Macrophage-Derived CXCL9 and CXCL10 Are Required for Antitumor Immune Responses Following Immune Checkpoint Blockade

CXCR3 blockade abrogates therapeutic efficacy of dual PD-1 and CTLA-4 blockade. A, C57BL/6 mice were injected subcutaneously with 5 × 10⁵ AT-3ova cells and allowed to establish for 14 days before dual treatment with anti-PD-1 (200 μg/mouse) and anti-CTLA-4 (150 μg/mouse) ± anti-CXCR3 (200 μg/mouse). CXCR3 blockade was given on days 13, 14, 18, 22, and 26. Tumor growth (B) and survival (C) was assessed. Data shown as the mean ± of 6–8 mice/group of a representative experiment of n = 3. Survival is shown as combined data of two independent experiments. **, P < 0.005; *, P < 0.05 (determined by two-way ANOVA and Mantel–Cox test). Data points for ISO and P+C treated are included in Fig. 1B. D and E, Representative sizes (D) and weights (E) of tumors harvested 7 days posttreatment. A representative experiment of n > 3 is shown. F–H, C57BL/6 mice were injected with 3 × 10⁵ MC38 cells (s.c.) and allowed to establish for 12 days (n = 6 mice per group). A total of 1.5 × 10⁵ E0771 cells (intra mammary fat pad) and allowed to establish for 3 days (n = 5–8 per group of representative experiment n = 2; G) or 2 × 10⁵ B16F10-OVA^dim cells were injected (s.c.; H) and allowed to establish for 4 days (n = 16–17 mice per group) before dual treatment with anti-PD-1/anti-CTLA-4 ± anti-CXCR3, as per B, except for MC38, which received only two doses of combined therapy.

CXCR3 blockade inhibits dual PD-1 and CTLA-4 blockade-induced CD8⁺ T-cell migration and activation. C57BL/6 mice were injected subcutaneously with 5 × 10⁵ AT-3ova cells and allowed to establish for 14 days before dual treatment with anti-PD-1 and anti-CTLA-4 ± anti-CXCR3 as per Fig. 2. On day 21 (7 days posttreatment) TIL infiltrate was analyzed as a percentage of CD45⁺ cells; CD8⁺ T cells (A), CD4⁺foxp3⁻ (B), NK1.1⁺ cells (C), and CD4⁺foxp3⁺ cells (D). E, T-cell infiltrate quantified by immunofluorescence staining. Shown are representative images and quantification of three tumors per group. F, IFNγ and TNFα expression in CD8⁺ and CD4⁺foxp3⁻ T cells was assessed. Shown are concatenated data from 6 mice (left) and quantification from two independent experiments (right). G, Ki67, PD-1, and TIM3 expression was determined in CD8⁺ and CD4⁺foxp3⁻ T cells. A–D, F, and G, Data represent combined data of two independent experiments with n = 11–15 per group. H, Where indicated, mice were dosed with either the S1PR1 inhibitor FTY720 (25 mg/kg) or DMSO control on days 13 and 14 post tumor inoculation, and every 2 days subsequently. Right, numbers of CD8⁺ and CD4⁺ T cells collected from blood three days post initial FTY720 administration. Data are represented as the mean ± SEM of 6–7 mice per group (****, P < 0.0001; ***, P < 0.001; **, P < 0.01 one-way; n.s., not significant, one-way ANOVA or two-way ANOVA).

CXCR3 and its ligands modulate T-cell priming within the tumor DLNs. DLNs were harvested from AT-3ova tumor bearing C57BL/6 mice following treatment with dual anti-PD-1 and anti-CTLA-4 ± anti-CXCR3 as per Fig. 2. A, Indicated immune cell populations were isolated using FACS and CXCL9, CXCL10, and CXCR3 expression was assessed by qRT-PCR. Expression is shown relative to the Rpl27 housekeeping gene and is representative of triplicate qRT-PCR reactions pooled from n = 4 mice per group. B–F, CD8⁺ and CD4⁺foxp3⁻ T cells were assessed by flow cytometry for expression of CXCR3 (B), IFNγ (C), Ki67 (D), Tbet (E), and PD-1 [F;*, P < 0.01; **, P < 0.01; n.s., not significant, one-way ANOVA (C–F) or unpaired t test (B)]. Data are representative of n = 12 per group from two pooled experiments.

Macrophages are the major source of intratumoral CXCL9 and CXCL10 following dual combination therapy. C57BL/6 mice were injected subcutaneously with either 5 × 10⁵ control AT-3ova (Mock KO) or CXCL9/CXCL10 double KO AT-3ova cells and allowed to establish for 14 days (A). Mice were then treated with either dual anti-PD-1 (200 μg/mouse) and anti-CTLA-4 (150 μg/mouse) or isotype control (200 μg/mouse). Treatment was repeated on days 18, 22, and 26. On day 21, tumors were collected from mice and CXCL9 and CXCL10 production assessed by chemokine bead array as per Fig. 1A and B. Data shown are the mean ± SEM of 5–6 mice per group, of a representative experiment (n = 2). *, P < 0.05, n.s., not significant assessed by one-way ANOVA. C, D, and F, Intracellular CXCL9 staining was performed ex vivo in the presence of golgi plug/golgi stop for 3 hours at 37°C. n = 3 mice per group, two representative experiments. C, CXCL9 production per cell type before and after therapy. D, CXCL9 staining in indicated populations shown as representative histograms. E, Tissue enrichment scores for the top 200 genes upregulated in tumors treated with anti-PD-1 and anti-CTLA-4 relative to isotype control treated mice. F, Pie charts indicating the source of CXCL9 by cell type. Each pie represents one mouse from a representative experiment. G, C57BL/76 mice were injected with MC38 subcutaneously and 19 days postinoculation expression of CXCL9 by tumor-infiltrating myeloid cells was determined. H, Expression of Cxcl9, Cxcl10, Adgrel (F4/80), Fcgr1 (CD64), and Itgae (CD103) in CT26 tumor-infiltrating immune cells as determined by single-cell RNA-seq (35). UMAP embedding of single cells in macrophage cluster (larger population) and fibroblast cluster (smaller population bottom left) identified as per the original study are shown, with color intensity representing normalized gene expression level.

IFNγ is sufficient to induce CXCL9 expression from intratumoral macrophages. A, Correlation of IFNγ or TNFα and CXCL9/CXCL10 production as determined by cytometric bead array of tumor samples obtained as per Fig. 1. Data is represented as n = 6 mice per group from two independent experiments. R² values determined by linear regression analysis. B, IFNγ and TNFα gene signature enrichment scores for CXCL9- and CXCL10-high tumors in patients with urothelial cancer treated with atezolizumab (ImVigor210 trial). C, AT-3ova tumor-bearing mice were treated as per Fig. 5A with the additional treatment of anti-IFNγ or anti-TNFα (250 μg/mouse) on days 14 and 18. Expression of CXCL9 in MoMacs and CD103⁺ dendritic cells determined in n = 6 mice per group. D and E, AT-3ova tumor-bearing C57BL/6 wild-type or BATF3^−/− mice were treated as per Fig. 5A. Seven days posttreatment, tumor-infiltrating lymphocytes were dissociated and stimulated for 3 hours with 1 ng/mL IFNγ and/or TNFα in the presence of Golgi Plug/Stop. Expression of CXCL9 in Mo/Macs or CD103⁺ DCs was determined by flow cytometry (*, P < 0.05; **, P < 0.01; ***, P < 0.001, not significant, one-way ANOVA).