Tuft cell acetylcholine is released into the gut lumen as an effector and regulator of type 2 immunity, and directly targets helminth parasites

Summary

Upon parasitic helminth infection, activated intestinal tuft cells secrete IL-25, which initiates a type 2 immune response during which lamina propria ILC2s produce IL-13. This causes epithelial remodelling, including tuft cell hyperplasia with an unknown function. We describe an unsuspected cholinergic effector function of tuft cells, which are the only epithelial cells expressing choline acetyltransferase (ChAT). During parasite infections, mice with epithelial-specific deletion of ChAT have increased worm burden and fitness, and faecal egg counts, although they are still able to mount a comparable type 2 immune response. Mechanistically, IL-13-amplified tuft cells release acetylcholine (ACh) into the gut lumen. Finally, we demonstrate a direct effect of ACh on worms, reducing their fecundity via helminth muscarinic ACh receptors. Thus, tuft cells are sentinels in naive mice, and their amplification upon helminth infections serves an additional type 2 immune response effector function.

Introduction

Tuft cells are a cellular subset mostly found in digestive and respiratory epithelia, and play critical roles in mucosal host defence. In the intestinal epithelium, tuft cells are primarily known for their critical sentinel function during parasite infections. The presence of helminth or protozoa in the gut triggers tuft cell secretion of the alarmin cytokine IL-25 that initiates a type 2 immune response ^1–3. Such a response is principally orchestrated by type 2 innate lymphoid cells (ILC2s) through the secretion of type 2 cytokines such as IL-4, IL-5 and IL-13. A critical aspect of IL-4 and IL-13 function is the profound remodelling they cause in the intestinal epithelium. This includes amplification of the mucus-producing goblet cells and the tuft cell lineages, as well as resistin-like beta (Retnlβ) ectopic expression by small intestinal goblet cells, in which it is usually absent ⁴. While Retnlβ directly interferes with worm physiology ^4–6, increased mucus production and smooth muscle hypercontractility, also known as the “weep and sweep” response, facilitates worm expulsion ⁷. In contrast, the physiological role of the dramatic increase in tuft cell numbers during type 2 immune responses is not yet understood. Increased IL-25 production following tuft cell lineage amplification is thought to lead to a more efficient type 2 immune response ^1,3, but alternative tuft cell functions also need to be considered. In particular, we note, firstly, that this amplification occurs downstream of the action of type 2 cytokines on epithelial cells; and secondly, that worm expulsion is significantly more delayed by the absence of tuft cells ¹ as compared to IL-25 deficiency alone. These points strongly suggest that, in addition to their alarm function ⁸, tuft cells are required, not only as initiators, but rather as an integral effector component of the type 2 immune response.

The acetylcholine (ACh) neurotransmitter, biosynthesized by the choline acetyltransferase (ChAT) enzyme, regulates a variety of neuronal and non-neuronal physiological functions ⁹. Studies with Chat-reporter mice revealed the presence of cholinergic epithelial cells in various tissues including the trachea ¹⁰, urethra ¹¹, gastro-intestinal and biliary tracts ¹², and thymus ¹³. These cells were identified as tuft (also called brush or solitary chemosensory) cells in the airway, where their actual ACh release was demonstrated ¹⁴. In the nasal cavity, cholinergic tuft cells regulate breathing and inflammation in the presence of irritants ¹⁵; and in the trachea, they control breathing reflexes ¹⁶ and muco-ciliary clearance in response to bacterial quorum sensing molecules ¹⁷. In addition, urethral tuft cells control micturition reflexes through cholinergic signalling to viscerosensory neurons in response to exogenous bitter compounds ¹¹. In contrast, the function of intestinal ChAT-expressing cells remains unknown.

Interestingly, some commonly used drugs against intestinal helminths, such as levamisole and pyrantel are cholinergic agonists. They target the worm acetylcholine receptors (AChRs), causing spastic paralysis which facilitates worm expulsion ¹⁸, and suggests potential direct effects of host ACh on parasites. In addition, available ACh for signalling results from the balance between its synthesis by the ChAT enzyme and its breakdown catalysed by acetylcholinesterase (AChE) enzymes. It is striking that, some parasitic nematodes that colonize mucosal surfaces, encode additional AChE isoforms, which are produced in specific secretory glands and secreted into the worm environment ¹⁹, most likely to avoid detrimental exposure to host ACh.

Here, we investigated the specific role of tuft cell-derived ACh in the context of type 2 immune responses. We confirmed that tuft cells are the only intestinal epithelial cells expressing the key Chat gene for ACh biosynthesis, and actually synthetize ACh. Worm clearance was delayed in mice with Chat-deficient intestinal epithelial cells, in spite of the establishment of a strong type 2 immune response. Mechanistically, we demonstrated in vivo tuft cell ACh release into the gut lumen and, ex vivo, a direct effect of ACh on worms, resulting in reduced fecundity via a muscarinic AChR-dependent pathway, as well as a transiently reduced mastocytosis in mice with Chat-deficient epithelial cells.

Results

The gene encoding the ChAT enzyme is specifically expressed by intestinal epithelial tuft cells

To confirm the specific potential of small intestinal tuft cells to biosynthesize the ACh neurotransmitter, we assessed the expression of the gene encoding the choline acetyltransferase (ChAT) enzyme that catalyses biosynthesis of ACh in FACS-enriched small intestinal EpCam+;Siglec-F+ tuft cells as compared to the EpCam+Siglec-F-non-tuft epithelial fraction. Efficiency of the cell sorting procedure was assessed by analysing the expression of the tuft cell marker Trpm5 mRNA in the tuft and non-tuft cell fractions. The Trpm5 mRNA was detected in the tuft cell fraction whereas it was below the threshold of detection in the fraction containing non-tuft cells (Figure 1A). Similarly, the Chat mRNA was detected in the tuft cell fraction but not in the fraction containing all non-tuft cell epithelial subsets (EpCam+;Siglec-F-) of the small intestinal epithelium, indicating specific expression of the Chat gene in tuft cells among intestinal epithelial cells in vivo (Figure 1B). Then, to assess whether Chat mRNA expression is a common property of small intestinal tuft cells or is limited to only a subset of these cells, in the context of an intact tissue, we coupled anti-Dclk1 immunofluorescence detection of the entire tuft cell population to in situ hybridisation to visualize Chat-expressing cells in naive C57BL/6 mice. This revealed specificity of the Chat probe signal in the vast majority of Dclk1-expressing tuft cells (Figure 1C). We then asked whether Chat expression is also a common feature of the amplified tuft cell population in the context of an helminth parasite infection and ongoing type 2 immune response ^1,3. Wild-type mice were thus infected by gavage with Heligmosomoides polygyrus infective L3 larvae which, after reaching the small intestine, penetrate the submucosa. There, they undergo two developmental moults before emerging again around day 10 post infection, as adult worms, into the gut lumen where they mate and produce eggs which are passed out in the faeces. The presence of H. polygyrus worms in the gut lumen is associated with a strongly polarised type 2 immune response ²⁰, which includes dramatic amplification of tuft cells ²¹. Chat mRNA expression was also restricted to tuft cells from H. polygyrus-infected mice (Figure 1D), indicating the absence of de novo Chat expression in non-tuft cells in the context of a type 2 immune response, and was detected in almost all tuft cells. Furthermore, a quantitative comparison revealed similar proportions of Chat⁺ Dclk1-expressing tuft cells in naive (574/614 examined cells) and infected (435/500 examined cells) mice (Figure 1E). Thus, Chat mRNA expression is present specifically in tuft cells in the mouse small intestinal epithelium and can be detected in almost all tuft cells regardless of their infection status.

