Dependence on Bcl6 and Blimp1 drive distinct differentiation of murine memory and follicular helper CD4+ T cells

Abstract

During the immune response, CD4+ T cells differentiate into distinct effector subtypes, including follicular helper T (Tfh) cells that help B cells, and into memory cells. Tfh and memory cells are required for long-term immunity; both depend on the transcription factor Bcl6, raising the question whether they differentiate through similar mechanisms. Here, using single-cell RNA and ATAC sequencing, we show that virus-responding CD4+ T cells lacking both Bcl6 and Blimp1 can differentiate into cells with transcriptomic, chromatin accessibility, and functional attributes of memory cells but not of Tfh cells. Thus, Bcl6 promotes memory cell differentiation primarily through its repression of Blimp1. These findings demonstrate that distinct mechanisms underpin the differentiation of memory and Tfh CD4+ cells and define the Bcl6-Blimp1 axis as a potential target for promoting long-term memory T cell differentiation.

Figure S1.

(A) Ox40-Cre⁺ Rosa26^YFP mice were infected with LCMV Arm, and spleen cells were analyzed at d7 or d30 p.i. Plots show Cre activity as reported by YFP expression on the indicated spleen T cell populations and is representative of more than three independent experiments. (B and C) Mixed BMCs made from Ctrl, Bcl6^AD, Blimp1^AD, or dKO tester and WT competitor were infected with LCMV Arm and analyzed at d7 p.i. (B) Ly6c versus PSGL1 expression on spleen GP66:I-A^b+ cells of the indicated genotype. (C) Bar graph shows the proportion of each subset as defined in B within the GP66:I-A^b+ populations in each tester genotype. (B and C) Data are from one experiment representative of three independent experiments with at least three animals per group. (D) Expression of Bcl6 and of YFP (indicative of Blimp1) on spleen GP66:I-A^b+ T cells (left plot) defines subsets analyzed for CCR7 versus CXCR5 expression (right three plots). (E) Graph shows the percentage of CCR7⁺ cells on the indicated GP66:I-A^b+ population. Each symbol represents a mouse. Data are from two pooled experiments representative of four independent experiments. **, P < 0.01; unpaired two-sided Welch’s t test. (F) UMAP plots from GP66:I-A^b+ T cells analyzed in Fig. 2. Each dot represents a cell and is color coded by genotype. Each plot includes cells of the experiment indicated at the top (defined in Table S1).

Figure 1.

Blimp1-independent impact of Bcl6 on CD4⁺ T cell responses. (A–D) Mixed BMC from Ctrl (Ox40-Cre⁺ Bcl6^+/+ Blimp1^+/+ or Ox40-Cre^–), Ox40-Cre⁺ Bcl6^fl/fl (Bcl6^AD), Blimp1^AD, or Ox40-Cre⁺ Bcl6^fl/fl Blimp1^fl/fl (dKO) tester, and WT competitor were infected with LCMV Arm and analyzed at d7 p.i. (A) Schematic representation of the experiment. (B) Tester/competitor ratios of GP66:I-A^b+ T cells normalized to that of naive CD4⁺ T cells in each animal. Data are from 3 independent transplantation experiments totaling 10 chimeras from each Ctrl, Blimp1^AD, Bcl6^AD, and dKO tester groups. Each symbol represents a mouse. (C) Flow cytometric analysis of CXCR5 versus PD1 expression on spleen GP66:I-A^b+ cells. Right graph shows the percentage of CXCR5⁺ PD1^hi among I-A^b-GP66–specific cells for each genotype. Data are from one experiment representative (three mice per group, each symbol representing a mouse) of three independent experiments with at least three animals per group. (D) Flow cytometric expression of CXCR5 versus CXCR6 on spleen tester GP66:I-A^b+ cells of the indicated genotype. Graph on right summarizes data from more than three independent experiments. In B–D, **, P < 0.01; ***, P < 0.001; unpaired two-sided Welch’s t test.

Figure 2.

Bcl6 promotes expression of the T_cmp transcriptome by repressing Blimp1. (A–C) Spleen GP66:I-A^b+ tester T cells from mixed BMC generated as in Fig. 1 A were analyzed by scRNAseq at d7 p.i. with LCMV Arm (Table S1 and Fig. S1 F). Data are from two replicates for each genotype, captured in experiments E1–E3. (A) Heatmap shows row-standardized signature scores (bottom rows) and expression of genes characteristic of Th1, Tfh, and T_cmp cells (top rows) among clusters (numbers below heatmap). Cells were clustered separately for each genotype. Top bar graph indicates the percentage of cells in each cluster within that genotype. (B) UMAP analysis of GP66:I-A^b+ T cells computed on cells of all genotypes from experiments E1–E3 and displayed separately for each genotype. Each dot represents a cell colored by cluster as in A. Outlines on the Ctrl plot circle clusters expressing genes characteristic of the T_cmp, Th1, and Tfh signatures and are projected on plots of mutant cells. (C) Violin plot shows relative gene expression across clusters and genotypes (key at bottom, color coded by cluster as in A). (D) Expression of CCR7 (top) and IL-7Rα (bottom) on spleen T cells from chimeric animals (left panels). Black lines show the expression on naive CD4⁺ T cells, and filled histograms show expression on GP66:I-A^b+ cells from the indicated genotype. Graphs on the right summarize the data; each symbol represents a mouse. Data are from one experiment representative of two independent ones, each with at least three mice per group. ***, P < 0.001; ****, P < 0.0001; unpaired two-sided Welch’s t test.

