Regulatory T Cells Restrain Interleukin-2- and Blimp-1-Dependent Acquisition of Cytotoxic Function by CD4+ T Cells

Abstract

In addition to helper and regulatory potential, CD4⁺ T cells also acquire cytotoxic activity marked by granzyme B (GzmB) expression and the ability to promote rejection of established tumors. Here, we examined the molecular and cellular mechanisms underpinning the differentiation of cytotoxic CD4⁺ T cells following immunotherapy. CD4⁺ transfer into lymphodepleted animals or regulatory T (Treg) cell depletion promoted GzmB expression by tumor-infiltrating CD4⁺, and this was prevented by interleukin-2 (IL-2) neutralization. Transcriptional analysis revealed a polyfunctional helper and cytotoxic phenotype characterized by the expression of the transcription factors T-bet and Blimp-1. While T-bet ablation restricted interferon-γ (IFN-γ) production, loss of Blimp-1 prevented GzmB expression in response to IL-2, suggesting two independent programs required for polyfunctionality of tumor-reactive CD4⁺ T cells. Our findings underscore the role of Treg cells, IL-2, and Blimp-1 in controlling the differentiation of cytotoxic CD4⁺ T cells and offer a pathway to enhancement of anti-tumor activity through their manipulation.

Keywords: Blimp-1; CD4-mediated anti-tumor response; IL-2; T-bet; Treg depletion; anti-CTLA-4; cytotoxic CD4(+) T cells; regulatory T cells.

Graphical abstract

Figure 1

Tumor-Reactive CD4⁺ T Cells Acquire Cytotoxic Phenotype following Lymphopenia-Induced Expansion (A–C) B16-tumor-bearing mice were left untreated or treated with Trp1 cells (Trp1 ctrl), Trp1 + GVAX + αCTLA-4, or RT + Trp1 + αCTLA-4 as per Figure S1A. (A) Tumor growth and survival (N = 5/group). (B) T-bet and IFN-γ expression by Trp1 TILs (N = 10–11/group in 2 experiments) and (N = 5–6/group in two experiments), respectively. (C) Representative plot and quantification of GzmB expression by Trp1 TILs (N = 13–17/group in four experiments). (D and E) WT and Ifngr1^−/− mice bearing B16 tumors were left untreated or treated with Trp1 or Trp1xPrf-1^−/− cells alone or in combination with RT + αCTLA-4. (E) Tumor growth and survival (N = 5/group). (F–H) Foxp3⁻ Trp1 cells were sorted from B16 tumors from mice treated with GVAX + αCTLA-4 or RT + αCTLA-4 as per Figure S1D. (G) Total differential gene expression between Th Trp1 cells and Th-ctx Trp1 cells (p ≤ 0.01) and differentially expressed transcription factors between Trp1 Th-ctx cells and Trp1 Th cells (p ≤ 0.01). (H) Reactome pathway enrichment analysis of immune-system-related pathways and cytokine signaling pathways (red) (NES > 2, p < 0.05). Right panel: gene set enrichment analysis of IL-2-dependent genes (Castro et al., 2012; GEO: GSE39110). All quantification plots show mean ± SEM (one-way ANOVA).

Figure 2

IL-2 Drives GzmB Expression in Both Murine and Human CD4⁺ T Cells In Vitro (A and B) OT-II cells were transferred to mice bearing B16-OVA tumors (as per Figure S2A) alone or in combination with RT + αCTLA-4. (A) Representative plots and quantification of GzmB-expressing OT-II TILs (N = 5/group). (B) Representative plots showing expression of GzmB and T-bet by OT-II cells in the Th-ctx condition. (C) OT-II cytotoxicity assay. Representative plots and quantification of specific lysis are shown. (D–F) Cell trace violet (CTV)-labeled OT-II cells were stimulated with indicated concentrations of OT-II peptide to asses GzmB expression within proliferating cells in the following conditions: (D) addition of IL-2 (two independent experiments), (E) addition of indicated cytokines, and (F) addition of 5 μg/ml of indicated antibodies. (G and H) CTV-labeled murine polyclonal CD4⁺ T cells were stimulated with (G) αCD3 and IL-2 to assess expression of GzmB in proliferating CD4⁺ T cells. A representative plot of perforin expression is also shown. (H) αCD3 and αCD28 and indicated antibodies. Representative plots and quantification of GzmB and T-bet-expressing cells (data are representative of two independent experiments). (I and J) CTV-labeled human polyclonal naive CD4⁺ T cells stimulated with αCD3 and αCD28. (I) After 24 h, either IL-2 or αCD25 antibody was added. Quantification of GzmB- and T-bet-expressing CD4⁺ T cells (cumulative data of two independent experiments). (J) Indicated ratios of autologous Treg cells added with or without IL-2. Representative plots and quantification of GzmB -expressing CD4⁺ T cells are shown. All quantification plots show mean ± SEM (one-way ANOVA) (B, two-way ANOVA).

