Glycocholic acid inhibits TRIB3-ID1 axis to acelerate colitis progression via suppressing intestinal stem cell renewal

Abstract

Intestinal stem cell number or their regeneration ability is crucial for attaining mucosal healing. Deciphering the molecular mechanisms responsible for the impairment of intestinal stem cells in inflammatory bowel disease could yield innovative therapeutic insights. Altered bile acid metabolism is a hallmark feature of inflammatory bowel disease, typically characterized by elevated fecal levels of primary bile acids, such as glycocholic acid. However, the relationship between glycocholic acid and inflammatory bowel disease remains unclear. Here, we report that glycocholic acid accelerates inflammatory bowel disease progression through downregulating TRIB3 expression to disrupt intestinal stem cells self-renewal. TRIB3 is highly expressed in crypt cells and sustains intestinal epithelial stemness by preventing ID1 palmitoylation and AP3D1-mediated lysosomal degradation. Glycocholic acid is found to suppress TRIB3-ID1 axis, thereby compromising intestinal epithelium integrity. Notably, we identify Bergenin, a natural compound, as a potential therapeutic agent against inflammatory bowel disease via upregulating TRIB3. These findings highlight the TRIB3-ID1 axis as a promising therapeutic target for inflammatory bowel disease therapy.

Glycocholic acid, a bile acid, accelerates IBD by downregulating the TRIB3-ID1 axis, impairing intestinal stem cells renewal. A natural compound—Bergenin, alleviates colitis via restoring TRIB3 and intestinal integrity.

Introduction

Inflammatory bowel disease (IBD), including ulcerative colitis (UC) and Crohn’s disease (CD), is a chronic and relapsing disorder with increased prevalence all over the world¹. Mucosal healing (MH) is universally recognized as the gold standard that must be achieved to induce long-term IBD remission^2–4. However, current available IBD treatments almost exclusively focused on promoting MH by suppression of inflammation, which are insufficient to completely cure IBD. Given the infectious risk associated with chronic immunosuppressive therapy, the goal of attaining MH without immunosuppression is an attractive strategy⁵.

The single layer of intestinal epithelial cells (IECs) forms the fundamental structural basis of the mucosal barrier, segregating the host from harmful contents in the intestinal lumen. Dysregulation or depletion of intestinal stem cells (ISCs), residing in the bottom of the stem cell niche—crypt, are major causes for IECs replenishment failure and intestinal mucosal barrier disruption^6–8. Methods to increase ISCs number or their regeneration ability to promote MH such as organoid transplantation, have shown long-term remission in the treatment of IBD within mouse models^9,10. However, the technical challenges, such as organoids delivering limited their further application. Activating pro-proliferative signals such as STAT3/STAT5 and Hippo-YAP/TAZ pathway can promote ISCs proliferation, consequently accelerating crypt regeneration following injury^11–13. These findings suggest that deciphering the molecular mechanisms responsible for ISCs impairment in IBD could yield innovative therapeutic insights, potentially leading to the identification of novel targets and strategies for achieving MH.

Multiple IBD patient cohorts have consistently demonstrated elevated fecal conjugated primary bile acids (BAs) alongside reduced secondary BAs, indicating a close relationship between BAs disorder and IBD progression^14–17. Primary sclerosing cholangitis (PSC) is a common chronic cholestatic liver disease characterized by cholestasis and impaired BA homeostasis¹⁸. Notably, up to 80% of PSC patients develop concurrent IBD, particularly UC¹⁹, whereas only approximately 1～2% of IBD patients progress to PSC^20,21. Such comorbidity disparity suggests that impaired BA homeostasis may represent a potential cause for IBD development and progression. Meanwhile, it has been reported that the dysregulation of lithocholic acid (LCA) and cholic acid (CA) contributes to IBD progression through regulating immune response or ISCs stemness traitsm^17,22. Glycocholic acid (GCA), a conjugated primary BA, was found with higher abundance in the faeces of IBD patients^16,23. However, the function of GCA in IBD has not been studied yet.

Tribbles homolog 3 (TRIB3) is considered as a stress sensor, which has been comprehensively studied by our group in neoplastic disease^24–27. These studies demonstrated that TRIB3 supports stemness traits of tumor cells in colorectal cancer (CRC), breast cancer, lung cancer and hematological malignancies. Other groups also reported the stemness-maintaining role of TRIB3 in normal somatic cells. For example, TRIB3 can inhibit adipocyte differentiation through the interaction with PPAR²⁸. Elevated TRIB3 expression contributes to the preservation of stem-like characteristics in mouse spermatogonial stem cells²⁹. However, the role of TRIB3 in maintaining ISC stemness and IBD progression remains unclear.

Here, we demonstrate that TRIB3 is highly expressed in intestinal crypts and maintains the ISCs stemness traits through interacting with and upregulating the protein stability of the inhibitor of differentiation protein 1 (ID1), a marker of intestinal stem/progenitor cells³⁰. GCA downregulates the expression of TRIB3 to accelerate the degradation of ID1, which results in the impairment of both reserve stem (RSC) cells and crypt base columnar (CBC) cells. Bergenin, a natural product, was found to enhance ISCs self-renewal and alleviate colitis progression via upregulating TRIB3 expression, emphasizing the potential of increasing TRIB3 expression as a potential strategy to promote ISC self-renewal and MH in IBD therapy.

Results

GCA impairs ISCs self-renewal to accelerate IBD progression

Elevation of conjugated bile acid—GCA, which is produced by both host and microbe, in fecal samples was reported in IBD patients compared to healthy controls (nearly 10-fold), as well as in active UC patients relative to those in remission stages^16,23,31. Therefore, we wanted to investigate the potential role of GCA in colitis progression. The dextran sulfate sodium (DSS) induced colitic mouse model provides a powerful tool for studying IBD³². Unlike the reported CA, the abundance of fecal GCA did not significantly increase after DSS treatment in C57BL/6 J mice (Supplementary Fig. 1a). Hence, exogenous supplementation of GCA were given to DSS-induced colitis mice via oral gavage. GCA treatment at 200 mg/kg yielded fecal GCA concentration approximating 1000 nmol/g (Supplementary Fig. 1b), closely mimicking the mean bile acid abundance observed in human IBD patients³¹. Based on this human-mimetic concentration, both 200 mg/kg and a sub-effective dose of 100 mg/kg GCA were selected for functional investigation in the UC model. GCA aggravated colitis in a dose dependent manner in DSS-treated male C57BL/6 J mice (Fig. 1a). Notably, mice receiving 200 mg/kg GCA exhibited a 10% extra body weight loss, higher disease activity index (DAI) and reduced colon length compared to Vehicle-treated controls (Fig. 1b–d). Besides, histopathological evaluation revealed larger areas of crypt loss, erosions and epithelial mucosa impairment in the colon after GCA administration (Fig. 1e). Given the critical role of mucosal barrier dysfunction in IBD, we assessed intestinal permeability using FITC-dextran in the colitis model. GCA administration significantly increased serum FITC-dextran levels, indicating exacerbated barrier leakage (Fig. 1f). Furthermore, GCA administration downregulated the expressions of tight junction genes in colon tissues, including Tight junction protein 1 (Tjp1) and Cadherin 1 (Cdh1), consistent with a leaky epithelial phenotype (Fig. 1g). Similar results were also got in female C57BL/6 J mice (Supplementary Fig. 1c–h). To test whether GCA supplementation alone could initiate colitis, 200 mg/kg GCA was given to healthy C57BL/6 J mice. Similar as observed in DSS conditions, GCA treatment induced a 4% reduction in body weight, elevated DAI, and marked colon shortening (Supplementary Fig. 1i–l). Moreover, mild damage of epithelial colon mucosa was also observed in the GCA treatment group (Supplementary Fig. 1m). These data collectively indicated that supplementary of GCA exacerbates intestinal injury.

Fig. 1. GCA supplementation impairs ISCs self-renewal and mucosal barrier integrity to accelerate IBD progression.

a Schematic diagram of GCA treatment. b, c Body weight change (b) and DAI (c). d Representative gross colon images and statistical data of colon length. e Representative H&E staining of colon tissues and statistical data of histology score. Scale bar = 1 mm (upper) or 50 μm (lower). f Relative serum FITC-dextran abundance. g Relative mRNA expression of tight junction-related genes in the colon tissues. h Representative EdU and Cdh1 staining in the colon tissues and the statistical data. Scale bar = 20 μm. i Representative Bmi1 immunofluorescence staining in the colon tissues and statistical data of Bmi1⁺ cell number per crypt (n = 5 independent biological experiments). Scale bar = 5 μm. j, k Representative histogram of Lgr5-FITC in the flow cytometry analysis towards colon tissues of Lgr5-EGFP-IRES-creERT2 mice treated with or without GCA (j), and statistical data of the Lgr5⁺ cell ratio in colonic mucosa (k). l Representative images (left), and the proliferation curve (right) of colon organoids (n = 10 independent biological experiments). Scale bar = 200 μm. m Representative EdU staining in the colonic organoids described in (l) and the statistical data. Scale bar = 20 μm. n FITC-dextran permeability was measured in mouse colon organoids treated with or without GCA (300 μM) or IL-1β (50 ng/ml) for 24 h. Scale bar = 50 μm. Data were presented as the mean ± SEM and analyzed by Kruskal-Wallis test (c, g-Cdh1 and n), Ordinary one-way ANOVA (b, d–f, g-Tjp1), unpaired two-tailed Student’s t test (h, j–m) and Mann-Whitney U test (i). NS, no statistical significance. n = 6 independent biological experiments for (a–d, f–h, m and n). n = 3 independent biological experiments for (e, j and k). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (a) are created in BioRender. shang, S. (2026) https://BioRender.com/tbfxh4l.

