IRF4 Regulates the Ratio of T-Bet to Eomesodermin in CD8+ T Cells Responding to Persistent LCMV Infection

Abstract

CD8+ T cell exhaustion commonly occurs in chronic infections and cancers. During T cell exhaustion there is a progressive and hierarchical loss of effector cytokine production, up-regulation of inhibitory co-stimulatory molecules, and eventual deletion of antigen specific cells by apoptosis. A key factor that regulates T cell exhaustion is persistent TCR stimulation. Loss of this interaction results in restoration of CD8+ T cell effector functions in previously exhausted CD8+ T cells. TCR stimulation is also important for the differentiation of Eomeshi anti-viral CD8+ effector T cells from T-bethi precursors, both of which are required for optimal viral control. However, the molecular mechanisms regulating the differentiation of these two cell subsets and the relative ratios required for viral clearance have not been described. We show that TCR signal strength regulates the relative expression of T-bet and Eomes in antigen-specific CD8+ T cells by modulating levels of IRF4. Reduced IRF4 expression results in skewing of this ratio in the favor of Eomes, leading to lower proportions and numbers of T-bet+ Eomes- precursors and poor control of LCMV-clone 13 infection. Manipulation of this ratio in the favor of T-bet restores the differentiation of T-bet+ Eomes- precursors and the protective balance of T-bet to Eomes required for efficient viral control. These data highlight a critical role for IRF4 in regulating protective anti-viral CD8+ T cell responses by ensuring a balanced ratio of T-bet to Eomes, leading to the ultimate control of this chronic viral infection.

Fig 1. IRF4 regulates the T-bet to Eomesodermin ratio in stimulated CD8⁺ T cells.

P14 WT, P14 Irf4 ^+/fl or P14 Irf4 ^fl/fl cells were stimulated with the indicated concentrations of GP33 and F6L peptides in-vitro. At (A, D) 24, (B, E) 48, or (C, F) 72 h, cells were stained and analyzed for T-bet and Eomes expression. The graphs show compilations of the ratios of MFIs for T-bet relative to Eomes, each normalized to the value for 1μM GP33-stimulated cells at each time-point. Data were generated from gated live CD8⁺CD45.2⁺CD44^hi T cells analyzed in four independent experiments. ☼, denotes statistically significant difference between 1μM GP33-stimulated cells and all other sample groups on the graph as determined by unpaired t test with Welch’s correction.

Fig 2. Irf4 gene dosage regulates CD8⁺ T cell clonal expansion in response to LCMV-clone 13 infection and the differentiation of T-bet^hi and Eomes^hi subsets.

(A) Splenocytes from LCMV-clone 13-infected WT, Irf4 ^+/fl and Irf4 ^fl/fl mice were harvested at D8 p.i. and stained with a viability dye, LCMV-specific H2-D^b-GP276 tetramer, and antibodies to CD8, T-bet and Eomes. Dot plots show CD8 versus H2-D^b-GP276 tetramer staining (left). Graphs show compilations of proportions and numbers from D8 p.i. (right). Each data point represents an individual mouse and data are a compilation of three independent experiments. (B) LCMV-clone 13 titers in serum at D12 post-infection. Dotted line indicates limit of detection. Each data point represents an individual mouse and data are a compilation of four independent experiments. (C) Representative dot plots show T-bet vs Eomes staining on gated CD8⁺ live H2-D^b-GP276 specific cells at D8 p.i (left). Graphs show the MFI of T-bet and Eomes, each normalized to the average of WT samples for live CD8⁺ H2-D^b-GP276 specific cells, and the ratio of normalized MFIs for T-bet relative to Eomes (right). Each data point represents an individual mouse and data are a compilation of three independent experiments. (D) Graphs show compilations of proportions and numbers of T-bet⁺ Eomes^- and T-bet⁺ Eomes⁺ cells. Each data point represents an individual mouse and data are a compilation of three independent experiments. Significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test.

