Acquisition of suppressive function by conventional T cells limits antitumor immunity upon Treg depletion

Abstract

Regulatory T (T_reg) cells contribute to immune homeostasis but suppress immune responses to cancer. Strategies to disrupt T_reg cell-mediated cancer immunosuppression have been met with limited clinical success, but the underlying mechanisms for treatment failure are poorly understood. By modeling T_reg cell-targeted immunotherapy in mice, we find that CD4⁺ Foxp3^- conventional T (T_conv) cells acquire suppressive function upon depletion of Foxp3⁺ T_reg cells, limiting therapeutic efficacy. Foxp3^- T_conv cells within tumors adopt a T_reg cell-like transcriptional profile upon ablation of T_reg cells and acquire the ability to suppress T cell activation and proliferation ex vivo. Suppressive activity is enriched among CD4⁺ T_conv cells marked by expression of C-C motif receptor 8 (CCR8), which are found in mouse and human tumors. Upon T_reg cell depletion, CCR8⁺ T_conv cells undergo systemic and intratumoral activation and expansion, and mediate IL-10-dependent suppression of antitumor immunity. Consequently, conditional deletion of Il10 within T cells augments antitumor immunity upon T_reg cell depletion in mice, and antibody blockade of IL-10 signaling synergizes with T_reg cell depletion to overcome treatment resistance. These findings reveal a secondary layer of immunosuppression by T_conv cells released upon therapeutic T_reg cell depletion and suggest that broader consideration of suppressive function within the T cell lineage is required for development of effective T_reg cell-targeted therapies.

Figure 1. T_reg cell ablation causes T_conv cells to acquire transcriptional features of T_reg cells.

(A) Tumor growth of B16-F10 tumors subcutaneously implanted into Foxp3^EGFP-DTR mice. Gray shading indicates time period over which PBS or DTx was administered (day 7-13 post-implantation, early disease; day 10-16 post-implantation, established). Dashed lines indicate individual mice. Solid line indicates average tumor area over time. Data representative of 4 individually repeated experiments, n > 5 **P < 0.01; two-tailed Mann–Whitney U-test. (B) Representative frequency of Foxp3⁺ T_reg cells among total CD4⁺ T cells within spleens (top) or tumors (bottom) of Foxp3^EGFP-DTR mice with established tumors administered PBS or DTx. (C) Heatmap showing the relative expression of transcripts upregulated in intratumoral T_reg cells compared with CD4⁺ T_conv cells (q<0.05; FC>4) in the indicated T cell subsets isolated at day 18 after implantation of B16-F10 tumors in Foxp3^EGFP-DTR animals administered PBS or DTx. Colors indicate expression normalized to row maxima. x-axis hierarchical clustering of intratumoral T_reg cell-expressed transcripts identifies 5 clusters of genes with distinct expression patterns. Gray bars to right of heatmap indicate expression greater than a third of the expression of given transcripts in intratumoral T_reg cells. (D) Average expression of genes within the 5 clusters identified in each T cell subset. (E) Heatmap showing pairwise Pearson distances between the global gene expression profiles of the indicated T cell subsets from B16-F10 tumor-bearing Foxp3^EGFP-DTR animals administered PBS or DTx. (F) Scatterplot comparing the global differences in gene expression between intratumoral T_reg cells and T_conv cells with transcriptional differences between CD4⁺ T_conv cells isolated from DTx versus PBS-treated animals. A highly significant correlation is observed indicating transcriptional convergence of intratumoral T_reg cells with T_conv cells in absence of T_reg cells. Data from 2-4 biological replicates isolated on independent days (C-F).

Figure 2. T_reg cell ablation promotes induction of suppressive function by CD4⁺ T_conv cells.

(A) Experimental schema. B16-F10 cells were subcutaneously implanted into Foxp3^EGFP-DTR CD45.2⁺ mice and administered PBS or DTx on day 7, 9, 11, and 13. Cells were harvested from tumors of PBS- and DTx-treated animals at day 16 post-implantation and used in suppression assays. (B and C) In vitro suppression assay. Proliferation of naïve CD45.1⁺ CD4⁺ T_conv responder (T_resp) cells 4 days cultured at a 4:1 ratio with indicated suppressor cell populations (Foxp3^EGFP− T_conv cells from tumors of T_reg-replete (PBS) or T_reg-depleted (DTx) mice, or GFP⁺ T_reg cells from tumors of T_reg-replete mice). Representative histograms and replicate measurements of proliferation dye dilution at day 4 post-stimulation gated on CD45.1⁺ T_resp cells shown. T_resp cell proliferation in the absence of a suppressor cell population was used as a control. Suppressor cells were co-cultured with responder T (Tresp) cells at a ratio of 1:4, with 2.5×10⁴ suppressor CD4⁺ T_reg or T_conv cells co-cultured with 1×10⁵ Tresp cells in the presence of 5.0×10⁴ antigen-presenting cells (APC). Data are representative of > 4 independently repeated experiments, n > 5 per group; ordinary one-way ANOVA, Tukey’s multiple comparisons. (D and E) Representative histograms and replicate measurements of CD44 expression by CD45.1⁺ T_resp cells incubated with indicated suppressor cell populations. (F-G) Representative flow cytometry and replicate measurements of the expression of the indicated proteins by intratumoral T_reg cells and CD4⁺ T_conv cells isolated at day 16 after implantation of B16-F10 tumors in Foxp3^EGFP-DTR mice treated with PBS or DTx. Data are representative of > 3 independently repeated experiments, n > 7 per group. ****P <0.0001; ns, not significant; one-way ANOVA Kruskal-Wallis, Dunn’s multiple comparisons test. Error bars show standard error of the mean (s.e.m.)