Histopathological analysis revealed concurrent immune cell infiltration and epithelial barrier disruption in the colons of GCA-treated mice. The underlying molecular mechanisms of GCA in exacerbating colitis was further explored. Given that both epithelial barrier dysfunction and immune hyperactivation independently contribute to IBD pathogenesis while engaging in mutual regulatory interactions^5,33, we sought to delineate whether GCA aggravates colitis through modulation of the immune microenvironment or disruption of epithelial homeostasis. The differentiation and function of intestinal macrophages replenished by blood monocytes are critical for the intestinal inflammation in IBD³⁴, therefore, we isolated bone marrow-derived monocytes/macrophages from wild-type C57BL/6 J mice to perform in vitro study. GCA treatment at 300 μM (equivalently 1000 nmol/g in feces of IBD patients) exhibits no discernible impact on the proliferation of CD11b⁺ myeloid cells, as evidenced by the carboxyfluorescein succinimidyl ester (CFSE) proliferation assay (Supplementary Fig. 1n). In addition, it does not influence the polarization of CD11b⁺F4/80⁺ macrophages towards either the M1 or M2 phenotypes (Supplementary Fig. 1o). In addition to macrophages, disturbance among T cell subsets particularly T_H1, T_H17 and T_reg cells also plays a critical role in driving IBD progression. As it has been reported, GCA has no obvious influence on the differentiation of CD4⁺ T cells toward T_H17 or T_reg cells³⁵. Here, GCA treatment does not affect the proliferation of naïve CD4⁺ T cells isolated from wild-type C57BL/6 J mice or their differentiation into T_H1 cells (Supplementary Fig. 1p, q). These data indicate that GCA does not exacerbate the progression of colitis by directly enhancing inflammatory responses.

Next, we explored the effect of GCA on epithelia mucosal barrier integrity. We found that GCA obviously decreased the EdU⁺ cells ratio as well as Cdh1 protein expression in the colon, indicating an inhibitory effect on intestinal cell proliferation and barrier integrity (Fig. 1h). In the intestinal crypt, two types of ISCs contribute to epithelium homeostasis and regeneration: crypt base columnar cells (CBCs, marked by Lgr5 and Olfm4) and reserve stem cells (RSCs, marked by Bmi1 and Dclk1)³⁶. We demonstrated that GCA treatment reduced the mRNA expression of all these four markers of ISCs (Supplementary Fig. 1r). Furthermore, GCA treatment reduced the number of Bmi1⁺ RSCs (Fig. 1i). Utilizing Lgr5-EGFP-CreERT2 transgenic mice, we confirmed diminished Lgr5 expression without alterations of Lgr5⁺ CBC counts in colonic sections after GCA administration (Supplementary Fig. 1s and Fig. 1j, k). A slower proliferation rate was observed in colonic and small intestinal (SI) organoids isolated from colitis mice treated with GCA compared to the untreated group (Fig. 1l, m, and Supplementary Fig. 1t, u). These data imply that GCA treatment leads to ablation of Bmi1⁺ RSCs, and results in functional impairment of Lgr5⁺ CBCs to inhibit the proliferation of intestinal cells. Furthermore, we analyzed whether GCA treatment impairs epithelium barrier integrity in vitro. The permeability of colon or intestinal organoids to 4 kDa FITC-dextran were very weak under normal conditions as reported before^37,38. While strong fluorescence was observed in the lumen of organoids treated with GCA (300 μM) for 24 hours, similar with the positive control of IL-1β treatment (Fig. 1n and Supplementary Fig. 2a). These findings demonstrate that GCA supplementation exacerbates colitis by reducing the number and function of intestinal stem cells (ISCs) and impairing epithelial barrier integrity.

GCA impairs ISCs proliferation through downregulating TRIB3

Our previous study has revealed the stemness maintaining role of TRIB3 in CRC cells²⁴. Here, we found that TRIB3 is also involved in the regulation of IECs self-renewal. Immunohistochemistry (IHC) staining showed higher TRIB3 protein expression in the crypt base regions compared to more differentiated epithelial cells within both human and murine colonic and small intestinal tissues (Supplementary Fig. 3a–c). In addition, we isolated crypt and villus cells from the small intestine of healthy C57BL/6 J mice and found that both the mRNA and protein abundance of TRIB3 were higher in the crypts than those in the villus (Supplementary Fig. 3d, e). Then we generated intestinal epithelial cell-specific Trib3 knockout mice (Trib3^IEC-KO) by crossing Vil1-Cre mice with Trib3-floxed (Trib3^f/f) mice (Supplementary Fig. 3f) to further explore the potential biological function of TRIB3 in ISCs. Trib3 depletion downregulated both Lgr5 and Bmi1 protein expression in the colon and small intestinal crypts (Supplementary Fig. 3g, h). Consistently, we observed a reduced number of Bmi1⁺ RSCs and diminished Lgr5 expression in the crypt of both colon and small intestinal in the Trib3^IEC-KO mice compared to the Trib3^f/f group (Supplementary Fig. 3i–l). Moreover, depletion of Trib3 inhibited the colon or SI organoids proliferation, indicating a weaker self-renewal capacity of ISCs (Supplementary Fig. 3m, n). Furthermore, depletion of Trib3 enhanced the permeability of intestinal organoids to 4 kDa FITC-dextran, an effect comparable to that of GCA treatment in vitro, suggesting that Trib3 is essential for maintaining the integrity of the intestinal epithelial barrier (Supplementary Fig. 3o, p). Collectively, our data suggested that TRIB3 is highly expressed in the intestinal crypts and maintains ISCs self-renewal.

We next investigated the role of TRIB3 in colitis progression. GEO datasets (GSE11223 and GSE105074) analysis revealed lower TRIB3 mRNA abundance in the inflamed intestinal tissues of UC patients compared to healthy ones (Fig. 2a, b). To determine the impact of intestinal epithelium TRIB3 on colitis, we assessed the effect of epithelial-specific Trib3 deletion on chronic and acute experimental colitis in mice. Depletion of Trib3 accelerated body weight loss, increased DAI, shortened colon length and upregulated the mRNA expression of inflammatory cytokines in the colonic mucosa from both the two animal models (Fig. 2c–g and Supplementary Fig. 4a–e). Besides, histopathological assessment further demonstrated that intestinal Trib3 depletion aggravated mucosal erosions, submucosal oedema, and substantial loss of crypt architectural integrity (Fig. 2h and Supplementary Fig. 4f). Moreover, Trib3 deficiency compromised mucosal barrier function, evidenced by elevated serum FITC-dextran levels and reduced expression of tight junction-associated genes (Fig. 2i, j and Supplementary Fig. 4g, h). Depletion of Trib3 reduced colonic epithelial cell proliferation and Cdh1 expression as evidenced by EdU and Cdh1 co-immunofluorescence staining (Fig. 2k and Supplementary Fig. 4i). Besides, depletion of Trib3 suppressed the mRNA abundance of Lgr5 and Olfm4 (markers for CBCs), as well as Bmi1 and Dclk1 (markers for RSCs) in the colonic mucosa (Fig. 2l and Supplementary Fig. 4j). In the in vitro culture system, reduced proliferation of intestinal organoids isolated from chronic UC model was also observed in the Trib3^IEC-KO group compared to the control ones (Fig. 2m, n). These data suggested that loss of Trib3 in the epithelial cells damaged ISCs self-renewal and mucosal barrier integrity to accelerate colitis progression.

Fig. 2. Intestinal Trib3 depletion accelerates while overexpression decelerates colitis progression.

a, b TRIB3 mRNA expression in normal uninflamed colon (abbreviated as HC) and UC inflamed colon (abbreviated as UC) tissues according to the dataset of GSE11223 (n = 69 in HC and 63 in UC group) (a) and GSE105074 (n = 28 in HC and 23 in UC group) (b). c Schematic diagram of chronic UC mouse model. d, e Body weight change (d) and DAI (e). f Representative gross colon images and statistical colon length. g Relative mRNA expressions. h Representative H&E staining and statistical histology score. Scale bar = 1 mm (upper) or 50 μm (lower). i Relative serum FITC-dextran abundance. j Relative mRNA expressions. k Representative EdU and Cdh1 staining. Scale bar = 20 μm. l Relative mRNA expressions. m Representative images (left) and proliferation curve (right) of colon organoids (n = 10 independent biological experiments). Scale bar = 200 μm. n Representative EdU staining in the colonic organoids. Scale bar = 20 μm. o Schematic diagram of acute severe UC mouse model. p Immunoblots of Trib3 in the colonic mucosa. q, r Body weight change (q) and DAI (r). s Representative gross colon images and statistical colon length. t Relative mRNA expressions. u Representative H&E staining and the statistical histology score. Scale bar = 1 mm (upper) or 50 μm (lower). v Relative FITC fluorescence in the serum of mice. w Relative mRNA expressions in the colon tissues. Data were presented as the mean ± SEM and analyzed by unpaired two-tailed Student’s t test (a, d, f–k left, m, p, q, t-Tnf-α, u and w), Mann-Whitney U test (e, n, r, s and t-IL-1β) and Welch’s t test (b, k right, l, t-IL-6 and v). n = 3 independent biological experiments in (h, p and u). n = 6 independent biological experiments in (c–g, i–l, n, o and q). n = 4 in the Ctrl group and 6 in Trib3^OE group for (r–t, v and w). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (c, o) are created in BioRender. Shang, S. (2026) https://BioRender.com/tbfxh4l.

Furthermore, Adeno-associated viruses (AAV)-Trib3 was applied in the DSS-induced UC model to validate the impact of Trib3 on colitis progression (Fig. 2o). Trib3 overexpression alleviated body weight loss, decreased the DAI, and accelerated the recovery from DSS-induced colitis (Fig. 2p–r). After sacrificing the mice, we found that overexpression of Trib3 relieved shortened colon length and decreased the expression of inflammatory cytokines in the colon tissues (Fig. 2s, t). Histopathological assessment further revealed that intestinal Trib3 overexpressing effectively alleviated erosions and intestinal oedema, retained more intact crypt structures in colitis mice (Fig. 2u). In addition, overexpression of Trib3 decreased the FITC-dextran content and upregulated the mRNA level of tight conjunction related genes in the colon mucosa compared to the Ctrl group, indicating a more intact mucosal barrier (Fig. 2v, w).