Fig 3. Persistent reduction in virus-specific T-bet⁺ Eomes^- CD8⁺ T cells in LCMV-clone 13-infected Irf4 ^+/fl mice.

(A, D) Splenocytes from LCMV-clone 13-infected WT, Irf4 ^+/fl and Irf4 ^fl/fl mice were were harvested at D21-24 p.i. and stained with a viability dye, LCMV-specific H2-D^b-GP276 and H2D^b-GP33 tetramers, and antibodies to CD8, T-bet and Eomes. Dot plots show CD8 versus H2-D^b-GP276 (A) or H2-D^b-GP33 (D) tetramer staining (left). Graphs show compilations of proportions and numbers from D21-24 p.i. (right). Each data point represents an individual mouse and data are a compilation of five independent experiments. (B) LCMV-clone 13 titers in serum at D26 post-infection. Dotted line indicates limit of detection. Each data point represents an individual mouse and data are a compilation of two independent experiments. (C, E) Dot plots show T-bet vs Eomes staining on live CD8⁺ H2-D^b-GP276 (C) or H2-D^b-GP33 (E) tetramer positive cells (left). Graph shows the ratio of MFIs of T-bet relative to Eomes, each normalized to the average value of WT samples (middle). Graphs show a compilation of proportions and numbers of T-bet⁺ Eomes^- cells for each population (right). Each data point represents an individual mouse and data are a compilation of five independent experiments. Significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test.

Fig 4. Cell-intrinsic role of IRF4 in regulating the balance of T-bet to Eomesodermin expression in CD8⁺ T cells responding to LCMV-clone 13 infection.

10⁴ CD90.1 WT P14 cells or CD90.1 Irf4 ^+/fl P14 cells were transferred into CD90.2 CD4-Cre⁺ host mice and one day later recipients were infected with LCMV-clone 13. (A) Survival curve showing % survival of recipient mice that received either WT or Irf4 ^+/fl P14 cells over 22 days. Data are a compilation of three independent experiments. (B) Serum was harvested from recipient mice at D15 post infection and virus titers were determined by plaque assay. Each data point represents an individual mouse and data are a compilation of two independent experiments. (C) Dot plots show CD90.1 (transferred cells) versus CD90.2 (host cells) staining on gated live CD8⁺ T cells (left). Graphs show compilations of proportions and numbers of transferred P14 cells at D22 post-infection (right). Each data point represents an individual mouse and data are a compilation of three independent experiments. (D) Dot plots show T-bet versus Eomes staining on gated live CD90.1⁺ CD8⁺ T cells (left). Graphs show the normalized MFI of T-bet and Eomes, each normalized to the average of P14 WT transferred samples and the ratio of the normalized MFIs for T-bet relative to Eomes, (right). Each data point represents an individual mouse and data are a compilation of three independent experiments. Statistical analysis was determined by Log-rank (Mantel-Cox) test (A) or unpaired t test with Welch’s correction (B-D).

Fig 5. High expression of IRF4 is essential for long-term control of LCMV-clone 13.

Kidney (A), livers (B) and sera (C) were harvested from LCMV-clone 13 infected WT, Irf4 ^+/fl and Irf4 ^fl/fl mice between D112-114 post-infection and virus titers were determined by plaque assay. Dotted line indicates the limit of detection. Each data point represents an individual mouse and data are a compilation of three independent experiments. (D) Serum was harvested from infected mice at various timepoints post-infection. Graph indicates the proportion of mice with viral titers above the limit of detection over time. Data are a compilation of three independent experiments; significant differences were determined by Log-rank (Mantel-Cox) test. (E) Anti-LCMV IgG antibody titers in sera at D40 p.i. Each data point represents an individual mouse and data are a compilation of three independent experiments; significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test. (F) Anti-LCMV IgG antibody titers in sera at D112-114 p.i. Data are segregated based on serum viral titers; at left are mice with undetectable virus in their sera (cleared) and at right are mice with persistent serum virus titers (persistent). Each data point represents an individual mouse and data are a compilation of three independent experiments; significant differences determined by unpaired t test with Welch’s correction (cleared) and Ordinary one-way ANOVA using Tukey’s multiple comparison test (persistent).

