Microbiota-derived butyrate restricts tuft cell differentiation via histone deacetylase 3 to modulate intestinal type 2 immunity

Abstract

Tuft cells in mucosal tissues are key regulators of type 2 immunity. Here, we examined the impact of the microbiota on tuft cell biology in the intestine. Succinate induction of tuft cells and type 2 innate lymphoid cells was elevated with loss of gut microbiota. Colonization with butyrate-producing bacteria or treatment with butyrate suppressed this effect and reduced intestinal histone deacetylase activity. Epithelial-intrinsic deletion of the epigenetic-modifying enzyme histone deacetylase 3 (HDAC3) inhibited tuft cell expansion in vivo and impaired type 2 immune responses during helminth infection. Butyrate restricted stem cell differentiation into tuft cells, and inhibition of HDAC3 in adult mice and human intestinal organoids blocked tuft cell expansion. Collectively, these data define a HDAC3 mechanism in stem cells for tuft cell differentiation that is dampened by a commensal metabolite, revealing a pathway whereby the microbiota calibrate intestinal type 2 immunity.

Keywords: HDAC3; butyrate; development; epigenetics; intestine; microbiota; organoid; stem cell; tuft cell; type 2 immunity.

Figure 1.. Tuft cell expansion in the intestine is suppressed by microbiota-derived butyrate.

(A) Fluorescence staining of tuft cells (DCLK1⁺, green) in ileum, (B) frequency of DCLK1⁺ tuft cells and (C) ILC2s by flow cytometry in GF and CNV mice treated with succinate for 7 days. (D) Fluorescence staining of tuft cells in ileum, (E) frequency of DCLK1⁺ tuft cells and (F) ILC2s by flow cytometry from control and vancomycin-treated mice receiving succinate for 7 days. (G) Fluorescence staining of tuft cells in ileum, (H) frequency of DCLK1⁺ tuft cells and (I) ILC2s by flow cytometry from GF and F. prausnitizii mono-colonized mice for 14 days then treated with succinate for 7 days. (J) Fecal concentration of butyrate in GF and F. prausnitizii mono-colonized mice. (K) Fluorescence staining of tuft cells in ileum, (L) frequency of DCLK1⁺ tuft cells, and (M) ILC2s by flow cytometry from WT mice treated with butyrate for 7 days. (N) Fluorescence staining, (O) quantification of DCLK1⁺ tuft cells per organoid, and (P) tuft cell-enriched gene expression by RNA-sequencing of WT organoids stimulated with vehicle (veh) or butyrate for 24 hrs, then treated with IL-13 for 72 hrs. Scale bars, 50μM. Tuft cells gated Live, CD45⁻, EpCAM⁺, DCLK1⁺. ILC2s gated Live, CD45⁺, Lineage (CD4, CD8a, CD11b, CD11c, B220, Ly6G)⁻, CD90.2⁺, CD127⁺, Sca-1⁺, KLRG1⁺. Data are representative of three independent experiments, 3-5 mice per group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 2.. Butyrate controls tuft cell dynamics through histone deacetylase 3.

(A) Fluorescence staining of tuft cells (DCLK1⁺, green) in ileum, (B) frequency of DCLK1⁺ tuft cells and (C) ILC2s by flow cytometry in GF and F. prausnitizii mono-associated HDAC3^FF and HDAC3^ΔIEC mice stimulated with succinate. Tuft cells gated Live, CD45⁻, EpCAM⁺, DCLK1⁺. ILC2s gated Live, CD45⁺, Lineage (CD4, CD8a, CD11b, CD11c, B220, Ly6G)⁻, CD90.2⁺, CD127⁺, Sca-1⁺, KLRG1⁺. (D) Experimental diagram. (E) mRNA expression of Hdac3, (F) Pou2f3, and (G) Dclk1, and (H) fluorescence staining and (I) quantification of tufts cells per murine organoid treated with tamoxifen (4-OHT) +/− butyrate. Scale bars, 50μM. Data are representative of at least three independent experiments, 3-4 mice per group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=not significant.

Figure 3.. Epithelial HDAC3 directs intestinal type 2 immunity and effective worm clearance.

(A) Fluorescence staining of tuft cells (DCLK1⁺, green), (B) frequency of DCLK1⁺ tuft cells by flow cytometry, (C) mRNA expression of IL-25 in SI IECs, and (D) frequency of ILC2s in ileum of WT mice naive and 7 days post-N. brasiliensis infection. (E) Fecal N. brasiliensis egg counts and (F) intestinal N. brasiliensis worm counts at day 10 in HDAC3^FF and HDAC3^ΔIEC mice infected with N. brasiliensis. (G) Fluorescence staining of tuft cells in ileum, (H) frequency of DCLK1⁺ tuft cells by flow cytometry, (I) mRNA expression of IL-25 from SI IECs, relative to naïve controls, (J) frequency of lamina propria ILC2s, and (K) serum concentration of IL-13 in HDAC3^FF and HDAC3^ΔIEC mice 10 days post N. brasiliensis infection. Tuft cells gated Live, CD45⁻, EpCAM⁺, DCLK1⁺. ILC2s gated Live, CD45⁺, lineage (CD4, CD8a, CD11b, CD11c, B220, Ly6G)⁻, CD90.2⁺, CD127⁺, Sca-1⁺, KLRG1⁺. Scale bars, 50μM. Data are representative of two independent experiments, 4-8 mice per group, per timepoint. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 4.. Active regulation by HDAC3 is necessary for IL-13-induced tuft cell expansion.

