Bifidobacterium stercoris KC84 attenuates IBS-D-like symptoms via modulation of serotonin-related pathways and dendritic cell-mediated IFN-β induction

Highlights

• •Bifidobacterium stercoris KC84 alleviates IBS-d-like symptoms in mouse and rat models.
• •KC84 modulates serotonin-related pathways associated with diarrhea.
• •KC84 induces IFN-β production from colonic dendritic cells.
• •IFN-β is associated with reduced intestinal smooth muscle contractility.

Abstract

Irritable bowel syndrome with diarrhea (IBS-D) is a prevalent disorder that significantly impairs quality of life, yet therapeutic advances remain limited. Serotonin dysregulation, primarily driven by enterochromaffin cells in the intestinal epithelium, is central to IBS-D pathogenesis. We hypothesized that targeted modulation of enterochromaffin cell activity through microbiome-based interventions could provide a novel treatment approach. Here, we screened 128 Bifidobacterium isolates and identified B. stercoris KC84 as a promising candidate. KC84 alleviated IBS-D-like symptoms in both chemically and stress-induced models, accompanied by changes in serotonin-related markers. Transcriptomic analysis revealed activation of type I interferon (IFN)-associated pathways, consistent with ex vivo evidence of KC84-induced IFN-β secretion, predominantly from CD11b^- dendritic cells. Furthermore, IFN-β treatment attenuated contractile activity in colonic smooth muscle cells. Collectively, these findings suggest that KC84 mitigates IBS-D-like symptoms, potentially through modulation of serotonin-related pathways and activation of a KC84-IFN-β-smooth muscle regulatory axis, supporting KC84 as a mechanism-guided probiotic candidate for IBS-D therapy.

Graphical abstract

Introduction

Irritable bowel syndrome (IBS) is a functional gastrointestinal disorder characterized by recurrent abdominal pain associated with defecation that is often linked to abnormal colonic smooth muscle contraction and dysregulated gut motility (Drossman and Hasler, 2016). Based on the Rome III or IV criteria, the estimated global prevalence of IBS is approximately 4–9 % (Oka et al., 2020; Sperber et al., 2021), with an annual economic burden reaching $20.8 billion in the United States (Drossman et al., 2002). Among its four subtypes, diarrhea-predominant IBS (IBS-D) is defined by frequent loose stools as the primary symptom. Only three drugs—alosetron, rifaximin, and eluxadoline—have received U.S. Food and Drug Administration (FDA) approval for the treatment of IBS-D, while several other medications are used off-label, highlighting a substantial unmet therapeutic need in this subtype. Considering its high prevalence, significant economic burden, and limited therapeutic options, development of new interventions for IBS-D remains a critical clinical priority.

Although the pathogenesis of IBS-D is incompletely understood, evidence implies that multiple factors, including serotonin dysregulation and immune responses, contribute to disease development (Enck et al., 2016). Serotonin plays a pivotal role in regulating intestinal motility, vasodilation, and neuronal signaling within the enteric nervous system. Excessive enteric serotonin can overstimulate the enteric nervous system, eliciting exaggerated intestinal motility and abdominal pain, thereby aggravating IBS-D symptoms (Mawe and Hoffman, 2013; Bayrer et al., 2023). Enterochromaffin cells, a subset of intestinal epithelial cells, produce the majority of serotonin through the enzymatic activity of tryptophan hydroxylase 1 (TPH1) and aromatic l-amino acid decarboxylase (AADC). TPH1 catalyzes the conversion of dietary tryptophan to 5-hydroxytryptophan (5-HTP), which is subsequently decarboxylated by AADC to form serotonin. Elevated serotonin levels act on receptors such as 5-HT3 that have been implicated in the exacerbation of IBS-D symptoms. Consistent with this, one FDA-approved drug for IBS-D, alosetron, exerts its therapeutic effects by antagonizing 5-HT3 receptors, highlighting serotonin signaling as a key target for IBS-D treatment. Mast cells also contribute to IBS-D pathophysiology by releasing serotonin, tryptase, and histamine through degranulation (Lee and Lee, 2016), thereby inducing an inflammatory response. Dysregulated immune activity, including the overproduction of pro-inflammatory cytokines (IL-1β, TNF-α, and IL-6) and the presence of low-grade inflammation, disrupts intestinal barrier integrity and tight junctions, leading to diarrheal symptoms (Barbara et al., 2011; Bischoff et al., 2014; Lazaridis and Germanidis, 2018; Ng et al., 2018).

Type I interferons (IFNs), primarily represented by IFN-α and IFN-β, are best known for their antiviral host defense functions (Stetson and Medzhitov, 2006). Since they are induced when microbial-associated or pathogen-associated molecular patterns (MAMPs and PAMPs) engage pattern recognition receptors, type I IFN responses can be triggered not only by viral but also by commensal or pathogenic microbial stimuli. Microbiota-driven type I IFN signaling has been shown to shape the tumor microenvironment, regulate immune tolerance, and remodel the metabolic state of dendritic cells (Schaupp et al., 2020; Lam et al., 2021; Vasquez Ayala et al., 2023). Beyond their immunologic roles, accumulating evidence indicates that type I IFNs also exert diverse biological functions (Wang et al., 2024b). Type I IFN signaling has been implicated in metabolic dysfunction, aging, and cardiovascular disease (Wieser et al., 2018; Lei et al., 2021; Oduro et al., 2022), and can influence bowel motility and cognitive performance (Wong et al., 2023; Zhang et al., 2025). These findings imply that microbiota-induced type I IFN responses may represent an additional mechanism through which the gut microbiota modulates host physiology.

The human gut harbors an estimated 10–100 trillion microorganisms, including bacteria, fungi, archaea, and protozoa, collectively known as the gut microbiota (Turnbaugh et al., 2007). The microbiota continuously and intimately interact with the host, influencing various aspects of health, metabolism, and immune regulation (Turnbaugh et al., 2006; Belkaid and Hand, 2014; Lynch and Pedersen, 2016). Several studies have demonstrated that the gut microbiota can regulate serotonin production. For example, metabolites derived from spore-forming bacteria and Edwardsiella tarda have been shown to stimulate enterochromaffin cells, leading to increased serotonin levels and enhanced gut motility (Yano et al., 2015; Ye et al., 2021). Another study reported that sphingolipids produced by Bacteroides fragilis are presented by CD1d molecules on enterochromaffin cells, which are subsequently recognized by TCRs of invariant natural killer T cells, resulting in elevated serotonin levels and accelerated intestinal transit (Luo et al., 2023). Beyond specific taxa, alterations in microbial composition have been reported in IBS patients compared with healthy individuals, with a consistent reduction in Bifidobacterium abundance (Kerckhoffs et al., 2009; Jeffery et al., 2020; Jacobs et al., 2023). This observation implies a potential association between Bifidobacterium and IBS pathophysiology. Accordingly, restoring Bifidobacterium populations has been proposed as a therapeutic strategy for IBS. Indeed, in vivo studies and randomized clinical trials have demonstrated that Bifidobacterium supplementation can reduce abdominal pain in IBS patients (Fukui et al., 2018; Pratt and Campbell, 2020; Zhou et al., 2020). However, the underlying mechanisms by which Bifidobacterium exerts its therapeutic effects in IBS-D remain poorly understood.

In this study, we screened 128 Bifidobacterium strains for their ability to reduce serotonin secretion and inhibit mast cell degranulation, identifying B. stercoris KC84 as a promising candidate. In two rodent models of IBS-D (5-HTP-induced and water avoidance stress-induced), KC84 alleviated disease symptoms and was associated with modulation of serotonin-related markers. Transcriptomic analysis revealed that KC84 upregulated colonic type I IFN signaling. Our data suggest that colonic dendritic cells can produce type I IFN and that this pathway may contribute to the modulation of smooth muscle contraction. These findings imply a type I IFN-mediated mechanism underlying the effects of KC84 on IBS-D-like symptoms, expanding our understanding of the immune contributions to gut motility disorders.