Then we wondered whether GCA treatment influences the expression of TRIB3. Either in the UC mouse model or in the NCM460 cells derived from the normal human colon mucosa, GCA administration downregulated TRIB3 expression (Fig. 3a, b). We next investigated the mechanism by which GCA regulates TRIB3. We found that GCA downregulated TRIB3 mRNA expression (Fig. 3c). Besides, even with the concurrent inhibition of both proteasomal and lysosomal degradation pathways by MG132 and bafilomycin, GCA treatment still reduced TRIB3 protein expression, indicating that GCA inhibits TRIB3 transcription but has no effect on its protein stability (Fig. 3d and Supplementary Fig. 5a). ATF4, β-catenin, CHOP, FOXO1, and SMAD3 have been reported as transcription factors of TRIB3^24,39–41. Using a TRIB3 promoter luciferase reporter in NCM460 cells, we found that only ATF4 knockdown rescued the GCA-induced reduction of TRIB3 transcription (Fig. 3e and Supplementary Fig. 5b–j). Consistent with this, knocking down ATF4 also abrogated the GCA-induced reduction of TRIB3 at the protein level. Furthermore, GCA was shown to downregulate ATF4 protein expression (Fig. 3f). Our data indicated that GCA attenuates TRIB3 transcription by downregulating the expression of its key transcriptional activator—ATF4.

Fig. 3. GCA impairs ISCs proliferation and destroys mucosal barrier integrity to exacerbate colitis through downregulating TRIB3.

a Immunoblots (left) of Trib3 in colonic mucosa of mice treated with 200 mg/kg GCA or not, and statistical relative protein level (right). b Immunoblots (left) of TRIB3 in NCM460 cells treated with different doses of GCA and statistical relative protein level (right). c Relative mRNA expression of TRIB3 in NCM460 cells treated with or without GCA. d Immunoblots of TRIB3 under the treatment of GCA with or without the addition of MG132 and bafilomycin in NCM460 cells. e, f Relative TRIB3-reporter luciferase activity (e) and immunoblots of TRIB3 and ATF4 (f) in NCM460 cells transfected with Ctrl or ATF4 specific siRNAs treated with or without 300 μM GCA (n = 4 independent biological experiments in e). g, h Schematic diagram of UC mouse model. Body weight change (g) and DAI (h) of mice. i Representative gross colon images and statistical data of colon length. j Relative mRNA expressions in the colonic mucosa of mice in the indicated groups. k Representative H&E staining and statistical data of histology score. Scale bar = 1 mm (upper) or 50 μm (lower). l Relative serum FITC-dextran abundance. m Relative mRNA expression of tight junction-related genes in the colon tissues. n Relative mRNA expressions. o Representative images and proliferation curve of colon organoids isolated from mice (n = 10 independent biological experiments). Scale bar = 200 μm. p Representative EdU staining and the statistical data (n = 6 independent biological experiments). Scale bar = 10 μm. Data were presented as the mean ± SEM and analyzed by Ordinary one-way ANOVA (b, d–i, j-IL-1β/IL-6, k, l, m, n-Lgr5/Olfm4, o and p), Welch’s t test (a), unpaired two-tailed Student’s t test (c), Brown-Forsythe and Welch ANOVA test (n-Bmi1/Dclk1) and Kruskal-Wallis test (j-Tnf-α). n = 3 independent biological experiments for (a–d, f and k). n = 6 independent biological experiments for (g–j, l–n). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (g) are created in BioRender. shang, S. (2026) https://BioRender.com/tbfxh4l.

We then investigated whether GCA treatment promotes colitis via downregulating TRIB3. Here, overexpression of Trib3 largely recovered the body weight loss and increased DAI caused by GCA administration (Fig. 3g, h). Consistent with these symptoms, ectopic expression of Trib3 attenuated colon shortening and decreased inflammatory cytokine expression in colon tissues induced by GCA treatment (Fig. 3i, j). Histopathological analysis revealed that intestinal Trib3 overexpression significantly alleviated mucosal erosions and oedema while preserving crypt structural integrity in colitis mice compared to the GCA-treated group (Fig. 3k). Furthermore, Trib3 overexpression reversed both the elevated serum FITC-dextran levels and the downregulated mRNA expression of tight junction-related genes in the colonic mucosa caused by GCA administration, indicating restoration of mucosal barrier integrity (Fig. 3l, m). Trib3 overexpression markedly rescued GCA-impaired proliferation of colonic epithelial cells and epithelial barrier integrity in DSS-induced UC mice, as evidenced by EdU and Cdh1 staining (Supplementary Fig. 6a). Moreover, ectopic expression of Trib3 also reversed the GCA-suppressed ISC markers expression in the colonic mucosa (Fig. 3n), and rescued the GCA-inhibited proliferation and permeability of colon organoids (Fig. 3o, p and Supplementary Fig. 6b). Collectively, our data suggested that GCA treatment exacerbates colitis by inhibiting ATF4-mediated TRIB3 transcription in the intestinal epithelium, thereby accelerating intestinal barrier disruption and colitis progression.

TRIB3 loss lowers ID1 stability to destroy ISCs regeneration

We next explored how GCA-mediated TRIB3 protein loss impairs ISC function and promotes IBD progression. Gene set enrichment analysis (GSEA) was applied to identify stemness-related signaling pathways associated with TRIB3. However, no significant enrichment of classical ISC-associated signaling pathways (Wnt, Notch, Hippo-YAP) were identified in TRIB3^high or TRIB3^OE groups through analyzing either UC patients mRNA microarray dataset (GSE105074) or our transcriptome profiling of Ctrl and TRIB3^OE NCM460 cells (Supplementary Fig. 7a–f). Given that TRIB3 typically serves as a scaffold protein influencing the stability and transcriptional regulation of its binding partners, we hypothesized that its role in this context might also be mediated through protein-protein interactions. Among the TRIB3-interacting proteins identified by yeast two-hybrid screening performed previously by our group, ID1 is the only one responsible for ISCs stemness maintenance, which is highly expressed in both intestinal RSCs and CBCs³⁰ (Fig. 4a). Using co-immunoprecipitation (Co-IP), proximity ligation assay (PLA) and multiple immunohistochemistry analysis, we confirmed the direct interaction between ID1 and TRIB3 in HEK293T cells, NCM460 cells and human colon tissues (Fig. 4b–d). Mapping the interaction region between TRIB3 and ID1, we found that the N-terminal region of TRIB3 kinase domain (KD), and the N-terminus or helix-loop-helix (HLH) domain of ID1 are critical for the binding of the two proteins (Supplementary Fig. 7g, h). Further analysis showed that depletion of TRIB3 downregulated ID1 protein expression both in the colon tissues from the acute UC model or in the NCM460 cells (Fig. 4e, f). Besides, GCA administration also downregulated Id1 protein expression, which was rescued by Trib3 overexpression (Fig. 4g, h). Mechanistically, we found that overexpression of TRIB3 had no effect on the ID1 mRNA expression in NCM460 cells, but prolonged its protein half-life in HEK293T cells (Fig. 4i and Supplementary Fig. 7i). Besides, knockdown of TRIB3 shortened the half-life of ID1 in the NCM460 cells (Fig. 4j). These data indicated that TRIB3 interacts with ID1 to enhance its protein stability in the intestinal epithelium.

Fig. 4. GCA-induced TRIB3 loss reduces ID1 protein stability to destroy ISCs regeneration and accelerate IBD progression.

a Schematic diagram of yeast two-hybrid screening for proteins that interact with TRIB3. b Cell extracts from HEK293T cells co-transfected with indicated expressing plasmids were immunoprecipitated with IgG or anti-HA antibody. c PLA for defining the interaction between TRIB3 and ID1 (n = 8 independent biological experiments). Scale bar = 5 μm. d Immunofluorescence staining of TRIB3 and ID1 in human colon tissue. Scale bar = 20 or 5 μm. e Immunoblots of Id1 and Trib3 in colon mucosa of Trib3^f/f and Trib3^IEC-KO mice in acute UC model. f Effect of silencing TRIB3 on ID1 expression in NCM460 cells. g Immunoblots of Id1 and Trib3 in the colon mucosa of the indicated mice in the UC model. h Immunoblots of ID1 and TRIB3 in indicated groups of NCM460 cells. i The effect of TRIB3 overexpression on ID1 protein stability in HEK293T cells. The dashed line intersects the degradation curves at the half-life. j Effects of silencing TRIB3 on ID1 protein stability in NCM460 cells. k Schematic diagram of the DSS-induced acute UC mouse model. l, m Body weight change (l) and DAI (m) of mice in the indicated groups. n Representative gross colon images and statistical data of colon length in the indicated groups. o Representative H&E staining and statistical data of histology score in the indicated groups. Scale bar = 1 mm (upper) or 50 μm (lower). p Relative serum FITC-dextran abundance. q Representative immunofluorescence staining of EdU and the statistical data. Scale bar = 20 μm. r Relative mRNA expressions. Data were presented as the mean ± SEM and analyzed by unpaired two-tailed Student’s t test (e, f, l–o, q and r), Ordinary one-way ANOVA (g and h), and Mann-Whitney U test (p). n = 3 independent biological experiments for (b, d–j, o and q). n = 5 independent biological experiments in Id1^KO-Ctrl group and n = 6 in Id1^KO-Trib3^OE group for (k–n, p and r). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (k) are created in BioRender. shang, S. (2026) https://BioRender.com/tbfxh4l.

Next, Trib3 was overexpressed in UC-bearing (induced by DSS and GCA) Id1-knockout (Id1^KO) mice to determine whether ID1 serves as a key downstream effector of TRIB3 in regulating ISCs functions and IBD progression (Fig. 4k and Supplementary Fig. 7j, k). Here, the role of Trib3 overexpression in alleviating body weight loss, reducing DAI, increasing colon length, as well as decreasing histology score and intestinal permeability were all abolished after Id1 depletion (Fig. 4l–p). Besides, after knocking out of Id1, overexpression of Trib3 no longer influence intestinal epithelium cells proliferation or the mRNA abundance of RSCs and CBCs markers (Fig. 4q, r). Collectively, these data demonstrate that TRIB3 stabilizes ID1 through direct protein interaction in intestinal epithelial cells. GCA-induced TRIB3 downregulation reduces ID1 protein stability, thereby impairing ISC self-renewal and epithelial barrier integrity, accelerating IBD progression.

