Human-derived gut microbiota modulates colonic secretion in mice by regulating 5-HT₃ receptor expression via acetate production

We found that gut microbiota alters serotonin (5-HT)-evoked intestinal secretion in a 5-HT₃ receptor-dependent mechanism and gut microbiota metabolite acetate alters 5-HT₃ receptor expression in colonoids.

Abstract

Serotonin [5-hydroxytryptamine (5-HT)], an important neurotransmitter and a paracrine messenger in the gastrointestinal tract, regulates intestinal secretion by its action primarily on 5-HT₃ and 5-HT₄ receptors. Recent studies highlight the role of gut microbiota in 5-HT biosynthesis. In this study, we determine whether human-derived gut microbiota affects host secretory response to 5-HT and 5-HT receptor expression. We used proximal colonic mucosa-submucosa preparation from age-matched Swiss Webster germ-free (GF) and humanized (HM; ex-GF colonized with human gut microbiota) mice. 5-HT evoked a significantly greater increase in short-circuit current (ΔI_sc) in GF compared with HM mice. Additionally, 5-HT₃ receptor mRNA and protein expression was significantly higher in GF compared with HM mice. Ondansetron, a 5-HT₃ receptor antagonist, inhibited 5-HT-evoked ΔI_sc in GF mice but not in HM mice. Furthermore, a 5-HT₃ receptor-selective agonist, 2-methyl-5-hydroxytryptamine hydrochloride, evoked a significantly higher ΔI_sc in GF compared with HM mice. Immunohistochemistry in 5-HT_3A-green fluorescent protein mice localized 5-HT₃ receptor expression to enterochromaffin cells in addition to nerve fibers. The significant difference in 5-HT-evoked ΔI_sc between GF and HM mice persisted in the presence of tetrodotoxin (TTX) but was lost after ondansetron application in the presence of TTX. Application of acetate (10 mM) significantly lowered 5-HT₃ receptor mRNA in GF mouse colonoids. We conclude that host secretory response to 5-HT may be modulated by gut microbiota regulation of 5-HT₃ receptor expression via acetate production. Epithelial 5-HT₃ receptor may function as a mediator of gut microbiota-driven change in intestinal secretion.

NEW & NOTEWORTHY We found that gut microbiota alters serotonin (5-HT)-evoked intestinal secretion in a 5-HT₃ receptor-dependent mechanism and gut microbiota metabolite acetate alters 5-HT₃ receptor expression in colonoids.

View this article's corresponding video summary at https://www.youtube.com/watch?v=aOMYJMuLTcw&feature=youtu.be.

serotonin [5-hydroxytryptamine (5-HT)] is an important neurotransmitter and a potent secretagogue in the gastrointestinal (GI) tract (2, 29). The majority (90%) of the body’s 5-HT is derived from hydroxylation of the essential dietary amino acid l-tryptophan in the enterochromaffin (EC) cells (10). 5-HT released from the EC cells activates multiple 5-HT receptor subtypes to modulate GI functions including peristalsis (8), intestinal secretion (26, 29, 36), and sensation (9, 22). Seven different families of 5-HT receptors and at least 14 pharmacologically distinct 5-HT receptor subtypes have been described (1). Among these receptor subtypes, 5-HT₃ and 5-HT₄ receptors play an important role in regulating intestinal secretion (3, 13, 40). 5-HT₃ receptors are the members of the Cys-loop ligand-gated ion channel family (4). In the GI tract, 5-HT₃ receptors are located on multiple cell types including submucosal neurons, interstitial cells of Cajal, and few enteroendocrine cells including the EC cells (11). 5-HT₄ receptors, on the other hand, are G protein-coupled receptors that stimulate cAMP production (30). 5-HT₄ receptors are localized in the enteric neurons (35), smooth muscles (32), and epithelial cells including the EC cells (15). Effects of 5-HT mediated secretion in Ussing chamber involve activation of both neuronal and nonneuronal 5-HT₃ and 5-HT₄ receptors (6, 26, 31).

The gut lumen is home to trillions of microorganisms that include over 1,000 different bacterial species (16). However, the dynamic interaction between the gut microbiota and the host and how it modulates GI physiology are largely unknown. We have recently shown that human gut microbiota increases host 5-HT biosynthesis through an effect of short-chain fatty acid (SCFA; 34) and modulates GI transit in part via 5-HT₃ and 5-HT₄ receptors (20). Since 5-HT₃ and 5-HT₄ receptors are important mediators of intestinal secretion, in this study we investigated the physiological effects of microbial colonization on ex vivo host secretory response to exogenously applied 5-HT and on 5-HT₃ and 5-HT₄ receptor expression.