Figure S2.

Analyses of adoptively transferred Cd4-ERt2-Cre transgenic Smarta cells undergoing Bcl6 and Blimp1 deletion before antigen-induced activation, and scRNAseq and scATACseq analyses of Ctrl and dKO d7 p.i. LCMV-specific CD4⁺ T cells, from mixed BMC or adoptive transfer experiments, as in Fig. 3. (A) Schematic representation of Smarta T cell adoptive transfer experiments. (B) Graphs summarize numbers of (left) and Ki-67 expression in (right) Smarta cells from tamoxifen-treated donors that were either Bcl6^fl/fl Blimp1^fl/fl Cd4-ERt2-Cre (dKO, red-filled symbols) or Cre^– littermates (Ctrl, open symbols), after transfer to LCMV Arm–infected WT recipients. (C) Expression of IL-7Rα (top) or CXCR5 (bottom) on Ctrl (gray filled) or dKO (red filled) adoptively transferred cells processed as in B. Dotted black lines represent the expression on naive CD4⁺ T cells from the same infected mice. Graph on right summarizes the data; symbol code as in B. MFI, mean fluorescence intensity. (D) Expression of CXCR5 versus CXCR6 on the same mice as in C. Graph on right summarizes the data with symbols as in B. (B–D) Data are from one experiment representative of two independent experiments with at least three animals per group. (E) UMAP plots display GP66:I-A^b+ T cells (tester cells from BMC; experiments E1, E2, and E4) or adoptively transferred Smarta cells (E5), captured at d7 p.i. LCMV Arm and processed for scRNAseq (top) or scATACseq (bottom) as in Fig. 3. Each dot represents a cell and is color coded by genotype. Each plot includes cells of a single experiment, as indicated at the top and defined in Table S1. (F) Bar plots indicate the distribution of scATACseq (top) or scRNAseq (bottom) clusters of d7 p.i. cells into Ctrl (gray) or dKO (red) genotypes. Data are from the three experiments shown in E. In B–D, **, P < 0.01; ***, P < 0.001; unpaired two-sided Welch’s t test.

Figure 3.

Bcl6 and Blimp1 regulate the epigenomic landscape of CD4⁺ T cells. Integrated scRNAseq and scATACseq analyses of Ctrl and dKO LCMV-specific CD4⁺ T cells at d7 p.i. For each analysis platform, UMAP projection and clustering were performed on the full set of cells (both genotypes, all captures), and data are from three replicates for each genotype (Table S1 and Fig. S2 E). (A) UMAP plots of scATACseq data, displayed per genotype with cells color coded by scATACseq cluster (Roman numerals, top) or by projected transcriptomic cluster (Arabic numerals, bottom). Putative T_cmp, Th1, and Tfh clusters are outlined on the Ctrl UMAP plot (bottom). (B) Heatmaps show row-standardized gene expression (scRNAseq) or chromatin accessibility (scATACseq) among clusters of Ctrl and dKO cells. Row-standardized scores of transcriptomic signatures is shown at the bottom of the scRNAseq heatmap (color code underneath). The top bar graph indicates the percentage of cells in each individual cluster within the indicated genotype. Cluster numbers and color code are shown underneath heatmaps. For each genotype, clusters comprising <4% of cells were omitted from the heatmap. (C) Traces show scATACseq peaks at relevant loci in cells from Ctrl clusters II–IV.

Figure 4.