TRIB3 hinders ID1 palmitoylation and lysosomal degradation

More efforts were made to explore the molecular mechanism of TRIB3 in regulating ID1 protein stability in IECs. Here we observed that bafilomycin treatment—but not MG132—rescued the TRIB3 depletion-induced ID1 protein reduction in NCM460 cells (Fig. 5a and Supplementary Fig. 8a). Besides, overexpression of TRIB3 abrogated the colocalization of ID1 with lysosomal associated membrane protein 1 (LAMP1) (Fig. 5b). These data indicated that TRIB3 inhibits the lysosomal degradation of ID1. In our previous study, TRIB3 was found to interact with the autophagic receptor p62, inhibiting p62-mediated lysosomal degradation of ubiquitinated substrates. Given that ubiquitinated ID1 is known to bind with p62 for autolysosomal degradation⁴², we hypothesized that TRIB3 might stabilize ID1 by either disrupting ID1-p62 interaction or impairing ID1 ubiquitination. However, Co-IP assays demonstrated that TRIB3 depletion affected neither process (Supplementary Fig. 8b, c). To identify alternative mechanisms by which TRIB3 regulates ID1 stability, we performed Co-IP coupled with semi-quantitative mass spectrometry (Co-IP/MS) in NCM460 cells. This analysis revealed adapter-related protein complex 3 subunit delta 1 (AP3D1)—essential for lysosomal cargo trafficking—in ID1 immunoprecipitates. Moreover, TRIB3 overexpression significantly reduced AP3D1 protein abundance in the immunoprecipitates of ID1 (Fig. 5c). We further demonstrated that AP3D1 directly interacted with ID1 and accelerated its degradation through the lysosomal pathway (Fig. 5d, e and Supplementary Fig. 8d). Overexpression of TRIB3 interrupts the binding of ID1 with AP3D1 in HEK293T cells (Fig. 5f). Whereas knockdown of TRIB3 enhances the binding of ID1 and AP3D1 in NCM460 cells (Fig. 5g).

Fig. 5. TRIB3 hinders palmitoylation of ID1 to inhibit its lysosomal degradation in IECs.

a Immunoblots of ID1 and TRIB3 in Ctrl or TRIB3^KD NCM460 cells under Bafilomycin or MG132 treatment. b Effects of TRIB3 overexpression on the interaction between ID1 and LAMP-1 detected by PLA in NCM460 cells. Scale bar = 5 μm. c Cell lysates of HEK293T cells were immunoprecipitated with anti-HA antibody and underwent semi-quantitative MS analysis (n = 4 independent biological experiments). d Cell lysates of HEK293T cells transfected with ID1-HA and AP3D1-DDK were immunoprecipitated with IgG or anti-HA antibody. e Effects of AP3D1-DDK overexpression on ID1-HA protein degradation were measured and quantified in HEK293T cells. f Effects of TRIB3-GFP overexpression on the interaction between ID1-HA and AP3D1-DDK were detected with Bafilomycin treatment in HEK293T cells. g Effects of TRIB3 overexpression on the colocalization of ID1 and AP3D1 detected by PLA in NCM460 cells. Scale bar = 5 μm. h, i Effects of different doses of PB (h) or 2-BP (i) treatment on ID1 protein expression were quantified in NCM460 cells. j Effects of PB on ID1 protein degradation was measured and quantified in HEK293T cells. k Effects of PB on the interaction of ID1 and AP3D1 detected by PLA in NCM460 cells. Scale bar = 5 μm. l Effects of TRIB3 overexpression on ID1 palmitoylation by Click-iT reaction. m Colocalization of ID1 and AP3D1 in NCM460 cells analyzed by PLA in the indicated groups. Scale bar = 5 μm. n Effects of 2-BP on ID1 degradation were measured and quantified in Ctrl or TRIB3^KD NCM460 cells. Data were presented as the mean ± SEM and analyzed by unpaired two-tailed Student’s t test (c), Ordinary one-way ANOVA (h and i), Mann-Whitney U test (b), Welch’s t test (g and k) and Kruskal-Wallis test (m). n = 3 independent biological experiments for (a, d–f, h–j, l and n). n = 10 independent biological experiments for (g and m). n = 20 independent biological experiments for (b and k). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (c) are created in BioRender. shang, S. (2026) https://BioRender.com/ogz3byb.

We next investigated how TRIB3 disrupts ID1-AP3D1 binding. Given that palmitoylation is critical for the recognition and lysosomal sorting of proteins such as protease-activated receptor-1 (PAR1) and IFNGR1 by adapter proteins (e.g., AP3D1), we investigated whether ID1 undergoes palmitoylation and is subsequently delivered to the lysosome in an AP3D1-dependent manner^43–45. Palmostatin B (PB), the inhibitor of depalmitoylation enzymes, downregulated the protein expression of ID1 in a dose and time-dependent manner (Fig. 5h and Supplementary Fig. 8e). In contrast, 2-bromopalmitate (2-BP), the inhibitor of palmitoylation enzymes showed the opposite effect (Fig. 5i and Supplementary Fig. 8f). PB and 2-BP also shortened or prolonged the half-life of ID1, respectively (Fig. 5j and Supplementary Fig. 8g). Besides, PB enhanced the interaction of ID1 and AP3D1 in NCM460 cells (Fig. 5k). These data suggested that palmitoylation might be essential for AP3D1-mediated ID1 lysosomal degradation. Then we tested if TRIB3 affects ID1 palmitoylation to interrupt its binding with AP3D1. ID1 palmitoylation was then confirmed in HEK293T cells using Click-iT assay, which was indeed reduced a lot after TRIB3 overexpression (Fig. 5l). Besides, 2-BP treatment abrogated the interaction between AP3D1 and ID1, which could not be reversed by TRIB3 depletion (Fig. 5m). Similarly, TRIB3 depletion didn’t accelerate ID1 degradation with addition of 2-BP in NCM460 cells (Fig. 5n). These data collectively indicated that TRIB3 interacts with ID1 to hinder its palmitoylation, which abolished the recognition and lysosomal degradation of ID1 mediated by AP3D1.

Bergenin upregulates TRIB3 to restore ISCs self-renewal

These above findings suggested that upregulating TRIB3 might be a potential therapeutic strategy against IBD. Screening of the CMAP database with the transcriptome signature of TRIB3^OE NCM460 cells suggested several potential compounds for upregulating TRIB3 (Fig. 6a). Among them, Bergenin, Isoxsuprine, Cilnidipine, JW-7-24-1 and Epothilone A were predicted with the highest connectivity score of 99.98. Bergenin, Isoxsuprine, and Cilnidipine have been reported with cytoprotection effects, while JW-7-24-1 and Epothilone A were anti-tumor agents with cell killing activity^46–50. Cytoprotective compounds were selected for further confirmation, and Bergenin was the only one upregulating TRIB3 and ID1 protein expressions in NCM460 cells (Fig. 6b). Furthermore, we found that Bergenin upregulated TRIB3 transcription but not protein stability (Supplementary Fig. 9a, b). We next explored how Bergenin upregulates ID1 abundance. Treatment of NCM460 cells with Bergenin significantly reduced the palmitoylation of ID1 (Fig. 6c). Furthermore, Bergenin extended the half-life of ID1. This effect was reversed by the autophagy-lysosome inhibitor Bafilomycin, indicating that Bergenin maintains ID1 protein stability by inhibiting its lysosomal degradation (Fig. 6d). Consistent with this mechanism, Bergenin also counteracted AP3D1-accelerated ID1 degradation (Fig. 6e).

Fig. 6. Bergenin was found to upregulate TRIB3 and inhibit ID1 degradation.

a Schematic diagram for screening compounds upregulating TRIB3. Differentially expressed genes identified by transcriptome sequencing in TRIB3^OE vs Ctrl NCM460 cells were submitted to the CMAP platform to identify compounds that increase TRIB3 levels. b Effects of different doses of Isoxsuprine, Cilnidipine and Bergenin on the protein expression of TRIB3 and ID1 in NCM460 cells. c The effect of 80 nM Bergenin on the palmitoylation of ID1-DDK in NCM460 cells. d The effect of Bergenin on ID1 degradation with or without treatment of Bafilomycin in NCM460 cells. e The effect of Bergenin on the AP3D1-mediated degradation of ID1. Data were presented as the mean ± SEM and analyzed by Ordinary one-way ANOVA (b, c). n = 3 independent biological experiments for (a–e). All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (a) are created in BioRender. shang, S. (2026) https://BioRender.com/ogz3byb.

Bergenin belongs to isocoumarins, which is the major bioactive ingredients of a traditional Chinese medicine—Bergenia purpurascens, and is clinically used for the treatment of chronic bronchitis in China. This compound has been reported to inhibit DSS and 2,4,6-Trinitrobenzene sulfonic acid (TNBS) induced mouse or rat colitis via regulating gut microbiota or macrophage activation, while its impact on ISCs self-renewal was still unknown^51–53. As reported, Bergenin treatment at 25 or 50 mg/kg alleviated body weight loss, decreased DAI, relieved shortened colon length, reduced histology score and mucosal barrier permeability in colitis mice (Fig. 7a–d and Supplementary Fig. 9c, d). Notably, Bergenin treatment enhanced the protein expression of Trib3 in the colonic mucosa, promoted the colonic epithelial cell proliferation as evidenced by EdU staining, and upregulated mRNA expression of both CBCs and RSCs markers in the colonic mucosa (Supplementary Fig. 9e–g). Besides, in the in vitro system, Bergenin administration enhanced the proliferation of colonic organoids isolated from the UC model (Fig. 7e). Whether Bergenin alleviates colitis progression through upregulating TRIB3 was then determined. Trib3 ablation abrogated the therapeutic effect of Bergenin in the UC model as evidenced by body weight change, DAI, colon length, histology score, serum FITC-dextran abundance and colonic epithelial cell proliferation (Fig. 7f–i and Supplementary Fig. 9h–j). These data indicated that Bergenin restores ISCs self-renewal to repair the damaged epithelial mucosal barrier and alleviates colitis by upregulating TRIB3. Then we tested whether a combination of Bergenin and Mesalazine (also known as 5-ASA)—a commonly used first-line drug for treating mild to moderate IBD has better therapeutic effects in 4% DSS-induced severe UC model (Fig. 7j). From the perspective of mouse survival, neither Mesalazine nor Bergenin alone showed therapeutic effects in the acute severe UC model. However, the combination of the two agents obviously reduced the mortality of UC mice (from 4/8 to 1/8) (Fig. 7k). Compared to single-agent therapy, Bergenin-Mesalazine combination alleviated body weight loss, reduced DAI and lowered histology scores in the acute severe UC mice (Fig. 7l, m and Supplementary Fig. 9k). While showing comparable anti-inflammatory efficacy to single-agent treatment, the combination therapy uniquely promoted IECs proliferation and enhanced mucosal barrier integrity (Fig. 7n and Supplementary Fig. 9l, m). These data suggest that pharmacological upregulating TRIB3 represents a potential strategy to relieve IBD by enhancing ISCs self-renewal and intestinal mucosal barrier integrity.

