Host-Protozoan Interactions Protect from Mucosal Infections through Activation of the Inflammasome

Abstract

While conventional pathogenic protists have been extensively studied, there is an underappreciated constitutive protist microbiota that is an integral part of the vertebrate microbiome. The impact of these species on the host and their potential contributions to mucosal immune homeostasis remain poorly studied. Here, we show that the protozoan Tritrichomonas musculis activates the host epithelial inflammasome to induce IL-18 release. Epithelial-derived IL-18 promotes dendritic cell-driven Th1 and Th17 immunity and confers dramatic protection from mucosal bacterial infections. Along with its role as a "protistic" antibiotic, colonization with T. musculis exacerbates the development of T-cell-driven colitis and sporadic colorectal tumors. Our findings demonstrate a novel mutualistic host-protozoan interaction that increases mucosal host defenses at the cost of an increased risk of inflammatory disease.

Keywords: IL-18; IL-1b; Tritrichomonas musculis; colon cancer; commensal protist; gut dendritic cells; gut macrophages; inflammasome; intestinal bacterial infection; microbiome.

Figure 1. Identification of a new protistic commensal in mice

(A) Colonic LP cells were isolated from B6 mice obtained from commercial sources or bred at the Mount Sinai animal facility (in-house). Cells were stained with anti-CD45 antibodies and analyzed by flow cytometry. Colored contour plots show staining for CD45 on gated live LP cells. Numbers adjacent to gates represent percentages (+/− SEM). (B) Bar graph shows quantification of CD45.2⁺ cells. (C) Bar graph shows IgA concentrations in sera (mg/ml, left) or colonic tissue explants (mg/g, right). (D) Surface staining of IgA bound to intestinal microbes. Numbers adjacent to gates represent percentages. (E) Cytospins of sorted IgA⁺ cells in (D) were stained with Giemsa. Micrograph shows a representative staining of colonic lavage (40x magnification). Bar indicates 10µm. (F) Scanning electron microscopy (SEM) of colonic tissues from in-house mice. (G) Protozoa per gram of cecum were quantified in five in-house B6 animals naturally colonized with protozoa (B6 Nat) or five animals gavaged with 2 × 10⁶ FACS sorted protozoa (B6 Gavage). Bar graph represents number of protozoa per gram of cecum. (H) DNA was isolated from FACS-purified protozoa and subjected to ITS PCR-DNA sequence analysis. Phylogenetic analysis was performed as described in Material and Methods. The sequence placed the rodent parabasalid, which we hereafter refer to as T. musculis (T.mu), as distinct and sister to a clade comprising the Tritrichomonadidae that shared 95% identity with T. muris. (I) Intestinal tissues were isolated from B6 mice 2 weeks after inoculation with purified T. mu, and paraffin-embedded sections were stained with H&E and PAS. (J) Pathological scoring of T. mu-inoculated cecums and colons was performed, assesing combined pathological score for colon tissue and colonic goblet cell hyperplasia (left) and cecum tissue and cecum goblet cell hyperplasia (right). Red dashed bar denotes borderline pathological score found in normal mice (score 0–2) (Barthel et al., 2003). Data shown represent 3 mice per group, whereof 4–5 high-power fields per tissue and animal were assessed. Statistical analysis was performed using unpaired Student’s t-test. Statistical significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001Scale bar 100 mm (A–C) Data shown is representative of at least two independent experiments with at least three animals per experimental group. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Fig. 1 and Supplementary Table 1.

Figure 2. Colonization with T.mu remodels the colonic mucosal immune tone

(A) Groups of B6 mice were gavaged with 2 × 10⁶ highly purified T.mu. Eight weeks later, colonic LP cells were stained with anti-CD45.2 antibodies and analyzed by flow cytometry. Colored contour plots show CD45⁺ cells. Numbers adjacent to gates represent percentages. (B) Bar graph shows absolute number of CD45.2⁺ cells in control (Ctrl) mice or mice inoculated with T.mu. (C) CyTOF analysis of colonic myeloid immune populations after colonization with purified T.mu. ViSNE dot plots show changes in immune cell composition after T.mu colonization. (D) Intracellular cytokine staining and flow cytometric analysis of TNFα-producing cells in the colonic tissue of T.mu colonized mice. Colored contour plots show staining for TNFα and MHCII within all CD45⁺ cells. Gates were drawn on TNFα⁺ MHCII⁺ cells and the expression of Ly6C and CD64 was analyzed. Numbers adjacent to gates represent percentages (E) Bar graphs show absolute numbers of all TNFα-producing cells and their distribution into the macrophage and monocyte population. (F–G) Mice were gavaged with purified T.mu and 14 days later, colonic LP CD4⁺ T cells were analyzed for the production of IL-17 and IFNγ (F) and for the production of IL-5 and IL-13 (G) by intracellular cytokine staining. Numbers adjacent to gates represent percentages. (H) Bar graphs show quantification of cytokine-producing cells. (I) Bar graphs show absolute numbers of IFNγ⁺, IL-17⁺ and IL-17⁺IFNγ⁺ (DP) Th cells. (J) IFNγ, IL-17 and IL-17⁺IFNγ⁺, production by CD4⁺ T cells isolated 7 days, 14 days, 28 days and 56 days after colonization with T.mu. Bar graphs show absolute numbers of Th cell-subsets. Data shown is representative of at least two independent experiments with at least 3 mice per group and experiment. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figures 2 and 3.