Figure 1. Chat gene expression is restricted to tuft cells in the intestinal epithelium in naive mice and during H. polygyrus infection.

(A-B) Expression of the Trpm5 (A) and Chat (B) transcripts in FACS-sorted tuft cells from C57BL/6J mice as compared to all other epithelial cells. Data represent median ± interquartile of the biological replicates (n=6 for tuft cell fractions, n=5 for all other epithelial cell fractions, with ** = p < 0.01 after a Mann-Whitney test). (C-D) Representative in situ hybridisation for the Chat mRNA (red), coupled with immunofluorescence detection of Dclk1 (green) in naive (C) and H. polygyrus-infected (D) mice. Scale bars = 10 μm for all panels. Dotted lines show Dclk1-expressing tuft cells positive for the Chat transcript signal. (E) Quantification of Dclk1 immunoreactive tuft cells expressing the Chat transcript in naive or H. polygyrus-infected mice. Data represent means ± SD of the biological replicates (based on quantification of 614 and 500 Dclk1-positive tuft cells, counted from n=8 and n=5 naive or infected mice, respectively, and analysed with Student T test. ns = p > 0.05).

Impaired host defence against helminth parasites in ChAT-deficient mice

We then investigated the function of tuft cell-produced ACh in the context of an in vivo type 2 immune response. We generated an inducible deletion of the Chat gene specifically in the intestinal epithelium by crossing Chat^{LoxP/LoxP 22} and Villin-Cre^ERT2 mice, to express the Cre recombinase in intestinal epithelial cells in a tamoxifen-inducible manner ²³. To assess the efficiency of the Villin-Cre^ERT2-mediated recombination at the Chat gene locus, we amplified by PCR the sequence of the LoxP-flanked exon 8 of the Chat gene in tamoxifen-treated Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. All mice were treated daily with tamoxifen during 5 days, and analysed 5 weeks later (Figure 2A). Chat Exon 8 sequence was readily amplified with genomic DNA from Chat^LoxP/LoxP mice enriched epithelial cells but was undetectable in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. Simultaneous amplification of the Villin-Cre^ERT2 sequence confirmed the presence of the Villin-Cre^ERT2 transgene uniquely in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice as compared to Chat^LoxP/LoxP mice, as well as the integrity of the genomic DNA purified from Chat^LoxP/LoxP;Villin-Cre^ERT2 mice (Figure 2B). This indicates highly efficient recombination of the Chat locus by the Villin-Cre^ERT2 transgene. Moreover, since the Villin gene promoter is active in all intestinal epithelial cells, including stem cells ²³, Cre^ERT2 activation by tamoxifen causes permanent gene deletion in the intestinal epithelium of compound Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. Indeed, exon 8 deletion was very stable in spite of the rapid renewal of the intestinal epithelial cells, as indicated by the absence of exon 8 sequence amplification from Chat^LoxP/LoxP;Villin-Cre^ERT2 mice intestinal epithelial cell genomic DNA five weeks after tamoxifen treatment (Figure 2A-B). Of note, Villin-1 and Chat expression was reported in subsets of small intestinal ILC2 and ILC3 immune cell populations (Immunological Genome Project (ImmGen) database). To exclude any confounding non-epithelial Cre^ERT2 contribution in the analysis of the Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, we investigated the expression of the Villin-Cre^ERT2 transgene product, more precisely its nuclear translocation upon tamoxifen treatment, in small intestinal lamina propria ILC1, ILC2 and ILC3 subsets. Following tamoxifen injection (Figure 2C), nuclear Cre^ERT2 could be detected in all epithelial cells but neither in CD45+ immune cells, nor, more specifically, in Tbx21+, Gata3+ and Rorγt+ immune subsets (Figure 2D), confirming the absence of Villin-Cre^ERT2 transgene expression in mouse ILC1, ILC2 and ILC3 subsets.

Figure 2. Permanent and intestinal epithelial cell-restricted recombination driven by the Villin-Cre^ERT2 transgene.

(A) Mice from either Chat^LoxP/LoxP or Chat^LoxP/LoxP;Villin-Cre^ERT2 genotypes were treated with tamoxifen during 5 consecutive days, and epithelial samples were recovered 5 weeks later. (B) PCR using primers specific to Chat exon 8 (red arrows in the scheme and electrophoresis), as well as primers specific for the Villin-Cre^ERT2 transgene (blue arrows in the scheme and electrophoresis). While a strong band for Chat exon 8 is detected in control, tamoxifen-treated, Chat^LoxP/LoxP mice (n=2, red arrow), this PCR band is completely missing in tamoxifen-treated, Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, 40 days after the initial induction (n=2). Upper bands (blue arrow), show specific signal related to presence of the Villin-Cre^ERT2 mice transgene. (C) Mice were treated with tamoxifen for 2 consecutive days, and immediately used for subsequent histological analyses. (D) Efficient nuclear translocation of the Cre^ERT2 fusion protein was monitored with IHC against the ER domain (red), and co-stained with the pan-leukocytes lineages marker (CD45, green, in combination with collagen-IV, grey, allowing us to discriminate between epithelial (epi) and stromal (lp) compartments), showing that Cre^ERT2 translocation only occurs in epithelial cells but not in intraepithelial lymphocytes. Cre^ERT2 translocation (red) was compared to expression of transcription factors of the ILCs population (Tbx21, Gata3 and RORγt, related to ILC1, ILC2 and ILC3, respectively, grey), showing that none of the ILC populations displays Cre^ERT2 translocation. Numbers of analysed cells were the following: 136 RORγt-, 187 Gata3- and 508 Tbx21-expressing cells, from 3 independent mice. Scales: 10μm. See also Figure S1 and S2.