Materials and methods

Bacteria preparation

A total of 128 Bifidobacterium strains were previously isolated from the feces of healthy Korean adults. For in vitro screening, each strain was inoculated into Lactobacillus MRS broth (BD Difco, Sparks, MD, USA) supplemented with 0.05 % l-cysteine hydrochloride at a 1 % (v/v) inoculation ratio, and cultured at 37 °C under anaerobic conditions using the AnaeroPack (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan) for 24 h. The bacterial cultures were subsequently subcultured into fresh MRS broth and incubated under the same conditions for an additional 24 h. Bacterial cells were harvested by centrifugation at 3434 × g, washed twice with sterile phosphate-buffered saline (PBS), and resuspended in PBS. Bacterial cells were then stained with 10 μM SYTO 9 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) for 15 min and cell counts were determined using a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, USA). For pasteurization, the resuspended bacterial suspensions were incubated in a water bath at 70 °C for 30 min. After centrifugation at 3434 × g, the supernatants were discarded and the pellets were stored at −80 °C until further use.

For in vivo experiments, the bacterial pellet of Bifidobacterium stercoris KC84 was mixed with cryoprotectant agents and lyophilized at −80 °C for 12 h. The lyophilized KC84 was stored at −20 °C until further use. Bacterial powder was resuspended in PBS supplemented with 0.05 % l-cysteine hydrochloride at 1 or 5 × 10⁹ CFU/mL and mixed using a Vortex-Genie 2 (Scientific Industries, Bohemia, NY, USA) for 30 min.

Cell lines

RIN14B rat pancreatic islet cell line (ATCC CRL2059, Manassas, VA, USA) was used as a surrogate for serotonin-secreting enterochromaffin cells. MC38 mouse colon carcinoma cell line (Kerafast, Boston, MA, USA) was used as a negative control for α-SMA staining. RIN14B and MC38 cells were maintained in RPMI 1640 medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10 % fetal bovine serum (FBS; Welgene, Gyeongsan, Republic of Korea), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C with 5 % CO₂. RBL-2H3 rat basophilic leukemia cell line (ATCC CRL2256, Manassas, VA, USA) was used to study mast cell degranulation. RBL-2H3 cells were maintained in DMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10 % FBS (Welgene, Gyeongsan, Republic of Korea), 1 × NEAA (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 0.1 % sodium bicarbonate (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C with 5 % CO₂.

In vitro serotonin secretion measurement

RIN14B cells were seeded into flat-bottomed 96-well cell culture plates at a density of 7.4 × 10⁴ cells/well and incubated for 48 h at 37 °C with 5 % CO₂. The cells were washed twice with 300 μL HBSS (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 2 μM fluoxetine (Sigma-Aldrich, St. Louis, MO, USA). Under light-protected conditions, pasteurized bacteria were added to cells in media with 2 μM fluoxetine at cell-to-bacterium ratios of 1:1, 1:10, and 1:100, followed by incubation for 24 h at 37 °C with 5 % CO₂. Telotristat ethyl (50 μM, MedChemExpress, Monmouth Junction, NJ, USA), a tryptophan hydroxylase inhibitor, was used as a positive control. After the incubation, the cell culture supernatants were harvested, filtered using a vacuum manifold (Millipore, Burlington, MA, USA) and a MultiScreen filter plate (Millipore, Burlington, MA, USA), then stored at −80 °C until analysis. Serotonin levels in the culture supernatants were quantified using LC-MS/MS without additional sample preparation. Values were normalized to the vehicle control group, which was set to 100 %.

In vitro mast cell degranulation measurement

RBL-2H3 cells were seeded into flat-bottomed 96-well cell culture plates at a density of 2 × 10⁵ cells/well and incubated for 24 h at 37 °C with 5 % CO₂. The cells were washed twice with 200 μL Siraganian buffer (BioSolution, Seoul, Republic of Korea). Pasteurized bacteria were added to the cells at cell-to-bacterium ratios of 1:1, 1:10, and 1:100, followed by incubation for 20 min at 37 °C. Quercetin (40 μM; Sigma-Aldrich, St. Louis, MO, USA) was used as a positive control. Calcium ionophore A23187 (5 μM; Sigma-Aldrich, St. Louis, MO, USA) was then added to the cells to induce degranulation for 20 min at 37 °C. After the incubation, the cell culture supernatants were harvested and filtered. For a β-hexosaminidase assay, 3.5 mg/mL p-nitrophenyl-N-acetyl-β-glucosaminide (PNAG; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in citrate buffer (pH 4.5) and sonicated to mix completely. 50 μL filtered sample was mixed with 50 μL PNAG solution and incubated for 2 h at 37 °C. The reaction was stopped by adding 50 μL of sodium bicarbonate buffer (pH 10). β-hexosaminidase activity was measured at 405 nm using a microplate spectrophotometer (Tecan, Männedorf, Switzerland), representing the level of mast cell degranulation. The results were normalized to the vehicle control group, which was defined as 100 %.

Animals and experimental design

Male ICR mice (35 ± 5 g) were obtained from BioLASCO (a Charles River Laboratory licensee; Taipei City, Taiwan), seven-week-old female Wistar rats were obtained from Orient Bio Inc. (Seongnam, Republic of Korea), and seven-week-old C57BL/6 mice were obtained from JA BIO, Inc. (Suwon, Republic of Korea). All animals were housed under controlled conditions with a 12 h light/dark cycle, a temperature of 20–24 °C and humidity of 30 %–70 %. Animals had free access to autoclaved tap water and standard chow diet. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Pharmacological Discovery Services Taiwan, Ltd. (for mice; IACUC No IN042–09282022–45783), Seoul National University Bundang Hospital (for rats; IACUC No BA-2408–397–002–05), and CHA University (for ex vivo; IACUC No IACUC250065).

To induce IBS-D symptoms, a 5-HTP mouse model was used (Roberts et al., 2020). 5-HTP (Alfa Aesar, Ward Hill, MA, USA), dissolved in PBS, was intraperitoneally injected into mice at a dose of 10 mg/kg daily for 7 consecutive days (D8–D14). A suspension of KC84 (5 × 10⁹ CFU/mL, 200 μL) was orally gavaged to mice once daily for 14 days (D1–D14). On day 14, fecal samples were harvested for an hour to assess diarrhea severity, defecation frequency, and fecal water content. Diarrhea score for each fecal pellet was evaluated based on its consistency and moisture level using a 0–3 scale (0: normal; 1: soft; 2: unformed; 3: watery), and the sum of the individual pellet scores was used to determine the diarrhea score for each mouse. Fecal pellets were weighed, then dried for 16 h at 60 °C and reweighed. Fecal water content (%) was calculated using the formula below.

$$\text{Fecal}\text{water}\text{content}(\%)=100\times \left(\text{Wet}\text{weight}-\text{dry}\text{weight}\right)/\left(\text{Wet}\text{weight}\right)$$

Mice were sacrificed by CO₂ inhalation 30 min after the fecal harvest. To measure serotonin levels, luminal contents of the colon were harvested, snap-frozen in liquid nitrogen, and stored at −80 °C until analysis. Colon tissues were dissected and divided into two portions for RNA extraction and histological analysis. For RNA extraction, the colonic tissues were immersed in RNAlater (Sigma-Aldrich, St. Louis, MO, USA) for 24 h at 4 °C, then transferred to −80 °C for storage. For histological analysis, the colonic tissues were fixed in 10 % neutral buffered formalin (CHIN I PAO Co., Taoyuan City, Taiwan), then used for further histological analysis.