Fig. 7. Bergenin restores ISCs self-renewal to repair damaged epithelial mucosal barrier and alleviates colitis by upregulating TRIB3.

a Schematic diagram of 2% DSS-induced UC model treated with Vehicle, 25 mg/kg and 50 mg/kg Bergenin. b DAI of mice. c Representative gross colon images and statistical data of colon length. d Relative serum FITC-dextran abundance of mice. e Effects of Bergenin treatment on the colonic organoid proliferation (n = 10 independent biological experiments). Scale bar = 200 μm. f Schematic diagram of 2% DSS-induced UC model in Trib3^IEC-KO mice treated with or without 25 mg/kg Bergenin. g DAI of mice. h Relative serum FITC-dextran abundance of mice. i Representative EdU staining in the colon tissues and the statistical data. Scale bar = 20 μm. j Schematic diagram of 4% DSS-induced acute severe UC mouse model using C57BL/6 J mice treated with Bergenin, Mesalazine or the combination. k Mice death rate. l DAI of mice. m Representative H&E staining of colon tissues and statistical data of histology score. Scale bar = 1 mm (upper) or 50 μm (lower). n Relative mRNA expression of tight junction-related genes in the colon tissues (n = 4 independent biological experiments in Vehicle, Bergenin and Mesalazine, n = 7 in the combine group). Data were presented as the mean ± SEM and analyzed by unpaired two-tailed Student’s t test (e, h and i), Mann-Whitney U test (g), Kruskal-Wallis test (l) and Ordinary one-way ANOVA (b–d, m and n). n = 3 independent biological experiments for (i, m). n = 4 independent biological experiments for (f–h). n = 6 independent biological experiments for (a–d). n = 8 independent biological experiments for j–l. All statistical tests were two-sided. Error bars represent 95% confidence intervals. Elements of (a, f and j) are created in BioRender. shang, S. (2026) https://BioRender.com/tbfxh4l.

Discussion

The mainstay of medical therapy in IBD centers is the suppression of the immune system. Several limitations persist with current immunosuppressive approaches, including restricted efficacy in only a subset of patients, considerable side effects, and frequent failure to achieve clinical MH. The concept of delivering MH without immunosuppression but promoting ISCs self-renewal and regeneration is an attracting therapeutic strategy. Therefore, understanding the mechanisms driving ISCs loss or dysfunction is in urgent need. Based on clinical BA metabolomics analysis, we investigated the impact of GCA—a conjugated BA component elevated in the feces of IBD patients—on IBD progression. Our findings reveal that GCA exacerbates UC by disrupting the intestinal mucosal barrier through reducing the number of RSCs and inhibiting the proliferative capacity of CBCs. GCA impairs ISCs self-renewal via suppressing the intestinal TRIB3-ID1 axis. Mechanistically, TRIB3 interacts with ID1 to inhibit its palmitoylation and AP3D1-mediated lysosomal sorting and degradation in ISCs, which maintains ISCs stemness traits and promotes MH (Supplementary Fig. 10).

Mounting evidence indicates that BAs critically regulate ISCs function. For example, secondary BAs—particularly LCA and deoxycholic acid (DCA)—at physiological concentrations were demonstrated to enhance intestinal organoid budding and improve stem cell function via TGR5-mediated YAP activation⁵⁴. In IBD conditions, disrupted BAs homeostasis manifests as increased host-derived primary BAs and decreased microbiota-derived secondary BAs⁵⁵. Therefore, it is rational to speculate that the reduced LCA/DCA abundance under pathological conditions may contribute to impaired ISCs function. Furthermore, a study confirmed that the primary BA—CA triggers Lgr5⁺ ISC dysfunction by suppressing PPARα-mediated fatty acid β-oxidation¹⁷. Our study revealed GCA as another primary BA that disrupts both RSCs and CBCs, which supplements current understanding of the relationship between BAs and ISCs in IBD. Furthermore, our findings corroborate previous reports⁵⁶ that DSS-induced colitis models inadequately recapitulate the multifactorial pathogenesis of human IBD. By integrating GCA into the DSS-induced murine colitis model, we establish a more suitable model in mimicking IBD pathophysiology conditions.

The recently expanding universe of microbially conjugated bile acids—such as tryptophan-conjugated cholic acid (Trp-CA) and GABA-bile acid hybrids—opens exciting frontiers. These novel species, which integrate amino acids or neuroactive metabolites into the bile acid scaffold, represent a newly uncovered layer of host-microbiome crosstalk with demonstrated roles in metabolic and immune signaling. A critical unanswered question is whether these microbial bile acid conjugates also engage the TRIB3-ID1 axis to influence intestinal barrier integrity and inflammation resolution in IBD. These findings will offer a solid theoretical foundation for treating IBD through modulation of the gut microbiota. Moreover, the identification of amino acid–conjugated bile acids that enhance TRIB3-ID1 axis activity could enable their conjugation with anti‑inflammatory small molecules. This integrated approach could thus achieve both anti-inflammatory and pro-regenerative effects within a single agent, providing a —two birds, one stone— strategy against IBD.

Our study establishes that GCA damages ISCs stemness traits and intestinal mucosal barrier integrity via inhibiting TRIB3 transcription, positioning TRIB3 augmentation as an uncovered therapeutic strategy for achieving MH in IBD. We revealed that TRIB3 maintains the cell stemness traits through distinct mechanisms in CRC stem cells (interacting with β-catenin to activate Wnt signaling) and in healthy ISCs (interacting with ID1 to enhance its protein stability), demonstrating the functional flexibility of the same protein under different pathophysiological contexts. Besides, our findings implicate the existence of a series of stemness regulators—including TRIB3, ID1, YAP, STAT3, and STAT5—that exert dichotomous effects on disease progression through stemness maintenance: promoting CRC progression while reducing the severity of IBD^{11,12,24,30,57–60}. Critically, therapeutic upregulation of these proteins in IECs for IBD treatment must be carefully controlled in a moderate extent to reduce the potential risk of CRC pathogenesis.

Bergenin upregulates Trib3 to synergistically enhance the anti-inflammatory potency of 5-ASA in the severe UC model, which might provide a broader therapeutic utility of 5-ASA besides mild to moderate UC. Bergenin has been found to alleviate UC progression by inhibiting macrophage activation⁵³. However, our study revealed that the knockout of intestinal Trib3 almost abrogated the therapeutic effects of Bergenin against IBD. In the context of IBD, intricate regulatory crosstalk occurs between IECs and immune cells. Hyperactivated proinflammatory immune cells contribute to epithelial barrier disruption, while IECs reciprocally modulate phenotype or polarization of immune cells through secreted immunoregulatory factors such as ISG15^61,62. Hence, Bergenin may indirectly regulate the proinflammatory phenotype of macrophages through its restorative effects on ISCs regeneration and proliferation. Our findings also highlight the repurposing potential of the natural product Bergenin. Moreover, dedicated high-throughput screening platforms targeting TRIB3 could be established in the future to facilitate the rapid identification of mucosa-restorative lead compounds for next-generation IBD therapies.

In summary, our findings elucidate a previously unrecognized mechanistic link between BA metabolism dysregulation and IBD progression. In addition, we provide proof of principle that upregulating TRIB3 expression to enhance ID1 stability is a potential therapeutic option for restoring ISCs stemness traits to promote MH of IBD.

An important limitation of this study is that the clinical relevance of the GCA-TRIB3-ID1 axis remains to be established. Specifically, it is still unclear whether patients with dysregulated GCA metabolism would represent the primary beneficiary subgroup of potential TRIB3‑targeted therapies. This critical translational gap highlights the need for validation in well‑defined clinical cohorts. An additional limitation of the current study is the lack of systematic pharmacological profiling of Bergenin as a lead compound. Further evaluations of its pharmacokinetic properties, structure–activity relationship, and toxicology will be essential to assess its translational potential.

Methods

Ethics

All procedures were performed according to the Declaration of Helsinki. All mice were maintained in the animal facility at the Institute of Materia Medica under specific-pathogen-free (SPF) conditions. Mice were housed under SPF conditions with a 12 h light/12 h dark cycle (lights on at 7:00 a.m.), at an ambient temperature of 22 ± 2 °C and a relative humidity of 50 ± 10%. Food and water were available ad libitum. For animal studies, the mice were earmarked before grouping and randomly separated into groups by an independent person. We used 4～8 mice per experimental group in all animal experiments. All animal procedures were conducted in accordance with the guidelines of the Animal Care and Welfare Committee of the Institute of Materia Medica, Chinese Academy of Medical Sciences & Peking Union Medical College for the Ethics of Animal Care and Treatment. The animal study also accorded with the ARRIVE guidelines. All animal study protocols have been approved by the Animal Care & Welfare Committee of the Institute of Materia Medica, with approval numbers: 2024-0051, IMM-N-25-0094, IMM-N-25-0165, IMM-N-25-0218, IMM-N-25-0261, IMM-N-25-0428, IMM-N-25-0663, IMM-N-25-0665, 00003752, 00004636, 00004796, 00006180, 00006183, 00007264, 00007290, 00007522, 00007550.

Mouse models

C57BL/6 J mice (male, 6 ~ 8 weeks old) were purchased from the BEIJING HFK BIOSCIENCE CO., LTD. Lgr5-EGFR-IRES-creERT2 mice, Trib3^f/f and Vil1-Cre mice were purchased from Cyagen Biosciences Inc. C57BL/6 JGpt-Id1^{em12Cd7375in8}/Gpt mice were provided by GemPharmatech CO., LTD.

For overexpression of colon Trib3 in C57BL/6 J mice, 3 ~ 4 week-mice were tail-vein injected with 1 × 10¹² infectious units (IU) AAV virus encoding villin promoter-mediated Trib3 overexpression or scramble Ctrl as previously described^58,63. After 3 weeks, DSS were given to induce the UC mouse model.

For the induction of acute UC mouse models, mice were given 2% or 4% (w/v) DSS (MP Biomedicals) drinking water freely for 7 days, followed by a 3-day recovery period on regular drinking water or not. For the induction of chronic UC mouse models, mice were freely given 1.5% (w/v) DSS drinking water for 5 days, followed by a 5-day recovery period on regular drinking water, which were repeated for 3 cycles.