Figure 3

CD4 TILs in the Th-ctx Condition Reduce GzmB Expression but Retain the Th1 Phenotype upon IL-2 Neutralization (A–C) B16-tumor-bearing mice were left untreated (Ctrl) or received RT + Trp1 + aCTLA-4 (Th-ctx) with or without αIL-2 or αIL-7 (see Figure S3A). Quantification of (A) Ki-67- (N = 7–18/group) and IL-2-expressing Trp1 cells (N = 7–13/group), (B) T-bet- and IFN-γ- expressing Trp1 cells (N = 7–13/group in two independent experiments), and (C) GzmB expression in Trp1 (CD45.1⁺) and endogenous CD4⁺ T cells in LN and tumors (N = 7-13/group in three experiments). (D and E) B16-OVA-tumor-bearing mice were treated as shown in Figure S3D. Tumors and dLNs were isolated at day 18 post-tumor inoculation for analysis. (D) Quantification of IFN-γ- (N = 5/group) and (E) GzmB-expressing cells within OT-II cells (N = 8–11/group in two experiments). All quantification plots show mean ± SEM (one-way ANOVA).

Figure 4

Increased IL-2 Availability after Treg Cell Depletion Contributes to Shaping the Th Cell Phenotype within Tumors (A and B) MCA205-tumor-bearing mice were treated with αCTLA-4 on days 6, 9, and 12 alone or with αIL-2, αCD8, and αMHC class II on days 6, 9, 12, and 15 after tumor implantation. (A) Individual tumor growth curves and (B) cumulative survival are shown. (C–G) MCA205-tumor-bearing mice were treated with αCTLA-4, αIL-2, or combination as in (A). TILs and dLNs were isolated 13 days post-tumor inoculation. (C) CD4eff/Treg cells in tumors (N = 10/group in two independent experiments). (D) TILs and dLNs from MCA205-bearing mice were re-stimulated with IL-2. Representative plots of pSTAT5 expression in CD4⁺ T cells and quantification of pSTAT5-expressing CD4eff T cells are shown (N = 5/group in two independent experiments). (E) Expression of GITR, CD69, and PD-1 by CD4eff TILs. Representative plots are shown with mean percentage of expression or mean fluorescence intensity (N = 5–10/group; two independent experiments). (F) Quantification of GzmB-expressing cells within CD4eff TILs and Treg TILs (N = 10/group in two independent experiments. (G) Quantification of GzmB-expressing cells within CD8 TILs and expression of Eomes and T-bet by CD4eff and CD8 TILs. Representative plots are shown with mean percentage of expression or mean fluorescent intensity (N = 10/group; two independent experiments). (H) dLN-infiltrating CD4⁺ T cells were cultured unstimulated or stimulated with MCA205-pulsed dendritic cells (DCs) or empty (np) DCs on αGzmB-coated ELISPOT plate for 24 h. Numbers represent GzmB spots per 10,000 responding CD4⁺ T cells. Graphical representation and quantification are shown. (I) Immunohistochemical analysis of GzmB expression by CD4⁺ T cells in human melanoma. Representative plots pre- and post-therapy are shown, with CD4 staining in brown, FOXP3 in green, and GZMB in blue. Quantification of CD4⁺GZMB⁺ cells within tumor pre- and post-treatment and ratio of CD4eff to CD4⁺FOXP3⁺ cells are shown (n = 10 patients, Wilcoxon matched-pairs signed rank test, one tailed). All other quantification plots show mean ± SEM (one-way ANOVA).