We then investigated the consequences of the epithelial Chat deficiency on the cellular composition of the gut mucosa. No gross alteration of the epithelial layer was perceptible following tamoxifen treatment of Chat^LoxP/LoxP;Villin-Cre^ERT2 mice as compared to Chat^LoxP/LoxP control mice, as identical representations of epithelial cells from the goblet, Paneth and enteroendocrine cell lineages were found in both mouse genotypes (Figure S1). Moreover, the overall distribution of the immune subsets (T cells, B cells and myeloid cell populations including monocytes, neutrophils, macrophages and mast cells) within the hematopoietic cell compartment in the intestine (intraepithelial and lamina propria fractions) as well as in more distant peripheral lymphoid organs (spleen, mesenteric and axillary/brachial/cervical lymph nodes) was equivalent between Chat^LoxP/LoxP;Villin-Cre^ERT2 and Chat^LoxP/LoxP mice (Figure S2A-B). More specifically, no difference was observed in the CD4 and CD8 T cell subpopulation as well as in their activation and polarization status (Figure S2A). Similarly, the relative percentage of ILC1, ILC2 and ILC3 subsets was comparable in the lamina propria and intraepithelial populations of Chat^LoxP/LoxP;Villin-Cre^ERT2 and Chat^LoxP/LoxP mice (Figure S2B). We then assessed the consequences of the ChAT deficiency during an infection with parasitic helminths. We first induced Chat gene deletion by tamoxifen treatment during five consecutive days. After 5 days of rest without tamoxifen, Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice were infected by gavage with H. polygyrus L3, and infection parameters were analysed from 10 to 40 days post-infection when adult worms are present in the gut lumen (Figure 3A). Although ChAT-deficient mice had no obvious phenotype, autopsy revealed increased numbers of adult worms in these mice as compared to Chat^LoxP/LoxP control mice (Figure 3B). Greater numbers of live worms were mirrored by the numbers of eggs in the faeces of infected mice, a dynamic readout of the type 2 immune response efficiency and worm persistence. Indeed, a significantly increased number of eggs was found in ChAT-deficient mice, as compared to Chat^LoxP/LoxP littermates, between 10 days post-infection, when adult worms start to be present in the gut lumen, and 40 days post-infection (Figure 3C). Increased numbers of adult worms and faecal eggs in ChAT-deficient mice thus revealed an essential role of tuft cell-derived ACh for an efficient defence against helminth parasites.

Figure 3. Tuft cell Chat deficiency delays worm expulsion.

(A) Scheme depicting the experimental design of H. polygyrus infections and mouse analysis time points. Control mice (Chat^LoxP/LoxP) or mice with an intestinal epithelium-targeted Chat gene deletion (Chat^LoxP/LoxP;Villin-Cre^ERT2) were first treated with 1 mg of tamoxifen between -10 and -5 dpi, and then orally infected with H. polygyrus L3 larvae at day 0. Infected animals were monitored and/or sacrificed from 10 to 40 days post-infection. (B) Kinetics of infection of Chat^LoxP/LoxP (white bars and circles) or Chat^LoxP/LoxP;Villin-Cre^ERT2 mice (dark bars and circles) with 200 H. polygyrus L3, monitored for numbers of adult worms found in their small intestines. Data represent mean ± SD of the biological replicates (n ranging from 8 to 10 mice per time point, ** = p < 0.01, ns = p > 0.05, after a one-way Anova test). (C) Kinetics of infection of Chat^LoxP/LoxP (white bars and circles) or Chat^LoxP/LoxP;VillinCre^ERT2 mice (dark bars and circles) with 200 H. polygyrus L3, monitored for numbers of eggs per gram of faeces. Data represent median ± interquartile of the biological replicates (n ranging from 8 to 10 mice per time point, * = p < 0.05, ns = p > 0.05, after a multiple Mann Whitney test for each time point). (D) Mass spectrometry-based ACh quantification (fmoles per tuft cell) in FACS-sorted tuft cells isolated from either control Chat^LoxP/LoxP or Chat^LoxP/LoxP;VillinCre^ERT2 mice, in naive and H. polygyrus-infected conditions. Bars represent medians ± interquartile of different biological samples (n=4, 4 and 5 replicates from naive, 20 dpi of Chat^LoxP/LoxP and Chat^LoxP/LoxP;VillinCre^ERT2 mice, respectively, * = p < 0.05, ns = p > 0.05, after Kruskal-Wallis analyse).

Increased ACh concentrations in tuft cells during type 2 immune responses

We then assessed the actual ACh concentration in tuft cells by mass spectrometry using lysates of FACS-sorted epithelial tuft cell fractions. We found a striking elevation of ACh in the EpCam+;Siglec-F+ tuft cell fractions of mice infected with H. polygyrus, as compared to the tuft cell fractions from naive mice, although high variability of ACh concentrations was present in infected mice, likely reflecting difference in the infection efficiency (Figure 3D). As this elevation occurred in cellular populations enriched in tuft cells, it likely reflects increased ACh synthesis per tuft cell in efficiently infected mice rather than a consequence of increased tuft cell numbers caused by the epithelial remodelling consequent to type 2 immune responses. As expected, ACh was not detected in the tuft cell fractions from Chat-deficient epithelial cells in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice (Figure 3D). Thus, the basal tuft cell ACh biosynthesis rate dramatically increases in the context of a type 2 immune response against H. polygyrus infection.

ChAT-deficient mice are able to mount a type 2 immune response

To understand the mechanisms leading to decreased parasite clearance in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, we assessed critical parameters of the type 2 immune response and subsequent epithelial remodelling in naive and H. polygyrus-infected mice. Quantification of type 2 immune responses such as numbers of Gata3+ ILC2s/Th2 cells, epithelial tuft and goblet cells, and expression of the Retnlβ peptide by small intestinal goblet cells indicated the presence of a strong type 2 immune response in both Chat^LoxP/LoxP control mice and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, with some of these parameters reaching even higher levels in infected ACh-deficient mice, possibly due to the presence of higher numbers of worms (Figure 4A-C). To determine whether increased type 2 immunity parameters in infected ChAT-deficient mice was more likely caused by higher worm burden, or directly consequent to the ChAT deficiency, we triggered a worm-independent activation of tuft cells and subsequent type 2 immune response by treating Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice with succinate. Succinate is a known activator of tuft cells, the only intestinal epithelial cells expressing the succinate receptor Sucnr1, and subsequent type 2 immune responses ^24,25. Nearly identical levels of type 2 immune response, as assessed by quantification of lamina propria Gata3+ cells, epithelial tuft cells and Retnlβ expression in goblet cells were observed in both mouse groups (Figure S3). This suggests that the increased type 2 immunity parameters in H. polygyrus-infected Chat^LoxP/LoxP;Villin-Cre^ERT2 mice are more likely caused by an increased worm burden than a direct effect of the ChAT deficiency. We also assessed the mRNA expression of tuft cell mediators of type 2 immune responses and factors known to be involved in ACh synthesis and transport, including the Alox5, Alox5ap, Ltc4s, Ptgs1, Ptgs2, Hpgds, Pou2f3, Dclk1, Sucnr1, Il25, Chat and VAChT genes, using intestinal epithelial cell extracts isolated from Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, in naive, 20 and 40 dpi with H. polygyrus. As expected, the Chat mRNA was absent in extracts from Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. With the exception of the Dclk1 mRNA, which was found elevated in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice at 40 days post infection, no significant difference was found between control and ChAT-deficient mice at the naive or infected states (Figure S4). This was then confirmed at the cellular level on tissue sections using immunofluorescence or in situ hybridisation for all markers for which antibodies or probes were available (Alox5ap, Alox5, Ltc4s, Ptgs2, Hpgds, Sucnr1). Again, no alteration in the expression patterns of these markers could be detected, suggesting that the ChAT deficiency does not alter significantly the expression of tuft cell alarmin molecules or their biosynthesis pathways (Figure S5), which is consistent with the observation of a strong type 2 immune response occurring in infected ChAT-deficient Chat^LoxP/LoxP;Villin-Cre^ERT2 mice.