Bcl6 and Blimp1 regulate the transcriptome and epigenome of memory CD4⁺ T cells. Mixed BMCs made from WT competitor and tester of indicated genotypes were infected with LCMV Arm and analyzed ≥30 d later. (A) Right: Tester/competitor ratios of spleen GP66:I-A^b+ T cells normalized to that of naive CD4⁺ T cells in each animal 30 d p.i. Data are from three independent transplantation experiments, totaling eight mice in each group and performed as schematized on the left. Each symbol represents a mouse. **, P < 0.01. (B–E) Integrated scRNAseq and scATACseq analyses of tester I-A^b-GP66–specific Ctrl and dKO CD4⁺ T cells performed as reported in Table S1 and Fig. S3 A and displayed as in Fig. 3. Data are from two independent experiments for each genotype. (B) scRNAseq UMAP plot displayed per genotype and color coded by cluster. (C) Heatmaps show row-standardized gene expression or chromatin accessibility among clusters of Ctrl and dKO cells as in Fig. 3 B. (D) scATACseq UMAP plots shown separately for each genotype and color coded by scATACseq cluster (top) or by projected transcriptomic cluster (bottom). (E) Violin plot shows the relative expression of indicated genes across clusters and genotypes.

Figure S3.

LCMV-specific T cell responses in mixed BMCs made from WT competitor and either Ctrl or dKO tester cells, infected with LCMV Arm and analyzed ≥ 30 d later as in Figs. 4 and 5. (A–C) Ctrl and dKO GP66:I-A^b+ tester spleen T cells from mixed BMC were analyzed at d30-40 p.i. with LCMV Arm. (A) UMAP plots display tester GP66:I-A^b+ T cells captured at d35 (E6) or d40 (E7) p.i. and processed for scRNAseq (left) or scATACseq (right) as in Fig. 4 B. Each dot represents a cell and is color coded by genotype. Each plot includes cells from the indicated experiment. (B) Bar plots indicate the Ctrl (gray) versus dKO (red) genotype distribution of scATACseq (top) or scRNAseq (bottom) clusters of d30–40 p.i. cells referred to in A. Data are from the two experiments shown in A. (C) Traces show chromatin accessibility at selected loci across all cells from indicated Ctrl scATACseq clusters defined in Fig. 4, C and D. (D) Expression of CCR7 (left) or IL-7Rα (right) on Ctrl (gray filled) or dKO (red filled) tester I-A^b-GP66–specific CD4⁺ T cells at d30 p.i. with LCMV Arm of BMC mice as in Fig. 4 A. Dotted black lines represent the expression on CD44^lo CD4⁺ T cells from the same infected mice. Graphs summarize the data, with symbol code on the right. Data are from one experiment representative of two independent experiments with at least three animals per group. Each symbol represents a mouse. ****, P < 0.0001; unpaired two-sided Welch’s t test. MFI, mean fluorescence intensity. (E) Graph summarizes the frequency of ex vivo Ctrl (black symbols) or dKO (red symbols) tester CD4⁺ CD44^hi GP66:I-A^b+ T cells from mixed BMC, analyzed as in Fig. 4 A at d80 p.i. LCMV Arm. (F) Filled symbols show intracellular expression of TNFα (left) or IL-2 (right) by d80 p.i. tester CD44^hi CD4⁺ T cells from BMC analyzed in E, as assessed by flow cytometry after 4-h in vitro culture with GP66 peptide. Open symbols summarize background staining as gated on CD44^lo cells in the same cultures. Data in E and F are from a single experiment with three mice for each genotype and are not corrected for chimerism.

Figure 5.

Bcl6 repression of Blimp1 mediates its impact on CD4⁺ T cell memory response. (A) Ctrl and dKO tester cells from mixed BMC generated and processed as in Fig. 4 A are analyzed by flow cytometry for (i) staining with anti-CD4 and I-A^b-GP66 tetramer (left column, ex vivo) and (ii) intracellular expression of TNFα and IL-2 after 4 h at 37°C with GP66 peptide (right columns). Bottom graphs summarize the percentage of GP66-specific cells (left, not corrected for chimerism to allow comparison with cytokine response; normalized tester/competitor ratio expressed as in Fig. 4 A was 0.5 in that experiment) and of cytokine-producing cells (right two plots). Data are from two pooled independent experiments (two mice per experiment) and are representative of a total of four experiments. (B) Mixed BMCs with Ctrl or dKO tester cells were infected with LCMV Arm and rechallenged 30 d later with LCMV clone 13 (left). Graph shows tester/competitor ratios of spleen GP66:I-A^b+ T cells in each animal 5 d after clone 13 rechallenge, normalized to the ratio of naive CD4⁺ T cells. Data are from three independent transplantation experiments, totaling eight Ctrl and seven dKO chimeras. (C) Histogram overlays (top) show, and bottom graph summarizes, Ki-67 expression on spleen GP66:I-A^b+ Ctrl and dKO tester cells from BMC analyzed 5 d after clone 13 rechallenge (as in B, filled histograms and symbols on graphs). Plain line histograms and open symbols on graphs show Ki-67 expression on memory cells from chimeras analyzed at d30 p.i. LCMV Arm with no rechallenge as in Fig. 4 A. Data are from one recall experiment with two or three animals per group (as shown on graph) and are representative of two other recall experiments. In A–C, **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired two-sided Welch’s t test.