Intestinal epithelial tuft cell induction is negated by a murine helminth and its secreted products

Abstract

Helminth parasites are adept manipulators of the immune system, using multiple strategies to evade the host type 2 response. In the intestinal niche, the epithelium is crucial for initiating type 2 immunity via tuft cells, which together with goblet cells expand dramatically in response to the type 2 cytokines IL-4 and IL-13. However, it is not known whether helminths modulate these epithelial cell populations. In vitro, using small intestinal organoids, we found that excretory/secretory products (HpES) from Heligmosomoides polygyrus blocked the effects of IL-4/13, inhibiting tuft and goblet cell gene expression and expansion, and inducing spheroid growth characteristic of fetal epithelium and homeostatic repair. Similar outcomes were seen in organoids exposed to parasite larvae. In vivo, H. polygyrus infection inhibited tuft cell responses to heterologous Nippostrongylus brasiliensis infection or succinate, and HpES also reduced succinate-stimulated tuft cell expansion. Our results demonstrate that helminth parasites reshape their intestinal environment in a novel strategy for undermining the host protective response.

Figure 1.

Gene expression in HpES- and cytokine-treated small intestinal organoids. Organoid cultures grown from duodenal crypt cells, taken from four individual C57BL/6 mice, were stimulated in the presence or absence of the type 1 cytokine IFN-γ, the type 2 cytokines IL-4 and IL-13 (IL-4/13), and/or the ES products of adult H. polygyrus parasites (HpES). After 24 h, RNA was extracted from each replicate, and all 24 samples were subjected to parallel RNA-seq analyses. Patterns of gene expression were then analyzed. (A) Principal component analysis plot showing transcriptomes of stimulated organoids projected onto two dimensions; each condition is represented by four replicates derived from four individual mice. Principal component analysis data were produced from log-transformed normalized counts in the DESeq2 package (Love et al., 2014). (B) Heat map of the top 1,000 DEGs, organized into seven gene clusters, and a group of unclustered genes. Z-scores of normalized count values are indicated by coloring from blue (lowest) to red (highest), based on data from four replicates per group, each individually presented. (C–E) MA plots showing log₂ fold change (M) plotted against log of mean normalized expression counts (A) for the comparisons of no stimulation versus HpES alone (C), IL-4/13 alone versus IL-4/13 + HpES (D), and IFN-γ alone versus IFN-γ + HpES (E). Data from four replicates per group were pooled; genes with adjusted P value < 0.05 are colored. The top 10 DEGs by log₂ fold change that have known function are annotated, with those increased in the presence of HpES shown in blue and those decreased in red. Additional genes of interest are shown in black. DEGs were calculated in DESeq2. (F) Expression levels of 12 genes from cluster 6 showing the most significant (by adjusted P value) differential expression in the presence of HpES. Graphs show normalized counts/log₂ fold change compared with no stimulation control from four independent biological replicates, for organoid cultures exposed to IL-4/13 alone (pink circles) or IL-4/13 + HpES (magenta triangles). Stim, stimulation.

Figure S1.

GO terms for DEGs. (A) HpES and cytokine treatment of small intestinal organoids. Top 12 GO terms, by gene ratio, for DEGs from HpES, IL-4/13, and IFN-γ individual treatments. Data are based on analyses of four biological replicates, composed of a total of 24 samples analyzed in parallel by RNA-seq. DEGs were selected as those with a P-adjusted value of <0.01. The enrichGO function from ClusterProfiler (Yu et al., 2012) was used to identify enriched GO terms, followed by ReViGO (Supek et al., 2011) to remove redundant GO terms. The top 12 GO terms were then selected using gene ratio (number of genes associated with GO term in list/total number of genes in list). (B) Responses of identified gene clusters to HpES, IL-4/13, and IFN-γ. Gene sets were split based on treatment (± HpES) and stimulation (IL-4/13, IFN-γ, or none). Clusters were identified using degPatterns, a part of the DEGreport package (Pantano et al., 2020). Adapt., adaptive; GTPase, guanosine triphosphatase; rec., reecombinant; reg., regulation; resp., response.

