recipient mice were infected with LCMV. [(A) to (E)] Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (A) After 10 days of infection, T-bet and GATA-3 protein expression were determined in splenic CD4⁺ Thy1.1⁺ donor cells by FACS. (B) Geometric mean indices + SEM of T-bet and GATA-3 staining within CD4⁺ Thy1.1⁺ donor T cells are depicted. IFN-γ production of CD4⁺ Thy1.1⁺ donor cells was measured upon restimulation of spleen cells with PMA/ionomycin (right bar graphs). Mean frequencies and geometric mean + SEM of IFN-γ⁺ cells within CD4⁺ Thy1.1⁺ donor cells are depicted. (C) RNA-seq analysis of selected 93 T_H1 and 87 T_H2 genes in in vivo–primed Tbx21^+/+, Tbx21^+/−, and Tbx21^−/− LCMV-specific CD4⁺ Thy1.1⁺ cells after expansion in vitro under neutral conditions. Data represent one experiment with two independent biological replicates pooled from seven to nine independent mice. [(D) and (E)] CD4⁺ Thy1.1⁺ donor T cells of the indicated genotypes were isolated and reactivated for two rounds of 5 days under T_H1 (IL-12, anti–IL-4, and anti–IFN-γ) or T_H2 conditions. T_H2 cells derived from WT naive LCMV-specific CD4⁺ Thy1.1⁺ cells served as control. (D) GATA-3 protein amounts were determined in CD4⁺ Thy1.1⁺ donor cells by FACS. Representative histograms are shown. Bar graph shows geometric mean indices + SEM of GATA-3 within CD4⁺ Thy1.1⁺ donor T cells isolated after reactivation for two rounds of 4 days under T_H2 conditions. (E) Quantitative PCR of the indicated cytokine genes after restimulation with PMA/ionomycin for 3 hours. Data are pooled from three independent experiments (n = 6 to 9).

Fig. 5.. Long-term maintenance of graded GATA-3 expression in IL-4–treated T_H1 cells.

(A to E) Naive Tbx21^+/+ and Tbx21^+/− LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into C57BL/6 mice, followed by LCMV infection. After 10 days, CD4⁺ Thy1.1⁺ donor T cells were isolated and reactivated in vitro for two rounds of 4 days under either neutral or T_H2 conditions. Subsequently, reactivated CD4⁺ Thy1.1⁺ cells were retransferred into naive C57BL/6 mice. Data represent the mean ± SEM of n = 4 to 6 mice per group and time point, with significance denoted as follows: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (A) Experimental scheme. (B) GATA-3 protein levels in CD4⁺ Thy1.1⁺ donor cells in the peripheral blood of recipient mice at various time points after retransfer. GATA-3 geometric mean index depicts the fold change of GATA-3 geometric mean in CD4⁺ Thy1.1⁺ donor cells compared to endogenous naive CD62L^hiCD4⁺Thy1.2⁺ T cells. Statistics: Two-way ANOVA with Tukey’s multiple comparisons test. (C) Seventeen days after secondary transfer, GATA-3 and T-bet expression in CD4⁺ Thy1.1⁺ donor cells isolated from the spleen were determined. GATA-3 relative protein expression in reprogrammed CD4⁺ Thy1.1⁺ cells compared to those treated under neutral conditions (left graph). Geometric mean indices indicate the fold increase in T-bet staining geometric mean compared to isotype control staining in CD4⁺ Thy1.1⁺ cells (right graph). (D) Cytokine production in splenocyte supernatants from CD4⁺ Thy1.1⁺ recipient mice after 24 hours of restimulation with GP_61–80 peptide. [(D) and (E)] Statistics: Mann-Whitney test. (E) GATA-3 expression in CD4⁺ Thy1.1⁺ donor cells from the spleen, 17 days after secondary transfer, followed by 5 days of culture with GP_61–80 and APCs. Geometric mean indices indicate fold change of GATA-3 staining geometric means compared to isotype control staining. Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test.