Figure 4. Chat gene deficiency does not impair establishment of a type-2 immune response.

(A) Characterisation of the intestinal mucosa of Chat^LoxP/LoxP (white bars and circles) and Chat^LoxP/LoxP;VillinCre^ERT2 mice (dark bars and circles) following H. polygyrus infection, at the indicated time points. Each cell type was quantified by IHC, using markers for tuft cells (Dclk1), ILC2s and Th2 immune cells (Gata3), and activated mucus-secreting goblet cells (resistin-like β). Bars represent means ± SD of the biological replicates (n ranging from 5 to 8; * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p<0.0001, ns = p > 0.05, after ANOVA analyse). (B) Representative co-immunostainings of Dclk1 (green), Gata3 (red) and nuclei (blue) in naive or infected Chat^LoxP/LoxP and Chat^LoxP/LoxP;VillinCre^ERT2 mice, 20 and 40 days post infection. (C) Representative co-immunostainings of Resistin-like β (red) and nuclei (blue) in naive or infected Chat^LoxP/LoxP and Chat^LoxP/LoxP;VillinCre^ERT2 mice, 20 and 40 days post infection. Scale bars = 10 μm for B and C. See also figure S3.

To complement these data, we also sought to assess the distribution of the main immune cell populations present in the gut mucosa and distant lymphoid organs in control and ChAT-deficient infected mice. Unfortunately, the low viability of the cells dissociated from the inflamed gut mucosa of infected mice precluded such analyses. However, immunophenotyping could be performed on distant peripheral lymphoid organs (mesenteric lymph nodes, axillary/brachial/cervical lymph nodes, and spleen), and no significant differences were detected in ChAT-deficient infected mice as compared to controls (Figure S6 and STAR Methods). This suggests that the effects of tuft cell ACh deficiency remained local and did not affect distant or local lymphoid structures.

Mast cells constitute another immune subset involved in the defence against helminth parasites ²⁶, which can be analysed by immunohistochemistry on inflamed tissue sections to circumvent the difficulty of studying dissociated cells from infected mice. We quantified the presence of mast cells in naive and infected mice using the mast cell protease 1 (Mcpt1) marker. In Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, the Mcpt1+ cell population was more heterogeneous as compared to controls. Therefore, to quantify the difference in mast cell populations in control versus ChAT-deficient mice, we characterized the mast cell population detected in infected mice both in terms of percentage of the gut tissue with high density of mast cells, and numbers of mast cells per microscopic field in regions of high mast cell densities (Movies S1-S6). The rare Mcpt1+ cells found in naive Chat^LoxP/LoxP mice strongly increased in H. polygyrus-infected mice 20 or 40 days post-infection. In ChAT-deficient mice, Mcpt1+ cell numbers also increased 20 days post infection but to a lesser extent as compared to control mice, and at 40 days post infection, both Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice had similarly elevated Mcpt1+ mast cell counts (Figure 5A-C). In both genotypes, Mcpt1+ mast cells almost always co-expressed other mast cell markers such as cKit and Granzyme B (GzmB), as well as CD63, suggesting an activated state (Figure 5D and S7A). Moreover, quantification of Mcpt1+ cells co-expressing CD63 revealed similar rates of mast cell activation between Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, with 92.9 % and 97.3 % double positive cells at 20 dpi, and 91.2 % and 92.4 % at 40 dpi, respectively (Figure S7B). Together, these data indicate that Chat^LoxP/LoxP;Villin-Cre^ERT2 ACh-deficient mice are able to mount a strong type 2 immune response in the absence of tuft cell-derived ACh, with the exception of mast cells, which were transiently less numerous at 20 days post infection in ChAT-deficient mice. Thus, higher worm persistence and egg production in ChAT-deficient mice is not due to a globally compromised type 2 immune response, indicating a yet unappreciated role of ACh in promoting worm expulsion.

Figure 5. Tuft cell-restricted Chat gene deficiency impairs intestinal mastocytosis establishment following H. polygyrus infection.

(A) Intestinal mast cell infiltration score during H. polygyrus infection at the indicated time points, based on numbers of crypt-villus axes displaying 2 or more mast cell protease 1-expressing mast cells, relative to the total number of crypt-villus axes of the small intestine section. Data represent median ± interquartile of the biological replicates, ranging from n=3 (naive), to n=5 (for infected control and ChAT-deficient mice at 20 and 40 dpi), with ** = p < 0.01, ns = p > 0.05, after a multiple Mann Whitney test). (B) Quantification of mast cell protease 1-expressing mast cells within small intestinal infiltrated areas. Data represent median ± interquartile of the biological replicates, ranging from n=3 (naive), to n=5 and n=5 (for infected control and ChAT-deficient mice at 20 dpi, respectively) and n=5 and n=5 (for infected control and ChAT-deficient mice at 40 dpi, respectively), with ** = p < 0.01, ns = p > 0.05, after a multiple Mann Whitney test). (C) Representative immunostainings of mast cell protease 1 (red) and nuclei (blue) in naive (left panel), 20 dpi (middle panel) or 40 dpi (right panel) H. polygyrus-infected mice, in either control (Chat^LoxP/LoxP) or mice with ChAT-deficient tuft cells (Chat^LoxP/LoxP;VillinCre^ERT2). (D) Representative immunostainings of c-Kit (grey), Mcpt1 (green), mast cell activation marker CD63 (red) and nuclei (blue) in naive or 20 dpi H. polygyrus-infected mice, in either control (Chat^LoxP/LoxP) or Chat^LoxP/LoxP;VillinCre^ERT2 mice. Scale bars = 10 μm for C and D. See also Figure S7.

Increased luminal ACh concentrations following H. polygyrus infection

Certain helminth parasites express secreted isoforms of the AChE enzyme ²⁷, suggesting that degrading neighbouring ACh is beneficial for the worms. In order to investigate the hypothesis of a direct effect of tuft cell-derived ACh on worms present in the host gut lumen, we next assessed luminal ACh concentrations in naive and H. polygyrus-infected mice. For this, an intestinal loop was surgically ligatured and filled with wash buffer, which was recovered after 30 minutes of incubation, and processed for detection of ACh by mass spectrometry. The resulting measurements revealed significantly more elevated ACh concentration in the gut lumen of control infected mice, as compared to naive animals (Figure 6A), which is likely due to contribution of both higher tuft cell numbers and increased ACh production per tuft cell, as indicated by ACh quantification from extracts of FACS-sorted tuft cells as compared to identical numbers of non-tuft intestinal epithelial cells (Figure 3D). We noted that ACh concentrations were variable in infected mice, probably reflecting infection efficiencies as well as the specific location of worms as regards the favourable intestinal loop for surgery. We then asked whether the actual presence of parasites was required for increased luminal ACh concentrations or if this augmentation was directly linked to tuft cell activation. We thus treated mice with succinate and assessed ACh concentration changes. Indeed, luminal ACh was also significantly increased in the gut lumen of succinate-treated wild type mice as compared to untreated mice, indicating luminal ACh release by tuft cells activated in the absence of worms (Figure 6B). Together, these data demonstrate the presence of tuft cell-derived ACh in the gut lumen of H. polygyrus-infected mice, and suggest the possibility of a direct impact of luminal ACh on worm physiology.