Figure 5

Treg Cell Depletion in the Absence of CTLA-4 Blockade Drives GzmB Expression by CD4⁺ T Cells (A–E) MCA205-bearing Foxp3^DTR mice were treated with DT alone or in combination with αIL-2 (schema in Figure S5F) from day 6 post-tumor inoculation. (A) Schema and quantification of Treg cells within CD4 T cells in dLNs and TILs (N = 10/ group in two experiments). Expression of (B) Ki67 (C) T-bet by CD4 TILs (representative plots and quantification). (D) Quantification of CD4⁺ IFN-γ- and GM-CSF-expressing cells (N = 10/group in two experiments). (E) Quantification of CD4⁺ GzmB-expressing cells within dLNs and TILs (N = 10/group in two experiments). All quantification plots show mean ± SEM (one-way ANOVA).

Figure 6

T-bet Is Not Required for CTLA-4-Mediated Rejection of MCA205 Sarcoma (A–C) WT and Tbx21^−/− MCA205-tumor-bearing mice were treated with αCTLA-4, αIL-2, or their combination. TILs and dLNs were isolated 13 days post-tumor inoculation for analysis of (A) Treg cells (N = 7–9/group in two independent experiments) and IFN-γ (N = 4–5/group) within CD4⁺ TILs and (B) GzmB-expressing cells within CD4eff and CD8⁺ TILs (N = 7–9/group in two independent experiments). (C) T-bet and Eomes expression by GzmB⁺ CD4eff and CD8⁺ TILs (N = 7–9/group in two experiments). (D) Tumor growth and survival in WT and Tbx21^−/− mice bearing MCA205 tumors and treated with αCTLA-4 alone or combined with depleting αCD8 or αCD4 antibodies on days 1, 3, 8, and 17 post-tumor implantation. (E) qRT-PCR for transcription factors in purified MCA205 CD4⁺Fopx3⁻ TILs and LNs at day 12 post-tumor inoculation (untreated versus αCTLA-4 treated mice). Results shown are expression relative to Hprt1 expression using the 2^−ΔΔC(t) method (N = 6/condition). All quantification plots show mean ± SEM (one-way ANOVA).

Figure 7

IL-2 Controls Cytotoxic CD4⁺ T Cell Differentiation in a Blimp-1-Dependent Manner (A) Prdm1^fl/fl and Prdm1^−/−Cd4^cre mice bearing MCA205 tumors were treated or not with αCTLA-4 and monitored for tumor growth and survival. (B and C) TILs and dLNs were isolated at day 12 post-tumor inoculation for quantification of (B) GzmB-expressing cells within CD4eff and CD8 TILs and (C) T-bet-expressing cells within CD4eff and CD8 TILs (N = 9–11/group in two experiments). (D) Purified CD4⁺ T cells from dLNs and tumors from MCA205-bearing Prdm1^fl/fl and Prdm1^−/−Cd4^cre mice were transferred to MCA205-bearing Rag1^−/− mice at day 3 post-inoculation followed by αCTLA-4. Overall survival is shown (N = 5/group). (E) Quantification of CD25- and IL-2-stimulated pSTAT5-expressing CD4eff TILs (N = 5–11 group in two experiments). (F) Prdm1^fl/fl and Prdm1^−/−Cd4^cre MCA205 tumor-bearing mice were treated with αCTLA-4 alone or in combination with high-dose intratumoral IL-2. TILs and dLNs were isolated at day 12 post-tumor inoculation for quantification of GzmB- and Prf-1-expressing cells within CD4⁺ TILs (N = 5/group). (G) Purified CD4⁺ T cells from WT (Prdm1^fl/fl) and Prdm1^−/−Cd4^cre mice were transduced with Trp1 TCR-expressing vector and transferred to B16-bearing mice alone or in combination with αCTLA-4 + RT. (H) Representative plots and quantification of GzmB- and IFN-γ-expressing cells within Trp1 effector TILs (N = 6/group in two independent experiments). (I) Transduced WT and Blimp-1-deficient Trp1 cells co-transferred 1:1 to the same host. Representative plots and quantification of GzmB-and Prf-1-expressing cells within Trp1 TILs are shown (N = 5/group). (J) Transduced cells as in (H) were transferred at day 8 to B16-bearing WT or Ifngr^−/− mice alone or in combination with αCTLA-4 + RT. Cumulative tumor growth and survival are shown (N = 5/group). All quantification plots show mean ± SEM (one-way ANOVA) (J, two-way ANOVA).