Fig. 6.. The magnitude of T-bet expression predicts T_H1 cell plasticity.

(A to C) Naive T-bet-ZsG (TBGR) mice were infected with LCMV. Ten days after infection, CD4⁺ CD44^hi CD62L^lo T cells were sorted based on ZsG expression intensity via FACS. Data were pooled from two independent experiments with three biological replicates each, derived from n = 20 mice. [(A) to (C)] Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (A) Experimental scheme and sorting controls. CD4⁺ CD44^hi CD62L^lo T cells were sorted according to ZsG expression, and intracellular T-bet protein expression was assessed by FACS. (B) ZsG intensity, T-bet, and GATA-3 protein expression were analyzed in sorted T cell fractions using FACS. Geometric mean indices + SEM of ZsG intensity and T-bet and GATA-3 staining in sorted T cell populations are indicated. Geometric mean + SEM of IFN-γ expression intensity within gated ZsG^hi, ZsG^int, and ZsG^lo CD4⁺ Thy1.1⁺ donor cells was determined after restimulation with GP_61–80 through intracellular cytokine staining. (C) Endogenous LCMV-specific CD4⁺ T cells were reactivated under the indicated conditions with GP_61–80 and APCs. T_H2 cells derived from naive LCMV-specific CD4⁺ cells served as a control. GATA-3 protein expression was determined in activated CD4⁺ CD154⁺ T cells by FACS. Representative histograms are shown, and a bar graph displays geometric mean indices + SEM of GATA-3 within activated CD4⁺ CD154⁺ T cells after reactivation under T_H2 conditions.

Fig. 7.. Graded T-bet expression is a stable property of T_H1 cells.

(A to C) Naive LCMV-specific TBGR CD4⁺ Thy1.1⁺ cells were adoptively transferred into untreated TBGR Thy1.2⁺ mice, followed by LCMV infection. Effector CD4⁺ Thy1.1⁺ T cells were FACS-sorted based on ZsG expression intensity. Equal numbers of isolated CD4⁺ ZsG⁺ Thy1.1⁺ cells were retransferred into uninfected TBGR Thy1.2⁺ recipient mice subsequently infected with LCMV. Data represent a compilation of two independent experiments with three biological replicates each, derived from n = 20 mice. (A) Experimental scheme. (B) T-bet and GATA-3 protein expression in donor CD4⁺ Thy1.1⁺ T cells 10 days after secondary infection. Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (C) CD4⁺ Thy1.1⁺ donor T cells were isolated after secondary infection and reactivated in vitro under T_H2 conditions. GATA-3 protein expression was analyzed by FACS, with representative histograms provided.

Fig. 8.. Continuous T-bet expression safeguards T_H1 cell stability.

Naive Tbx21^fl/fl and Tbx21^fl/+ Rosa26CreERT2⁺ LCMV-specific CD4⁺ Thy1.1⁺ cells were adoptively transferred into Thy1.2⁺ Rosa26CreERT2⁺ C57BL/6 mice. Recipient mice were infected with LCMV. On days 7, 8, and 9 after infection, recipient mice were treated with tamoxifen, while a control group was left untreated. On day 10 after infection, CD4⁺ Thy1.1⁺ donor T cells were isolated and directly analyzed or reactivated for 5 days under T_H2 conditions. Data are representative of two independent experiments (n = 6 to 7 per group and experiment). (A) Experimental scheme. (B) IFN-γ expression and (C) T-bet and GATA-3 expression were determined in CD4⁺ Thy1.1⁺ donor cells by FACS on day 10 after infection. (D) GATA-3 and T-bet protein expression were determined by FACS in CD4⁺ Thy1.1⁺ donor cells after re-activation for 5 days under T_H2 conditions. T_H2 cells derived from naive LCMV-specific CD4⁺ Thy1.1⁺ cells served as control. (E) Quantitative PCR of the indicated cytokine genes after restimulation with PMA/ionomycin for 3 hours. Statistics: Kruskal-Wallis test with Dunn’s multiple comparisons test; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.