Figure S2.

Cluster analysis of HpES modulated gene expression. (A) Top 5 or 10 GO terms by gene ratio for each cluster. Only clusters 1–4 were identified as having corresponding GO terms using enrichGO from ClusterProfiler, and only five GO terms were returned for cluster 2. Redundant GO terms were removed using ReViGO before plotting. (B) Mean normalized counts for the top 10 genes from each cluster by P-adjusted values. Adapt., adaptive; reg., regulation; resp., response.

Figure 2.

GSEA of HpES modulated gene expression. (A) GSEA of the genes expressed by organoids exposed to IL-4/13, HpES, or neither, for gene sets of tuft cells, goblet cells, and Paneth cells as described by Haber et al. (2017). Graphs depict the enrichment score (y axis, green line) with positive values where gene sets are induced, and negative values where they are inhibited. Each vertical bar on the x axis represents an individual gene within the gene set for the stated cell type, and its relative ranking against all genes analyzed. NES and false discovery rate (FDR) are indicated on each graph. Gene expression data are pooled from four independent biological replicates, which were analyzed in parallel by RNA-seq. (B) Heat map of NES for cell type gene sets from Haber et al. (2017) in organoids treated with combinations of IL-4/13, HpES, or neither. (C) Heat map of tuft cell gene set expression, with key genes indicated on the y axis, showing log₂ fold change of genes in organoids treated with combinations of IL-4/13, HpES, or neither. Stim, stimulation.

Figure 3.

HpES and H. polygyrus larvae repress tuft cell expansion in small intestinal organoids. (A and B) Expression of canonical tuft cell and goblet cell genes in organoids treated with combinations of IL-4/13, HpES, or neither. Log₂ fold change of qRT-PCR values for Dclk1 and Muc2 compared with nonstimulated control in four independent biological replicates analyzed by RNA-seq. Statistical analysis was by ordinary one-way ANOVA with Tukey’s multiple comparisons test. *, P < 0.05; ***, P < 0.001. (C) Representative images of organoids stained for tuft cells (anti-DCLK1), shown in yellow, in in organoids treated with combinations of IL-4/13, HpES, or neither. Nuclear staining (DAPI) shown in cyan. Scale bar is 100 µm. (D and E) Expression of Dclk1 and Muc2 in H. polygyrus L3 larvae exposed organoids, Log₂ fold change of qRT-PCR values shown compared with nonstimulated control. Data are pooled from four experiments each with two to four replicates, total n = 8–10 per group. Statistical analysis was by ordinary one-way ANOVA with Tukey’s multiple comparisons test. *, P < 0.05; ***, P < 0.001. Stim, stimulation.

Figure S3.

Cell type–specific inhibition by HpES and H. polygyrus. (A) Pou2f3 gene expression in HpES and IL-4/13 treated organoids. Log₂ fold change shown compared with nonstimulated control in four separate biological repeats analyzed in parallel by RNA-seq. (B) Goblet cell counts in mice singly or coinfected with H. polygyrus and/or N. brasiliensis, counted villus tip to tip after periodic acid–Schiff staining; 25 counts villus-crypt units were analyzed per mouse. Data are from one experiment with four mice per group, representative of three similar experiments. Kruskal–Wallis test with Dunn’s multiple comparisons test was used; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (C) Representative images of goblet cell staining using periodic acid–Schiff staining. Scale bar is 100 µm. (D) Lysozyme gene expression in intestinal samples. Data are from one experiment with four mice per group, representative of three similar experiments. Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used; *, P < 0.05; **, P < 0.01. (E) Paneth cell counts, number of cells per crypt after staining with anti-lysozyme. Data are from one experiment with four mice per group, representative of three similar experiments. Kruskal–Wallis test with Dunn’s multiple comparisons test was used; *, P < 0.05. (F) Representative images of Paneth cell staining with anti-lysozyme. Scale bar is 100 µm. (G) Expression of Sca1 from organoid cultures under the indicated conditions. Change shown compared with nonstimulated control in four independent biological replicates analyzed in parallel by RNA-seq. Hp, H. polygyrus; Nb, N. brasiliensis; Stim, stimulation.