MATERIALS AND METHODS

Animals.

Swiss Webster germ-free (GF) breeders, used to produce the GF and the humanized (HM; ex-GF colonized with human gut microbiota) mice in this study, were obtained from Taconic Farms (Germantown, NY) and maintained in gnotobiotic isolators in the Mayo Clinic gnotobiotic facility. Humanization of GF mice was done by colonizing GF mice with 300 µl of prereduced PBS and fecal microbiota (a 1:1 suspension) from a healthy human donor (a 49-yr-old man) at 4–5 wk of age by gavage. All HM mice were colonized with human microbiota for 4–5 wk before experiments. GF and HM male mouse tissues were collected between 10 and 12 wk. Age-matched conventionally raised (CR) mice were purchased from Taconic Farms and maintained outside the isolator. All studies were conducted after approval from Mayo Clinic Institutional Animal Care and Use Committee. Tissue from 5-HT_3A-green fluorescent protein (GFP) mice was obtained from University of Colorado, where conventionally housed FVB/N-Swiss Webster mice [Stock Tg(Htr3a-EGFP)DH30Gsat/Mmnc; RRID:IMSR_MMRRC:000273] were crossed with C57BL/6 line for two to four generations. Mice were intraperitoneally injected with 5-hydroxy-l-tryptophan (50 mg/kg; Sigma-Aldrich, St. Louis, MO) 1 h before euthanasia to increase 5-HT levels.

Tissue preparation.

Mice were asphyxiated with CO₂, and euthanasia was confirmed by cervical dislocation. Proximal colon segments 2–3 cm below the cecum were removed and opened along the mesenteric border. The tissues were pinned flat on the Sylgard with mucosa side facing up. The luminal contents were then washed with chilled Krebs solution (composition in mM/l: 11.5 d-glucose/d-mannitol, 120.3 NaCl, 15.5 NaHCO₃, 5.9 KCl, 1.2 NaH₂PO₄, 2.5 CaCl₂·2H₂O, and 1.2 MgCl₂; pH 7.3–7.4). The colon segments were then flipped, and the underlying muscle layer was peeled away under a stereomicroscope with fine forceps. A 0.2-cm² segment of mucosa-submucosa sheet was stored in RNAlater Stabilization Solution (Ambicon) for 1 h at 4°C and then moved to −20°C until ready for real-time PCR analysis. The remaining larger mucosa-submucosa epithelial sheets were used for Ussing chamber studies.

Colonoid preparation.

GF mice were asphyxiated with CO₂, and euthanasia was confirmed by cervical dislocation. Entire colon segment below the cecum was removed in a sterile manner and collected in a cold, sterile Ca²⁺-Mg-free PBS (21-040-CV; Corning). The colon was cut along the mesenteric border, and the pellet was cleaned using a 10-ml sterile syringe filled with sterile PBS. With the use of sterile scissors, the colon segment was minced into ~1-mm² pieces. The tissue chunks were then washed with cold Ca²⁺-Mg-free PBS until the supernatant was clear of any pellet debris. The clean tissue pieces were incubated in gentle cell dissociation media (7174; Stem Cell Technologies) for 15 min and rinsed vigorously using a 10-ml Pasteur pipette filled with freshly prepared filter-sterilized 0.1% PBS-BSA solution. The solution was then filtered through 70-µm nylon mesh (22363548; Fisher Scientific), and the filtrate-containing crypt cells were centrifuged in a 15-ml conical tube at 290 g at 4°C for 5 min. The centrifuged cell pellets were incubated in 6–7 ml of DMEM-F12 media (11330-032; Life Technologies) and recentrifuged at 290 g at 4°C for 5 min. The cell pellets were then broken down and mixed with 150 µl of IntestiCult Organoid Growth Medium (IOGM; 06000; Stem Cell Technologies) and 150 µl of reduced Matrigel (1:1 ratio, 356231; Corning). Seventy-five microliters of cell-Matrigel mixture were plated in a warm 24-well plate and supplemented with 750 µl (1:10 of Matrigel dome-IOGM mix) of warm IOGM. The colonoids were allowed to grow for 1 wk and passaged four times before they were incubated with sodium acetate (S2889; Sigma-Aldrich) and sodium butyrate (B5887; Sigma-Aldrich). Three technical replicates from five mice were plated uniformly in a warm 24-well plate. Quantitative real-time (qRT)-PCR was performed as described in materials and methods, Quantitative real-time PCR for 5-HT3 and 5-HT4 receptor mRNA expression.

