The histone methyltransferase EZH2 primes the early differentiation of follicular helper T cells during acute viral infection

Abstract

Epigenetic modifications to histones dictate the differentiation of naïve CD4⁺ T cells into different subsets of effector T helper (T_H) cells. The histone methyltransferase enhancer of zeste homolog 2 (EZH2) has been implicated in the mechanism regulating the differentiation of T_H1, T_H2 and regulatory T (T_reg) cells. However, whether and how EZH2 regulates follicular helper T (T_FH) cell differentiation remain unknown. Using a mouse model of acute lymphocytic choriomeningitis virus (LCMV) infection, we observed abundant EZH2 expression and associated H3K27me3 modifications preferentially in the early committed virus-specific T_FH cells compared to those in T_H1 cells. Ablation of EZH2 in LCMV-specific CD4⁺ T cells leads to a selective impairment of early T_FH cell fate commitment, but not late T_FH differentiation or memory T_FH maintenance. Mechanistically, EZH2 specifically stabilizes the chromatin accessibility of a cluster of genes that are important for T_FH fate commitment, particularly B cell lymphoma 6 (Bcl6), and thus directs T_FH cell commitment. Therefore, we identified the chromatin-modifying enzyme EZH2 as a novel regulator of early T_FH differentiation during acute viral infection.

Fig. 1

Chromatin states of the virus-specific T_FH and T_H1 cells in response to an acute viral infection. a Numbers of chromatin peaks with differential accessibility (FDR < 0.05; FC > 4) between SMARTA T_FH cells and SMARTA T_H1 cells at the indicated time points after LCMV Armstrong infection. b PCA plot of the peak accessibilities in naïve SMARTA CD4⁺ T cells, SMARTA T_FH cells (days 2, 5 and 8 postinfection) and SMARTA T_H1 cells (days 2, 5 and 8 postinfection). Each dot represents a replicate of the indicated group. c Chromatin accessibility heat map of differential peaks from a. Each row represents one of the 15,600 differential peaks that was center-aligned and extended upstream and downstream by 1 kb from the center. The peaks are K-means clustered. d ATAC-Seq signal profiles of T_FH lineage-associated gene loci. e ATAC-Seq signal profiles of T_H1 lineage-associated gene loci. Differential peaks are highlighted in gray (d and e). The data were obtained from one experiment with three biological replicates (pooled from at least five mice per group) (a–e)

Fig. 2

Dynamic changes in EZH2 expression and the H3K27me3 modification in virus-specific T_FH cells. a, b Comparison of EZH2 (a) and H3K27me3 (b) levels between SMARTA T_FH cells (CD25^loCXCR5⁺) and SMARTA T_H1 cells (CD25^hiCXCR5⁻) from the spleens of CD45.2⁺ wild-type mice that underwent adoptive transfer of CD45.1⁺ SMARTA cells and analyzed on day 2 after LCMV Armstrong infection and in naïve (CD44^loCD62L^hi) SMARTA cells (N). c, d Flow cytometry analysis of EZH2 (c) and H3K27me3 (d) levels in T_FH cells derived from the mice listed in a on days 2, 5, 8, 15 and 30 after the LCMV Armstrong infection and in naïve SMARTA cells (N). The EZH2 or H3K27me3 mean fluorescence intensity (MFI) ratio at each time point was calculated as the MFI of T_FH cells / MFI of CD4⁺CD44^lo T cells in an identical mouse. e Flow cytometry analysis of the EZH2^hi, EZH2^inter and EZH2^lo subsets of T_FH cells and T_H1 cells from the mice shown in a on day 2 after the LCMV Armstrong infection. The proportions of T_FH cells and T_H1 cells and the MFI of Bcl-6 in each population are summarized in f, g and h, respectively. i The MFI of H3K27me3 in the EZH2^hi, EZH2^inter and EZH2^lo subsets of T_FH cells described in e. The proportions of T_FH cells (j) and T_H1 cells (k) among the H3K27me3^hi, H3K27me3^inter and H3K27me3^lo subsets of SMARTA cells from mice shown in a on day 2 after the LCMV Armstrong infection. *P < 0.05, ***P < 0.001 and ****P < 0.0001 (unpaired two-tailed t-test). The data are representative of two independent experiments with at least four mice per group (a–d and f–k; error bars in a–d and f–k indicate the s.d.)