(A) Fluorescence staining of tuft cells (DCLK1⁺, green) in ileum, (B) frequency of DCLK1⁺ tuft cells and (C) ILC2s by flow cytometry in HDAC3^ΔIEC-IND mice treated with succinate for 5 days then given vehicle or tamoxifen 1x/day i.p. for 5 days, and harvested 5 days after the final dose. Tuft cells gated Live, CD45⁻, EpCAM⁺, DCLK1⁺. ILC2s gated Live, CD45⁺, lineage (CD4, CD8a, CD11b, CD11c, B220, Ly6G)⁻, CD90.2⁺, CD127⁺, Sca-1⁺, KLRG1⁺. (D) mRNA expression of Hdac3 in HDAC3^FF and HDAC3^ΔIEC-IND organoids treated with tamoxifen (4-OHT) +/− IL-13. (E) Fluorescence staining of tuft cells in organoids, (F) quantification of DCLK1⁺ tuft cells per organoid, (G) Pou2f3 and (H) Dclk1 mRNA expression in HDAC3^FF and HDAC3^ΔIEC-IND organoids treated with 4-OHT +/− IL-13. Scale bars, 50μM. Data are representative of at least three independent experiments, 3-4 mice per group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=not significant.

Figure 5.. Human tuft cell differentiation is restricted by butyrate inhibition of HDAC3.

(A) Fluorescence staining of tuft cells (p-EGFR⁺, green) in human crypt-derived intestinal organoids treated with IL-13. Scale bars, 50μM or 20μM for in-lay images. (B) mRNA expression of POU2F3 and (C) mRNA expression of TRPM5 in organoids treated +/− butyrate for 24 hrs then IL-13 for 72hrs, n=3 patients. (D) Fluorescence staining of tuft cells in human intestinal organoids treated +/− butyrate for 24 hrs then IL-13 for 72hrs. (E) mRNA expression of POU2F3 and (F) mRNA expression of TRPM5 in organoids treated +/− HDAC3i (RGFP966) for 24 hrs then IL-13 for 72hrs, n=4 patients. (G) Fluorescence staining of tuft cells in human intestinal organoids treated +/− HDAC3i (RGFP966) for 24 hrs then IL-13 for 72hrs. (H) Fluorescence staining of tuft cells in human intestinal organoids treated with IL-13 +/− butyrate and/or HDAC3i. (I) Quantification of human p-EGFR⁺ tuft cells, n=4 patients. Data are from organoids derived from 3-4 different patients, each color in the panel represents distinct patients. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=not significant.

Figure 6.. Stem cell-intrinsic HDAC3 promotes tuft cell differentiation.

(A) Experimental design. (B) Fluorescence staining and quantification of tuft cells (DCLK1⁺, green) per 5 crypt or 5 villi, and (C) frequency of DCLK1⁺ tuft cells by flow cytometry in the crypt or (D) villi-associated IECs from the ileum of HDAC3^FF and HDAC3^ΔIEC-IND mice stimulated with succinate then treated with tamoxifen and harvested 24 hrs after final tamoxifen dose. (E) Diagram of HDAC3^ΔLGR5 mice. (F) Frequency of TdTomato⁺ (TdT) cells in EpCAM⁺ gate. (G, H) Frequency of DCLK1⁺ tuft cells from TdT⁺ gated cells and (I) frequency of DCLK1⁺ tuft cell from TdT⁻ gated cells in ileum of HDAC3^WT and HDAC3^ΔLGR5 mice. (J) Fluorescence staining and quantification of tuft cells (DCLK1⁺, green) in ileum in HDAC3^WT and HDAC3^ΔLGR5 mice. (K) Experimental design. (L) Fluorescence staining of tuft cells in ileum, (M) frequency of DCLK1⁺ tuft cells, and (N) frequency of ILC2s by flow cytometry in HDAC3^ΔLGR5 mice treated with succinate for 5 days, then treated with vehicle or tamoxifen and harvested five days after the final dose as depicted in (K). Tuft cells gated Live, CD45⁻, EpCAM⁺, DCLK1⁺. ILC2s gated Live, CD45⁺, lineage (CD4, CD8a, CD11b, CD11c, B220, Ly6G)⁻, CD90.2⁺, CD127⁺, Sca-1⁺, KLRG1⁺. Scale bars, 50μM. Data are representative of three independent experiments, 3-5 mice per group. *p<0.05, **p<0.01, ns=not significant.

Figure 7.. Commensal bacteria epigenetically instruct tuft cell differentiation.

(A) Fluorescence staining of tuft cells (DCLK1⁺, green) and (B) frequency of DCLK1⁺ tuft cells by flow cytometry in ileum of Spry2^FF and Spry2^ΔIEC mice. (C) Chip-qPCR for H3K9Ac in Spry2 promoter in HDAC3^FF and HDAC3^ΔIEC IECs. (D) mRNA expression of Spry2 in SI crypts from tamoxifen-treated HDAC3^WT and HDAC3^ΔLGR5 mice. (E) Frequency of DCLK1⁺ tuft cells by flow cytometry in ileum of Spry2^FF and Spry2^ΔIEC mice treated with butyrate for 7 days. (F) Chip-qPCR for Spry2 H3K9Ac and (G) mRNA expression in GF and F. prausnitizii mono-associated IECs. (H) Chip-qPCR for Spry2 H3K9Ac and (I) mRNA expression in murine intestinal organoids stimulated with butyrate for 96 hrs. (J) Chip-qPCR for H3K9Ac at hSprouty2 (SPRY2) promoter in human crypt-derived intestinal organoids treated with butyrate or HDAC3i (RGFP966) for 96 hrs, n=3 patients, each color in the panel represents distinct patients. (K) SPRY2 mRNA expression in human intestinal organoids treated with butyrate or HDAC3i (RGFP966) for 96 hrs, n=4 patients, each color in the panel represents distinct patients. Scale bars, 50μM. Data are representative of at least three independent experiments, 3-4 mice per group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=not significant.