Ussing chamber for measuring in vitro epithelial transport.

To measure 5-HT-induced in vitro epithelial transport, a proximal colon section 1–2 cm below the cecum was mounted in an Ussing chamber (Physiologic Instruments, San Diego, CA) with an aperture of 0.31 cm². The chamber on the submucosal side was bathed with 4 ml of glucose-containing Krebs solution while the mucosal side was bathed with 4 ml of mannitol Krebs solution. The Krebs solution was bubbled with 97% O₂ and 3% CO₂ mixture. The tissues were short circuited by a voltage-clamp at zero potential using Ag/AgCl electrode that was inserted into a pipette tip filled with a salt-filled agar bridge (P2020-S; Physiologic Instruments) and connected to a voltage-clamp amplifier (VCC MC8; Physiologic Instruments). Transepithelial resistance (TER) was measured and averaged over a period of half an hour before application of drugs. The final TER (Ω·cm²) was calculated by multiplying the sample resistance with effective membrane area (0.31 cm²). Baseline short-circuit current (I_sc) and change in short-circuit current (ΔI_sc) were continuously recorded using Acquire & Analyze software (Physiologic Instruments). ΔI_sc values were measured after application of drugs to the submucosal side. ΔI_sc values were normalized to the tissue area for final calculations.

Quantitative real-time PCR for 5-HT₃ and 5-HT₄ receptor mRNA expression.

To detect whether there were changes in 5-HT₃ and 5-HT₄ receptor mRNA from mice proximal colon mucosa-submucosa preparation and in SCFA-treated GF colonoids, qRT-PCR was used. RNA was extracted and isolated using a tissue homogenizer and the RNeasy Mini Kit (Qiagen, Valencia, CA) with DNase I (Qiagen) treatment. RNA isolates were reverse transcribed with the SuperScript VILO cDNA Synthesis Kit (Invitrogen, Carlsbad, CA). The cDNA thus obtained was used for real-time PCR analysis. The qRT-PCR assays were performed in 25-µl reactions containing 1× LightCycler 480 SYBR Green I Master (Roche, Basel, Switzerland) and 300-nM gene-specific primers (100 nM for housekeeping gene). Samples were analyzed in triplicate using a LightCycler 480 Instrument II (Roche). All cDNA synthesis reactions were set up with additional samples lacking reverse transcriptase as a control for genomic DNA contamination. Total mRNA expression was calculated after normalization to L32 mRNA (housekeeping) using the comparative cycle threshold (ΔΔC_T) analysis method. Relative mRNA expression (fold difference) was calculated on the basis of the mRNA expression in humanized control after normalization. 5-HT₃ and 5-HT₄ primer sequences were synthesized at the Integrated DNA Technologies facility at Mayo Clinic (Rochester, MN). The primer sequences are listed in Table 1.

Table 1.

5-HT₃ and 5-HT₄ primer sequences

Sequence No.	Receptor	Primer Sequence
1	5-HT3B	Forward: 5′-CTGCATGGCCCTCTTGGTT-3′
		Reverse: 5′-AATCTAGACTCCTCGGCATC-3′
2	5-HT4	Forward: 5′-AGTTCCAACGAGGGTTTCAGG-3′
		Reverse: 5′-CAGCAGGTTGCCCAAGATG-3′
3	L32	Forward: 5′-CCTCTGGTGAAGCCCAAGATC-3′
		Reverse: 5′-TCTGGGTTTCCGCCAGTTT-3′

Western blot analysis for 5-HT₃ subunit protein expression.

