Multi-omics analysis identifies a microbiota-bile acid-TLR signaling axis driving bladder injury in interstitial cystitis

Abstract

Hunner-type interstitial cystitis/bladder pain syndrome (HIC) is a debilitating condition defined by bladder pain and urinary urgency, yet its upstream drivers remain poorly understood. To identify upstream mechanisms that exacerbate urothelial injury, here we apply an integrative multi-omics framework combining metagenomic sequencing, targeted metabolomics of urine and serum, and single-cell RNA sequencing. This approach reveals a microbial signature enriched in Enterococcus avium and a marked alteration in bile acid metabolism, including increased taurochenodeoxycholic acid (TCDCA). Single-cell analysis indicates that these changes converge on Toll-like receptor 3 (TLR3) activation in urothelial cells. Further validations show that a microbiota-bile acid-TLR3 axis disrupts epithelial barrier integrity and triggers inflammatory responses in experimental models. Transplantation and metabolite administration confirm the causal role of E. avium and TCDCA, while TLR3 inhibition ameliorates injury. These findings uncover an upstream pathway linking gut-derived metabolites to bladder pathology and suggest opportunities for biomarker development and targeted therapies for HIC.

Fig. 1. Microbial composition and functional alterations in patients with Hunner-type interstitial cystitis/bladder pain syndrome (HIC).

A β-diversity analysis based on 16S rRNA sequencing, assessed by principal coordinates analysis (PCoA) using Bray–Curtis distances (n = 30 control vs. 32 HIC; two-sided PERMANOVA with 1000 permutations; multiple testing not applicable). B Differential bacterial genera identified by LEfSe analysis (LDA score ≥ 3 and p < 0.05; n = 30 control vs. 32 HIC; multiple testing not applicable). C Comparison of α-diversity based on metagenomic sequencing (n = 30 vs. 30; data show median (IQR); two-sided Mann–Whitney U test; ns not significant; multiple testing not applicable). D Comparison of β-diversity based on metagenomic sequencing. E Differential species identified by LEfSe. F Functional enrichment analysis of differential species in the HIC group. G Correlation network based on differential species and metabolic pathways. Pairwise associations were calculated using two-sided Spearman correlation. Significant correlations were defined as p < 0.05 and are represented as edges in the network. No multiple-testing correction was applied, as these analyses were exploratory. H Diagnostic performance (AUC) of the top three differential species for HIC classification. Source data are provided as a Source Data file.

Fig. 2. Blood and urine metabolomic analysis reveals bile acids (BA)-related metabolic alterations.

Visualization of metabolite distribution in blood (A) and urine (B) using orthogonal partial least-squares discriminant analysis (OPLS-DA; exploratory analysis; multiple testing not applicable). Categorization of differential metabolites in blood (C) and urine (D). Exploration of differential metabolic pathways in blood (E) and urine (F; two-sided hypergeometric test with FDR correction (Benjamini–Hochberg method); pathways with adjusted p < 0.05 are selected). BA-related metabolites in blood (G) and urine (H) (n = 30 control vs. 50 HIC; metabolites with VIP ≥ 1 from OPLS-DA and p < 0.05 from two-sided Mann–Whitney U test; exploratory analysis; p-values not adjusted). Source data are provided as a Source Data file.

Fig. 3. Single-cell RNA sequencing identifies TLR signaling as a key mediator of urothelial injury in HIC.

A Differentially expressed genes (DEGs) in full-thickness bladder tissues from HIC patients and controls identified via RNA sequencing (|log₂FC|≥2; adjusted p < 0.05, two-sided Wald test using DESeq2). Gene Ontology (GO) (B) and KEGG (C) enrichment analyses of DEGs (two-sided hypergeometric test with FDR correction). D UMAP visualization of single-cell transcriptomic landscape of urothelium, with urothelial cells (UCs) identified by KRT19 and UPK1A expression. GO (E) and KEGG (F) analyses of upregulated DEGs in UCs (two-sided hypergeometric test with FDR correction). GO (G) and KEGG (H) analyses of downregulated DEGs in UCs (two-sided hypergeometric test with FDR correction). I, J Gene Set Enrichment Analysis (GSEA) of altered pathways in UCs (two-sided permutation test with 1000 permutations, with multiple testing corrected using the FDR). The exact p value for TLR signaling of UCs in panel I is 0.0009. K Expression of Toll-like receptor (TLR) subtypes in UCs from HIC bladders. mRNA (L) and protein (M) expression levels of TLR3 in isolated urothelium from control and HIC patients (n = 7 control vs. 10 HIC; data show median (IQR); two-sided Mann–Whitney U test; ns not significant; Bar for panel M: 1 cm). N Immunostaining showing the distribution of TLR3 protein in the urothelium of patients with HIC (n = 7 control vs. 10 HIC; one section and field per patient; Bar: 100 μm). IHC immunohistochemistry, HL Hunner lesions, U urothelium, LP lamina propria (blue line), M muscularis. Source data are provided as a Source Data file.