Figure 6. Tuft cell-derived ACh is released into the host gut lumen, and ACh directly impairs worm physiology.

(A) Quantification of intestinal lavage fluid ACh concentration, as described in the methods section, in naive and H. polygyrus-infected mice, 20 days and 40 days post infection. Data represent median ± interquartile of the biological replicates (n ranging from 7 to 10, * = p < 0.05, **** = p<0.0001, ns = p > 0.05, after Kruskal-Wallis analyse). (B) Quantification of intestinal lavage fluid ACh concentration, as described in the methods section, in naive or succinate-treated mice (one week treatment with 100 mM succinate in the drinking water). Bars represent median ± interquartile of the biological replicates (n ranging from 3 to 6, with * = p < 0.05, after a Mann-Whitney test). For A and B, ACh concentrations were determined in the intestinal lavage fluid recovered per animal. (C) ATP levels (μM) in worms recovered from either Chat^LoxP/LoxP (white bars and circles) or Chat^LoxP/LoxP;Villin-Cre^ERT2 mice (dark bars and circles) infected with H. polygyrus at different time points. Data represent mean ± SD of the different biological replicates (n ranging from 5 to 8 groups of 25 worms per time point, * = p < 0.05, ns = p > 0.05, after ANOVA analysis). (D) Percentage of eggs released in the culture medium during a 24h time frame, after treatment of adult H. polygyrus worms with 10 μM ACh, as compared to untreated control worms. Bars represent mean ± SD of the biological replicates (n ranging from 11 to 12 groups of 25 female worms per condition, **** = p < 0.0001 after Student t-test). (E) Percentage of eggs released in the culture medium during a 24h time frame, after treatment of adult H. polygyrus worms with 10 μM ACh (ACh, n=10), or 10μM ACh combined with 5μM mecamylamine (ACh Meca, n=8) or 5μM atropine (ACh Atrop, n=8), or 5μM mecamylamine or 5μM atropine alone (Meca and Atrop, n=9 and 8, respectively), as compared to untreated control worms (n=10). Bars represent mean ± SD of the biological replicates, ns = p > 0.05, * = p < 0.05, ** = p < 0.01, *** = p < 0.001 after a one-way ANOVA. See also table 1.

Increased worm fitness in mice with ACh-deficient tuft cells

To directly assess the consequences of H. polygyrus exposure to ACh, we used two complementary approaches. Noteworthily, intestinal goblet cells produce the Retnlβ molecule, specifically in the context of type 2 immune responses, and this directly interferes with H. polygyrus physiology ⁴. Subsequently, it was found that worms exposed to Retnlβ had decreased ATP levels, reflecting altered viability and fecundity, as compared to unexposed worms ^5,6. Thus, using a similar approach, we quantified ATP concentrations in worms recovered from the intestines of Chat^LoxP/LoxP and Chat^LoxP/LoxP;Villin-Cre^ERT2 mice, as a proxy of their global fitness. No significant difference was found in ATP levels of worms from mice of either genotype recovered at 20 days post infection. In contrast, by day 40, significantly increased ATP levels were found in worms from mice with ACh-deficient tuft cells as compared to control Chat^LoxP/LoxP littermates, suggesting decreased fitness of the worms obtained from an environment containing ACh for several weeks (Figure 6C).

ACh exposure directly decreases worm fecundity

To confirm this finding, we also assessed worm fecundity ex vivo, in the presence of 10µM ACh, which is consistent with the micromolar range ACh concentrations found in the intestinal lumen lavage fluid obtained from H. polygyrus-infected Chat^LoxP/LoxP control mice. Of note, it should also be taken into account the certainly partial recovery of luminal ACh using a single 30 min lavage, as well as an estimated 15 times dilution factor corresponding to the lavage fluid volume injected into the intestinal loop (Table 1). We thus quantified egg production per H. polygyrus adult female worm, freshly recovered from infected Chat^LoxP/LoxP mice 14 days post-infection to ensure that the worms are present in the gut lumen and not yet strongly altered by luminal ACh. When ACh was added to the culture medium, egg production was significantly decreased, as compared to untreated worms (Figure 6D). These data indicate that worm exposure to physiologically relevant ACh concentrations directly decreases helminth fecundity.

Table 1. Estimation of small intestinal luminal content volumes prior to surgeries.

The table shows recovered volumes of small intestinal contents prior to surgeries. Data were acquired from 6 mice, means and SD are shown.

Mouse ID	Recovered volume of intestinal content (μl)
1	20
2	18
3	8
4	9
5	13
6	18
Mean	14.33
SD	4.64

ACh targets worm physiology via their muscarinic ACh receptors

To identify the mechanisms underlying the effect of ACh on worm physiology, we assessed worm fecundity ^5,6 in the presence of ACh combined with different inhibitors of cholinergic signalling. As shown previously in Figure 6D, exposure to ACh alone decreased worm fecundity as compared to untreated worms. Exposure to ACh combined with mecamylamine, an antagonist of the nicotinic AChRs, caused a comparable significant drop in worm fecundity. In contrast, the inhibitory effect of ACh on egg production was not observed with worms treated with a combination of ACh and atropine, an antagonist of the muscarinic AChRs (Figure 6E). This implies that the inhibitory effects of ACh on worm fecundity are mediated by muscarinic AChRs expressed by the worms. Finally, none of the drugs had significant effects on egg production when used alone (Figure 6E). Altogether, these data argue that, in addition to their sentinel function in initiating type 2 immune responses, intestinal tuft cells act as effectors of such responses by releasing into the host lumen non-neuronal ACh, in a range of concentrations capable of directly altering helminth fecundity through their muscarinic AChRs as assessed in mechanistic ex vivo experiments (Figure 7). In addition, luminally- or basolaterally-secreted ACh likely contributes to the increased presence of activated mast cells following H. polygyrus infection.

Figure 7. Scheme illustrating the new tuft cell effector function during helminth parasite infection.

After initial detection of helminths, tuft cells secrete alarmins to initiate a type-2 immune response (left panel), as previously described, eventually leading to mucosal remodelling and drastic tuft cell amplification. Newly produced tuft cells are able to release ACh (green dots) into the lumen to target helminths through their muscarinic receptors (upper right panel). Tuft cell-derived ACh also enhances mastocytosis and triggers fluid secretion from neighbouring epithelial cells (lower right), as shown in Ref ⁴⁵.