Mucosa-submucosa from the proximal colon was quickly removed in cold Krebs solution to prevent protein degradation. Tissues were stored at −80°C until ready to be processed and homogenized. Homogenized tissues were lysed with 1× lysis buffer (9803s; Cell Signaling Technology) that contained protease inhibitor cocktail (539131; Calbiochem). Protein extracts (50 µg per lane) were loaded for SDS-PAGE along with the full-range rainbow molecular weight marker (MagicMark, LC5602; Invitrogen). Proteins were separated at 105 V on a 12% Mini-PROTEAN TGX gel (456-1044s; Bio-Rad) and electrotransferred for 45 min onto a polyvinylidene difluoride membrane (IPVH00010; Millipore) at 15 V. Tissues were later blocked for 1 h at room temperature with 5% milk in 1× Tris-buffered saline (170-6435; Bio-Rad) and 0.1% Tween 20 (TBS-T; 9005-64-5; Fisher Bioreagents) to prevent nonspecific binding. Membranes were then incubated overnight at 4°C with primary antibodies 5-HT_3B (1:1,600, PA5-41033; Invitrogen) and GAPDH (1:10,000, G9545; Sigma-Aldrich). Membranes were then washed (3 × 10 min) with TBS-T and blotted with the horseradish peroxidase-conjugated secondary antibodies (anti-5-HT_3B, 1:800; anti-GAPDH, 1:2,000) for 1 h at room temperature in blocking solution. Blots were detected with enhanced chemiluminescence (ECL) Western Blotting Substrate (33209; Thermo Fisher Scientific) and analyzed with ImageJ software (version 1.50i; National Institutes of Health). Relative protein abundance (fold difference) was determined relative to HM group after normalization to GAPDH.

Immunohistochemistry for 5-HT₃ receptor expression.

Proximal colon tissues 2–4 cm below the cecum were harvested from conventionally housed 5-HT_3A-GFP mice immediately following perfusion fixation and postfixed with 4% paraformaldehyde for 4 h. Tissues were cryoprotected in 20% sucrose and shipped, on ice, to Mayo Clinic for embedding in optimum cutting temperature compound. Cryostat sections were cut at 10 μm. The sections were washed (3 × 10-min intervals) with PBS and permeabilized with 0.1% Triton X-100. Tissues were blocked with 5% normal donkey serum and 0.5% BSA in PBS for 30 min at room temperature and immunoreacted with chicken anti-GFP antibody (1:500 dilution, A10262; Life Technologies) and goat anti-5-HT antibody (1:2,000 dilution, AB66047; Abcam) for 2 h at room temperature. Tissues were then washed (2 × 5-min intervals) with PBS. Antibody binding was visualized using Alexa Fluor 488-conjugated donkey anti-chicken antibody (1:500 dilution, 703-545-155; Jackson ImmunoResearch) and Alexa Fluor 594-conjugated donkey anti-goat antibody (1:500 dilution, 705-585-147; Jackson ImmunoResearch), respectively. Slides were mounted with a 1:1 mixture of Citifluor AF1 (17970-25; VWR) and SlowFade Gold Antifade Mountant with 4′,6-diamidino-2-phenylindole (DAPI; S36938; Thermo Fisher Scientific). Fluorescent images were obtained using an epifluorescence Olympus (BX51WI) microscope using ×20 and ×40 objectives.

Statistics.

Data are presented as means ± SE. Potential differences were tested either using ANOVA or Student’s t-test as appropriate. P < 0.05 was considered significant. Analysis was done in GraphPad Prism 6.0 software (San Diego, CA), and n stands for number of animals from which tissues were harvested.

Drugs.

The following drugs were used: serotonin (5-HT), forskolin (10 µM), ondansetron (1 µM), GR-113808 (100 µM), 2-methyl-5-hydroxytryptamine hydrochloride (10 µM), and tetrodotoxin (TTX; 500 nM). 5-HT hydrochloride was obtained from Sigma-Aldrich (CAS no. 153-98-0). Forskolin was obtained from Sigma-Aldrich (CAS no. 66575-29-9). Ondansetron hydrochloride (CAS no. 99614-01-04), GR-113808 (CAS no. 144625-51-4), and 2-methyl-5-hydroxytryptamine hydrochloride (CAS no. 845861-44-6) were all obtained from Tocris Bioscience (Bristol, United Kingdom). Drugs were dissolved and diluted in distilled water.

RESULTS

Microbial colonization does not affect intestinal permeability and baseline secretion.

To test whether microbial colonization affects tissue integrity, baseline secretory response, and forskolin-evoked secretory response, we measured TER, baseline I_sc, and forskolin-induced ΔI_sc in GF and HM mice. We found that there was no significant difference in TER (47.5 ± 3.20 vs. 48.4 ± 2.95 Ω·cm², n = 10, unpaired t-test; P > 0.05; Fig. 1A), baseline I_sc (94.28 ± 12.47 vs. 77.16 ± 12 µA/cm², n = 10, unpaired t-test; P > 0.05; Fig. 1B), or forskolin-evoked ΔI_sc response (269 ± 20.9 vs. 247 ± 17.6 µA/cm², n = 5, unpaired t-test; P > 0.05; Fig. 1C) between GF and HM mice.

