Modulation of the gut microbiota engages antigen cross-presentation to enhance antitumor effects of CAR T cell immunotherapy

Abstract

Several studies have shown the influence of commensal microbes on T cell function, specifically in the setting of checkpoint immunotherapy for cancer. In this study, we investigated how vancomycin-induced gut microbiota dysbiosis affects chimeric antigen receptor (CAR) T immunotherapy using multiple preclinical models as well as clinical correlates. In two murine tumor models, hematopoietic CD19⁺-A20 lymphoma and CD19⁺-B16 melanoma, mice receiving vancomycin in combination with CD19-directed CAR T cell (CART-19) therapy displayed increased tumor control and tumor-associated antigens (TAAs) cross-presentation compared with CART-19 alone. Fecal microbiota transplant from human healthy donors to pre-conditioned mice recapitulated the results obtained in naive gut microbiota mice. Last, B cell acute lymphoblastic leukemia patients treated with CART-19 and exposed to oral vancomycin showed higher CART-19 peak expansion compared with unexposed patients. These results substantiate the role of the gut microbiota on CAR T cell therapy and suggest that modulation of the gut microbiota using vancomycin may improve outcomes after CAR T cell therapy across tumor types.

Keywords: CAR T cell; antigen cross-presentation; gut microbiota; immunotherapy; vancomycin.

Graphical abstract

Figure 1

Administration of oral vancomycin boosts CAR T cell therapy and enhances antitumor response in vivo (A) Experimental schema of A20 mouse tumor model and treatment administration. (B) A20 tumor growth in mice treated with untransduced T cells (UTD), UTD plus oral vancomycin (UTD + Vanco), CART-19 only (CAR T), CART-19 plus oral vancomycin (CAR T + Vanco), and untreated (CTR). n = 5–10 mice per group. Data are representative of three independent experiments. (C and D) Percentage of CART-19 cells (CD45.1 cells gated from live, CD45+CD3+ cells) tumor (C) and spleen (D) tissue at 5 days post UTD/CART-19 injection. (E and F) Percentage of CD45⁺CD3⁺CD8⁺ T cells in the tumor (E) and spleen (F) of mice at 5 days post UTD/CART-19 injection. (G) qPCR analysis of genes involved in T cell activation in A20 tumor tissue collected 5 days post UTD/CART-19 IV injection. Graphs show the mean ± SEM by one-tailed t test. Data are representative of three independent experiments.

Figure 2

Oral vancomycin treatment enables antitumor effects of CART-19 therapy in solid organ tumors (A) Experimental schema of B16-CD19 mouse tumor model and treatment administration. (B) B16-CD19 tumor growth in mice treated with UTD, UTD plus oral vancomycin (UTD + Vanco), CART-19 only (CAR T), CART-19 plus oral vancomycin (CAR T + Vanco), and untreated (CTR). n = 5–10 mice per group. Data are representative of three independent experiments. (C and D) Percentage of CART-19 cells (CD45.1) in tumor (C) and spleen tissue at 5 days post UTD/CART-19 injection. Data are representative of three independent experiments.

Figure 3

Vancomycin administration promotes the intratumoral antigen-presenting pathway (A) Volcano plot represents the differentially expressed genes in the tumor of mice treated with CAR T or CAR T + Vanco. The cutoff criteria used to generate the plot are false discovery rate (FDR)-adjusted −log10(p value) >1.5 and |log2FC| of 1.2. FDR correction for multiple hypothesis testing was done using the Benjamini-Hochberg method as described in the section “materials and methods.” (B) NanoString analysis of PanCancer Immune Profiling genes in tumors from mice treated with CAR T or CAR T + Vanco 5 days post CART-19 injection. The heatmap shows the genes with an FDR-adjusted −log10(p value) >1.5. Gene expression values are represented as Z scores. (C) Pathway analysis shows the fold enrichment (x axis), enriched terms (y axis), and significant genes (color) within an enriched term (bubble size). The enrichment chart shows the top 10 enriched gene ontologies. (D and E) qPCR analysis of antigen-presenting molecules in B16-CD19 (D) and A20 (E) tumor tissue collected 5 days post UTD/CART-19 injection. Graphs show the mean ± SEM by one-tailed t test.

