Programmable bacteria synergize with PD-1 blockade to overcome cancer cell-intrinsic immune resistance mechanisms

Abstract

Interferon-γ (IFN-γ) is a potent cytokine critical for response to immunotherapy, yet conventional methods to systemically deliver this cytokine have been hindered by severe dose-limiting toxicities. Here, we engineered a strain of probiotic bacteria that home to tumors and locally release IFN-γ. A single intratumoral injection of these IFN-γ-producing bacteria was sufficient to drive systemic tumor antigen-specific antitumor immunity, without observable toxicity. Although cancer cells use various resistance mechanisms to evade immune responses, bacteria-derived IFN-γ overcame primary resistance to programmed cell death 1 (PD-1) blockade via activation of cytotoxic Foxp3^-CD4⁺ and CD8⁺ T cells. Moreover, by activating natural killer (NK) cells, bacteria-derived IFN-γ also overcame acquired resistance mechanisms to PD-1 blockade, specifically loss-of-function mutations in IFN-γ signaling and antigen presentation pathways. Collectively, these results demonstrate the promise of combining IFN-γ-producing bacteria with PD-1 blockade as a therapeutic strategy for overcoming immunotherapy-resistant, locally advanced, and metastatic disease.

Fig. 1.. Generation and characterization of SLIC–IFN-γ.

(A) Schematic showing that SLIC–IFN-γ drives production and release of IFN-γ when reaching quorum. Briefly, when bacterial growth reaches a critical population density, the genomically integrated SLC triggers population-wide lysis events and subsequently the release of genetically encode payloads. (B) Quantification of IFN-γ in culture lysates of SLIC or SLIC–IFN-γ. Data are representative of three independent experiments (***P = 0.0002, unpaired Student’s t test). (C) Quantification of NO₂⁻ production by BMDMs in response to the indicated lysate conditions with or without rIFN-γ. Data are representative of three independent experimental replicates (****P < 0.0001, one-way ANOVA with Holm-Šidák post hoc test). (D to G) Flow cytometric analysis of MHC-I and PD-L1 up-regulation on the surface of MC38-GFP cells after coincubation with the indicated conditions. (D and E) Histogram and quantification of PD-L1 on MC38-GFP cells. Data are representative of three independent experimental replicates (****P < 0.0001, one-way ANOVA with Holm-Šidák post hoc test). (F and G) Histogram and quantification of MHC-I on MC38-GFP cells. Data are representative of three independent experimental replicates (****P < 0.0001, one-way ANOVA with Holm-Šidák post hoc test). ns, not significant; MFI, mean fluorescence intensity.

Fig. 2.. A single injection of SLIC–IFN-γ promotes antitumor immunity.

(A) Tumor growth curves from mice (n ≥ 6 tumors per group; n = 5 mice per group) implanted with MC38 cells on both hind flanks. Data are representative of three independent experiments (****P < 0.0001, two-way ANOVA). (B) Tumor growth curves from mice (n ≥ 9 tumors per group; n = 5 mice per group) implanted with MC38 cells on both hind flanks. Data are representative of three independent experiments (***P < 0.001, two-way ANOVA). (C and D) ELISA quantification of IFN-γ concentration in (C) tumor homogenates and (D) sera of MC38-bearing mice 3 days after a single IT injection. Data are representative of two independent experiments (n ≥ 6 mice per group, **P < 0.01, one-way ANOVA). (E and F) Flow cytometric analysis of lymphocytes isolated from MC38 tumors 9 days after a single IT injection. (E) Histogram and quantification of TNFα⁺ CD8⁺ T cells and (F) quantification of TNFα⁺CD4⁺Foxp3⁻ T cells. Data are representative of four independent experiments (n ≥ 9 tumors per group; *P < 0.05, ***P < 0.001, and ****P < 0.0001, one-way ANOVA). (G) Tumor growth curves from mice (n ≥ 8 tumors per group) implanted with CT26 cells on both hind flanks. Data are representative of two independent experiments (****P < 0.0001, two-way ANOVA). (H) Kaplan-Meier survival curve for CT26 tumor–bearing mice, as in (G) (n ≥ 4 mice per group, *P < 0.05, log rank test for trend). (I to K) Flow cytometric analysis of immune cells isolated from CT26 tumors (n ≥ 3 mice per group) 9 days after a single IT injection. (I) Representative flow plots and quantification of TNFα⁺IFN-γ⁺CD4⁺Foxp3⁻ T cells. (J) Quantification of TNFα⁺IFN-γ⁺CD8⁺ T cells. (K) Ratio of M1 (CD11b⁺F4/80⁺CD40⁺MHCII⁺) over M2 (CD11b⁺F4/80⁺CD206⁺) macrophages. For (I) to (K), data are representative of two independent experiments (n ≥ 3 tumors per group; *P < 0.05, **P < 0.01, and ***P < 0.001, one-way ANOVA). See also figs. S1 to S4. APC, allophycocyanin; PE, phycoerythrin.