Fig. 1.

Microbial colonization does not affect intestinal permeability and baseline secretion. TER (A), baseline I_sc (B), and forskolin (10 µM)-evoked ΔI_sc (C) in GF and HM mice.

Host secretory response to exogenously applied 5-HT is increased in the absence of gut microbiota.

Since we did not observe changes in baseline secretion with microbial colonization, we next wanted to determine whether gut microbiota colonization modulates 5-HT-evoked secretory response. Hence we measured 5-HT-evoked ΔI_sc in GF and HM mice. Application of cumulative concentrations of 5-HT increased ΔI_sc in both GF and HM mice (Fig. 2). 5-HT-evoked ΔI_sc was significantly higher in GF compared with HM mice (n = 6, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 2A). 5-HT-evoked maximum ΔI_sc (E_max) was significantly higher in GF than HM mice (227.2 ± 21.4 vs. 149.3 ± 18.4 µA/cm², n = 6, unpaired t-test; P < 0.05; Fig. 2B).

Fig. 2.

The maximal secretory response to 5-HT is greater in GF than HM mice. A: change in short-circuit current in response to cumulative concentration of 5-HT in HM and GF mice. B: 5-HT-induced E_max in GF and HM mice. *P < 0.05.

5-HT₃ receptor mRNA and protein expression is significantly increased when gut microbiota is absent.

To determine whether increased secretory response to 5-HT seen in GF mice was due to an increase in 5-HT (5-HT₃ and 5-HT₄) receptor expression, we compared mRNA and protein expression in GF and HM mice (Fig. 3). We found a significant increase in 5-HT_3B receptor mRNA expression in GF compared with HM mice (2.0 ± 0.38 vs. 1.0 ± 0.22, n = 8–9, unpaired t-test; P < 0.05; Fig. 3A). This difference was not observed in 5-HT₄ receptor expression between GF and HM mice (0.78 ± 0.29 vs. 1.00 ± 0.17, n = 8–9, unpaired t-test; P > 0.05; Fig. 3B). To verify whether increased 5-HT_3B receptor mRNA expression translated to higher protein expression in GF mice, we quantified 5-HT₃ receptor protein using Western blot (Fig. 4A). The 5-HT_3B antibody identified two bands at 50 kDa (expected size for 5-HT₃) and 40 kDa (Fig. 4A; 25). Both bands were absent in the presence of a specific blocking peptide. GF mice had significantly higher expression of the 50-kDa band compared with HM mice (4.65 ± 1.05 vs. 1.15 ± 0.17, n = 12, unpaired t-test; P < 0.05; Fig. 4C); however, no difference in 40-kDa band intensity was observed (1.05 ± 0.20 vs. 1.22 ± 0.15, n = 13, unpaired t-test; P > 0.05; Fig. 4D).

Fig. 3.

5-HT_3B receptor mRNA expression is higher in GF mice than HM mice. A: relative expression of 5-HT_3B mRNA in HM and GF mice. B: relative expression of 5-HT₄ mRNA in HM and GF mice. *P < 0.05.

Fig. 4.

5-HT_3B receptor protein expression is higher in GF mice than HM mice. A: 5-HT_3B subunit antibody detected two bands at ~50 and 40 kDa. Controls include blockade of antibody binding by preincubating the primary antibody with its competing peptide (1°+B.P/2°) and secondary-only controls (2° Only). B: representative blot of 5-HT_3B subunit and GAPDH expression in GF and HM mouse proximal colon. C and D: quantification of the 5-HT_3B subunit band intensity present at 50 kDa (top band) and 40 kDa (bottom band), respectively, in HM and GF mice. *P < 0.05.

Colonization-dependent difference in secretory response to 5-HT in GF mice is mediated by altered 5-HT₃ receptor expression.