Discussion

Thus far, the importance of tuft cell-produced ACh have been mostly studied in the airways. In the mouse tracheal epithelium, tuft cells are capable of sensing bitter compounds present in the airway lining fluid and use cholinergic signalling to neighbouring nerve endings to cause an aversive reflex consisting of a reduced breathing frequency ¹⁰. Tuft cells also participate in the regulation of bacterial populations present in the airway. These cells can be activated by bacterial quorum sensing molecules (QSM) - used by microbes to evaluate their own population density - such as 3-OxoC₁₂-HLS and use ACh signalling to cause respiratory changes ¹⁶. In addition to respiratory reflexes, mucociliary clearance is an important innate protective process to eliminate inhaled pathogens from the airway. Paracrine cholinergic signalling from tracheal tuft cells, activated by bitter compounds, QSM, or other bacteria-derived products, was found to increase mucociliary clearance by airway ciliated cells, as assessed by the transport speed of particles on the tracheal surface, thereby directly linking chemosensation of bacterial signals with innate defence ^14,17. Previous studies with reporter mice also reported expression of Chat ¹¹ and of the Tas2r143 bitter taste receptor ²⁸ in a subset of tuft cells from the urogenital tract as well as from the gastrointestinal tract ¹². Furthermore, intraurethral application of a bitter compound regulated bladder activity in rats ¹¹. We confirm intestinal tuft cell-restricted expression of the Chat gene for biosynthesis of ACh. Previous studies using Chat-GFP reporter mice reported heterogeneity in GFP expression in trachea and urethra Villin-immunoreactive tuft cells ^10,11. Interestingly, we found co-expression of the endogenous Chat mRNA and tuft cell Dclk1 protein in nearly all analysed cells, suggesting either different regulation of Chat expression in the tuft cells from distinct organs, or incomplete detection of the Chat gene product expression in genetically engineered reporter mice.

We present here two facets of a cholinergic function of small intestinal tuft cells. Firstly, we identified a regulator role on mastocytosis, the specific impact of which on the dynamics of H. polygyrus remains, currently, difficult to delineate. Secondly, we report an effector function of tuft cells through the production of intestinal luminal ACh that may have direct effects on helminth parasites. Indeed, we report the presence of ACh in the gut lumen, with a dramatic concentration increase during type 2 immune responses in H. polygyrus-infected mice, reaching a micromolar range. Interestingly, the increase in luminal ACh concentration can also be triggered by succinate treatment-driven tuft cell activation, suggesting that luminal ACh production is an integral component of the type 2 immune response downstream to tuft cell activation. Tuft cell-derived acetylcholine directly or indirectly decreased worm viability 40 days post-infection, but not 20 days post-infection, as assessed by worm global ATP levels. Moreover, we demonstrated ex vivo a direct effect on worm fecundity following ACh exposure during 24h, acting through H. polygyrus muscarinic AChRs. It is possible that in vivo quantification of global worm ATP levels, on the one hand, and ex vivo egg production, on the other hand, reflect ACh-dependent alterations of distinct physiological functions. Consistent with this, endogenous worm-derived ACh was reported to inhibit the egg laying behaviour through worm muscarinic AChRs in the Caenorhabditis elegans helminth ²⁹. It is therefore plausible that host-derived acetylcholine directly interferes with a related H. polygyrus muscarinic ACh receptors-regulated pathway, to decrease its fecundity. The reduced worm fitness observed 40 days post infection may, in turn, rely on the alteration of distinct pathways, possibly involving nicotinic AChRs and related functions such as neuromuscular activity and feeding ³⁰.

Luminal ACh release by tuft cells in the context of a type 2 immune response illustrates how different epithelial subsets constitute integral parts of this process. Signalling through epithelial IL-4Rα receptors causes specific amplification of two epithelial cell lineages, tuft cells and goblet cells. In addition to the effector function of tuft cells discussed above, goblet cells not only contribute to the so-called weep-and-sweep response by increased mucus production; they also express de novo the Retnlβ peptide that directly interacts with the helminth’s ability to feed on host tissues, and thus also contribute to decrease their fitness and fecundity ^4–6. As parasitic helminths have co-evolved with their hosts, it is plausible that the use of different defence mechanisms, relying on distinct host cell types, such as ACh and Retnlβ, produced by tuft and goblet cells, respectively, renders the task more challenging for parasites to counteract host immunity. In this view, it might be interesting to combine both ACh and Retnlβ host deficiencies to investigate possible synergistic action of these two molecules on different helminth parasite species.

Cholinergic regulation of immune responses is well described ³¹, including in the context of helminth infections ^32–35. In the current study we did not investigate the consequences of tuft cell ACh deficiency on immune subsets in more detail, with the exception of intestinal mucosal mast cells. This limited exploration was justified because, despite a strong type 2 immune response, no additional alterations in the expression of tuft cell-derived immune mediators nor significant alterations of immune populations were identified in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. We however report that the mastocytosis observed following H. polygyrus infection was partially inhibited in mice with ChAT-deficient tuft cells. Although the precise links between mast cells and type 2 immune responses remain only partially understood, multiple studies reported decreased anti-parasitic defence in mouse lines with mast cell-deficiencies, including after infection with H. polygyrus ^36,37. Thus, although mastocytosis following helminth infection occurs in the absence of tuft cell-derived ACh, it is quantitatively lower as compared to infected control mice, which might contribute to the delayed control of H. polygyrus infection observed in Chat^LoxP/LoxP;Villin-Cre^ERT2 mice. It is interesting to note that several studies on the role of mast cells during helminth infections used cKit gene loss of function models ^38,36,39 or its pharmacological inhibition ^37,40 to cause mast cell deficiencies. cKit is also expressed in hematopoietic progenitors ⁴¹, as well as in epithelial Paneth ⁴², goblet ⁴³ and tuft cells ⁴⁴, which might be confounding if a tuft cell deficiency is induced besides the mast cell deficiency.

Luminal release of ACh and fine tuning of mucosal mastocytosis may not be the unique modes of action of tuft cell ACh in the small intestine. Additional cholinergic mechanisms, for instance paracrine signalling to other epithelial cells or nerve cells, as reported in the airway ^10,14, have recently been reported in the small intestine ⁴⁵. Moreover, additional ACh sources, such as ILC2s, exist in the gut mucosa, and also contribute to ILC2 numbers inflation and anti-helminth immunity ³⁵.

Our study on the cholinergic functions of intestinal tuft cells thus reveals additional functions for these cells and a novel luminal mode of action of ACh. Hence, we propose that tuft cells play two distinct but complementary roles in the context of the host defence against helminth infections. In the naive mucosa, tuft cells are a rare epithelial subset functioning as a sentinel to initiate type 2 immune responses upon parasitic infections. Our new findings now document how the amplification of the tuft cell lineage, occurring in the context of an ongoing type 2 immune response, underlies a second function as a cholinergic regulator of mastocytosis and effector that likely directly contributes to parasitic helminth clearance by compromising their global fitness. Moreover, as demonstrated ex vivo, ACh concentrations measured in the gut lumen can directly alter H. polygyrus fecundity. This not only highlights the role of tuft cell hyperplasia during type 2 immune responses, but may also provide a missing link to understand why IL-25 deficiency, a key tuft cell cytokine involved in their sentinel function ⁸, causes a less severe phenotype than the complete absence of tuft cells in Pou2f3-deficient mice ¹.