Figure 3. T_reg cell ablation promotes the expansion of tumor-infiltrating CCR8⁺ T_conv cells.

(A) Uniform manifold approximation and projection (UMAP) of scRNA-Seq of TCRβ⁺ cells isolated at day 16 after implantation of B16-F10 tumors in Foxp3^EGFP-DTR mice treated with PBS or DTx on days 7, 9, 11 and 13. (B) Density plots showing change in distribution of cells within tumors of T_reg-replete (PBS) and T_reg-depleted (DTx) animals. (C) Relative frequency of cells within each cluster normalized to their average ratio among PBS animals. N = 3 biological replicates per group. (D) Average enrichment of expression of the genes in Cluster D from Fig. 1C across scRNA-Seq clusters, (n=3, unpaired two-tailed Student’s t-test, *P < 0.05, **P < 0.01). (E) Heatmap showing the expression of differentially upregulated genes in each cluster identified. (F) UMAP plots showing expression of indicated genes within T cells of tumors from tumor-bearing PBS- or DTx-treated Foxp3^EGFP-DTR animals. (G) Representative flow cytometry and (H) replicate measurements of total counts (top) and the frequency (bottom) of CCR8⁺ of CD4⁺ T_conv cells from spleens, draining lymph nodes (dLN) and tumors of B16-F10 tumor-bearing Foxp3^EGFP-DTR mice treated with PBS or DTx. Data are representative of 3 independently repeated experiments (G and H). Numbers in gates show percentages. n > 10. one-way ANOVA Kruskal-Wallis, Dunn’s multiple comparisons test. **P <0.01, ***P < 0.001, ****P < 0.0001; Error bars show standard error of the mean (s.e.m.)

Figure 4. CCR8 marks highly suppressive T_conv cells.

(A) Heatmap showing the relative expression of differentially expressed genes between intratumoral CCR8⁺ and CCR8⁻ CD4⁺ T_conv cells (q<0.05; |FC|>3) isolated at day 16 after subcutaneous implantation of B16-F10 tumors of T_reg-depleted Foxp3^EGFP-DTR animals treated with DTx at days 7, 9, 11 and 13. Data from 4 biological replicates isolated on the same day. (B) Gene-set enrichment analysis (GSEA) demonstrating a negative enrichment of genes upregulated in Foxp3⁻ T_conv cells vs Foxp3⁺ T_reg cells among CCR8⁺ T_conv cells compared with CCR8⁻ T_conv cells isolated from tumors of DTx-treated Foxp3^EGFP-DTR mice. (C) Scatterplot comparing global changes in gene expression between intratumoral T_reg and T_conv cells with transcriptional differences between CCR8⁺ and CCR8⁻ CD4⁺ T_conv cells. (D) Representative flow cytometry and (E) replicate measurements of the expression of the indicated proteins from intratumoral T_reg cells and CCR8⁺ and CCR8⁻ CD4⁺ T_conv cells from tumors of B16-F10 tumor-bearing Foxp3^EGFP-DTR mice treated with PBS or DTx. Data are representative of 3 independently repeated experiments n > 4 one-way ANOVA Kruskal-Wallis, Dunn’s multiple comparisons test. *P < 0.05, **P <0.01, ***P < 0.001, ****P < 0.0001, ns, not significant. (F) Representative histograms and (G) replicate measurements of T_resp cells (naïve CD45.1⁺ CD4⁺ T_conv cells) incubated with intratumoral CD45.2⁺ TCRβ⁺ CD4⁺ GFP⁻ CCR8⁻ or CD45.2⁺ TCRβ⁺ CD4⁺ GFP⁻ CCR8⁺ suppressor T_conv cells isolated at day 16 after implantation of B16-F10 tumors in Foxp3^EGFP-DTR mice. Suppressor cells were incubated with responders at a ratio of 1:8, with 1.25×10⁴ suppressor CD4⁺ T_conv cells co-cultured with 1×10⁵ Tresp cells in the presence of 5×10⁴ APC. Cell proliferation of T_resp cells was analyzed after 4 days. Data are representative of 2 independently repeated experiments n > 3. ordinary one-way ANOVA, Tukey’s multiple comparisons. *P < 0.05; ***P < 0.001, ****P < 0.0001; Error bars show standard error of the mean (s.e.m.)

Figure 5. CCR8⁺ T_conv cells expressing high levels of CD25 are found within the tumors of NSCLC patients.