For GCA treatment, C57BL/6J mice were administered 100 or 200 mg/kg GCA (p.o., MCE) once a day, and were concurrently given 2% DSS in drinking water for 7 days or not.

For the therapeutic studies with Bergenin. C57BL/6 J mice were administered with 25 mg/kg Bergenin (p.o., MCE) once a day, and concurrently treated with 2% DSS and 200 mg/kg GCA for 7 days. For the combinational therapy with Mesalazine, C57BL/6 J mice were treated with 25 mg/kg Bergenin (p.o.), or 50 mg/kg Mesalazine (p.o.), or in the combination of the two agents once a day, and concurrently treated with 4% DSS for 7 days.

Mice were euthanized by cervical dislocation after anesthesia.

Human tissue samples

Tumor adjacent normal colon tissues were obtained from human colorectal cancer (CRC) tissue microarray (HColA160Su02), which were purchased from Shanghai Outdo Biotech Co., Ltd. The use of these samples in this study was approved by the Ethics Committee of Shanghai Outdo Biotech Company (approval No. SHYJS-CP-1704001). The requirement for individual informed consent was waived by the Ethics Committee. Patient age and sex information is publicly accessible via the manufacturer’s product webpage. Donors did not receive any financial compensation for providing their tissue samples. The tissue microarray slides were subsequently destroyed after the completion of the analyses.

Cell lines

Human cell lines NCM460 and HEK293T were purchased from the cell culture center of Peking Union Medical College, where cells were authenticated by STR profiling. NCM460 cells were cultured in RPMI-1640 medium (GIBCO) supplemented with 10% fetal bovine serum (FBS). HEK293T cells were cultured in DMEM medium (GIBCO) supplemented with 10% FBS. All cells were maintained at 37 °C in an incubator with a humidified atmosphere of 5% CO₂. Cells were regularly tested and verified to be mycoplasma negative using Myco-Blue Mycoplasma Detector (D101-02, Vazyme).

Crypt isolation and organoid culture

Crypts were isolated and cultured as reported before⁶⁴. Briefly, SI or colon were opened longitudinally and flushed with cold PBS for 3 times. Then, intestinal tissues were cut into small pieces (5 mm³) and collected in a total volume of 30 mL PBS with 10 mM EDTA solution (colon) or 2 mM EDTA solution (SI) for 1 hour (colon) or 30 min (SI) on ice. During incubation, place the conical tube on a vertical rotator to provide gentle agitation. The EDTA solution buffer were aspirated off after tissue fragments settling down for 30 sec, and the tissues were washed with PBS. Crypts were got in the supernatant after pipetting the colon or SI tissues up and down with 3 mL (colon) or 10 mL (SI) PBS by 20 times for 3 cycles, and the crypt fraction was filtered by a cell strainer (70 μm) to remove debris. After centrifugation of filtered crypts at 300 × g for 5 min, they were resuspended using Matrigel (Corning, 356231) at a density of ~ 400 crypts per well, and the Matrigel was allowed to polymerize in a 5% CO₂ incubator at 37 °C. Next, 500 μL organoid culture medium (Stem cell, 06005) will be added and replaced every two days. Images were captured on day 6 after organoids seeding.

Evaluation of intestinal permeability in organoids

Small intestinal or colon crypts were isolated from healthy C57BL/6 J mice, Trib3^f/f mice, Trib3^IEC-KO or C57BL/6 J mice infected with Trib3^OE virus, and cultured for organoids as described above. The organoids were maintained in complete culture medium for four days. To detect the effect of GCA on organoids permeability, GCA was added to the culture medium of the organoids for 24 hours at a concentration of 300 μM. To exacerbate the permeability defect induced by IL-1β, the organoids were treated with IL-1β for 24 h at a concentration of 50 ng/ml, and subsequently treated for 1 h with culture medium containing 1 mM EDTA. The organoids were subjected to three washes with PBS. The permeability assay was then conducted by adding 300 μg/ml of FITC-conjugated dextran (4 kDa) and incubating for 1 h. After incubation, FITC was removed by gently washing the samples three times with PBS to minimize background autofluorescence. Fluorescence images were acquired using an Olympus microscope (FV3000). For quantitative analysis, Image J software was utilized to measure the mean fluorescence intensity specifically within the luminal area of each organoid.

Proximity ligation assay

Cells on coverslips were fixed in 4% paraformaldehyde solution for 20 min, permeabilized with 0.5% Triton-X100 for another 20 min, and blocked with Duolink® blocking solution (Merck, DUO82007). The following primary antibodies were used: Anti-ID1 antibody (Abcam, ab168256, AB_2923370), Anti-TRIB3 antibody (Abcam, ab137526, AB_2876352), Anti-DDDDK-tag mAb (MBL International, M185-3L, AB_11123930), HA tag Polyclonal antibody (Proteintech, 51064-2-AP, AB_11042321), LAMP1 (Abcam, ab25630, AB_470708). Cells were then probed with the indicated primary antibodies and treated with the Duolink PLA probes using the Duolink® In Situ Red Starter Mouse/Rabbit kit (Merck, DUO92008) according to the manufacturer’s instructions. Images were captured using a confocal fluorescent microscope (Olympus, FV3000).

Click-iT assay

ID1-HA, together with or without TRIB3-DDK plasmids were transfected into HEK293T cells for 24 h. 100 μM Click-iT palmitic acid-azide was added into the cells and incubated for 6 hours at 37 °C. The cells were washed with PBS before the addition of lysis buffer containing protease and phosphatase inhibitors. The cell lysates were incubated for 30 min on ice, sonicated with a probe sonicator, vortexed for 1 min, and centrifuged at 15,294 x g at 4 °C for 5 min. Then, we transferred the supernatants to a tube and determined the protein concentration using the BCA protein assay kit. The protein samples were reacted with biotin alkyne using the Click-iT protein reaction buffer kit. The biotin alkyne-azide-palmitic acid–protein complexes were pulled down by streptavidin. The pellets were subjected to immunoblotting for HA detection.

DAI analysis

DAI is calculated according to the body weight change, stool consistency and intestinal bleeding, ranging from 0 to 12⁶⁵. Weight loss of 0, 1 ~ 5%, 6 ~ 10%, 11 ~ 18%, and > 18% were scored as 0, 1, 2, 3, and 4, respectively. For stool consistency, normal stool, soft but still formed stool, soft stool, very soft or wet stool and watery diarrhea were scored as 0, 1, 2, 3, and 4, respectively. For the intestinal bleeding, negative hemoccult was scored as 0 ~ 1, positive hemoccult was scored as 2, visible blood traces in stool was scored as 3, and gross rectal bleeding was scored as 4.

Histological analysis

Histology score was calculated according to the percentage of area involved, crypt loss, erosions, number of follicle aggregates, oedema, infiltration of mononuclear and polymorphonuclear cells, ranging from 0 to 20⁶⁶. For the percentage of area involved and crypt loss, normal tissues were scored as 0, < 10% were scored as 1, 10% were scored as 2, 10 ~ 15% were scored as 3, and 15 ~ 50% were scored as 4. For the erosions, intact epitheliums were scored as 0, involvement of the intestinal lamina propria were scored as 1, involvement of the submucosa were scored as 2, and transmural ulceration were scored as 3. For the number of follicle aggregates, oedema, infiltration of mononuclear and polymorphonuclear cells, absent of the conditions were scored as 0, weak were scored as 1, moderate were scored as 2, and severe were scored as 3.

FITC–dextran intestinal permeability assay⁶⁷

Mice were treated with 100 μL FITC-dextran (average of 4 kDa, 600 mg/kg, Merck, FD4-250MG) intragastrically after food deprivation for 1 h. Hemolysis-free serum was collected after another 3 h of starvation. Intestinal FITC-dextran concentration was calculated according to the fluorescence intensity (excitation, 488 nm; emission, 520 nm; Multi-Mode Microplate Reader).

EdU assay

To measure the proliferation of intestinal epithelial cells, mice were intraperitoneally injected with EdU (Beyotime, ST067-250 mg) at 100 mg/kg 2 h prior to killing. Intestines were collected and made into pathological paraffin sections, and detected according to the EdU staining kit (Beyotime, C0071S). To measure the proliferation of colon or SI organoids, EdU (10μM) was added to the organoids cultured in an 8-well plate with polymer coverslip bottom (ibidi, 80806) and incubated under 37 °C for 2 h. EdU-positive cells were detected according to the EdU staining kit.

Isolation of crypt and villus

Mice subjected to different treatments were killed and cleansed with 75% ethanol for surface sterilization. Intestines were isolated, and then flushed with cold D-Hanks buffer (containing DTT and antibiotics) using a 20 mL syringe for 3 times. The intestines were cut into small fragments (1～2 cm), everted, and washed with D-Hanks buffer to remove luminal contents. 50 mL EDTA buffer (10 mM) were then added into the intestinal fragments, which were rotated at 150 rpm for 15 min, and shaked at 60 shakes/min for 2 min on ice. Following this step, the supernatant was discarded, and the intestines were added with fresh EDTA buffer (10 mM). After 3 cycles, the supernatants were centrifuged at 150 g for 3 min to get the villus. 50 mL EDTA buffer (10 mM) was then added to the remnant intestines, and repeated the shaking procedure described above for another 3 cycles to clearly remove the remained villus. 10 mL EDTA buffer (10 mM) was then added to the intestines and rotated softly, supernatants were collected every 2 min after addition of fresh EDTA buffer for 4 times. The combined supernatants were left to stand for 2 min to exclude the precipitant villus, the crypts were got after centrifuging for 5 min at 300 × g.

Isolation of intestinal epithelial cells and flow cytometry

Lamina propria cells were harvested by scraping with a microscope slide and incubated in digestion buffer (IMDM supplemented with 1% FCS, 10 mM HEPES, penicillin/streptomycin, 50 mg/ml DNase I, 0.4 mg/ml Liberase TL) for 30 min at 37 °C with 200 rpm shaking. Single cell suspensions were obtained by filtering through a 70 μm strainer. After red blood cell lysis, cells were pelleted at 425 x g for 5 min. The cells were then blocked with CD13/CD26 (Biolegend, 101302, AB_312801) for 15 min at room temperature, followed by staining with APC anti-mouse CD45 antibody (Biolegend, 103112, AB_312977) and PerCP-Cy5.5 anti-mouse Epcam antibody (Biolegend, 118219, AB_2098647) for 15 min at room temperature to make further flow cytometry analysis. Gating strategy was summarized in Supplementary Fig. 16a.

