The Taste Receptor TAS1R3 Regulates Small Intestinal Tuft Cell Homeostasis

Mice

Wild-type (WT) C57BL/6J mice were bred and housed in microisolator cages in the specific-pathogen-free (SPF) barrier facility at the Harvard T.H. Chan School of Public Health and at Stanford University. BALB/c, Gfi1b^eGFP/+, Plcb2^−/−, and Stat6^−/− mice were obtained from The Jackson Laboratory (Bar Harbor, ME). C57BL/6J Tas1r3^−/− and Trpm5^−/− mice (10, 11) were generously provided by Dr. R. Margolskee (Monell Chemical Senses Center), and Gpr91^−/− mice were generously provided by Dr. J. Peti-Peterdi with permission from Amgen (12). All genetically modified mouse lines were bred on a C57BL6/J background and were cohoused for at least 2 wk prior to the experiments. Gfi1b^eGFP/+ Gpr91^−/−, Gfi1b^eGFP/+ Plcb2^−/−, Gfi1b^eGFP/+ Stat6^−/−, Gfi1b^eGFP/+ Tas1r3^−/−, and Gfi1b^eGFP/+ Trpm5^−/− mice were generated by breeding Gfi1b^eGAQFP/+ mice to the relevant knockout mice. Germ-free (GF) WT C57BL/6J mice were bred and maintained in semirigid gnotobiotic isolators in the Harvard T. H. Chan Gnotobiotic Center for Mechanistic Microbiome Studies. Mice were used experimentally between 6 and 12 wk of age. Animal studies and experiments were approved and carried out in accordance with Harvard Medical School’s Standing Committee on Animals, Stanford’s Institutional Animal Care and Use Committee, and the National Institutes of Health guidelines for animal use and care.

Infection with protozoa and helminths

T. muris was isolated and cultured from bred-in-house mice as described previously (1). A total of 5 × 10⁶ T. muris were orally administered to mice. Mice were then sacrificed 16–18 d postinfection.

H. polygyrus maintenance and infection was performed as previously described (13). Mice were orally infected with 200 L3 larvae and sacrificed 6 wk later. The proximal 15 cm of small intestine was excised, and worms were counted with a dissection microscope.

T. muris enumeration

T. muris load in the distal small intestine was enumerated as previously described (1). The distal 10 cm of small intestine was removed and flushed with ice-cold sterile PBS using a 19-gauge feeding needle. The intestinal contents were then pelleted by centrifugation and stored at −20°C. Genomic DNA was isolated from the stool with QIamp Fast DNA Stool Mini Kit (Qiagen) according to the manufacturer’s directions. To detect and enumerate T. muris, quantitative PCR (qPCR) was performed using KAPA SYBR Fast Universal PCR (KAPA Biosystems) with the following primers recognizing the T. muris 28S rRNA gene: 59-GCTTTTGCAAGCTAGGTCCC-39and 59-TTTCTGATGGGGCG TACCAC-39. These qPCR values were converted to T. muris numbers using a standard curve generated using known amounts of T. muris.

Metabolite/taste receptor ligand feeding

For succinate treatment, mice were provided with 100 mM succinic acid disodium salt (no. 224731; Sigma-Aldrich) in the drinking water for 1, 2, or 4 wk as indicated. Two hundred millimolar sodium chloride was used as a control to match sodium molarity. Sucralose was fed at concentrations of 2, 20, or 100 mM in the drinking water, and rebaudioside A was fed at a concentration of 1 mM for 8 d.

IL-25 injection

Mice were i.p. injected daily with 0.5 μg of rIL-25 (endotoxin level, 1.0 endotoxin unit per 1 μg; R&D Systems) or equivalent volume of sterile PBS for 7 d before harvesting the distal small intestine.