Fig. 3

Role of EZH2 in early T_FH commitment during an acute viral infection. a Experimental setup. A retrovirus overexpressing iCre was introduced into CD45.1⁺Ezh2^+/+ SMARTA (SM) cells and CD45.1⁺Ezh2^fl/fl SM cells, which were transferred into CD45.2⁺ WT recipients. Then, the recipients were infected with the LCMV Armstrong strain and analyzed on day 2 postinfection. b Flow cytometry analysis of Ezh2^+/+ and Ezh2^fl/fl SM cells transduced with a retrovirus expressing iCre. The numbers adjacent to the outlined areas indicate the percentages of CD25^loCXCR5⁺ T_FH cells, which are summarized in c. d Experimental setup. Ezh2^+/+Cd4-Cre SMARTA cells (Ezh2^+/+ SM-Cre; CD45.1⁺CD45.2⁻) and Ezh2^fl/flCd4-Cre SMARTA cells (Ezh2^fl/fl SM-Cre; CD45.1⁺CD45.2⁺) were cotransferred into WT recipients (CD45.1⁻CD45.2⁺), which were infected with the LCMV Armstrong strain and assessed on day 2 postinfection. e Flow cytometry analysis of Ezh2^+/+ SM-Cre and Ezh2^fl/fl SM-Cre SMARTA cells. The numbers adjacent to the outlined areas indicate the percentages of CD25^loCXCR5⁺ T_FH cells, which are summarized in f (left panel). Total numbers of CD25^loCXCR5⁺ T_FH cells in e are presented (f, right panel). g Quantification of TCF-1, Bcl-6, ICOS and CXCR5 levels in the CD25^loCXCR5⁺ T_FH cells shown in e. h Experimental setup. After treatment with either EPZ6438 or vehicle for 3 days, CD45.1⁺SMARTA cells were transferred into WT CD45.2⁺ recipients that were subsequently infected with the LCMV Armstrong strain. The adoptively transferred SMARTA cells were analyzed on day 2 postinfection. i Flow cytometry analysis of EPZ6438-treated and vehicle-treated SMARTA cells. The numbers adjacent to the outlined areas indicate the proportions of CD25^loCXCR5⁺ T_FH cells, which are summarized in j (left panel). Total numbers of CD25^loCXCR5⁺ T_FH cells analyzed in i are presented in j (right panel). k Quantification of TCF-1, Bcl-6, ICOS and CXCR5 levels in the CD25^loCXCR5⁺ T_FH cells shown in j. NS not significant; *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (paired two-tailed t-test (c, f and g) or unpaired two-tailed t-test (j and k)). The data are representative of two independent experiments with at least three mice (c, f, g, j and k) per group (error bars in j and k indicate the s.d.)

Fig. 4

Requirement for EZH2 expression in endogenous virus-specific T_FH cell differentiation. a Flow cytometry analysis of GP66–77 tetramer-positive CD4⁺ T cells in the spleens of Ezh2^fl/flCd4-Cre mice and Ezh2^fl/fl mice on day 8 after LCMV Armstrong infection. The numbers adjacent to the outlined areas indicate the proportions of ICOS^hiCXCR5⁺ T_FH cells. b The percentage (left panel) and number (right panel) of ICOS^hiCXCR5⁺ T_FH cells in a. c Quantification of Bcl-6, TCF-1, PD-1 and CXCR5 levels in the ICOS^hiCXCR5⁺ T_FH cells shown in a. d Setup of the BM chimera experiment. Irradiated CD45.1⁺ WT recipients underwent the adoptive transfer of CD45.1⁺ WT BM cells (60%) and CD45.2⁺ Ezh2^fl/flERT2-Cre BM cells (40%). Two months after reconstitution, the recipients were treated with tamoxifen and then infected with the LCMV Armstrong strain. e Flow cytometry analysis of GP66–77 tetramer-positive CD4⁺ T cells in the spleens of recipient mice shown in d on day 8 after the LCMV Armstrong infection. The numbers adjacent to the outlined areas indicate the proportions of Bcl-6^hiCXCR5⁺ T_FH cells, which are summarized in f. g Quantification of Bcl-6, TCF-1, PD-1 and CXCR5 levels in the Bcl-6^hiCXCR5⁺ T_FH cells shown in e. NS not significant; *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (unpaired two-tailed t-test (b and c) or paired two-tailed t-test (f and g)). The data are representative of two independent experiments with at least four mice (b, c, f and g) per group (error bars in b and c indicate the s.d.)