Since we observed an increase in 5-HT₃ but not 5-HT₄ receptor expression in GF mice, we next wanted to determine the role of 5-HT₃ and 5-HT₄ receptors in differential secretory response to 5-HT in GF and HM mice using specific 5-HT₃ and 5-HT₄ receptor antagonists (Fig. 5). Ondansetron (100 nM), a 5-HT₃ receptor antagonist, significantly inhibited 5-HT-evoked ΔI_sc in GF mice (n = 6–8, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 5A) but not in HM mice (P > 0.05; Fig. 5B). GR-113808 (30 nM), a 5-HT₄ receptor antagonist, significantly inhibited 5-HT-evoked ΔI_sc in both GF and HM mice (n = 6–8, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 5, D and E). 5-HT-evoked E_max in GF was significantly reduced by both ondansetron and GR-113808 (221 ± 16.3 vs. 121 ± 14.9 vs. 116 ± 15.3 µA/cm², n = 6–8, 1-way ANOVA; P < 0.05; Fig. 5G), whereas 5-HT-evoked E_max in HM was significantly inhibited by GR-113808 (163 ± 18.5 vs. 69 ± 14.0 µA/cm², n = 6, 1-way ANOVA; P < 0.05) but not by ondansetron (163 ± 18.5 vs. 120 ± 13.0 µA/cm², n = 6, 1-way ANOVA; P > 0.05; Fig. 5H). We found that similar to HM mice, 5-HT-evoked ΔI_sc in CR mice was significantly reduced following GR-113808 application (n = 6, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 5F) but not with ondansetron (n = 6, 2-way ANOVA with Bonferroni post hoc test; P > 0.05; Fig. 5C), suggesting that the effect is common to both mouse and human-derived gut microbiota. To further determine whether the increased secretory response in GF mice can be attributed to increased expression of 5-HT₃ receptors, we used a 5-HT₃ receptor-specific agonist, 2-methyl-5-hydroxytryptamine hydrochloride (100 μM), which evoked a significantly higher ΔI_sc in GF compared with HM mice (86.6 ± 18.1 vs. 32.0 ± 5.6 µA/cm², n = 5, unpaired t-test; P < 0.05; Fig. 6).

Fig. 5.

5-HT₃ receptor antagonist inhibits 5-HT-evoked ΔI_sc only in GF mice whereas 5-HT₄ receptor antagonist significantly blocks ΔI_sc in both HM and GF mice. A: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of ondansetron in GF mice. B: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of ondansetron in HM mice. C: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of ondansetron in CR mice. D: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of GR-113808 in GF mice. E: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of GR-113808 in HM mice. F: change in short-circuit current in response to cumulative concentration of 5-HT and 5-HT in the presence of GR-113808 in CR mice. 5-HT-induced E_max for GF (G), HM (H), and CR (I) mice in the presence of ondansetron (Ond) and GR-113808 (GR). *P < 0.05.

Fig. 6.

Increased secretory response to 5-HT in GF mice is mediated by 5-HT₃ receptor. 5-HT₃ receptor-selective agonist shows a significantly greater secretory response in GF mice compared with HM mice. *P < 0.05.

Epithelial 5-HT₃ receptors likely contribute to increased host secretory response to 5-HT in GF mice.

An earlier study identified 5-HT₃ receptor expression in both the nerve fibers and the EC cells of rat small intestine (11). To determine the distribution of 5-HT₃ receptors in mouse colon, we used 5-HT_3A-GFP mouse tissues. Immunohistochemistry revealed 5-HT_3A-GFP-positive cells in the mouse colonic epithelium and nerve fibers within the lamina propria (Fig. 7A). In the colonic epithelium, 5-HT_3A-GFP colocalized with 5-HT and DAPI representing EC cells (Fig. 7, B and C) confirming mucosal expression of 5-HT₃ receptors. Next, to ascertain whether the difference in ΔI_sc response to 5-HT between GF and HM mice was in part due to the epithelial 5-HT₃ receptors, we measured 5-HT-evoked ΔI_sc in the presence of TTX (500 nM) to block the neuronal contribution. We found that the ΔI_sc was significantly higher in GF compared with HM mice (n = 5, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 8A) in the presence of TTX suggesting a nonneuronal epithelial component contributing to the difference in 5-HT-evoked ΔI_sc. To determine whether epithelial 5-HT₃ receptors play a role in the differential 5-HT-evoked response, we measured ΔI_sc in the presence of TTX and ondansetron. The above-noted difference in ΔI_sc between GF and HM in the presence of TTX was lost in the presence of TTX and ondansetron (n = 5, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 8B) suggesting that epithelial 5-HT₃ receptor likely contributes to the differential secretory response in GF and HM mice. Interestingly, the difference in 5-HT-evoked ΔI_sc between GF and HM mice in the presence of TTX persisted in the presence of GR-113808 and TTX (n = 5, 2-way ANOVA with Bonferroni post hoc test; P < 0.05; Fig. 8C) suggesting that epithelial 5-HT₄ receptors are likely involved in the differential secretory response to 5-HT.