Epithelial cell isolation and flow cytometry

The small intestine and/or colon were removed and flushed with ice-cold sterile PBS using a 19-gauge feeding needle. The small intestine was further subdivided into the duodenum (proximal 7 cm), jejunum (next 10 cm), and ileum (distal 10 cm), and Peyer patches were removed. Both small intestinal regions and colons were then opened longitudinally and gently agitated at 4°C in PBS, 2% FBS, 5 mM HEPES, and 1 mM DTT for 10 min. The tissue was then transferred into prewarmed PBS, 2% FBS, 5 mM HEPES, and 5 mM EDTA and rotated at 37°C for 15 min followed by vigorous shaking to remove epithelial cells. This was repeated and epithelial cells from both fractions were combined and washed with PBS. The epithelium was then digested in DMEM containing 10% FBS, 0.5 U/ml Dispase II (StemCell Technologies), and 50 μg/ml DNase (Roche) for 10 min at 37°C. The resulting solution was passed through 40-μm filters and washed with PBS, 2% FBS, and 1 mM EDTA. The resulting single-cell suspension was initially Fc blocked with anti-CD16/CD32 (clone 93; BioLegend) and then stained with the following Abs: PacBlue-conjugated anti-CD45 (clone 30-F11; BioLegend) and allophycocyanin-conjugated anti-EpCam (clone G8.8; BioLegend). Cell viability was assessed by propidium iodide (BioLegend) staining.

Histology and fluorescence microscopy

The small intestine was removed and divided into proximal and distal sections before overnight fixation in 4% paraformaldehyde. The tissue was then embedded in paraffin and cut into 5-μm thick sections. For both histology and immunofluorescence, sections were initially deparaffinized and rehydrated. Goblet cells were identified by Alcian blue/nuclear red staining and enumerated along the crypt-villus axis by quantitative microscopy. For immunofluorescence, heat-mediated Ag retrieval was performed in Tris-EDTA buffer 0.05% Tween-20 (pH 9) for 20 min. Afterwards, the slides were washed in PBS and blocked in PBS containing 3% BSA, 3% donkey serum, 0.1% Triton X-100, and 0.1% saponin for 1 h at room temperature. Primary Abs were incubated overnight at 4°C and secondary Abs were applied for 1.5 h at room temperature. Primary Abs included the following: rabbit anti-DCLK1 (1:250 dilution, ab37994; Abcam) and mouse anti–E-cadherin (1:400 dilution, 36/E-cadherin; BD Biosciences). DNA was labeled with DAPI (0.5 μg/ml).

RNA isolation and RT-PCR

Epithelial cells from Gfi1b^eGFP/+ mice were isolated and stained for FACS as previously detailed. Tuft cells were sorted as GFP⁺EpCam⁺CD45⁻PI⁻, whereas the remaining epithelial cells were GFP⁻EpCam⁺CD45⁻PI⁻. RNA was then extracted fromtuft cells and the remaining epithelium using RNeasy Micro Kit (Qiagen). For RNA isolation from total epithelium, the epithelial fraction was collected after the EDTA wash (as described above) and lysed in Qiazol (Qiagen) for RNA extraction following manufacturer’s instructions. cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad), and quantitative RT-PCR was performed using the PowerUp SYBR Green Master Mix qPCR Kit (Applied Biosystems). The following primers were used: Dclk1, 5′-CAGCCTGGACGAGCTGGTGG-3′, 5′-TGACCAGTTGGGGTTCACAT-3′; Trpm5, 5′-CCTCCGT GCTTTTTGAACTCC-3′, 5′-CATAGCCAAAGG TCGTTCCTC-3′; Klf4, 5′-ATCCTTTCCAACTCGCTAACCC-3′, 5′-CGGATCGGATA GCTGAAGCTG-3′; Muc2, 5′-CCATTGAGTTTGGGAACATGC-3′, 5′-TTCGGCTCGGTGTTCAGAG-3′; Chga, 5′-CAGGCTACAAAG CGATCCAG-3′, 5′-GCCTCTGTCTTTCCATCTCC-3′; Tph1, 5′-AACAAAGACCATTCCTCCGAAAG-3′, 5′-TGTAACAGG CTCACATGATTCTC-3′; Def20, 5′-TGTAGAAAAGGAGGC TGCAATAG-3′, 5′-AGAACAAAAGTCGTCCTGAGC-3′; Lyz1, 5′GCCAAGGTCTACAATCGTTGTGAGTTG-3′, 5′-CAGTCA GCCAGCTTGACACCACG-3′; Lgr5, 5′-CCTACTCGAAGACT TACCCAGT-3′, 5′-GCATTGGGGTGAATGATAGCA-3′; Ascl2, 5′-GCCTACTCGTCGGAGGAA-3′, 5′-CCAACTGGAAAA GTCAAGCA-3′; Gapdh, 5′-CCTCGTCCCGTAGACAAAATG-3′, 5′-TCTCCACTTTGCCACTGCAA-3′;Tas1r3, 5′-AGGTGGCT CACAGTTCTGCT-3′, 5′-GAGGTGAGCCATTGGTTGTT-3′.