Figure 4

Improved antitumor response in mice receiving CART-19 cells in combination with oral vancomycin requires TAA cross-presentation (A) Percentage of CD11b⁺, CD11c⁺, and CD103⁺ DCs in lymph nodes tissue of A20 tumor-bearing mice receiving with UTD, UTD + Vanco, CAR T, and CAR T + Vanco. (B) qPCR analysis of Batf3 and Irf8 gene expression in A20 tumor samples collected from mice treated with CAR T or CAR T + Vanco. Graphs show the mean ± SEM by one-tailed t test. (C) IFN-γ ELISpot assay plated with splenocytes obtained from A20 tumor-bearing mice treated with UTD, UTD + Vanco, CAR T, and CAR T + Vanco stimulated with AH1 peptide. (D) IFN-γ ELISpot assay from co-culture of A20 tumor cells and AH1-specific CD3⁺ T cells isolated from mice immunized with gp70 DNA vaccine incubated overnight with AH1 peptide with or without mouse aMHC-I antibody. (E) Co-culture of tumor isolated DCs from CAR T and CAR T + Vanco treated A20 tumor-bearing mice with T cells isolated from mice immunized with gp70 DNA vaccine incubated overnight with AH1 peptide in IFN-γ ELISpot plate. (F) Percentage of AH1-specific CD8⁺ T cells in A20 tumor tissue collected from mice receiving CAR T or CAR T + Vanco treatment. Graphs show the mean ± SEM. (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).

Figure 5

Adoptive transfer of CD3+ T cells from mice treated with CAR T + Vanco confers tumor protection in a colorectal tumor model via shared TAAs (A) Experimental schema of adoptive T cell therapy from A20 tumor-bearing mice treated with CAR T or CAR T + Vanco into recipient CT26 tumor-bearing mice. (B) Percentage of AH1-specific CD8⁺ T cells upon 72 h of in vitro culture. (C) CT26 tumor growth in mice adoptively transferred with T cells isolated from spleens of A20 tumor-bearing mice treated with CAR T or CAR T + Vanco or untreated (CTR). n = 5–10 mice per group. Data are representative of two independent experiments. (D) Percentage of AH1-specific CD8⁺ T cells in tumor and spleen tissue of CT26 tumor-bearing mice adoptively transferred. (E) IFN-γ ELISpot assay with splenocytes obtained from CT26 tumor-bearing stimulated overnight with AH1 peptide. Graphs show the mean ± SEM. (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001).

Figure 6

Human-to-mouse FMT confirms enhanced cross-presentation and increased CAR T and non-CAR T antitumor immunity (A) Experimental schema of human FMT to recipient A20 tumor-bearing mice. (B) A20 tumor progression in mice receiving FMT from human donor stool treated with CAR T or CAR T + Vanco. (C) Percentage of AH1-specific CD8⁺ T cells in tumor and spleen tissue collected from mice receiving human FMT and treated with CAR T or CAR T + Vanco. n = 5–10 mice per group. Data are representative of two independent experiments. (D) Percentage of CART-19 cells (CD45.1) in spleen and tumor tissue collected from mice receiving human FMT and treated with CAR T or CAR T + Vanco. (E) Percentage of CD11b⁺, CD11c⁺, and CD103⁺ DCs in lymph nodes tissue. (F) qPCR analysis of Ifng, Granzyme b, and Perforin 1 genes. Graphs show the mean ± SEM by one-tailed t test.

Figure 7

Vancomycin exposure time and CART-19 expansion (A) Orange bars shows the time of oral vancomycin treatment after CART-19 infusion. Red triangles highlight the days that patients had their CART-19 peak of expansion. (B) Peak CART-19 expansion in the peripheral blood of the patients exposed to oral vancomycin. Statistical significance was assessed by Mann-Whitney test (∗p < 0.05).