Future studies will be needed to evaluate the potential therapeutic perspectives consecutive to these findings, in refining the common anti-helminth treatments of humans or livestock, such as cholinergic agonists or inhibitors of AChEs, with a focus on a luminal mode of action of these drugs.

Star Methods

Resource Availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Philippe Jay (Philippe.jay@igf.cnrs.fr).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Experimental Model

Animal models

The conditional Chat^LoxP/LoxP allele, (Chat^{tm1Mlt, 22}) was kindly provided by Sylvie Berrard and crossed with the intestinal epithelium-specific, tamoxifen-inducible, Villin-Cre^ERT2 mouse strain (Tg(Vil1-cre/ERT2)23Syr ²³. All the mice were bred and maintained in an SOPF animal facility. All animal experiments were conducted in accordance with the French Ministry for Education and Research, regarding the care and use of animals for experimental procedures, under the 2022020111554986, and 2019031411197134 Apafis references, according the ARRIVE Guidelines. Mice were analysed between 8 and 12 weeks of age, regardless of the sex. Cohorts of controls and deficient mice were obtained from littermates. No statistical method was used to predetermine sample size and the experiments were not randomized. Unless otherwise stated, the investigators were not blinded to allocation during experiments and outcome assessment.

Method Details

Animal procedures

Cre-mediated recombination was achieved with a daily intraperitoneal injection of 1 mg of tamoxifen (Sigma), for 5 consecutive days. For H. polygyrus infection experiments, mice were inoculated orally with 200 infective L3. Infection parameters were monitored according to the numbers of faecal eggs, or numbers of intestinal adult worms from day 10 to day 40 post infection. Samples of luminal intestinal contents were obtained from naive or infected mice, according to the following surgery procedure. After 4 hours of fasting, mice received subcutaneously a single dose of buprenorphine (0.1 mg/kg body weight), and were anesthetized 30 min later with 3% isoflurane. After local lidocaine treatment, the abdominal cavity was opened, and an intestinal loop was ligatured. This intestinal loop was filled with up to 400 μl of a PBS wash solution containing 5 μM of Neostigmine Bromide (Sigma). The loop content was recovered 30 min later, after which mice were euthanized.

Collection of adults H. polygyrus and in vitro culture for treatment with cholinergic drugs

Adult H. polygyrus worms were collected from the intestinal lumen of infected mice at 14 dpi to evaluate ex vivo the effects of cholinergic drugs on egg release. Worms were washed with PBS containing penicillin (5U/ml)/streptomycin (5µg/ml) /gentamicin (1%), counted, and 20 to 25 female worms per well were incubated at 37°C, 5% CO2, in 200μL RPMI containing antibiotics. Adult H. polygyrus worms were then treated with the following drugs: acetylcholine chloride (Sigma, A6625) at 10μM, atropine (Sigma, A0132), and mecamylamine hydrochloride (Sigma, M9020) at 5μM. After 24h, culture media were mixed with a saturated NaCl solution (in a 50/50 volume ratio) to count the eggs. For global worm fitness evaluation, worms were collected from mice at 20 and 40 dpi, washed in PBS, and directly used as substrate for ATP level quantification, using the ATPLite kit (6016943; Perkin Elmer), according to the manufacturer’s instructions.

Cell sorting experiments

Single intestinal epithelial cells were obtained from small intestines after incubation in ice cold 30mM EDTA (Sigma) in HBSS pH 7.4 (Life Technologies) for 20 min. Tissues were then vigorously shaken in DMEM (Life Technologies) supplemented with 10% FBS (Sigma), with 100μl of Dispase (Corning), and 100μl of DNase I at 2,000 Kunitz (Sigma). After filtration on a 40 μm mesh, single cell suspensions were incubated with phycoerythrin rat anti-mouse Siglec-F antibody (BD Pharmigen, 552126), and FITC rat anti mouse EpCam antibody (17-5791-82, ebiosciences) for 30min at 4°C, and washed with HBSS and resuspended in appropriate volume of HBSS pH 7.4 supplemented with 5% FBS before staining with 7-aminoactinomycin D (Life Technologies) to exclude dead cells. Siglec-F+ live cells were sorted using a FACSAria (Becton Dickinson), directly in RLT lysis buffer (Qiagen) for subsequent RNA extraction, or methanol for ACh quantification assays. After removing epithelial cell fractions as described above, the remaining tissues were proceeded for lamina propria cell preparation. Tissues were washed twice with RPMI 1640 (Life technologies) supplemented with 5% FCS, and minced into small pieces and digested with a solution composed with RPMI 1640, 100μg/mL of Liberase TM (Roche) and 50μg/mL of DNase I (Sigma-Aldrich) for 30 minutes at 37°C under gentle shaking. Supernatants were filtered onto a 40μm mesh. After a 5 minutes centrifugation at 2000RPM, 4°C, cell pellets were resuspended in 500μL PBS-5% FCS-2mM EDTA for further stainings.

Immunophenotyping and flow cytometric analysis

Cells isolated from peripheral lymph nodes, mesenteric lymph nodes, spleen and intestinal tissue (lamina propria and intra-epithelial lymphocytes) were stained with Live/dead fixable viability dye (Ebioscience/ Thermofisher) together with the appropriate conjugated anti-TCRβ, CD45, CD19, CD4, CD8, CD11b, Ly6G, Ly6C, FcεRIa, F4/80, CD117 (eBioscience/Thermofisher or Becton Dickinson) at a 1:200 dilution (mAb), in a total volume of 50-100 μls as previously described. Cells were incubated in the dark for 20 minutes in PBS containing 2% FBS at 4°C and then washed once in the same medium at 300 g for 5 min prior to be evaluated by flow cytometry. For ILC staining, cells were stained with a lineage cocktail and lineage-negative CD45+ cells were assessed for expression of CD127, and intracellular expression of Gata-3 (clone L50-823), RoRγt and Tbx21 was performed following fixation/permeabilization (eBioscience/Thermofisher). Stained cells were assessed by flow cytometry (LSR Fortessa, Becton Dickinson, San Jose, CA) and a minimum of 10,000 events were recorded for each staining. Data analyses were performed using FlowJo software (Tree Star, Ashland, OR).