Data are presented as relative expression normalized to Gapdh.

Small intestine organoid culture and flow cytometry

Distal small intestinal organoids were prepared as previously described (14). When indicated, IL-13 (10 ng/ml, endotoxin level < 0.01 ng/μg; BioLegend) was added to organoid media for 48 h. To perform flow cytometry, organoids were liberated from the matrigel matrix as described (14) and digested in DMEM containing 10% FBS, 0.5 U/ml Dispase II (StemCell Technologies), and 50 μg/ml DNase (Roche) for 8 min at 37°C. The resulting solution was filtered through 40-μm mesh and stained for flow cytometry with allophycocyanin-conjugated anti-EpCam (clone G8.8; BioLegend) with cell viability assessed with propidium iodide (BioLegend).

Statistics

The Mann–Whitney U test was used to compare two samples. One-way ANOVA was used for multiple-group comparisons followed by Holm–Sidak post hoc testing. The p values are indicated in the figure legends.

RNA sequencing

C57BL6/J and BALB/c mice differ in tuft cell response to the protozoa T. muris but not the helminth H. polygyrus

Given the critical role of chemosensation in tuft cell–mediated antiparasite immunity in the gut, we considered if inbred mouse strains with their well-characterized genetics and immune responses to parasites could be used to further elucidate the potential connection between parasite surveillance and taste receptors. Inbred mouse strains such as BALB/c and C57BL/6J can differ in their immune responses to parasites, which results in marked differences in clearing these infections (19, 20). Furthermore, BALB/c and C57BL/6J mice carry genetic polymorphisms in the taste receptor Tas1r3, with C57BL/6J having a high preference for the TAS1R3 ligand saccharin and BALB/c having a low preference (21, 22). To investigate if these differences between mouse strains could alter the tuft cell response to parasites, we colonized both C57BL/6J and BALB/c mice with the protozoa T. muris. T. muris induced a significant increasein tuft cell frequency at the site of colonization, the distal small intestine (ileum) (1), of C57BL6/J mice, which contrasted with an absence of tuft cell expansion in T. muris–colonized BALB/c mice (Fig. 1A, 1B). This difference in tuft cell responses was not due to variation in T. muris colonization, as protozoal loads were similar between C57BL/6J and BALB/c mice (Fig. 1C).

Intestinal helminth–induced immunity is also accompanied by a significant increase in tuft cell frequency (1–3). Consistent with these prior results, BALB/c and C57BL/6J mice both showed robust expansion of tuftcellsinthe proximal small intestineduring H. polygyrus infection (Fig. 1D, 1E), confirming that BALB/c mice can generate tuft cells in response to some parasitic stimuli. To determine if the defect in response to T. muris was specific to the protozoa or a more general feature of the ileum, we injected BALB/c mice with rIL-25. This cytokine bypasses microbial signal–dependent pathways to directly expand tuft cells via induction of IL-13 production by ILC2s (1–3). IL-25 injection resulted in tuft cell expansion in the distal small intestine (Supplemental Fig. 1), suggesting that the lack of BALB/c response to T. muris was not due to an inability to increase tuft cells in the ileum.