Fig 6. The proportions of T-bet⁺ Eomes^- cells correlate with viral control at late timepoints of LCMV-clone 13 infection.

(A, D) Splenocytes from LCMV-clone 13-infected WT, Irf4 ^+/fl and Irf4 ^fl/fl mice were stained with a viability dye, LCMV-specific H2-D^b-GP276 and H2D^b-GP33 tetramers, and antibodies to CD8, T-bet and Eomes, and analyzed between D112-114 p.i. Dot plots show CD8 versus H2-D^b-GP276 (A) or CD8 vs H2D^b-GP33 (D) tetramer staining on live CD8⁺ T cells. Graphs show compilations of proportions and numbers of tetramer-specific cells from D112-114 post infection. (B, E) Graphs show compilations of the numbers and proportions of T-bet⁺ Eomes^- H2-D^b-GP276 (B) or H2-D^b-GP33 (E) specific cells. (C, F) Graphs show compilations of the numbers and proportions of T-bet^hi PD-1^lo H2-D^b-GP276 (C) or H2-D^b-GP33 (F) specific cells. Symbols outlined in bold represent mice with undetectable titers of virus in sera at D112-114 p.i. Each data point represents an individual mouse and data are a compilation of three independent experiments; significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test. $ denotes statistically significant difference between WT and Irf4 ^+/fl samples when analyzing only mice with undetectable serum viral titers (bold outlined symbols). Significant differences between outlined samples were determined by unpaired t test with Welch’s correction.

Fig 7. Irf4-Eomes compound haplodeficiency restores the T-bet to Eomes ratios in virus-specific CD8⁺ T cells.

Splenocytes from LCMV-clone 13 infected WT, Irf4 ^+/fl, Eomes^+/fl and Irf4 ^+/flEomes^+/fl mice were stained with a viability dye, LCMV-specific H2-D^b-GP276 and H2D^b-GP33 tetramers, and antibodies to CD8, T-bet and Eomes and analyzed between D77-82 p.i. (A, C) Dot plots show CD8 vs H2-D^b-GP276 (A) or CD8 vs H2D^b-GP33 (C) tetramer staining on live CD8⁺ T cells. Graphs show compilations of proportions and numbers of virus-specific CD8⁺ T cells. (B, D) Dot plots show T-bet vs Eomes staining of H2-D^b-GP276 (B) or H2-D^b-GP33 (D) specific cells. Graphs show the MFI of T-bet and Eomes each normalized to average value for WT samples in each experiment, and the ratio of normalized MFIs for T-bet relative to Eomes, and the compilation of proportions of T-bet⁺ Eomes^- cells for each virus-specific subset. Each data point represents an individual mouse and data are a compilation of three independent experiments; significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test

Fig 8. Irf4-Eomes compound haplodeficiency restores virus control during persistent LCMV-clone 13 infection.

Kidney (A), livers (B) and sera (C) were harvested from LCMV-clone 13 infected WT, Irf4 ^+/fl, Eomes^+/fl and Irf ^+/flEomes^+/fl mice between D77-82 post infection and virus titers were determined by plaque assay. Dotted lines indicate the limit of detection. Each data point represents an individual mouse and data are compilations of two independent experiments. (D) Serum was harvested from infected mice at various timepoints post infection. Graph indicates the proportion of mice with viral titers above the limit of detection in serum over time. Each data point represents an individual mouse and data are compilation of two independent experiments; significant differences as determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test (A) and Log-rank (Mantel-Cox) test (D). E) Anti-LCMV IgG antibody titers in sera at D40 and D78-82 p.i. were determined. For D40 timepoint, data points corresponding to 3 WT and 4 Irf4 ^+/fl mice were previously shown in Fig 5E. Each data point represents an individual mouse and data are a compilation of two independent experiments; significant differences determined by Ordinary one-way ANOVA using Tukey’s multiple comparison test.