To induce IBS-D symptoms via stress, a water avoidance stress (WAS) rat model was used (Lee et al., 2017). Rats were placed on a block (5.8 cm length × 5.8 cm width × 6 cm height) in the middle of a plastic tank (26.7 cm length × 48.3 cm width × 20.3 cm height) which was filled with warm water (25 °C) to 1 cm below the top of the block. The rats were maintained on the block for an hour (between 8:00 and 10:00 a.m.) daily for 10 consecutive days (D4–D13). The control group was placed in a tank without water for an hour daily during the same period and time. A suspension of KC84 (1 × 10⁹ CFU/mL, 1 mL) was orally gavaged to rats once daily for 13 days (D1–D13). During the stress induction period, fecal pellets were counted daily to determine defecation frequency. On Day 11, fecal pellets were collected for 2 h to assess diarrhea severity. Diarrhea scoring was performed as described in the 5-HTP-induced model. Colon tissues were dissected and divided into two portions for RNA extraction and serotonin measurement. For RNA extraction, the colonic tissues were immersed in RNAlater (Sigma-Aldrich, St. Louis, MO, USA) for 24 h at 4 °C, then transferred to −80 °C for storage. For serotonin measurement, colon tissues were snap-frozen in liquid nitrogen and stored at −80 °C until analysis.

Histology and immunohistochemistry

Fixed colonic tissues were embedded in paraffin blocks. Tissue sections were cut (4 μm thickness) and stained with hematoxylin and eosin (H&E) for colitis scoring (Dieleman et al., 1998) under a light microscope DM2700 M (LEICA, Wetzlar, Germany). Colitis scoring was evaluated by a veterinary pathology expert (BioLASCO, Taiwan). Histological criteria included abnormalities of mucosal architecture, the extent of inflammation, erosion or ulceration, epithelial regeneration, and the percentage of involved in the disease process (supplementary Table S1). The scoring was based on the observer’s assessment of three representative sections per animal. Total scores were summed to yield a combined histological score ranging from 0 to 20.

For immunohistochemistry (IHC) analysis, the slides were stained with anti-serotonin antibody ab315150 (1:5000; Abcam, Cambridge, UK). Samples were examined microscopically by the veterinary pathologist. Serotonin-positive areas were highlighted, and the stained area was quantified using ImageJ (National Institutes of Health, Bethesda, MD, USA).

Quantitative PCR

RNA was extracted using RNeasy Plus Mini Kit (QIAGEN, Hilden, Germany). Colonic tissues (20 mg) were homogenized in RLT Plus buffer supplemented with β-mercaptoethanol and a 5 mm stainless steel bead (QIAGEN, Hilden, Germany) using the TissueLyser II (30 Hz, 5 min, twice; QIAGEN, Hilden, Germany). The remaining procedures were performed according to the kit instructions. An aliquot of the extracted RNA was stored at −80 °C for RNA-seq analysis. Complementary DNA (cDNA) was synthesized using the High-Capacity RNA-to-cDNA kit (Applied Biosystems, Foster City, CA, USA) and a thermal cycler (Applied Biosystems, Foster City, CA, USA). Quantitative PCR was performed using the Power SYBR green PCR Mix (Applied Biosystems, Foster City, CA, USA) on a QuantStudio 5 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Gapdh and Actb were used as housekeeping genes for normalization using the 2^-ΔΔCt method in the 5-HTP and WAS models, respectively. Primers used for each gene are listed in supplementary Table S2.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS)

For in vitro samples, serotonin levels were measured using a 1290 Infinity II LC system (Agilent, Santa Clara, CA, USA) coupled with an LC-QTOF 6546 mass spectrometer (Agilent, Santa Clara, CA, USA). An ACQUITY UPLC HSS T3 Column (100 Å, 1.8 μm, 2.1 mm × 50 mm; Waters, Milford, MA, USA) was used for sample separation (injection volume: 2 μL; flow rate: 0.25 mL/min) at 40 °C. Mobile phases consisted of 0.1 % formic acid in distilled water (A) and 0.1 % formic acid in acetonitrile (B), with a 7.6-min gradient: 2 % B at 0–2 min; 2–98 % B, 2–4.5 min; 98 % B, 4.5–5.6 min; 98–2 % B, 5.6–6.6 min; 2 % B, 6.6–7.6 min. Electrospray ionization (ESI) was operated in positive ion mode with a source voltage of 3.5 kV. The nebulizing gas was supplied at 50 psi, with a flow rate of 11 L/min at 300 °C. The scan rate and range were 1.5 spectra/s and m/z 100–300, respectively.

For in vivo samples, luminal contents of mice and colon tissues of rats were homogenized in ice-cold methanol (Burdick & Jackson, Muskegon, MI, USA) with a 5 mm stainless steel bead (QIAGEN, Hilden, Germany) using the TissueLyser II (30 Hz, 5 min, twice; QIAGEN, Hilden, Germany). The homogenates were incubated for an hour at −20 °C and centrifuged at 15,000 × g for 10 min at 4 °C. The supernatants were collected and filtered through a 0.22 μm syringe filter (Sartorius, Göttingen, Germany). The samples were transferred to screw-cap vials (Agilent, Santa Clara, CA, USA). Serotonin levels were measured using the same LC-MS/MS system (1290 Infinity II LC coupled with LC-QTOF 6546; Agilent, Santa Clara, CA, USA). An ACQUITY UPLC BEH Amide Column (130 Å, 1.7 μm, 2.1 mm × 50 mm; Waters, Milford, MA, USA) was used for sample separation (injection volume: 2 μL) at 40 °C. Mobile phases consisted of 40 % acetonitrile with 10 mM ammonium formate (A) and 90 % acetonitrile with 10 mM ammonium formate (B), with a 6.4-min gradient: 100–95 % B, 0–0.7 min; 95–85 % B, 0.7–1.6 min; 85–20 % B, 1.6–3.6 min; 20 % B, 3.6–4.4 min; 20–100 % B, 4.4–5.4 min; 100 % B, 5.4–6.4 min. The flow rate was set to 0.3 mL/min from 0 to 4.4 min, and reduced to 0.2 mL/min from 4.4 to 6.4 min. The scan rate and range were 1 spectra/s and m/z 85–500, respectively. ESI and nebulizing gas conditions were identical to those used in the in vitro analysis.

RNA-sequencing

RNA was extracted from colonic tissues and submitted to Macrogen (Seoul, Republic of Korea) for mRNA sequencing (Martin and Wang, 2011). Raw FASTQ files were trimmed using Trimmomatic (Usadel Lab, Aachen, Germany) to remove adapters and low-quality reads. Trimmed reads were aligned to the mouse reference genome (mm10) using HISAT2 (Johns Hopkins University, Baltimore, MD, USA). Transcript assembly and gene-level quantification were performed using StringTie (Pertea Lab, Baltimore, MD, USA). Gene expression levels were reported as raw read counts and transcripts per million (TPM). Differentially expressed gene (DEG) analysis was conducted in R (R Foundation for Statistical Computing, Vienna, Austria) using a DESeq2 package, and genes with adjusted p ≤ 0.05 and fold change ≥ 1.5 were considered differentially expressed.

For visualization, variance-stabilizing transformation (VST) was applied to the raw counts using DESeq2, and z-score normalization was performed across genes (Love et al., 2014). Heatmaps were generated from the normalized matrix to illustrate expression patterns of DEGs. Gene set enrichment analysis (GSEA) was performed using the clusterProfiler package in R (Subramanian et al., 2005; Wu et al., 2021). Genes were ranked based on fold change, and the mouse-specific GO Biological Process gene set (m5.go.bp.v2024.1.Mm.Symbols.gmt) was used as the reference. Significance was determined by permutation testing and enrichment results were reported as normalized enrichment scores (NES) and false discovery rate (FDR). Representative pathways were visualized using gseaplot2 and ggplot2 (Wickham, 2016). For individual gene visualization, expression data were presented as log₂(TPM+1), and statistical significance was determined based on the DESeq2-adjusted p-value.