(A) CCR8 and CD25 expression among FOXP3⁻ CD4⁺ T cells from representative samples from NSCLC patients (n=48). (B) Frequency of FOXP3⁻ CCR8⁺ CD25^bright cells among CD4⁺ T cells from the indicated patients’ samples. Lines indicate paired samples. (C) Correlation of the frequency of CD8⁺ T cells (of CD3⁺ T cells) with FOXP3⁻ CCR8⁺ CD25^bright cells (of CD4⁺ T cells) in tumors from patient samples. (D) UMAP analysis of concatenated CD4⁺ FOXP3⁻ T_conv cells. Colors depict cell clusters identified by Phenograph (k=500). Separate UMAP plots of relative marker expression by concatenated CD4⁺ T cells from tumors. (E) Representative frequency and (F) replicate measurements of indicated markers from patient samples, box plots show median and interquartile range (IQR). Bars indicate standard deviation. Dots depict values of a single tumor sample. **P < 0.01, ***P < 0.001, ****P < 0.0001; two-tailed Mann–Whitney U-test.

Figure 6. IL-10 production by CD4⁺ T_conv cells limits anti-tumor immunity upon T_reg cell depletion.

(A) Screen to identify mechanisms of suppression by CD4⁺ T_conv cells from tumors of T_reg-depleted animals. Proliferation of CTV-labeled naïve splenic CD4⁺ T_conv responder cells (T_resp) cultured alone or at a ratio of 8:1 with CD4⁺ Foxp3^EGFP− T_conv suppressor cells (T_supp) isolated at day 16 from B16-F10 tumors of DTx-treated Foxp3^EGFP-DTR animals. Cells were cultured alone (gray) or with suppressors (purple) and with indicated reagents. AG, aminoglutethimide. CD45.2⁺ GFP⁻ T_conv suppressor cells were co-cultured with 1.25×10⁴ suppressor CD4⁺ T_conv cells in the presence of 5.0×10⁴ antigen-presenting cells (APC). Data are representative of 2 independently repeated experiments n > 3. P values show significance of difference between no suppressor and suppressor (Student t test) and are Bonferroni corrected. (B) Representative frequency of dividing T_resp cells incubated with tumor CD4⁺ GFP⁻ T_conv cells in the presence of anti-IL-10R antibodies or vehicle. T_resp cells without tumor T_conv cells were used as a control. (C) Co-correlation between the expression of indicated genes within single cell gene expression profiles of T cells from tumors of PBS- or DTx-treated B16 tumor-bearing Foxp3^EGFP-DTR animals. Pearson correlation co-efficient values are indicated by color scale and genes are hierarchically clustered to identify clusters of co-expressed transcripts within T cell populations. scRNA-Seq data are representative of 3 biological replicates per group. (D) Measurement of Ccr8 and Il10 mRNA expression within CCR8⁻ and CCR8⁺ T_conv cells from PBS- and DTx-treated animals. Data representative of 3-4 biological replicates per group. ordinary one-way ANOVA, Tukey’s multiple comparisons. ns, not significant. (E) Tumor area of heterotopic B16-F10 melanoma tumors at indicated time-points following implantation into Il10^flox/flox Cd4^Cre Foxp3^EGFP-DTR or Il10^+/+ Cd4^Cre Foxp3^EGFP-DTR control mice administered DTx or PBS from day 10-16 post-implantation. Data are representative of 2 independently repeated experiments, n > 7. ordinary one-way ANOVA, Tukey’s multiple comparisons. (F) Representative histograms (top) and replicate measurements (bottom) of the frequency of CD8⁺ CD44⁺ T cells and Foxp3⁻ CD4⁺ CD44⁺ T_conv cells from tumors of animals within indicated treatment groups. (G) Representative frequency and (H) replicate measurements of the frequency (top) and total counts (bottom) of CD8⁺ IFN-γ⁺ TNF⁺ T cells within tumors. Data are representative of 2 independently repeated experiments, n > 4, one-way ANOVA, Kruskal-Wallis, Dunn’s multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001, ****P < 0.0001. Error bars show standard error of the mean (s.e.m.)

Figure 7. Blockade of IL-10 signaling synergizes with T_reg cell-depletion to drive potent anti-tumor immune responses.

(A) Tumor area of B16-F10 melanoma tumors at indicated time-points following implantation into Foxp3^EGFP-DTR animals administered indicated combinations of DTx and anti-IL-10R or control reagents from days 10-16 after tumor implantation. Data are representative of 2 independently repeated experiments. n > 10, ordinary one-way ANOVA, Tukey’s multiple comparisons. *P < 0.05 **P < 0.01; (B) Representative frequency and (C) replicate measurements of the frequency and total counts of CD8⁺ IFN-γ⁺ TNF⁺ T cells from tumors. n > 3, one-way ANOVA, Kruskal-Wallis, Dunn’s multiple comparisons test., *P < 0.05. (D) Representative frequency and (E) replicate measurements of the frequency and total counts of Foxp3⁻ CD4⁺ IFN-γ⁺ TNF⁺ T cells from tumors. Data from > 2 independent biological replicates. n > 3, Kruskal-Wallis, Dunn’s multiple comparisons test., *P < 0.05; Error bars show standard error of the mean (s.e.m.)