Fig. 7.

5-HT₃ receptors are expressed in the epithelial cells of colonic mucosa and in submucosal neuronal fibers. A: arrows highlight 5-HT_3A-GFP-positive cells in the mouse colonic epithelium. A subset of 5-HT₃-GFP-positive cells (B) colocalize with 5-HT and DAPI (C) in the proximal colon.

Fig. 8.

Epithelial 5-HT₃ receptor is responsible in part for the differential secretory response to 5-HT. Change in short-circuit current in response to cumulative concentration of 5-HT in the presence of TTX (A), change in short-circuit current in response to cumulative concentration of 5-HT in the presence of TTX and ondansetron (B), and change in short-circuit current in response to cumulative concentration of 5-HT in the presence of TTX and GR-113808 (C) in GF and HM mice. *P < 0.05.

Incubation of colonoids with acetate significantly decreases 5-HT₃ receptor expression in GF mice.

To determine the microbial mediators that affect 5-HT₃ receptor expression, we treated GF mouse-derived colonoids with acetate and butyrate. We found that incubation of epithelial colonoids with 10 mM sodium acetate for 24 h significantly decreased 5-HT₃ receptor expression compared with untreated controls (2.6 ± 0.26 vs. 1.33 ± 0.44, n = 5, repeated-measures 1-way ANOVA with Holm-Sidak’s multiple-comparisons test; P < 0.05; Fig. 9A). Although incubation with sodium butyrate did not significantly decrease 5-HT₃ receptor expression, sodium butyrate (0.5 mM) showed a trend toward reduced 5-HT₃ receptor expression (1.62 ± 0.16 vs. 1.07 ± 0.17, n = 5; P = 0.065; Fig. 9B).

Fig. 9.

Incubation of GF colonoids with acetate significantly decreases 5-HT_3B receptor mRNA expression. A: effect of acetate on 5-HT_3B receptor expression in GF colonoids. B: effect of butyrate on 5-HT_3B receptor expression in GF colonoids. *P < 0.05.

DISCUSSION

The aim of this study was to investigate the effect of microbial colonization on 5-HT-evoked secretory response and the receptor subtype involved in this response. We chose to use human-derived gut microbiota for colonization given the significant differences in mouse and human gut microbiota and the greater translational applicability of our findings for human conditions. We found no difference in TER, baseline I_sc, and ΔI_sc in response to forskolin in GF and HM mice, suggesting that microbial colonization status does not affect tissue permeability or baseline secretion or cause defects in cAMP-dependent secretory response in mice. We next examined whether microbiota colonization status alters evoked secretory response to exogenously applied 5-HT, a potent secretagogue, which has been previously shown to mediate Cl⁻ and HCO₃⁻ secretion (12, 19). We found that both GF and HM mice showed increased ΔI_sc in response to exogenously applied 5-HT in a concentration-dependent manner, suggesting that 5-HT mediates the secretory response irrespective of microbial colonization. Interestingly, we found that GF mice had a significantly higher secretory response to exogenously applied 5-HT than HM mice, which implies that the presence of gut microbiota and associated changes result in a dampening of 5-HT-induced secretion. We hypothesized that the increase in secretory response to exogenously applied 5-HT in GF mice was due to increased 5-HT receptor expression in GF mice. Among various families of 5-HT receptors (5-HT_1P, 5-HT₂, 5-HT₃, and 5-HT₄) that potentially mediate secretion (26), 5-HT₃ and 5-HT₄ receptors have been best studied (29). We therefore investigated whether gut microbiota alters 5-HT₃ or 5-HT₄ receptor mRNA expression. The 5-HT₃ receptor family comprises five different receptor subtypes—5-HT_3A, 5-HT_3B, 5-HT_3C, 5-HT_3D, and 5-HT_3E—among which only 5-HT_3A and 5-HT_3B subtypes are well characterized (5). Studies suggest that although 5-HT₃ receptors can exist as either a homopentameric 5-HT_3A or a heteropentameric 5-HT_3A/B receptor, only heteropentameric 5-HT_3A/B receptors replicate functional native mammalian 5-HT₃ channels (5, 7, 24, 27, 37, 38). Among 5-HT_3A and 5-HT_3B receptor subtypes, we chose to study 5-HT_3B receptor expression for two reasons: first, 5-HT_3B, and not 5-HT_3A, receptor subunit mRNA expression undergoes adaptive changes in mice (21, 23), and second, 5-HT_3B receptor expression modulates the biophysical and pharmacological properties (increased single-channel conductance, low permeability to calcium ions, and faster recovery from desensitization) of heteromeric 5-HT_3A/B receptors to closely mimic native 5-HT₃ receptors (5, 18). We found that 5-HT_3B subunit mRNA expression was significantly higher in GF mice compared with HM mice whereas 5-HT₄ receptor expression remained unchanged between the two groups of mice. We verified that increased 5-HT_3B mRNA expression leads to increased 5-HT_3B protein expression using Western blot. The protein band at ~50 kDa (expected molecular weight for 5-HT_3B subunit in mice), was significantly higher in GF mice than HM mice. An additional protein band, which has been previously described, was also detected at ~40 kDa and was neutralized using control peptide. However, it remains unclear whether this additional band represents a posttranslational modification (25, 28, 33) or a nonspecific protein recognized by this particular antibody. The 5-HT_3B subunit is a major determinant of 5-HT receptor function (5, 14), and our data show an increase in 5-HT_3B receptor subunit in GF mice, which correlates with increased secretory response to 5-HT in GF mice. We further confirmed that increase in 5-HT₃ receptor expression in GF mice contributes to increased colonic secretion as evidenced by a significantly greater response to specific 5-HT₃ receptor agonist and significant inhibition by specific 5-HT₃ receptor antagonist. We found that although 5-HT₄ receptors also mediate secretion in mice colon, microbial colonization does not affect 5-HT₄ receptor expression or 5-HT₄-dependent secretory response to 5-HT. The effect on 5-HT₃ receptor-dependent secretory response was seen in both HM and CR mice suggesting that a relatively conserved microbial function drives this process.