Isolation of colonic smooth muscle cells and lamina propria

Single cell suspensions from the colonic lamina propria (cLP) were prepared as previously described (Kim et al., 2016; Ko et al., 2020), with minor modifications. In brief, colon tissues were dissected and opened longitudinally. To remove epithelial cells, colons were cut into small pieces and incubated in PBS buffer containing 3 % FBS, 10 mM EDTA, 20 mM HEPES, 100 U/mL penicillin, 100 μg/mL streptomycin, and 1 mM sodium pyruvate with gentle stirring at 37 °C for 20 min. The tissue segments were digested with 400 Mandl units/mL collagenase D (Roche, Basel, Switzerland), 10 μg/mL DNase I (Roche), and Dispase II (Roche) in RPMI 1640 medium containing 3 % FBS, 20 mM HEPES, 100 U/mL penicillin, 100 μg/mL streptomycin, 1 mM sodium pyruvate, and 1 mM non-essential amino acids (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) at 37 °C for 40 min with stirring. The enzyme reaction was stopped by adding 10 mM EDTA, followed by thorough mixing to release cells from the tissue. Erythrocytes were removed by treatment with RBC lysis buffer (Sigma-Aldrich, St. Louis, MO, USA) for 5 min.

For colonic smooth muscle cell isolation, cell suspension was cultured in RPMI medium supplemented with 10 % FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin for 7–14 days. During this period, non-smooth muscle cells, including immune cells and endothelial cells, were gradually lost due to their inability to survive long-term under these conditions. Adherent smooth muscle cells remaining after this period were harvested and used for collagen gel contraction assay.

For lamina propria cell isolation, the cell suspension was subjected to 40–80 % Percoll (Cytiva, Uppsala, Sweden) density gradient centrifugation. Isolated lamina propria cells were seeded at 2 × 10⁵ cells per well in 96-well plates and incubated for 6 h or 24 h at 37 °C with 5 % CO₂ at cell-to-KC84 ratios of 1:10 or 1:100. Culture supernatants were collected at 24 h, filtered, and used for measuring IFN-β levels by an ELISA kit (R&D Systems, Minneapolis, MN, USA). Cells harvested at 6 h were used for flow cytometric analysis of immune cell populations.

Flow cytometry

To measure intracellular IFN-β levels, colonic lamina propria cells were stimulated with a Cell Stimulation Cocktail containing PMA, ionomycin, brefeldin A, and monensin (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) for 4 h at 37 °C. Cells were pre-incubated with Fc block [anti-CD16/CD32 (clone 93, Cat. No 14–0161–85, Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA)] for 5 min at 4 °C, followed by addition of a surface antibody cocktail and incubation for a further 30 min at 4 °C. The antibody cocktail included anti-CD45-BUV395 (clone 30-F11, Cat. No 564279, BD Biosciences, San Jose, CA, USA), anti-CD11c-BUV737 (clone HL3, Cat. No 612796, BD Biosciences), anti-I-A/I-E-BB700 (clone M5/114.15.2, Cat. No 746197, BD Biosciences), anti-CD11b-BV786 (clone M1/70, Cat. No 740861, BD Biosciences), anti-PDCA-1-BV421 (clone 927, Cat. No 566431, BD Biosciences), anti-Ly6C-BV605 (clone AL-21, Cat. No 563011, BD Biosciences), and anti-F4/80-PE (clone BM8, Cat. No 123110, BioLegend, San Diego, CA, USA). Stained cells were fixed and permeabilized using Transcription Factor Staining Buffer Set (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) for 20 min at 4 °C. Intracellular IFN-β was stained with anti-IFN-β-APC (polyclonal, Cat. No 32183–05161, ASSAYPRO, St. Charles, MO, USA) for 1 h at 4 °C. Stained cells were acquired on a FACSymphony A3 flow cytometer (BD Biosciences) and data were analyzed using FlowJo software (Tree Star, Ashland, OR, USA). Gating strategy is depicted in supplementary Fig. S5.

To assess the purity of the smooth muscle cells, intracellular αSMA staining was performed using an anti-αSMA-Alexa Fluor 488 antibody (clone 1A4, Cat. No 53–9760–82, Invitrogen). Cells were fixed and permeabilized prior to staining. The MC38 cell line was used as a negative control (Supplementary Fig. S6).

Collagen gel contraction assay

To investigate the contractile activity of colonic smooth muscle, we used a Collagen Gel Contraction Kit (Cell Biolabs, Inc., San Diego, CA, USA) following the manufacturer’s instructions. Briefly, isolated colonic smooth muscle cells were mixed with a collagen solution at a cell-to-collagen ratio of 2:8 and incubated for 1 h at 37 °C with 5 % CO₂. After collagen polymerization, cell culture medium with or without 10 ng/mL recombinant mouse IFN-β (R&D Systems, Minneapolis, MN, USA) was added on top of the collagen gels and incubated for 24 h at 37 °C with 5 % CO₂. The collagen gels were then released using a spatula and photographed at 15 min or 16 h after release. Gel size was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA) and the change in gel size was calculated relative to the 15 min value (%).

Statistical analysis

Data are expressed as the mean ± standard error of the mean (SEM). Statistical analysis and visualization were performed using Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Comparisons between two groups were made using an unpaired two-tailed Student’s t-test. A p-value < 0.05 was considered statistically significant. For correlation analyses, p-values were adjusted for multiple comparisons using the Benjamini-Hochberg method to control the false discovery rate (FDR), and corresponding adjusted p-values were calculated. Statistical analysis of RNA-seq data was conducted as described above.

Results

B. stercoris KC84 was identified as a therapeutic candidate based on dual in vitro screening

To identify Bifidobacterium strains with potential therapeutic activity against IBS-D, we performed in vitro screening of 128 strains derived from nine species (B. adolescentis, B. animalis subsp. lactis, B. bifidum, B. breve, B. faecale, B. longum subsp. infantis, B. longum subsp. suillum, B. pseudocatenulatum, and B. stercoris) from our human-origin bacterial bank. We focused on two major features of IBS-D: excessive serotonin production and increased mast cell degranulation (Fig. 1A). Given the established role of serotonin in promoting visceral motility (Mawe and Hoffman, 2013), we hypothesized that serotonin-reducing bacteria could provide therapeutic benefits in IBS. To test this, we evaluated all 128 Bifidobacterium strains for their effects on serotonin secretion in RIN14B cells, a rat pancreatic cell line that is widely used as a model for enterochromaffin cells. The 128 strains demonstrated serotonin-reducing capacities of 0–99 % compared with the negative control. Among these, B. stercoris strain KC84 exhibited the seventh-highest serotonin-reducing ability, achieving a 92 % reduction (Fig. 1B). A dose-dependent decrease in serotonin levels was confirmed for KC84 (Fig. 1C). Species-level analysis revealed that strains of B. adolescentis, B. faecale, and B. stercoris showed superior serotonin-reducing capacity compared with other species (supplementary Fig. S1A), with KC84 ranking third among B. stercoris strains (supplementary Fig. S1B).

Fig. 1.