Immunohistochemistry staining

Mice subjected to different treatments were killed, and the SI or colons were fixed in 4% paraformaldehyde. The tissues were sliced to 10 μm thickness and deparaffinized with xylene, rehydrated through graded ethanol, followed by an antigen retrieval via microwave heating in 10 mM citrate buffer (pH 6.0). Endogenous peroxidase activity was then quenched by 3% hydrogen peroxide. Tissue sections were permeabilized with 0.5% Triton X-100 in PBS, blocked with 3% BSA in PBS and incubated with rabbit polyclonal antibody against TRIB3 (ThermoFisher Scientific, PA5-114501) at 4 °C overnight. Then the sections were immunostained by the DAB chromogenic kit (ZSGB-BIO, ZLI-9018). Subsequently, sections were counterstained with hematoxylin and mounted in non-aqueous mounting medium.

mIHC staining

For the immunofluorescence staining, the formalin-fixed, paraffin-embedded (FFPE) sections were stained using a four-color labeled kit (TSA-RM) (PANOVUE Biotechnology Co., LTD, 10080100100) according to the manufacturer’s protocol⁶⁸. Briefly, the slides were deparaffinized, rehydrated, and antigen retrieved using Trilogy buffer by autoclaving for 15 min. The slides were blocked using PBS plus 3% bovine serum albumin (BSA) and 0.1% Triton X-100 for 30 min at room temperature and incubated with the relevant primary antibody (1:100) at 4 °C overnight. The following primary antibodies were used: Trib3 (ThermoFisher Scientific, PA5-114501, AB_2890514), Bmi1 (Abcam, ab269678), Cdh1 (Cell Signaling Technology, 3195T), ID1 Polyclonal antibody (Proteintech, 18475-1-AP, AB_2248812), Fluorescent mounting medium with DAPI (ZSGB-BIO, ZLI-9557). After a 1 h incubation with the relevant secondary antibody (1:100) at room temperature, the slides were incubated in the amplification diluent containing a tyramide-conjugated fluorophore for 10 minutes. Prior to the next primary antibody incubation, the slides were heated in 10 mM citric acid (pH 6.0) for 10 min at 95 °C to strip the antibodies of the previous staining round. The protocol was repeated from the blocking step until a total of two markers were co-stained. After the last staining round, the slides were washed, incubated with 0.5 μg/ml DAPI for 5 min, washed again, and mounted using Prolong Diamond medium. Slides were scanned using Olympus (FV3000).

Isolation of BMDM

BMDM were isolated as reported before^69,70. Bone-marrow-derived cells were aseptically collected from 6 to 8-week-old mice by flushing the leg bones of euthanized mice with DMEM. Cell suspensions were filtered through a 70-μm cell strainer to exclude cell aggregates. Red blood cells were lysed by re-suspending the pellet in 2 mL red blood cell lysis buffer and incubated at room temperature for 2 min. Approximately 1 × 10⁷ bone-marrow-derived cells were purified by gradient centrifugation from the femurs and tibias of a single mouse. BMDMs were cultured in DMEM with 10% FBS and treated with supernatant of L929 cells every other day.

CFSE assay

For the CFSE assay of bone marrow derived myeloid cells, cells isolated from bone marrows were spined down and resuspend at 1 × 10⁷ cells/mL in 5 μM CFSE (Biolegend, 423801) working solution. Incubate cells for 20 min at 37 °C and keep protected from light. Then quench the staining by adding 5 times the original staining volume of cell culture medium containing 10% FBS. Finally, cells were cultured as described above for 48 h and were stained with CD11b (Biolegend, 101228, AB_893232) to make further flow cytometry analysis.

For the CFSE assay of CD4⁺ T cells, mice spleen was homogenized, and single cells were collected by centrifugation. Red blood cells were lysed by re-suspending the pellet in 2 mL red blood cell lysis buffer and incubated at room temperature for 2 minutes. CD4⁺ T cells were sorted by a CD4⁺ T cell isolation kit and labeled with CFSE as described above. The cells were then resuspended at 1 × 10⁵ cells/mL in RPMI-1640 culture medium containing 10% FBS, 10 ng/mL mouse recombinant IL-2 (PeproTech, 212-12-5UG), 10 μM 2-mercaptoethanol, 4 mM glutathione and 2 μg/mL CD28 antibody (BioGems, 10312-25-500). Then the cells were seeded onto culture dishes precoated with mouse CD3 antibody (BioGems, 05112-25-500). The cells were cultured for 5 days and stained with CD4-APC (Biolegend Cat# 100412, AB_312697) antibody for further flow cytometry analysis.

Flow cytometry analysis of macrophage polarization

Bone-marrow-derived macrophages were polarized with recombinant murine IFN-γ (PeproTech, 315-05-20UG) at a concentration of 20 ng/mL, together with LPS (100 ng/mL, Merck, 916374) for 24 hours to induce pro-inflammatory (M1-like) phenotype. BMDMs were treated with IL-4 (20 ng/mL, PeproTech, 214-14) for 48 hours to induce an anti-inflammatory (M2-like) phenotype. M1-like BMDMs were stained with CD45-FITC (Biolegend, 103108, AB_312973), CD11b-PerCp/Cyanine5.5 (Biolegend, 101228, AB_893232), F4/80-PE (Biolegend, 123110, AB_893486) and CD86-APC antibodies (BD Biosciences, 561964, AB_10898000), and M2-like BMDMs were stained with CD45-FITC, CD11b-PerCp/Cyanine5.5, F4/80-PE and CD206-APC (Biolegend, 141708, AB_10900231) antibodies to make further flow cytometry analysis. Gating strategy was summarized in Supplementary Fig. 16b.

Flow cytometry analysis of Th₁ cells polarization

Mice spleen was homogenized, and single cells were collected by centrifugation. Red blood cells were lysed by re-suspending the pellet in 2 mL red blood cell lysis buffer and incubated at room temperature for 2 min. CD4⁺ T cells were sorted by a CD4⁺ T cell isolation kit (Stem cell, 19765) and then resuspended at 5 × 10⁵ cells/mL in RPMI-1640 culture medium containing 10% FBS, 10 ng/mL mouse recombinant IL-2, 10 μM 2-mercaptoethanol, 4 mM glutathione and 2 μg/mL CD28 antibody. Then the cells were seeded onto culture dishes precoated with mouse CD3 antibody for 3 days. Then the cells were induced with 10 ng/mL IL-12 and 10 μg/mL IL-4 antibody (BioGems, 81112-25-500) for another 3 days, and stained with CD45-APC (Biolegend, 103112, AB_312977), CD4-FITC and IFN-γ-PE-Cy7 (Biolegend, 505826, AB_2295770) for further analysis. The gating strategy was summarized in Supplementary Fig. 16c.

Luciferase assay

NCM460 cells were transfected with different siRNAs described above. After 24 h, the plasmid of p-TRIB3 luciferase reporter were transfected. 12 h later, 300 μM GCA was added and cultured for 24 h. Cell supernatants were collected and analyzed for luciferase activity (Genecopoeia, catalog LF033) according to the manufacturer’s protocol.

RNA extraction and real-time quantitative PCR

Total RNA from cells or intestinal mucosa was extracted using RNA-Quick purification kit (Shang Hai Yishan Biotechnology Co., LTD, ES-RN001) guided by the manufacturer’s instructions. RNA was quantified using a NanoDrop spectrophotometer. Reverse transcription of the total cellular RNA was carried out using oligo (dT) primers and MMLV reverse transcriptase. Polymerase chain reaction was performed using the Trans-Scrip@ One-Step RT-PCR SuperMix (AT411-02, Transgene, AT411-02) according to the manufacturer’s instructions. Quantitative polymerase chain reaction (qPCR) primers were all synthesized by Ruibotech company, and the sequences are as follows: mouse Lgr5 primers, Forward: CCTACTCGAAGACTTACCCAGT, Reverse: GCATTGGGGTGAATGATAGCA; mouse Bmi1 primers, Forward: ACTACACGCTAATGGACATTGCC, Reverse: CTCTCCAGCATTCGTCAGTCCA; mouse Olfm4 primers, Forward: AAAGTGACCTTGTGCCTGCC, Reverse: AGGGTTCTCTCTGGATGCTGA; mouse Dclk1 primers, Forward: ATGTGGACCAGAGAAGTTCCG, Reverse: CCGCCATGCTGAGAGATCC; mouse Tjp1 primers, Forward: GCCGCTAAGAGCACAGCAA, Reverse: GCCCTCCTTTTAACACATCAGA; mouse Cdh1 primers, Forward: CAGGTCTCCTCATGGCTTTGC, Reverse: CTTCCGAAAAGAAGGCTGTCC; mouse TNF-α primers, Forward: CTGGGACAGTGACCTGGACT, Reverse: GCACCTCAGGGAAGAGTCTG; mouse IL-6 primers, Forward: TTGGGACTGATGCTGGTGAC, Reverse: TTGCCATTGCACAACTCTTTTC; mouse IL-1β primers, Forward: GATCCACACTCTCCAGCTGCA, Reverse: CAACCAACAAGTGATATTCTCCATG; mouse Trib3 primers, Forward: CTGCGTCGCTTTGTCTTCAGCA, Reverse: CTGAGTATCTCTGGTCCCACGT; Human TRIB3 primers, Forward: CACGAGACTCGCAGCGGA, Reverse: TTGTCATCCAACTCCAACCGC; mouse Id1primers, Forward: TTGGTCTGTCGGAGCAAAGCGT, Reverse: CGTGAGTAGCAGCCGTTCATGT; Human ID1primers, Forward: GTTGGAGCTGAACTCGGAATCC, Reverse: ACACAAGATGCGATCGTCCGCA; mouse Gapdh primers, Forward: CATCACTGCCACCCAGAAGACTG, Reverse: ATGCCAGTGAGCTTCCCGTTCAG; Human GAPDH primers, Forward: GTCTCCTCTGACTTCAACAGCG, Reverse: ACCACCCTGTTGCTGTAGCCAA.