ACh quantification

Known numbers of sorted tuft cells (recovered in pure methanol) and intestinal lavage samples (0.2 mL, immediately mixed with 0.8 mL pure methanol), were kept at -80C° for 48h to allow protein precipitation. Next, 2-morpholinoethansulfonic acid (Cat. # 341-01622; Dojindo, Tokyo, Japan) was added as quality control (QC) standard at 0.5 μM final concentration (FC) and the samples were then supplemented with formic acid to 1 % FC. The samples were vortexed vigorously for 1 min and then centrifuged for 10 min at 20.000xg to precipitate proteins. A volume of 0.9 mL supernatant was then loaded on the Captiva EMR plate (Agilent, Santa Clara, USA, Cat.# 5190-1001) assembled on the Vacuum Manifold (Agilent, Cat.# A796) together with the Deep Well collection plate (Agilent, Cat.# A696001000). The flow-through fractions were first dried using a Speedvac concentrator and then resuspended in 200 μL of water. Calibration curve was established by diluting defined amounts of ACh in the matrix (generated by intestinal lavage of KO mice). Protein quantities from each sample were measured using the BCA kit (Thermo Scientific) according to manufacturer’s recommendations. Briefly, protein pellets were solubilized in 0.3 mL of 2 % sodium dodecylsulfate solution in water. Two μL of sample was then used to estimate the protein content of each sample.

One microliter of resuspended Captiva flow-through fraction or 1 μL of standard were injected on LC-MS systems consisting of UHPLC (Agilent, 1200 Infinity II Biocompatible) coupled to triple quadrupole MS (Agilent, 6495C). The samples were analysed using in MRMs acquisition mode. Following transitions were used to quantify ACh: 147.1 → 43.0, 147.1 → 88.1, 147.1 → 87.1 (positive mode). The transitions for the 2-morpholinoethansulfonic acid were: 196.2 → 100.0 (positive) and 194 → 80.15 (negative). The retention times were set as follows: 2.04 min and 8 min, for 2-morpholinoethansulfonic acid and ACh respectively. The analytical column was UPLC Discovery™ column HS F5-3 (Cat. # 567503-U; Sigma Aldrich). Mobile phases were composed as follows: A: 99.9 % water, 0.1 % formic acid (Sigma Aldrich, Cat. # 33015); B: 99.9 % acetonitrile (Biosolve BV, Cat.# 001204102BS), 0.1 % formic acid. The gradient was: 0 min (100 % A), 2 min (100 % A), 5 min (75 % A), 11 min (65 % A), 15 min (5 % A), 25 min (5 % A), 25.10 min (100 % A), 35 min (100% A). Flow rate was 0.25 mL/min and the temperature column was set to 40 °C. The following source parameters were used: Gas Temperature: 150 °C ; Gas Flow: 11 L/min ; Nebulizer: 40 psi ; Sheath Gas Temperature: 400 °C ; Sheath Gas Flow: 12 L/min ; Capillary Voltage (neg. mode): 4000 V ; Capillary Voltage (pos. mode): 4000 V ; Nozzle Voltage: 500 V ; IFunnel High Pressure RF Pos: 100 V Neg: 50 V ; IFunnel Low Pressure RF Pos: 100 V Neg: 50 V.

Following the analysis, peak integration was conducted using Agilent Masshunter Quantitative Analysis software (version 10.1.733.0). The absolute quantification of ACh was calculated as previously demonstrated, by applying the peak area from each sample to the ACh calibration curves ⁴⁶.

RNA extraction and PCR

Total RNAs from intestinal tissues were isolated using TRIzol (Life Technologies). RNeasy Micro and Mini Kit columns (Qiagen) were also used for RNA purification from cell-sorted experiments. Sorted-cell RNA were further amplified using the Arcturus RiboAmp Plus kit (ThermoFischer Scientific, KIT0501) according to the manufacturer’s instructions. Reverse transcription was performed with 1 μg of purified RNA using Transcriptor First Strand cDNA synthesis KIT (Roche) according to the manufacturer’s instructions. Real-time quantification was performed in triplicate with a LightCycler480 (Roche) using LightCycler 480 SYBR Green I Master (Roche) on 5ng of RT product using the average Ct of Gapdh and Hprt as internal loading controls, and the ΔΔCt method was used for calculating relative expression.

Fluorescent immunohistochemistry or in-situ hybridisation on paraffin-embedded tissue

Tissue dissection, fixation, and immunohistochemistry on thin sections of paraffin-embedded tissue were performed essentially as described previously ⁴⁷. Epitope retrieval was achieved by boiling in 10 mM in sodium citrate (pH 6.4) during 20 minutes. Primary antibodies used in this study were incubated ON at 4°C, and are listed in the key resource table. Slides were then washed twice with 0.1% PBS-Tween (Sigma-Aldrich) before incubation with fluorescent dyes-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, Inc.) and DAPI at 2 μg/ml (Sigma-Aldrich) in TBS–Triton X-100 0.1% (Sigma-Aldrich), or HRP-conjugated secondary antibodies, revealed with DAB (Sigma). Slides were mounted in FluoroMount (Sigma) or Pertex (Histolab), for fluorescent or visible imaging, respectively. For mRNA in situ hybridisation, tissues were hybridised with a probe targeting the Chat mRNA (Cat No. 408731, ACDBio), SucnR1 (Cat No 437721), and Ltc4s (Cat No 1046751). Slides were processed according to the manufacturer’s instructions until probe revelation, after which immunofluorescence detection of Dclk1 was performed following the methodology described above.

Microscopy and imaging

Fluorescent pictures were acquired at room temperature on an AxioImager Z1 microscope (Carl Zeiss, Inc.) equipped with a camera (AxioCam MRm; Carl Zeiss, Inc.), EC Plan Neofluar (5X NA 0.16; 10X NA 0.3; 20X 0.5 NA; 100X NA 1.3) and Plan Apochromat (40X NA 0.95; 63X NA 1.4) lenses, apotome Slider system equipped with an H1 transmission grid (Carl Zeiss, Inc.), and Zen software (Carl Zeiss, Inc.). Post-treatment of pictures (level correction), annotations, and panel composition were performed using the Photoshop (Adobe) or Zen (Carl Zeiss) softwares. Bright-field immunohistochemistry pictures were taken at room temperature on an Eclipse 80i microscope (Nikon) with Plan Fluor (10X NA 0.3; 20X NA 0.5; 40X NA 0.75; and 60X NA 0.5–1.25) lenses (Nikon) and a digital camera (Q-Imaging Retiga 2000R with a Q-Imaging RGB Slider), with Q-Capture Pro software (Nikon). Stained slides were also imaged with a Nanozoomer device (Hamamatsu), visualised and annotated with the NdpView Software (Hamamatsu).

Statistical analyses

The Prism software was used for descriptive statistical analyses. After a normality distribution test, datasets were analysed with either a two-tailed Student T test (normal distribution, pairwise comparisons) or with the non-parametric two-tailed Mann–Whitney U-test (non-normal distribution, pairwise comparisons). For multiple comparisons, datasets were proceeded with one-way ANOVA analyses (normal distribution, with Benjamini-Yekutieli post-hoc correction) or with Kruskal-Wallis analyses (non-normal distribution, with Benjamini-Yekutieli post-hoc correction). For some analyses, multiple Mann-Whitney analyses were performed. Depending on the normality distribution, datasets are represented in each graph as individual values and bars corresponding to mean ± SD (normal distribution) or either as individual values and bars corresponding to median ± interquartile (non-normal distribution). Details for statistical analyses are mentioned in each figure legend.