Tuft cells express the type 1 taste receptor TAS1R3, which affects their response to expansion stimuli

To understand the mechanism underlying the lack of tuft cell response to T. muris in BALB/c mice, we returned to the observation that polymorphisms in Tas1r3 severely reduce the sensitivity of this receptor in BALB/c mice (21, 22). In lingual taste, TAS1R3 activates a signal transduction pathway involving PLCb2 and TRPM5, which are chemosensory components used by tuft cells to respond to T. muris (1). Previous reports found that tuft cells in the stomach, trachea, thymus, gallbladder, and urethra express Tas1r3 (6, 23, 24), but the level of Tas1r3 expression by intestinal tuft cells was unclear. We sorted ileal tuft cells from C57BL/6J mice and used RNA sequencing to measure the expression of type 1 and type 2 taste receptors, which signal through the transduction components PLCβ2 and TRPM5. Tas1r3 had the highest expression, whereas most other receptor genes were not detected (Fig. 2A). Quantitative RT-PCR analysis confirmed this result (Fig. 2B), consistent with single-cell sequencing analysis of small intestinal epithelial cells that identified Tas1r3 as a tuft cell marker (25).

We hypothesized that if the inactive BALB/c TAS1R3 contributed to impaired BALB/c responses to T. muris, then Tas1r3^−/− C57BL/6J mice should recapitulate this lack of response. Indeed, tuft cell frequency in Tas1r3^−/− mice did not significantly increase after T. muris colonization (Fig. 3A, 3B). Similar to what we observed in BALB/c mice, this defect was not due to lower T. muris colonization (Fig. 3C). Furthermore, the effect of TAS1R3 was restricted to T. muris, as Tas1r3^−/− mice had a similar ability to expel H. polygyrus as WT mice, whereas Trpm5^−/− mice had an increased worm burden as previously described (Fig. 3D) (1).

Although tuft cell frequency remained unchanged in the majority of Tas1r3^−/− mice after T. muris colonization, we noted that a subset of mice exhibited a strong tuft cell response. This bimodal response was not observed in mice lacking other tuft cell–regulating components, including PLCβ2, STAT6, (a transcription factor downstream of IL-13R signaling), and GPR91 (a succinate receptor reported to be involved in sensing of Tritrichomonas species) (6, 7) (Fig. 3A).These observations led us to question whether TAS1R3 regulates tuft cell expansion through sensing of a T. muris–associated signal, such as succinate, or potentially another mechanism.

Feeding mice succinate in the drinking water induces a significant expansion of tuft cells in the distal small intestine, which highly resembles the tuft cell response to Tritrichomonas colonization (6–8). WT mice fed 100 mM succinate in the drinking water for 1 wk exhibited tuft cell hyperplasia (Fig. 3E). However, similar to Gpr91^−/− and Plcb2^−/−mice, feeding succinate to Tas1r3^−/− mice for 1 wk failed to increase tuft cell levels in the distal small intestine (Fig. 3E). To assess the kinetics of tuft cell response tosuccinate, we also fed mice 100 mM succinate for 2 and 4 wk. In contrast to 1 wk of succinate feeding, these longer succinate exposures increased the proportion of Tas1r3^−/− mice with enhanced tuft cell abundance. These findings suggest that TAS1R3 is not required to sense succinate but may act to increase the response kinetics of tuft cells to expansion stimuli.