Immunoprecipitation and immunoblotting

A co-immunoprecipitation experiment was performed as described previously. In brief, cells were collected and lysed on ice for 30 minutes. The cell lysates were incubated with the indicated antibody-conjugated magnetic microbeads at 4 °C overnight. The immunocomplex was washed 5 times and boiled in 2 × SDS sample buffer for 10 min. The co-precipitates were resolved using SDS-PAGE and immunoblotted with specific antibodies. Western blot images were captured by a Tanon 5200 chemiluminescent imaging system with super-sensitive ECL luminescence reagent. The following primary antibodies were used: TRIB3 Polyclonal Antibody (ThermoFisher Scientific, PA5-15480, AB_2209090), TRIB3 antibody [EPR3151Y] (Abcam, ab75846, AB_1310768), ID1 Polyclonal antibody (Proteintech, 18475-1-AP, AB_2248812), PCNA (D3H8P) XP® Rabbit mAb (Cell Signaling Technology, 13110, AB_2636979), Mouse Anti-GAPDH mAb (ZSGB-Bio, TA-08, AB_2747414), Rabbit monoclonal [EPR3065Y] to LGR5 (Abcam, ab75850, AB_1523716), Bmi1 (Abcam, ab269678), Anti-GFP (Green Fluorescent Protein) pAb (MBL International, 598, AB_591819), Anti-DDDDK-tag mAb (MBL International, M185-3L, AB_11123930), HA tag Polyclonal antibody (Proteintech, 51064-2-AP, AB_11042321), Mouse Anti-β actin mAb (ZSGB-BioTA-09, AB_2636897), MYC tag Polyclonal antibody (Proteintech, 16286-1-AP, AB_11182162), anti-rabbit IgG HRP-linked Antibody (ZSGB-Bio, ZB-2301, AB_2747412), anti-mouse IgG HRP-linked Antibody (ZSGB-Bio, ZB-2305, AB_2747415). Uncropped blots have been provided in the Source Data file and in Supplementary Fig. 11–15.

GSEA analysis

Patient-related gene set enrichment analysis for IBD was carried out using publicly available gene expression data of 27 human IBD colon samples (accession number GSE105074). Genes were divided into 2 groups according to the mRNA levels of TRIB3. Pre-ranked gene set enrichment analysis was performed using gene sets of Wnt, Notch and Hippo-YAP signaling. Besides, we also ranked the sequenced genes by their association with Ctrl (n = 3) and TRIB3^OE (n = 3) groups using the signal-to-noise measure in the GSEA. A gene set was considered significantly enriched when the FDR was less than 0.25.

RNA-sequencing assay

NCM460 cells were infected with Ctrl or TRIB3-3×Flag overexpressing lentivirus (HanBio Co., Ltd), and screened by puromycin, which were sent for RNA-sequencing. Total RNA was extracted using the Trizol reagent according to the manufacturer’s protocol. RNA purity and quantification were evaluated using the NanoDrop 2000 spectrophotometer (Thermo-Fisher Scientific). RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies). Then the libraries were constructed using the VAHTS Universal V6 RNA-seq Library Prep Kit according to the manufacturer’s instructions. The transcriptome sequencing and analysis were conducted by OE Biotech Co., Ltd.

Mass spectrometry

For protein identification by mass spectrometry, cell lysates of HEK293T cells transfected with ID1-HA or together with TRIB3-DDK were collected and immunoprecipitated with anti-HA antibody. The HA peptide eluents were separated with the EASY-nLC 1000 system, which was directly interfaced with the mass spectrometer. Data of MS/MS were searched against the human fasta from UniProt using an in-house Proteome Discoverer.

Metabolomics detection

Faeces of each mouse were collected for metabolomics detection. The hindgut content samples (20 mg) were first ground using a ball mill. Subsequently, 10 μL of an internal standard mixture working solution (1 μg/mL) and 200 μL of methanol/acetonitrile (v/v = 8:2) were added for homogenization. After homogenization, the samples were shaken at 664 x g for 10 min, and then the samples were kept at − 20 °C for 10 min to precipitate protein, followed by centrifugation for 10 min at 15,294 x g and 4 °C. The resulting supernatant was transferred to clean plastic microtubes and concentrated using a concentrator (CentriVap, LABCONCO, USA). Upon completion of the concentration step, the samples were reconstituted with 100 μL of 50% methanol–water solution for further LC-MS/MS.

The sample extracts were analyzed using an LC-ESI-MS/MS system (UHPLC, ExionLCTM AD, https://sciex.com.cn/; MS. Applied Biosystems 6500 Triple Quadrupole, https://sciex.com.cn/). The analytical conditions were as follows, HPLC: column, Waters ACQUITY UPLC HSS T3 C18 (100 mm × 2.1 mm i.d., 1.8 μm); solvent system, water with 0.01% acetic acid and 5 mmol/L ammonium acetate (A), acetonitrile with 0.01% acetic acid (B); The gradient was optimized at 5% to 40% B in 1 min, then increased to 50% B in 6 min, then increased to 75% B in 5 min, and then 75–95% in 2 min, washed with 95% B for 2 min, finally ramped back to 5% B (16 ~ 17.5 min); flow rate, 0.35 mL/min; temperature, 40 °C; injection volume: 3 μL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

Linear ion trap (LIT) and triple quadrupole (QQQ) scans were acquired on a triple quadrupole-linear ion trap mass spectrometer (QTRAP), QTRAP® 6500 + LC-MS/MS System, equipped with an ESI Turbo Ion-Spray interface, operating in negative ion mode and controlled by Analyst 1.6.3 software (Sciex). The ESI source operation parameters were as follows: ion source, ESI-; source temperature 550 °C; ion spray voltage (IS)—4500 V; curtain gas (CUR) was set at 35 psi, respectively. Bile acids were analyzed using scheduled multiple reaction monitoring (MRM). Data acquisitions were performed using Analyst 1.6.3 software (Sciex). Multiquant 3.0.3 software (Sciex) was used to quantify all metabolites. Mass spectrometer parameters, including the declustering potentials (DP) and collision energies (CE) for individual MRM transitions, were done with further DP and CE optimization. A specific set of MRM transitions were monitored for each period according to the metabolites eluted within this period.

Plasmid construction

Full-length ID1 and its truncated mutants, M1 (amino acids 1 ~ 52), M2 (amino acids 53 ~ 104), and M3 acids 105 ~ 155) were constructed into the pEGFP-C1 vector (CLONTECH Laboratories, Inc.). Full-length of TRIB3 and its truncated mutants, M1 (amino acids 1 ~ 179), M2 (amino acids 180 ~ 359) and M3 (amino acids 72 ~ 315) were constructed into the pEGFP-C1 vector. ID1-HA and TRIB3-HA was constructed into the pcDNATM3.1 vector.

The RNAi sequences

si-ATF4-1: 5’-CCTGAAAGATTTGATAGAA-3’. si-ATF4-2: 5’-CAGATTGGATGTTGGAGAA-3’. si-SMAD3-1: 5’-TGGTGCGAGAAGGCGGTCA-3’. si-SMAD3-2: 5’-GCAACCTGAAGATCTTCAA-3’. si-FOXO1-1: 5’-CCAGATGCCTATACAAACA-3’. si-FOXO1-2: 5’-CTCAAATGCTAGTACTATT-3’. si-CHOP-1: 5’-GGCTCAAGCAGGAAATCGA-3’. si-CHOP-2: 5’-GAACCAGGAAACGG AAACA-3’. si-β-catenin-1: 5’-GCCACAAGATTACAAGAAA-3’. si-β-catenin-2: 5’-GACTACCAGTTGTGGTTAA-3’.

Statistical & reproducibility

Analyses were performed using GraphPad Prism 10.0 software with significance set to P  <  0.05. For comparisons between two groups: If data passed both normality and variance homogeneity tests, an unpaired or paired Student’s t test was used, as appropriate. For data with unequal variances (normally distributed), Welch’s t test was applied. For non-normally distributed data, the Mann-Whitney U test (unpaired) or Wilcoxon matched-pairs signed rank test (paired) was used. For comparisons among three or more groups: If data were normally distributed with equal variances, a one-way ANOVA was performed, followed by Tukey’s post hoc test for multiple comparisons if the overall ANOVA was significant. For normally distributed data with unequal variances, Brown-Forsythe and Welch ANOVA with Games-Howell post hoc test was used. For non-normally distributed data, the Kruskal-Wallis test was employed, followed by Dunn’s post hoc test with Bonferroni correction for multiple comparisons. Two-sided tests were used in the statistics. A Shapiro-Wilk test was used to test the data normality. The F test or the Levene test was used to test the variance homogeneity. All quantitative data are expressed as means ± SEM. One outlier value (8.04742) in the UC group of Fig. 2b was excluded from the data according to Tukey’s rule. The experiments were randomized; The investigators were not blinded to allocation during experiments and outcome assessment.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

F.H. and S.S. conceived the project and designed all experiments; S.S. and J.L. performed most of the experiments and analyzed data; S.-Y. D., X.-X.L., Y.-X.L, R.L, Y.-X.W., R.-K.X., Q.-Y.S. and C.-J.S. assisted in the biological experiments. F.H. and S.S. wrote the manuscript, and the other authors revised it.

Peer review

Peer review information

Nature Communications thanks Ipsita Mohanty, Shiro Yui and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by grants from National Natural Science Foundation of China (82473990), Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM609), the National Key Research and Development Program of China (2022YFC2504002, 2022YFC2504003), CAMS Innovation Fund for Medical Sciences (CIFMS) (2023-I2M-2-009, 2025-I2M-KJ-015) and 2024 China Industrial Technology Infrastructure Public Service Platform Project (GN2024-31-4700).

Data availability

The RNA sequencing data generated in this study have been deposited on the NCBI Gene Expression Omnibus under accession number GSE276541. The metabolome data generated in this study have been deposited on the Metabolights under MTBLS14031 and MTBLS14032. The mass spectrometry proteomics data generated in this study have been deposited on ProteomeXchange Consortium (iProX) under accession number PXD072472 (https://www.iprox.cn//page/project.html?id=IPX0014758000). No original code has been reported in this manuscript. Source data are provided with this paper. Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Fang Hua (huafang@imm.ac.cn). Source data are provided in this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-74713-6.