Microbiota-dependent activation of CD4+ T cells induces CTLA-4 blockade-associated colitis via Fcγ receptors

Abstract

Immune checkpoint inhibitors can stimulate antitumor immunity but can also induce toxicities termed immune-related adverse events (irAEs). Colitis is a common and severe irAE that can lead to treatment discontinuation. Mechanistic understanding of gut irAEs has been hampered because robust colitis is not observed in laboratory mice treated with checkpoint inhibitors. We report here that this limitation can be overcome by using mice harboring the microbiota of wild-caught mice, which develop overt colitis following treatment with anti-CTLA-4 antibodies. Intestinal inflammation is driven by unrestrained activation of IFNγ-producing CD4⁺ T cells and depletion of peripherally induced regulatory T cells through Fcγ receptor signaling. Accordingly, anti-CTLA-4 nanobodies that lack an Fc domain can promote antitumor responses without triggering colitis. This work suggests a strategy for mitigating gut irAEs while preserving antitumor stimulating effects of CTLA-4 blockade.

Fig. 1.. Immune checkpoint blockade induces colitis in C57BL/6 mice harboring microbiota of wild-caught mice.

(A) GF C57BL/6 (WT) mice were colonized with SPF JAX microbiota or WildR microbiota followed by treatment with anti-CTLA-4 antibodies and anti-PD-1 antibodies (ICB) or isotype (ISO) every 3 days. Colitis was assessed by fecal lipocalin-2. N.S., P > 0.05; ***, P < 0.001; ****, P < 0.0001 versus WildR ISO, Dunnett’s multiple comparisons test. Data presented as mean ± SEM. (B) Hematoxylin and eosin (H&E)-stained sections of cecal tissues 15 days after starting ICB treatment, and histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage, black arrowheads point to regenerating epithelia, the gray arrow indicates a crypt abscess, and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. n = 4 to 6 mice per group. Representative data of two independent experiments. N.S., P > 0.05; **, P < 0.01, Kruskal-Wallis with Dunn’s test. (C) GF C57BL/6 and GF Rag1^−/− mice were colonized with WildR microbiota and treated with ICB or ISO. Colitis was assessed by fecal lipocalin-2. N.S., P > 0.05; *, P < 0.05; ***, P < 0.001 versus Rag1^−/− ICB, Dunnett’s multiple comparisons test. Data presented as mean ± SEM. (D) H&E-stained sections of cecal tissues 12 days after starting ICB treatment and histology scores. Scale bars, 200 mm. White arrowheads point to epithelial damage, black arrowheads point to regenerating epithelia, and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. n = 3 to 5 mice per group. Results were confirmed using SPF mice. N.S., P > 0.05; *, P < 0.05, Kruskal-Wallis with Dunn’s test.

Fig. 2.. Anti-CTLA-4 antibodies elicit IFNγ and CD4⁺ T cell-mediated colitis.

(A) FACS analysis of lymphocytes isolated from cecal tissues of WildR microbiota mice 9 days after starting anti-CTLA-4 antibody or isotype treatment. Pseudocolor plots depict IFNγ and IL-17 expression by CD4⁺ T cells and numbers of cytokine-expressing CD4⁺ T cells. Events displayed in flow plots: 2808, isotype and 6388, anti-CTLA-4 antibody-treated. (B) Numbers of Foxp3⁻ CD4⁺ T cells expressing Tbet and RORγt isolated from cecum. (C) Pseudocolor plots of Foxp3 expression by cecal T cells and percentage expression of Foxp3 by T cells. Events displayed in flow plots: 5906, isotype and 14803, anti-CTLA-4 antibody-treated. (D) Pseudocolor plot of Helios and RORγt expression by Foxp3⁺ T cells and ratio of Helios⁺ T_reg (tT_reg) to RORγt⁺ T_reg (pT_reg) numbers. Events displayed in flow plots: 940, isotype and 1039, anti-CTLA-4 antibody-treated. Each dot represents an individual animal. n = 4 mice per group. Representative data of two independent experiments. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired two-tailed t-test. (E) WildR microbiota mice undergoing CTLA-4 blockade were treated with isotype control antibodies or cell-depleting antibodies against CD4 or CD8. H&E-stained cecal sections 12 days after starting ICB treatment and corresponding histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage, black arrowheads point to regenerating epithelia, and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. Data were combined from three independent experiments. N.S., P > 0.05; ****, P < 0.0001; Kruskal-Wallis with Dunn’s test. (F) WildR microbiota mice receiving anti-CTLA-4 antibodies were treated with neutralizing anti-IFNγ antibodies or isotype control. H&E-stained cecal sections 12 days after starting ICB treatment, and histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage, the black arrowhead points to regenerating epithelium, and open arrows indicate inflammatory infiltrates. Each dot represents an individual mouse. Representative data of two independent experiments. **, P < 0.01; Mann-Whitney test.

Fig. 3.. Gene expression and clonotype analyses of CD4⁺ T cells in ICB-induced colitis.