TAS1R3 regulates homeostatic tuft cell abundance

Tuft cells function in a feed-forward loop in which tuft cell–derived IL-25 leads to increased IL-13 production by ILC2s and expansion of Th2 cells, which results in additional tuft cells and further production of IL-25. Therefore, we hypothesized that the response kinetics of tuft cells to a stimulus would be affected by the number of tuft cells that are available to initiate the feed-forward loop. To determine if TAS1R3 influences tuft cell frequency at steady-state, we profiled tuft cell populations across different regions of the gut and found that Tas1r3^−/− mice had fewer tuft cells compared with WT, with the most pronounced defect in the ileum (Fig. 4A). We further verified the relative reduction of tuft cells in the ileal epithelium of Tas1r3^−/− animals by measuring RNA expression of the tuft cell markers Dclk1 and Trpm5 and by immunofluorescence microscopy, which revealed markedly fewer DCLK1⁺ tuft cells in the absence of TAS1R3 (Fig. 4D, Supplemental Fig. 2). Because IL-13 stimulates epithelial progenitor cells to differentiate into both tuft cells and goblet cells (1–4), we also examined the number of goblet cells in the ileum of WT and Tas1r3^−/− mice at steady-state but found no significant difference (Fig. 4B, 4C). We confirmed that other epithelial populations remained largely unchanged in Tas1r3^−/− mice by measuring the expression of representative gene markers for intestinal epithelial cell subsets (25) (Fig. 4D). One notable exception was the stem cell–regulating gene Ascl2. Although Ascl2 is a signature gene associated with intestinal epithelial stem cells, it is also expressed by tuft cells (26) and therefore the decrease in Ascl2 expression in the epithelium of Tas1r3^−/− mice may reflect the reduced number of tuft cells rather than changes in the stem cell population. Therefore, we surmised that TAS1R3 is required for maintenance of homeostatic levels of tuft cells and that the few ileal tuft cells present in Tas1r3^−/− mice cannot trigger the IL-25–and IL-13–dependent feedback loop required for expansion with the same kinetics seen in WT mice.

To understand how TAS1R3 regulates intestinal tuft cells, we next measured the expression of Tas1r3, Dclk1, and Trpm5 in epithelial cells harvested from the duodenum, jejunum, ileum, and colon of WT mice. Although the expression of the tuft cell markers Dclk1 and Trpm5 was relatively consistent across the small intestine, Tas1r3 expression was elevated in the ileum (Fig. 5). This pattern of Tas1r3 expression inversely correlated with the severity of the tuft cell defect in Tas1r3^−/− mice. To identify intestinal TAS1R3 ligands that alter tuft cell frequency, we generated organoids from the ileal epithelium of Gfi1b^eGFP/+ and Gfi1b^eGFP/+ × Tas1r3^−/− mice. In this system, TAS1R3 did not affect baseline tuft cell abundance or the IL-13–induced increase in tuft cell frequency (Fig. 6A). This observation led us to consider potential sources of TAS1R3 ligands in vivo. The microbiota is a rich source of ligands for host receptors, so we compared the number of tuft cells in the ileum of conventionally raised SPF WT and Tas1r3^−/− mice with GF mice. GF mice had a similar frequency of tuft cells as conventional WT mice but significantly more tuft cells than Tas1r3^−/− mice (Fig. 6B), suggesting that the microbiota is not an important source of TAS1R3 ligands in the ileum.

In lingual taste cells, TAS1R3 forms a heterodimer either with TAS1R1 to detect umami flavor or TAS1R2 to detect sweet flavor (27). However, in the ileum, we observed considerably lower expression of Tas1r1 and Tas1r2 compared with Tas1r3 (Fig. 2A), suggesting that many common sweet- and umami-taste agonists may not stimulate intestinaltuft cells. Therefore, tomimic adietary agonist of TAS1R3, we fed mice the artificial sweetener sucralose, which has been reported to target TAS1R3 directly (28). After feeding mice 2, 20, or 100 mM sucralose in the drinking water for 8 d, we did not observe a significant change in tuft cell abundance in either the duodenum or the ileum (Supplemental Fig. 3). Similarly, rebaudioside A, a steviol glycoside that potentiates TRPM5 signaling (29), did not increase tuft cell frequency (Supplemental Fig. 3). Because the TAS1R3 receptor putatively activates the canonicaltastesignal transductionpathway involving PLCβ2, we measured the frequency of tuft cells in the ileum of Plcb2^−/− mice at homeostasis. Surprisingly, there were significantly fewer tuft cells in Tas1r3^−/− mice than Plcb2^−/− mice (Fig. 6C). This defect was not observed in mice lacking another tuft cell–regulating GPCR, GPR91. Additionally, there were fewer tuft cells in Tas1r3^−/− than in Stat6^−/− mice (Fig. 6C). Together, these findings imply that TAS1R3 may not regulate tuft cell abundance through currently established taste chemosensory pathways or through IL-13/IL-4 activation of STAT6 signaling.