Bifidobacterium stercoris KC84 was identified as a therapeutic candidate based on dual in vitro screening. (A) Schematic overview of two screening assays: serotonin reduction and inhibition of mast cell degranulation. (B and C) RIN14B cells were treated with bacterial strains for 24 h, and serotonin levels in the culture supernatant were quantified using LC-MS/MS. Serotonin reduction was calculated relative to the vehicle control, which was set at 100 %. 128 Bifidobacterium strains were tested at a 1:100 cell-to-bacterium ratio (n = 2–3) (C). KC84 was evaluated at 1:1, 1:10, and 1:100 ratios, with telotristat ethyl used as a positive control (n = 3) (C). (D and E) RBL-2H3 cells were treated with bacterial strains for 20 min, and a β-hexosaminidase assay was performed to assess mast cell degranulation. Reduction of mast cell degranulation was calculated relative to the vehicle control, which was set at 100 %. 128 Bifidobacterium strains were tested at a 1:100 cell-to-bacterium ratio (n = 2–3) (D), and KC84 was evaluated at 1:1, 1:10, and 1:100 ratios, with quercetin as a positive control (n = 3) (E). (F) All 128 strains were plotted on a two-dimensional graph, with mean values from (B) and (D) representing the X and Y axes, respectively. KC84 is highlighted as a blue dot labeled with its strain name. Data are presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001 vs. vehicle group.

Concurrently, we assessed the ability of the 128 strains to inhibit mast cell degranulation using RBL-2H3 cells, a basophilic cell line commonly used as a model for mast cells. Degranulation contributes to the exacerbation of IBS-D symptoms; thus, suppression of this process is therapeutically relevant. The tested strains reduced degranulation by 0–68 % compared with the negative control, with KC84 achieving a 59 % reduction and ranking fourth overall (Fig. 1D). Dose-dependent inhibition of degranulation was also observed (Fig. 1E). Unlike the serotonin assay, species-specific variation in degranulation suppression was minimal (supplementary Fig. S1C), and KC84 again ranked third among B. stercoris strains (supplementary Fig. S1D). When the results of the two screening systems were plotted, a positive correlation was observed (Fig. 1F; r = 0.550, p = 1.85 × 10⁻¹¹). We then ranked all 128 strains by summing their serotonin-reduction and degranulation-reduction ranks, which identified KC84 as a top therapeutic candidate.

KC84 alleviated 5-HTP-induced IBS-D-like symptoms

To evaluate the in vivo efficacy of KC84, we administered the strain to a mouse model of 5-HTP-induced IBS-D. The mice received daily oral gavage of KC84 for 14 days, and from day 8 until sacrifice, 5-HTP was injected intraperitoneally to induce diarrhea (Fig. 2A). Fecal samples were collected at sacrifice, and diarrhea severity was assessed for each pellet using a standardized scoring system. 5-HTP administration significantly increased three diarrhea-related parameters, diarrhea score (Fig. 2B), defecation frequency (Fig. 2C), and fecal water content (Fig. 2D), while 14-day KC84 treatment significantly alleviated all three symptoms (Fig. 2B–D).

Fig. 2.

KC84 alleviated 5-HTP-induced IBS-D symptoms. (A) Schematic of the experimental design for the 5-HTP-induced IBS model and oral administration of KC84. Mice were divided into three groups: Group 1, vehicle/vehicle; Group 2, 5-HTP/vehicle; Group 3, 5-HTP/KC84. (B–D) Diarrhea score, defecation frequency, and fecal water content were evaluated on D14. (E) Representative H&E-stained colonic sections showing inflammatory cell infiltration in the submucosa (indicated by red asterisks). Scale bar = 120 μm. (F) Total colitis scores were quantified from H&E-stained colon sections. (G and H) Relative gene expression levels of Tjp1 and Il1b were measured by qPCR. Data are presented as mean ± SEM (n = 10/group). *p < 0.05; **p < 0.01; ***p < 0.001 vs. 5-HTP/vehicle group.

Colon tissues were stained with H&E, and the colitis index was scored based on five histopathological criteria. Among these, 5-HTP administration aggravated three parameters—mucosal architecture abnormalities, inflammation extent, and percent of involvement—all of which were ameliorated by KC84 treatment (supplementary Fig. S2C–G). Accordingly, KC84 significantly reduced the total colitis score, representing the sum of all five criteria (Fig. 2E and F). The histological improvements prompted further examination of molecular markers related to barrier function and inflammation, both of which are closely associated with IBS-D (Liebregts et al., 2007; Zhou et al., 2009). RNA extracted from colon tissues was subjected to qPCR analysis. KC84 treatment increased Tjp1 expression and decreased Il1b expression, implying enhanced barrier integrity and attenuated pro-inflammatory signaling (Fig. 2G and H). These molecular alterations may underlie the observed improvements in IBS-D phenotypes. Collectively, these findings suggest that KC84 mitigates IBS-D-like symptoms, accompanied by improvements in mucosal inflammation and barrier-related markers, thereby supporting its therapeutic potential.

KC84 was associated with changes in serotonin-related markers

To validate the serotonin-lowering effect of KC84 in vivo, mouse colon tissues were stained with an anti-serotonin antibody. KC84 treatment significantly reduced the serotonin-positive area that was elevated by 5-HTP induction (Fig. 3A and B). Consistently, targeted LC-MS analysis revealed a decreasing trend in luminal serotonin levels in the KC84-treated group (p = 0.06; Fig. 3C). Correlation analysis demonstrated that the serotonin-positive area was positively associated with the total colitis score (r = 0.485, p. adj = 0.019; Fig. 3D). Similarly, luminal serotonin levels were positively correlated with diarrhea score (r = 0.396, p. adj = 0.033; Fig. 3E), defecation frequency (r = 0.470, p. adj = 0.019; Fig. 3F), and total colitis score (r = 0.450, p. adj = 0.019; Fig. 3G). Given the significant associations between serotonin levels and disease severity, we next examined the expression of serotonin-related genes in colonic tissues. The expression of Slc6a4, encoding the serotonin reuptake transporter, was significantly downregulated by 5-HTP and restored by KC84 (Fig. 3H). In addition, Tph1, encoding a serotonin-synthesizing enzyme, was significantly downregulated by KC84 treatment (Fig. 3I). These findings suggest that KC84 may modulate intestinal serotonin levels by influencing both reuptake and biosynthesis. Collectively, these findings are consistent with a potential role for serotonin modulation in KC84-mediated improvement of IBS-D-like symptoms.

Fig. 3.

KC84 suppressed serotonin levels and modulated serotonin-related gene expression. (A) Representative immunohistochemistry images of colonic sections (from the same cohort as in Fig. 2) stained with an anti-serotonin antibody. Representative serotonin-positive areas are indicated by yellow arrows. Scale bar = 60 μm. (B) Quantification of serotonin-positive areas using ImageJ (n = 10/group). (C) Serotonin levels in colonic luminal contents measured using LC-MS/MS (n = 9–10/group). (D–G) Correlations between serotonin levels (serotonin-positive area or luminal serotonin concentration) and phenotypic features (total colitis score, diarrhea score, or defecation frequency), plotted on the X and Y axes, respectively. (H and I) Relative gene expression levels of Slc6a4 and Tph1 were determined by qPCR (n = 10/group). For correlation analyses (D–G), p-values were adjusted using the Benjamini-Hochberg method to control the false discovery rate and corresponding adjusted p-values were indicated. Data are presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001 vs. 5-HTP/vehicle group.

KC84 mitigated stress-induced IBS-D phenotypes

To assess the therapeutic potential of KC84 under stress conditions further, we used a water avoidance stress (WAS)-induced IBS-D rat model. Rats were subjected to WAS for 1 hour daily over 10 consecutive days, while oral gavage of KC84 began 3 days before the first stress session and continued throughout the experimental period (Fig. 4A). Daily defecation frequency was consistently higher in the WAS group compared with sham controls but was attenuated by KC84 administration (Fig. 4B). Averaged across the entire experiment, the defecation frequency demonstrated the same trend: an increase induced by WAS and a reduction by KC84 treatment (Fig. 4C). On day 11, WAS significantly elevated diarrhea scores, whereas KC84 significantly mitigated this stress-induced increase (Fig. 4D). LC-MS analysis of colon tissues revealed a trend toward lower serotonin levels in the KC84-treated group (p = 0.06; Fig. 4E), mirroring the reduction observed in the 5-HTP-induced model.