(A) Sort-purified intestinal CD4⁺ T cells from WildR mice treated with isotype or anti-CTLA-4 antibodies underwent scRNA-seq analyses; each sample contained pooled cells from tissues of 3 mice per treatment group. CD4⁺ T cell clusters were visualized by supervised capacity preserving mapping (supCPM). (B) Circos plot showing relative contribution to each cluster identity by CD4⁺ T cells from each treatment. (C) Cluster 5 T_reg cells were segregated and underwent further clustering and supCPM rendering and (D) circos plots showing relative abundance of each T_reg subcluster by treatment group. (E) Doughnut plots showing relative abundance of clonotypes in each cluster classified by three levels of clonal expansion. CD4⁺ T cells with moderate clonal expansion include clonotypes with frequencies between two and four, and CD4⁺ T cells with marked clonal expansion include clonotypes with frequencies of five or greater.

Fig. 4.. Colitis induced by anti-CTLA-4 antibodies requires Fcγ receptors.

(A) WT and Fcer1g^−/− (FcRγ KO) mice were colonized with WildR microbiota and treated with anti-CTLA-4 antibodies (ICB) or isotype control antibodies (ISO). Colitis was assessed by fecal lipocalin-2. Statistical significance between ICB-treated WT and FcRγ KO mice was determined by Sidak’s multiple comparisons test and indicated by N.S., P > 0.05; *, P < 0.05; ***, P < 0.001. Data presented as mean ± SEM. (B) H&E-stained sections of cecal tissues 12 days after starting indicated treatment and histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. N.S., P > 0.05; *, P < 0.05; Kruskal-Wallis with Dunn’s test. Representative data from two independent experiments. (C) Pseudocolor plots of cytokine expressing CD4⁺ T cells isolated from ceca of WildR microbiota-colonized WT and FcRγ KO mice treated with isotype or anti-CTLA-4 antibodies. Events displayed in flow plots: 2949, WT (ISO); 2515, FcRγ KO (ISO); 8060, WT (ICB); 1165, FcRγ KO (ICB). (D) Numbers of cytokine-expressing T cells. Each point represents an individual mouse. Combined data from two independent experiments for each treatment group. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; Dunnett’s multiple comparisons test. (E) Pseudocolor plots of Foxp3 expression by CD4⁺ cells isolated from cecal tissues, and percentages of Foxp3 expression by T cells. Events displayed in flow plots: 5949, WT (ISO); 6487, FcRγ KO (ISO); 6664, WT (ICB); 4870, FcRγ KO (ICB). (F) Subsetting Helios⁺ tT_regs and RORγt⁺ pT_regs, and the ratios of tT_regs to pT_regs. Events displayed in flow plots: 1272, WT (ISO); 1293, FcRγ KO (ISO); 567, WT (ICB); 1024, FcRγ KO (ICB). Each point represents an individual mouse. Representative data from two independent experiments. ***, P < 0.001; ****, P < 0.0001; Dunnett’s multiple comparisons test.

Fig. 5.. Anti-CTLA-4 nanobodies stimulate anticancer immunity without inducing intestinal inflammation.

(A to C) WildR microbiota-colonized WT mice were injected subcutaneously with MC38 tumor cells, then treated with isotype, anti-CTLA-4 antibodies, or anti-CTLA-4 H11 nanobodies with half-life extender (H11-HLE). Tumor volumes were tracked (A), and intestinal inflammation was assessed by fecal lipocalin-2 (B). (C) H&E-stained cecal sections 15 days after starting ICB treatment, and histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. *, P < 0.05; **, P < 0.01; ***, P < 0.001; Dunnett’s multiple comparisons test (A and B), or Kruskal-Wallis with Dunn’s test (C) versus anti-CTLA-4 group. Top symbols denote P-values of comparison with isotype treatment group and bottom symbols denote P-values of comparison with H11-HLE treatment group (B). Data presented as mean ± SEM (A and B). Representative data of two independent experiments. (D to F) BALB/c mice colonized with WildR microbiota were injected subcutaneously with CT26 tumor cells, then treated with isotype, anti-CTLA-4 antibodies and anti-PD-1 antibodies, or H11-HLE and anti-PD-1 antibodies. Tumor volumes were tracked (D), and intestinal inflammation was assessed by fecal lipocalin-2 (E). (F) H&E-stained cecal sections 18 days after starting ICB treatment and histology scores. Scale bars, 200 μm. White arrowheads point to epithelial damage and open arrows indicate inflammatory infiltrates. Each point represents an individual mouse. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; Dunnett’s multiple comparisons test (D, E), or Kruskal-Wallis with Dunn’s test (F) versus anti-CTLA-4 and anti-PD-1 group. Top symbols denote P values of comparisons with isotype treatment group, and bottom symbols denote P-values of comparisons with H11-HLE and anti-PD-1 treatment group (E). Data presented as mean ± SEM (D and E). Representative data of two independent experiments.