Figure 4.

H. polygyrus inhibits tuft cell expansion in vivo. Mice were infected with 200 H. polygyrus L3 for 28 d before infection with N. brasiliensis for 7 d (A–D) or treatment with 100 mM succinate in drinking water for 7 d. Intestinal tissues were taken at day 7 for mRNA isolation and immunohistological analysis. (A and B) Expression of canonical tuft cell genes Dclk1 (A) and Pou2f3 (B) measured by qRT-PCR in intestinal samples taken from the singly and coinfected mice, presented as log₂ fold change compared with uninfected controls. Statistical analysis was by ordinary one-way ANOVA with Tukey’s multiple comparisons test. Data are from one of three replicate experiments, each with four or five mice per group. ***, P < 0.001; ****, P < 0.0001. (C and D) Tuft cell counts (C) and representative images of tuft cell staining (D) in singly and coinfected mice. Scale bar is 100 µm. Tuft cell counts from ≥20 villus/crypt units were averaged per mouse, and the means for each of five mice per group are presented. Experiments were performed three times with similar results, and data from one representative experiment are shown. Kruskal–Wallis test with Dunn’s multiple comparisons test was used as the discrete data gathered will not be normally (Gaussian) distributed. *, P < 0.05; ****, P < 0.0001. (E and F) Expression of canonical tuft cell genes Dclk1 (E) and Pou2f3 (F) measured by qRT-PCR in intestinal samples taken from succinate and H. polygyrus–infected mice, presented as log₂ fold change compared with untreated and uninfected controls. Statistical analysis was by ordinary one-way ANOVA with Tukey’s multiple comparisons test. Data are from one of two replicate experiments, each with four mice per group. *, P < 0.05; ****, P < 0.0001. (G and H) Tuft cell counts (G) and representative images of tuft cell staining (H) from succinate-treated mice. Scale bar is 100 µm. Tuft cell counts are given as mean for each of four mice per group. Data are from one of two replicate experiments, each with four mice per group. Kruskal–Wallis test with Dunn’s multiple comparisons test was used. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (I) Tuft cell counts from mice administered with succinate and/or HpES; 5 µg HpES was given at days −1, 0, +1, +2, +3 and +4, i.p. Results are pooled from two independent experiments each with five mice per group. Kruskal–Wallis test with Dunn’s multiple comparisons test was used. *, P < 0.05; ***, P < 0.001; ****, P < 0.0001. Hp, H. polygyrus; Nb, N. brasiliensis; Succ, succinate.

Figure 5.

H. polygyrus infection alters organoid morphology and the intestinal stem cell niche. Developmental changes in the intestinal epithelium were investigated by analysis of organoid morphology and key developmental gene expression following HpES exposure, and expression of the same key genes in intestinal tissue following infection with H. polygyrus. (A) Images 16 h after culture of control organoids and organoids incubated with HpES. Scale bar is 100 µm. (B and C) Quantification of organoid architecture after addition of HpES, showing in the distribution of organoids classed as budding or spheroid in the control (top) and HpES (bottom) treatment conditions (B); and quantification by the area of organoid images over 48 h (C). Unpaired t tests were used for statistical analysis; ****, P < 0.0001. (D–F) Expression of intestinal development–related genes, Atoh1, Neurog3, and Hes1, from organoid cultures under the indicated conditions. Change shown compared with nonstimulated control in four independent biological replicates analyzed in parallel by RNA-seq. One-way ANOVA with Tukey’s multiple comparisons test was used; *, P < 0.05; **, P < 0.01; ***, P < 0.001. (G–I) qRT-PCR on in vivo intestinal samples for the same intestinal development–related genes. Data shown are from five individual mice in one of three replicate experiments. One-way ANOVA with Tukey’s multiple comparisons test was used; *, P < 0.05; **, P < 0.01; ***, P < 0.001. Hp, H. polygyrus; Nb, N. brasiliensis; Stim, stimulation.