Fig. 4.

KC84 mitigated WAS-induced IBS-D phenotypes. (A) Schematic of the experimental design for the WAS-induced IBS model and oral administration of KC84. Rats were divided into three groups: Group 1, no stress/vehicle; Group 2, WAS/vehicle; Group 3, WAS/KC84. (B–E) Defecation frequency during WAS induction was recorded from D4 to D13, excluding D9 and D11. Defecation frequency is shown as a time-series graph (B) and as the mean defecation frequency (C). Diarrhea scores were assessed on D11 (D). (E) Serotonin levels in colonic tissues were measured using LC-MS/MS. (F and G) Relative gene expression levels of Prss1 and Prss2 in colonic tissues were measured by qPCR. Data are presented as mean ± SEM (n = 6/group). Statistical significance for the time-series graph (B) was analyzed using two-way ANOVA. *p < 0.05; **p < 0.01; ***p < 0.001 vs. WAS/vehicle group.

At the molecular level, KC84 also modulated the expression of trypsin-like serine proteases implicated in IBS-D pathophysiology (Lee et al., 2017; Rolland-Fourcade et al., 2017). Colonic gene expression of Prss1 and Prss2, which were elevated by WAS induction, was significantly downregulated following KC84 treatment (Fig. 4F and G). Together, these findings show that KC84 mitigates stress-induced IBS-D phenotypes and associated molecular alterations.

KC84 upregulated type I IFN signaling as revealed by transcriptomic analysis

Given the marked improvement in symptoms, colonic gene expression, and serotonin levels, we next sought to delineate transcriptomic changes to gain mechanistic insights into how KC84 exerts its protective effects. RNA-seq analysis was performed on colonic tissues collected from the 5-HTP mouse model (Fig. 2A). Differentially expressed gene (DEG) analysis identified 28 upregulated and 14 downregulated genes in the KC84-treated group compared with the vehicle group (fold change ≥ 1.5, Benjamini-Hochberg-adjusted p ≤ 0.05), as illustrated in the volcano plot (Fig. 5A, supplementary Table S3). Among the upregulated genes were canonical IFN-stimulated genes (e.g., Oas1b, Dhx58, Herc6, Slfn4, and Irgm1), as well as Trim30a and Rnf213, both previously reported to be type I IFN-responsive genes (Fig. 5B) (Wang et al., 2015; Li et al., 2025).

Fig. 5.

KC84 upregulated type I IFN signaling as revealed by transcriptomic analysis. RNA sequencing was performed using colon tissues from the same cohort as in Fig. 2 (n = 5 for the 5-HTP/vehicle group and n = 4 for the 5-HTP/KC84 group). (A) Volcano plot showing differentially expressed genes (DEGs) based on the thresholds |fold change| > 1.5 and Benjamini-Hochberg-adjusted p < 0.05. In total, 28 genes were upregulated and 14 genes were downregulated following KC84 administration. (B) Heatmap of variance-stabilizing transformation (VST)-normalized expression for the 28 upregulated and 14 downregulated DEGs. (C) Gene ontology (GO)-based gene set enrichment analysis (GSEA) highlighting representative enriched terms; type I IFN-related gene sets are marked in blue. (D) Bar plots showing the expression levels of representative type I IFN-related genes on a log₂(TPM + 1) scale. Data are presented as means ± SEM, and statistical significance was determined using Benjamini-Hochberg-adjusted p-values from DESeq2 analysis. *p < 0.05; **p < 0.01 vs. 5-HTP/vehicle group.

To obtain a comprehensive overview, Gene Ontology (GO)-based gene set enrichment analysis (GSEA) was conducted. Consistent with the findings from the volcano plot, this analysis revealed enrichment of type I IFN-related gene sets in the KC84-treated group (Fig. 5C). Among these, the “response to IFN-β” gene set exhibited the highest normalized enrichment score (NES = 2.2035, p = 0.0063; supplementary Fig. S3A). The KC84-induced increase in IFN-related responses was further supported by enrichment of additional pathways, including the “interferon-mediated signaling pathway” (NES = 2.1364, p = 0.0014; supplementary Fig. S3B) and “response to type I IFN” (NES = 1.9578, p = 0.0212; supplementary Fig. S3C). A similar trend toward higher expression was also observed for other IFN-stimulated genes (Oasl2 and Ddx60) and the transcription factor Stat2 in the TPM plots (Fig. 5D). These data indicate that KC84 activates a type I IFN response in the colon, which may underlie its protective effects against IBS-D pathology.

In addition to IFN-related pathways, altered expression of several circadian rhythm-related genes was observed in the KC84-treated group (supplementary Fig. S3D). Arntl and Npas2 were reduced, whereas Per1–3 and Cry1 were increased, along with coordinated changes in D box-related and RORE-related genes. However, due to the time-dependent nature of circadian gene expression, the use of a single time point does not allow full characterization of overall circadian expression patterns.

KC84 induced IFN-β production from colonic dendritic cells, with potential effects on reducing muscle contractility

To determine whether KC84 directly induces type I IFN production, we treated isolated colonic lamina propria cells ex vivo with KC84 and subsequently measured IFN-β levels in supernatants, along with immune cell subset profiling by flow cytometry (Fig. 6A). Consistent with the RNA-seq results showing enhanced type I IFN signaling, KC84 treatment for 24 h significantly increased IFN-β concentrations in supernatants (Fig. 6B). Flow cytometric analysis further revealed a higher proportion of IFN-β-producing cells among CD45⁺ leukocytes (Fig. 6C and D).

Fig. 6.

KC84 induced IFN-β production from colonic dendritic cells, leading to reduced muscle contractility in an IFN-β-dependent manner. (A) Schematic of the ex vivo experimental design for IFN-β production from the colonic lamina propria and the collagen gel contraction assay. (B) IFN-β levels in the culture supernatant quantified by ELISA (n = 6/group). (C–F) Colonic lamina propria cells were incubated with KC84 for 6 h and analyzed by flow cytometry (n = 6/group). Representative plots are shown for IFN-β-producing CD45⁺ leukocytes (C) and CD11b^- dendritic cells (E), with corresponding quantifications in (D) and (F), respectively. (G–I) Primary colonic smooth muscle cells were used for a collagen gel contraction assay (n = 3/group). Representative collagen gel images at 16 h (G), together with quantification of gel size (H) and percent changes between 15 min and 16 h (I), are shown. Data are presented as mean ± SEM. **p < 0.01; ***p < 0.001 vs. vehicle group.

This increase was also evident in dendritic cells (DCs defined as CD11c⁺MHC class II⁺; supplementary Fig. S4A and B). Among DC subtypes, plasmacytoid DCs (pDCs; PDCA-1⁺CD11c⁺MHC class II⁺) are regarded as professional producers of type I IFNs (Colonna et al., 2004); however, KC84-induced IFN-β production was observed predominantly in CD11b^- DC subset (CD11b⁻PDCA-1⁻CD11c⁺MHC class II⁺; Fig. 6E and F), with no consistent changes detected in the CD11b⁺ DC (CD11b⁺PDCA-1⁻CD11c⁺MHC class II⁺) or pDC subsets (supplementary Fig. S4C–F). Additionally, KC84 treatment significantly increased the frequency of IFN-β⁺ monocytes (CD11b⁺Ly6C⁺MHC class II⁻), although event counts were low (supplementary Fig. S4G and H). These findings suggest that KC84 stimulates CD11b^- DCs in the colonic lamina propria to produce IFN-β.