Fig. 3.. Treatment with SLIC–IFN-γ generates a systemic adaptive immune response that restrains growth of untreated tumors.

(A) Flow cytometric analysis of tetramer⁺CD8⁺ T cells isolated from MC38-OVA tumors 9 days after a single IT injection with the indicated treatment (n = 5 tumors per group, **P < 0.01, one-way ANOVA). (B) Tumor growth curves from mice (n ≥ 6 tumors per group; n ≥ 6 mice per group) implanted with MC38 cells on both hind flanks. Mice received a single IT injection with the indicated treatment into a single tumor. Data were combined from two independent experiments (***P < 0.001 and ****P < 0.0001, two-way ANOVA). (C) CFU analysis in C57BL/6 mice (n = 4 mice) treated as in (B). (D) Flow cytometric analysis of CD8⁺ T cells isolated from treated and untreated MC38-OVA tumors 9 days after a single IT injection with the indicated treatment into a single tumor. Data are combined from two independent experiments (n ≥ 7 mice per group, *P < 0.05 and **P < 0.01, Kruskal-Wallis test). (E and F) Flow cytometric analysis of DCs isolated from tumor-draining LNs 9 days after a single IT injection into a single tumor, as in (D). Data are representative of two independent experiments (n ≥ 3 mice per group, *P < 0.05, one-way ANOVA). (G) Tumor growth curves from mice (n ≥ 15 tumors per group) implanted with MC38 cells on both hind flanks. Mice received a single intravenous injection with the indicated treatment 12 days after implantation. Data are combined from two independent experiments (***P < 0.001 and ****P < 0.0001, two-way ANOVA). (H) Kaplan-Meier survival curve (n ≥ 9 mice per group, **P < 0.01, log rank test for trend) and (I) respective body weight change for MC38 tumor–bearing mice, as in (G). See also figs. S5 to S8.

Fig. 4.. SLIC–IFN-γ overcomes PD-1 blockade primary resistance mechanisms in the MC38 model.

(A) Tumor growth curves from C57BL/6 mice (n ≥ 12 tumors per group; n ≥ 7 mice per group) subcutaneously implanted with MC38 cells on both hind flanks. Data were combined from two independent experiments (**P < 0.01, ***P < 0.001, and ****P < 0.0001, two-way ANOVA with Holm-Šidák post hoc test at final measurement time point). (B) Respective body weight change (n ≥ 7 mice per group). (C) Tumor growth curves from C57BL/6 mice (n ≥ 9 tumors per group; n = 5 mice per group) subcutaneously implanted with MC38 cells on both hind flanks. Data are representative of two independent experiments (****P < 0.0001, two-way ANOVA with Holm-Šidák post hoc test at final measurement time point). (D) Respective body weight change and (E) Kaplan-Meier survival curve for MC38 tumor–bearing mice (n = 5 mice per group, **P < 0.01, log rank test for trend). See also fig. S9.