Fig. 5

EZH2 is not essential for the late differentiation and maintenance of virus-specific T_FH cells during an acute viral infection. a Setup of the BM chimera experiment. Irradiated CD45.1⁺ WT recipients underwent adoptive transfer of CD45.1⁺ WT BM cells (60%) and CD45.2⁺ Ezh2^fl/flERT2-Cre BM cells (40%). Two months after reconstitution, the recipients were infected with LCMV Armstrong, followed by the administration of tamoxifen from days 4 to 7 after infection and an analysis on day 8. b Expression of EZH2 in virus-specific Bcl6^hiCXCR5⁺ T_FH cells from the BM chimeric WT and Ezh2^fl/flERT2-Cre mice shown in a. c Flow cytometry analysis of GP66–77 tetramer-positive CD4⁺ T cells in the spleens of the chimeras described in a on day 8 after the LCMV Armstrong infection. Numbers adjacent to outlined areas indicate the percentage of Bcl-6^hiCXCR5⁺ T_FH cells, which were summarized in d. e Levels of Bcl-6, TCF-1, PD-1 and ICOS in the Bcl-6^hiCXCR5⁺ T_FH cells shown in c. f The BM chimeras described in a were infected with LCMV Armstrong, treated with tamoxifen on days 9 to 12 after infection and analyzed on day 60. g Quantification of EZH2 expression in virus-specific CXCR5⁺ T_FH cells originating from BM chimeric WT and Ezh2^fl/flERT2-Cre mice shown in f. h Flow cytometry of GP66–77 tetramer-positive CD4⁺ T cells in the spleens of the chimeras described in f on day 60 after the LCMV Armstrong infection. Numbers adjacent to outlined areas indicate the percentage of CXCR5⁺ T_FH cells, which were summarized in i. NS not significant; *P < 0.05 and ***P < 0.001 (paired two-tailed t-test (b, d, e, g and i)). The data are representative of two independent experiments with at least three mice (b, d, e, g, h and i) per group

Fig. 6

Role of EZH2 in the remodeling of T_FH lineage-associated chromatin accessibility during viral infection. a Numbers of chromatin peaks with differential accessibility (FDR < 0.05; FC > 1.5) in naïve CD4⁺ T cells (N), WT T_FH cells and KO T_FH cells. b PCA plot of the peak accessibilities in naïve CD4⁺ T cells (green), WT T_FH cells (red) and KO T_FH cells (blue). Each dot represents a replicate of the group. c ATAC-Seq signal profiles of T_FH lineage-associated gene loci in naïve CD4⁺ T cells (N), WT T_FH cells and KO T_FH cells. d Tracks of H3K27me3 ChIP-Seq data from T_FH lineage-associated gene loci in naïve CD4⁺ T cells (N), T_FH cells and T_H1 cells. The data were obtained from two independent experiments with one biological replicate (pooled from at least five mice per group) in each experiment (a–c) and from one experiment with one biological replicate (d)

Fig. 7

Bcl-6 overexpression rescues compromised H3K27me3 modification-induced T_FH cell differentiation. a Heat map of the canonical T_FH lineage-associated genes in WT T_FH cells and EZH2 KO T_FH cells based on data obtained from the microarray analysis. b Quantitative RT-PCR of selected genes in a (normalized to their expression in WT T_FH cells). c Experimental setup. After treatment with either EPZ6438 or vehicle for 3 days, CD45.1⁺SMARTA (SM) cells transduced with an empty retrovirus or retroviral vector overexpressing Bcl6 were transferred into WT CD45.2⁺ recipients that were subsequently infected with LCMV Armstrong. The adoptively transferred SM cells were analyzed on day 5 postinfection. d Flow cytometry analysis of vehicle/EPZ6438-treated SM cells that were transduced with (hCD2⁺) or without (hCD2⁻) a retrovirus expressing Bcl-6. The numbers adjacent to the outlined areas indicate the percentages of SLAM^loCXCR5⁺ T_FH cells, which are summarized in e. NS not significant; *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (unpaired two-tailed t-test (b); paired two-tailed t-test (e)). The data were obtained from one experiment with two biological replicates pooled from at least ten mice per group (a), are representative of two independent experiments with three technical replicates pooled from at least ten mice per group (b) or are representative of two independent experiments with at least three mice (e) per group (b; error bars indicate the s.d.)