5-HT-evoked intestinal secretion could occur either via activation of 5-HT₃ neural reflex arch or via direct autocrine mucosal activation of 5-HT₃ receptors (13). We found that 5-HT₃ receptors are expressed in EC cells in the colonic epithelium. An increased secretory response to 5-HT in GF mice was also seen following blockade of neural activity. Furthermore, the secretory response was inhibited by specific 5-HT₃ receptor antagonist following blockade of neural activity, suggesting a role for epithelial 5-HT₃ receptors in microbiota-mediated modulation of 5-HT secretory response. Our results support previously published studies that show that 5-HT₃ receptor is not only present in the GI epithelium (11, 39) but also involved in mediating intestinal secretion via a nonneuronal pathway (17). Gut microbiota has been shown to increase 5-HT biosynthesis through an effect of SCFA (34), which may, in turn, affect 5-HT₃ receptor expression. Our study suggests that bacterial fermentation product acetate can modulate 5-HT_3B receptor expression in vitro in colonoids.

In conclusion, our study suggests that gut microbiota can alter host secretory response to 5-HT, in part by an effect on 5-HT₃ receptor expression via acetate production. Future studies elucidating the mechanisms that lead to acetate-driven altered mucosal 5-HT₃ receptor expression will allow a better understanding of the role of gut microbiota in pathogenesis of functional bowel disorders and the development of novel live biotherapeutic and microbial products to alter intestinal secretion.

GRANTS

This work was made possible by funding from National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant 100638 (P. C. Kashyap), Global Probiotics Council (P. C. Kashyap), and Center for Individualized Medicine, Mayo Clinic (P. C. Kashyap). M. Grover is supported by NIDDK Grant 103911. A. Beyder is supported by NIDDK Grant 106456.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.B., B.A.S., and P.C.K. conceived and designed research; Y.B. and B.A.S. performed experiments; Y.B., B.A.S., and P.C.K. analyzed data; Y.B., B.A.S., D.R.L., M.G., A.B., G.F., and P.C.K. interpreted results of experiments; Y.B. and P.C.K. prepared figures; Y.B. drafted manuscript; Y.B., B.A.S., D.R.L., E.D.L., M.G., A.B., G.F., and P.C.K. edited and revised manuscript; Y.B., B.A.S., D.R.L., E.D.L., M.G., A.B., G.F., and P.C.K. approved final version of manuscript.