We next investigated whether IFN-β could directly influence gut tissue physiology, particularly colonic smooth muscle contractile activity. Recombinant mouse IFN-β was applied to primary colonic myofibroblasts, and contractile function was assessed using collagen gel contraction assays (Fig. 6A). Representative images at 16 h showed markedly reduced gel contraction in IFN-β-treated cultures compared with controls, which was corroborated by quantitative analysis of gel size (Fig. 6G and H). When normalized to the gel area at 15 min, IFN-β-treated samples retained a higher percentage of their initial size after 16 h, indicating attenuated contractile activity (Fig. 6I). These results suggest that IFN-β produced by KC84-stimulated immune cells can attenuate colonic muscle contractility.

Collectively, these findings suggest that B. stercoris KC84 alleviates IBS-D-like symptoms not only by modulating serotonin-related markers but also by activating colonic type I IFN signaling, which may contribute to the suppression of excessive colonic muscle contractility.

Discussion

IBS-D remains a highly prevalent condition that significantly impairs quality of life, yet therapeutic innovation in this field has been limited, with only three FDA-approved drugs currently available (Andrae et al., 2013). Serotonin dysregulation, primarily mediated by enterochromaffin cells within the gut epithelium, is a central driver of IBS-D pathophysiology (Mawe and Hoffman, 2013). We therefore hypothesized that targeted modulation of enterochromaffin cell activity through microbiome-based interventions could provide a novel therapeutic avenue. This hypothesis guided our search for a probiotic strain capable of alleviating IBS-D by modulating serotonin production and gut motility.

Based on previous reports linking the genus Bifidobacterium to IBS-D (Pratt and Campbell, 2020), we screened 128 strains within this genus and identified B. stercoris KC84 as a promising candidate. KC84 alleviated IBS-D-like symptoms in both chemical-induced and stress-induced models, accompanied by modulation of serotonin-related markers. Transcriptomic analysis revealed that KC84 treatment upregulated genes associated with type I IFN signaling. To further examine this finding, we conducted ex vivo experiments and found that KC84 induced IFN-β production, with a significantly increased proportion of IFN-β-producing cells observed within the CD11b^- DC population. Furthermore, IFN-β was shown to reduce colonic smooth muscle contractility directly, suggesting a KC84-IFN-β-smooth muscle axis that contributes to the attenuation of IBS-D-like symptoms.

These findings have several important implications. First, the mechanism underlying the therapeutic effect of KC84 may be explained by two complementary actions: modulation of serotonin-related pathways, which are known to regulate colonic smooth muscle contraction (Mawe and Hoffman, 2013); and enhancement of IFN-β secretion from colonic immune cells, which may contribute to the attenuation of muscle contractility. Notably, Zhang et al. demonstrated that suppression of type I IFN signaling exacerbates intestinal hypercontractility and worsens IBS-D symptoms, suggesting a protective role of IFN signaling in disease regulation (Zhang et al., 2025). In addition, Wong et al. showed that increased type I IFN signaling can drive serotonin reduction through impaired tryptophan metabolism, supporting a mechanistic link between type I IFN signaling and serotonin regulation (Wong et al., 2023). While these studies highlight roles for type I IFN signaling in gut physiology and serotonin regulation, respectively, our work differs in that we identify a probiotic strain capable of activating this pathway in a therapeutically relevant context. To our knowledge, this is the first study to show that a probiotic can activate type I IFN signaling in a therapeutically beneficial manner, supporting KC84 as a mechanism-based live biotherapeutic candidate for IBS-D.

Second, our ex vivo studies showed that KC84 can stimulate colonic immune cells to produce IFN-β. ELISA results confirmed increased IFN-β secretion upon direct exposure of lamina propria cells to KC84, implying that MAMPs from this strain likely engage host pattern recognition receptors. Given that myeloid cells are typically the primary responders to bacterial MAMPs, we performed flow cytometric analysis and identified CD11b^- DC as a major subset contributing to this response. Although the frequency of IFN-β-producing CD11b^- DCs was relatively low, type I IFNs are known to exert biological effects even at low levels through tonic signaling and downstream amplification mechanisms (Pucella et al., 2025). In addition, previous studies have reported that conventional dendritic cell subsets can produce type I IFNs through innate sensing pathways, supporting the plausibility of our observations (Jneid et al., 2023; Wang et al., 2024a). Meanwhile, dendritic cell activation has been associated with exacerbation of IBS-D symptoms through the release of proinflammatory cytokines (Zhao et al., 2019). In our system, however, KC84 activated CD11b^- DCs in a manner that enhanced IFN-β production rather than proinflammatory outputs, suggesting that dendritic cell activation can be therapeutically advantageous in IBS-D when mediated through the IFN-β pathway.

Third, in our data, IFN-β production was not detectably increased in pDCs following KC84 stimulation (supplementary Fig. S4G and H). Although pDCs are canonical producers of type I IFNs, particularly in response to viral nucleic acids, their activation by bacterial stimuli appears to be highly stimulus- and strain-dependent. For example, a commensal-derived polysaccharide from Bacteroides fragilis mediated anti-inflammatory pDC responses without inducing detectable type I IFN secretion from isolated pDC in vitro (Dasgupta et al. 2014). In addition, a broad screening of lactic acid bacteria showed that type I IFN-inducing activity was absent in 90 rod-shaped strains and restricted to a subset of spherical strains (Jounai et al., 2012). These findings suggest that bacterial stimulation does not necessarily result in a robust or detectable type I IFN response in pDCs. In this context, the lack of detectable IFN-β in pDCs following KC84 stimulation may reflect the stimulus- and strain-dependent nature of pDC responsiveness.

Fourth, this study highlighted that IFN-β can exert physiological effects beyond classical immune responses. Recent evidence indicates that type I IFN signaling extends beyond its canonical antiviral and immunomodulatory roles, contributing to physiological and metabolic regulation (Wang et al., 2024b). Hoang et al. demonstrated that in adult mice, cytosolic mitochondrial RNA triggers a type I IFN-IRF7 axis that suppresses mitobiogenesis and thermogenesis, thereby promoting obesity (Hoang et al., 2022). These findings support the emerging concept that IFN signaling regulates non-immune processes, such as energy metabolism and tissue function. In this context, our study adds to the expanding scope of IFN biology by showing its involvement in gut neuromuscular modulation.

Fifth, the therapeutic potential of KC84 is further supported by its favorable safety profile. Unlike systemic administration of recombinant type I IFNs, which often leads to adverse effects such as flu-like symptoms and cardiovascular complications (Russo and Fried, 2003; Sakabe et al., 2013), KC84 is associated with activation of this pathway in a microbiota-driven manner. In contrast, oral administration of Lactobacillus acidophilus, engineered to constitutively express IFN-α, unexpectedly worsened colitis in a dextran sulfate sodium-induced inflammatory bowel disease model (McFarland et al., 2011), illustrating the risks associated with excessive type I IFN activation. KC84, a human-derived commensal strain, induced IFN-β signaling in a host-compatible fashion and attenuated IBS-D phenotypes without causing body weight loss or systemic inflammation (supplementary Fig. S2B), suggesting a potentially safer and more targeted therapeutic approach than conventional cytokine delivery.

Sixth, given that KC84 induces IFN-β production in colonic myeloid cells, bacterial components may contribute to this response. It is therefore plausible that pattern recognition receptor-mediated pathways, such as TLR-associated MyD88 signaling or NOD/RIP2-mediated sensing, may be involved (Li and Wu, 2021). Future studies using pathway-specific inhibitors or genetic approaches may help to further define the mechanisms underlying KC84-mediated immune activation.