Figure 3. Distinct DC subsets control T.mu associated immune responses

(A–G) Mice were gavaged with purified T.mu 14 days prior to analysis. Colonic LP cells were isolated from (A) Ccr7^−/−, (B) Batf3^−/−, (E) Irf8^ΔDC, (H) Irf4^ΔDC. Cells were incubated in the presence of brefeldin A, Ionomycin and PMA for 4h and stained with anti-CD45, CD3, CD4, IL-17 and IFNγ. IL-17 and IFNγ-producing CD4⁺ T cells were analyzed using flow cytometry. Bar graphs show quantification of absolute numbers of cytokine-producing colonic Th cells. Colonic LP cells isolated from indicated knock out mice were stained with anti-CD45, MHCII, CD11c, CD11b, CD103, CD64 and Ly6C. DC populations were analyzed by gating on CD45⁺ MHCII⁺ CD11c⁺ CD64⁻ Ly6c⁻ cells. (C, D, F and G) contour plots show the abundance of CD103⁺ DC, CD103⁺CD11b⁺ DC and CD103⁻CD11b⁺ DC within the colonic LP of WT, Batf3^−/−, Irf8^ΔDC, Irf4^ΔDC mice. Data shown is representative of at least 3 independent experiments with at least 3 mice per experimental group for each genotype. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figure 4.

Figure 4. Inflammasome activation promotes T.mu associated immune responses

(A) Tissue culture supernatants of tissue explants or RNA, isolated from total colonic tissue were analyzed. Supernatants were analyzed for the presence of IL-18 protein using ELISA. Quantitative RT-PCR was performed to measure IL-18 RNA on whole colonic tissue. Bar graphs show pg/ml of IL-18 or relative expression of Il18 normalized to Hprt. (B–E) Mice were gavaged with purified T.mu 14 days prior to analysis. Colonic LP cells were isolated from (B) Il18^−/−, (C) Lethally irradiated bone marrow chimeric mice reconstituted with Il18^+/+ or Il18^−/−bone marrow cells, at least 8 weeks prior to inoculation with T.mu (D) Asc^−/−, (E) Il1r1^−/−, mice. Cells were incubated in the presence of brefeldin A, Ionomycin and PMA for 4h and then stained with anti-CD45, CD3, CD4, IL-17 and IFNγ. IFNγ and IL-17 production by CD4⁺ T cells was analyzed using flow cytometry. Bar graphs show quantification of colonic Th cells producing IFNγ only, IL-17 only or IL-17 and IFNγ (DP). Data shown is representative of at least 3 independent experiments with at least 3–4 mice per experimental group for each genotype. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figures 4 and 6.

Figure 5. T.mu protects against mucosal infection through inflammasome activation

(A) Groups of mice were either gavaged with purified T.mu or left untreated 14 days prior to Salmonella typhimurium infection. Cecum weight was measured and pathological scoring was performed on H&E stained sections. Data shown is representative of 3 individual experiments with 5 mice per experiment and experimental condition. (B) S. typhimurium colony-forming units (CFU) were measured in the fecal pellet, the cecum, spleen and MLN 48h after infection. (C) Groups of mice were gavaged with purified T.mu 14 days prior to S. typhimurium infection and injected for three consecutive days with 200 ug of either isotype control IgG or anti-IL18 neutralizing antibody right before S. typhimurium infection. Cecum weight was measured and pathological scoring was performed on H&E stained sections. (D) CFU of S. typhimurium were measured in the MLN and spleen 48h after infection. Data shown is representative of at least 2–3 independent experiments with at least 5 mice per group and experiment. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. Scale bar 100 mm.

Figure 6. T.mu exacerbates pathogenic inflammation

(A) Groups of GF Rag2^−/− mice were left untreated (black), conventionalized (pink) or inoculated with quadruple sorted T.mu (blue). Groups of conventional Rag2^−/− mice were left untreated (grey) or inoculated with purified T.mu (purple). Conventionalization of mice was carried out 4 to 6 weeks prior to adoptive transfer of sorted splenic CD45RB^hi CD4⁺ T cells. Gavage with T.mu was performed 2 weeks prior to T cell transfer. Weight loss was monitored weekly and end-point assessment of the disease score was performed using histology assessing degree of inflammation, epithelial erosion, ulceration and hyperplasia, mucin depletion and goblet cell numbers. Pie charts and histology show the clinical assessment of colitis. Data shown is representative of at least 3–4 individual experiments with at least 3–4 mice per experimental group. Scale bar 100 mm. (B and C) Epithelial cell death and epithelial proliferation in colonic tissues of naïve or T.mu inoculated mice. Representative staining of DNA (DAPI blue) and TUNEL (green) in naïve (top row) or T.mu colonized (bottom row) groups. (C) Ki67 staining in colonic tissue of naïve or T.mu inoculated mice. Representative staining of DNA (DAPI blue) and Ki67 (red) in naïve (top row) or T.mu colonized (bottom row). Scatter dot plots show quantification of TUNEL⁺ cells and Ki67⁺ cells. (D) Quantitative RT-PCR on T.mu-driven intestinal target genes (Nos2, Tnfa, Reg3g and Reg3b), normalized to Hprt. (E and F) Apc^+/min mice were left untreated or inoculated with purified T.mu. (E) Tumor burden and size was measured 8–12 weeks later. Micrographs of formalin-fixed colonic tissue counterstained with methylene blue (0.5% PBS) show representative tumor burden in Apc^+/min mice untreated or inoculated with T.mu. (F) Bar graphs show tumor counts and size. Data shown is representative of at least 3 individual experiments with 3–5 mice per experimental condition. All data are shown as mean ± SEM. Student’s t-test or One-way ANOVA Bonferroni’s multiple comparison T test was performed for (A) Statistical significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001. See also Supplementary Figure 5.