Fig. 5.. SLIC–IFN-γ overcomes tumor cell–intrinsic MHC-I deficiency and promotes systemic antitumor immunity.

(A) Tumor growth curves from mice (n ≥ 9 tumors per group) implanted with MC38 B2M KO cells on both hind flanks. Data are representative of two independent experiments (**P < 0.01 and ****P < 0.0001, two-way ANOVA). (B) Kaplan-Meier survival curve for MC38 B2M KO tumor–bearing mice (n ≥ 4 mice per group). (C to E) Flow cytometric analysis of lymphocytes isolated from MC38 B2M KO tumors. Frequencies of (C) TNFα⁺NK cells, (D) IFN-γ⁺CD4⁺Foxp3⁻ T cells, and (E) IFN-γ⁺CD8⁺ T cells. Data are representative of two independent experiments (n ≥ 4 mice per group; *P < 0.05, **P < 0.01, and ***P < 0.001, one-way ANOVA). (F) Tumor growth curves of untreated MC38 tumors from mice (n ≥ 5 tumors per group; n ≥ 5 mice per group) implanted with MC38 B2M KO cells on the left flank and WT MC38 cells on the right. Data are representative of two independent experiments (**P < 0.01 and ***P < 0.001, two-way ANOVA). (G to I) Flow cytometric analysis of lymphocytes isolated from the untreated MC38 tumors, as in (F). Frequencies of (G) TNFα⁺CD4⁺Foxp3⁻ T cells, (H) IFN-γ⁺CD4⁺Foxp3⁻ T cells, and (I) TNFα⁺CD8⁺ T cells. Data are representative of two independent experiments (n ≥ 4 mice per group, *P < 0.05 and **P < 0.01, one-way ANOVA). (J and K) Tumor growth curves of (J) treated MC38 B2M KO tumors and (K) untreated MC38 WT tumors from from mice (n ≥ 3 tumors per group; n ≥ 3 mice per group) implanted with MC38 B2M KO cells on the left flank and MC38 WT cells on the right. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, two-way ANOVA. (L) Tumor growth curves of untreated MC38 B2M KO tumors from mice (n ≥ 7 tumors per group; n ≥ 7 mice per group) implanted with MC38 WT cells on the left flank and MC38 B2M KO cells on the right. Data are representative of two independent experiments. See also figs. S10 to S12.

Fig. 6.. SLIC–IFN-γ overcomes genetic resistance mechanisms to PD-1 blockade in MC38 JAK1 and JAK2 KO models.

(A) Tumor growth curves from C57BL/6 mice (n ≥ 8 tumors per group; n ≥ 5 mice per group) subcutaneously implanted with MC38 B2M KO cells on both hind flanks. Data are representative of two independent experiments (**P < 0.01, ***P < 0.001, and ****P < 0.0001, two-way ANOVA). (B) Kaplan-Meier survival curve for MC38 B2M KO tumor–bearing mice (n ≥ 5 mice per group, ****P < 0.0001, log rank test for trend). (C) Tumor growth curves from C57BL/6 mice (n ≥ 6 tumors per group; n ≥ 3 mice per group) subcutaneously implanted with MC38 JAK1 KO cells on both hind flanks. **P < 0.01, ***P < 0.001, and ****P < 0.0001, two-way ANOVA. (D) Kaplan-Meier survival curve for MC38 JAK1 KO tumor–bearing mice (n ≥ 3 mice per group, ***P < 0.001, log rank test for trend). (E) Tumor growth curves from C57BL/6 mice (n ≥ 4 tumors per group; n ≥ 3 mice per group) subcutaneously implanted with MC38 JAK2 KO cells on both hind flanks. **P < 0.01 and ****P < 0.0001, two-way ANOVA. (F) Kaplan-Meier survival curve for MC38 JAK2 KO tumor–bearing mice (n ≥ 3 mice per group, ****P < 0.0001, log rank test for trend). See also fig. S13.