Seventh, circadian rhythm disruption has been implicated in IBS (Hong et al., 2025; Lu et al., 2025). In our transcriptomic analyses, KC84 treatment was associated with changes in the expression of multiple circadian clock genes, including Ciart and Per3 (supplementary Fig. S3D). These findings raise the possibility that circadian regulation may contribute to the observed improvement of IBS-D-like symptoms. However, circadian gene expression is highly time-dependent, and cross-sectional transcriptomic data are limited in capturing circadian dynamics. Therefore, the functional relevance of these changes remains unclear and requires further investigation, including time-course analyses.

Building on our finding that KC84 modulates the IFN-smooth muscle-IBS-D axis, several questions remain to be addressed. First, it remains unclear whether the therapeutic effects require live bacteria or can be mediated by specific metabolites or bacterial components. In our in vitro assays, pasteurized KC84 retained serotonin-suppressive activity, suggesting that heat-stable components may contribute to this effect. However, as all in vivo experiments were performed using live bacteria, the relevance of non-viable components in vivo remains to be determined. Second, all in vivo data were generated using rodent models; thus, validation in human cohorts or patient-derived datasets will be essential for clinical translation. In this regard, given the complexity of IBS-D pathophysiology, the 5-HTP-induced model does not fully recapitulate the multifactorial nature of the disease, including stress, low-grade inflammation, and visceral hypersensitivity. Third, additional physiological approaches to assess smooth muscle function would be helpful in further characterizing contractile responses. In addition to the collagen gel contraction assay used in this study, more quantitative methods such as intestinal smooth muscle tension assays may provide a more precise evaluation of contractile function. Fourth, although our data suggest an association between KC84-induced IFN-β production and reduced muscle contractility, a causal role of IFN-β signaling in mediating these effects has not been definitively established. Future studies employing IFN pathway-deficient models, such as IFNAR1 knockout mice, or neutralization approaches using anti-IFN-β antibodies will be important to further validate this mechanism. Finally, although our data suggest that KC84-induced IFN-β production may contribute to the attenuation of smooth muscle contractility, the relatively low frequency of IFN-β-producing CD11b^- DC and the lack of direct functional validation limit the strength of this interpretation. Future studies using functional approaches, such as co-culture systems using KC84-treated dendritic cells and colonic smooth muscle cells, will be required to establish a causal link.

In conclusion, this study identified B. stercoris KC84 as a mechanism-based live biotherapeutic candidate for IBS-D that alleviates IBS-D-like symptoms through dual actions: modulation of serotonin-related pathways and IFN-β-mediated attenuation of colonic smooth muscle contractility. By targeting both neurochemical and immune pathways, KC84 may offer a novel and potentially safer therapeutic approach for IBS-D. These findings not only expand the mechanistic understanding of type I IFN biology but also underscore the promise of human-derived probiotics as targeted interventions for gut motility disorders.

Data availability

Processed RNA-seq data have been deposited in the NCBI Gene Expression Omnibus (GEO) under accession number GSE17065, and raw sequencing data are available in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1402738

Use of AI-assisted tools

During the preparation of this work, the authors used ChatGPT (OpenAI, GPT-4) to improve the clarity and grammar of the manuscript. The AI tool was not used for data analysis, data interpretation, figure generation, statistical analysis, study design, or determination of the scientific direction of the study. All scientific content and conclusions were developed, reviewed, and approved by the authors.

Disclosure of potential conflicts of interest

This study was funded by Celltrion Inc. Some authors are affiliated with the funding organization or collaborating institutions. No other potential conflicts of interest were reported.

Funding

This study was supported by Celltrion Inc. (Incheon, Republic of Korea).

CRediT authorship contribution statement

Seokcheon Song: Writing – original draft, Conceptualization, Project administration, Investigation, Formal analysis, Data curation, Visualization. Myeonghwan Han: Conceptualization, Investigation, Data curation. Kwanwoo Moon: Investigation, Data curation. Haeseong Lee: Investigation, Data curation. Sungchan Ha: Investigation. Juseok Seo: Funding acquisition, Conceptualization. Seungki Lee: Funding acquisition. Soyong Jang: Funding acquisition. Sang Kyun Lim: Supervision. Nayoung Kim: Supervision. GwangPyo Ko: Supervision.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

GwangPyo Ko reports financial support was provided by Celltrion, Inc. Sang Kyun Lim, Seokcheon Song, Juseok Seo, Seungki Lee, and Soyong Jang has patent pending to KoBioLabs, Inc. and Celltrion, Inc. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. Supplementary materials

Supplementary Fig. S1. Bifidobacterium species specificity for serotonin reduction and inhibition of mast cell degranulation. (A) Reduction of serotonin levels by bacterial strains was calculated relative to the vehicle control, which was set at 100 %. 128 Bifidobacterium strains were tested at a 1:100 cell-to-bacterium ratio. Mean values for individual bacterial strains are presented as dots, with strains grouped by species. (B) Serotonin reduction results for B. stercoris strains, arranged in ascending order (n = 3). KC84 is indicated in blue. (C) Mast cell degranulation inhibition by bacteria was calculated relative to the vehicle group and plotted using the same format as the serotonin analysis. (D) Degranulation inhibition results for B. stercoris strains, arranged in ascending order (n = 3). KC84 is indicated in blue. Data are shown as mean ± SEM. Supplementary Fig. S2. Body weight and individual colitis score data in a 5-HTP-induced mouse model. (A) Schematic of the experimental design for the 5-HTP-induced IBS model and oral administration of KC84 on D14. (B) Body weight was measured daily, and statistical significance between groups was analyzed using two-way ANOVA. (C–G) Individual colitis scores were determined for abnormalities of mucosal architecture (C), extent of inflammation (D), epithelial regeneration (E), erosion or ulceration (F), and percent involvement (G). Data are shown as mean ± SEM (n = 6/group). *p < 0.05; **p < 0.01; ***p < 0.001 vs. 5-HTP/vehicle group. Supplementary Fig. S3. Type I IFN-related gene set enrichment analysis and expression of circadian rhythm-related genes. (A–C) Gene ontology (GO)-based gene set enrichment analysis (GSEA) was performed, and representative enriched terms are shown: response to IFN-β (A), IFN-mediated signaling pathway (B), and response to type I IFN (C). (D) Bar plots showing the expression levels of representative circadian rhythm-related genes on a log₂(TPM + 1) scale. Data are presented as means ± SEM, and statistical significance was determined using Benjamini-Hochberg-adjusted p-values from DESeq2 analysis. *p < 0.05; **p < 0.01; ***p < 0.001 vs. 5-HTP/vehicle group. Supplementary Fig. S4. KC84 induced IFN-β production from colonic immune cells. (A–H) Colonic lamina propria cells were incubated with KC84 for 6 h and analyzed by flow cytometry (n = 6/group). Representative plots are shown for IFN-β-producing dendritic cells (A), monocytes (C), CD11b⁺ dendritic cells (E), and plasmacytoid dendritic cells (G). Corresponding quantifications are presented in (B), (D), (F), and (H), respectively. Data are shown as mean ± SEM. *p < 0.05; **p < 0.01 vs. vehicle group. Supplementary Fig. S5. Gating strategy. Supplementary Fig. S6. α-SMA expression in cultured colonic smooth muscle cells. (A–D) Colonic smooth muscle cells (SMCs) and MC38 cells were stained with α-SMA antibody and analyzed by flow cytometry (n = 4/group). Representative histogram (A) and scatter plots (B) are shown for α-SMA expression in colonic smooth muscle cells and MC38 cells Corresponding quantifications are presented in (C) and (D). Data are shown as mean ± SEM. ***p < 0.001 vs. MC38 cell line group.