Fig. 4. Crosstalk among gut microbiota alterations, bile acid metabolism, and bladder injury-related pathways in HIC.

Spearman correlation between differential microbial species and altered blood (A) and urinary (B) metabolites (No. matched participants, n = 60). Correlation between differential urinary metabolites and gene expression in the Toll-like receptor (C) and cell junction (D) signaling pathways based on single-cell RNA sequencing (No. matched participants, n = 17). Correlation between differential microbial species and genes involved in the Toll-like receptor (E) and cell junction (F) signaling pathways (No. matched participants, n = 17). G Spearman correlation of key microbial species and urinary metabolites with clinical symptom scores, including interstitial cystitis symptom index (ICSI), problem index (ICPI), and visual analogue scale (VAS) (No. matched participants, n = 60). Statistical analysis: two-sided Spearman correlation was used for all panels (A–G). *p < 0.05; **p < 0.01; ***p < 0.001. Source data are provided as a Source Data file.

Fig. 5. HIC-derived fecal microbiota alters host metabolism and increases bladder vulnerability.

A Fecal microbiota transplantation (FMT) in antibiotic-pretreated mice (Abx-mice), using stool samples from healthy controls (Group A, No. of controls = 5) and HIC patients (Group B, No. of patients = 5). B Bladder weight comparison post-FMT. C Assessment of voiding behavior in FMT-recipient mice (each mouse measured once). D Evaluation of mechanical pain threshold following FMT. Comparisons: (1) Control vs. Autoimmune cystitis (AC); (2) Control vs. Control + A; (3) Control vs. Control + B; (4) AC vs. AC + A; (5) AC vs. AC + B. E Representative histological analysis of bladder tissues after FMT (one section and field per mouse). F Quantification of urinary TCDCA and TUDCA in recipient mice. For panels (B–F): group sizes are n = 7 for Control + A and n = 8 for all other groups; data are presented as median (IQR); statistical analysis was performed using the Kruskal–Wallis test; ns not significant. IHC immunohistochemistry, Red arrow: urothelial thinning, detachment, and exposure, Blue line: LP lamina propria, U urothelium, M muscularis. Source data are provided as a Source Data file.

Fig. 6. E. avium transplantation induces urothelial injury via TCDCA upregulation.

A Experimental workflow of E. avium colonization in antibiotic-pretreated mice (Abx-mice, n = 6 per group). Created in BioRender. Chen, J. (2025) https://BioRender.com/3lccwo1. B Increased abundance of E. avium confirmed by metagenomic sequencing (n = 6 per group). C Evaluation of voiding behavior following E. avium transplantation (n = 6 per group; one measurement per mouse). D Measurement of mechanical pain threshold post-transplantation (n = 6 per group). E Representative histological analysis of bladder tissues after E. avium transplantation (n = 6 per group; one section and field per mouse). F, G Quantification of blood and urinary bile acids using comprehensive targeted bile acid profiling (n = 6 per group). H Cell viability of human urothelial cells (HUCs) following TCDCA or TUDCA treatment (n = 3 independent experiments). I Expression of ZO-1, TNF-α, and TLR3 following TCDCA (200 µM) exposure (n = 3 independent experiments). J Experimental workflow of intravesical instillation of TCDCA in rats (n = 5 per group). Created in BioRender. Chen, J. (2025) https://BioRender.com/3lccwo1. K Evaluation of voiding function after TCDCA instillation (n = 5 per group; each rate measured once). L Assessment of mechanical pain threshold following TCDCA exposure (n = 5 per group). M Representative histology of bladder tissues post-instillation (n = 5 per group; one section per rat, one field quantified per section). For panels (B–H, K, L): data are presented as median (IQR). Statistical analysis was performed using the two-sided Mann–Whitney U test. IHC immunohistochemistry, Red arrow: urothelial thinning, detachment, and exposure, Blue line: LP lamina propria, U urothelium, M muscularis. Source data are provided as a Source Data file.

Fig. 7. TCDCA-induced urothelial injury via TLR3 signaling.

A, B Expression of tight junction protein ZO-1 and inflammatory marker TNF-α after TLR3 intervention in TCDCA-pretreated human urothelial cells (HUCs) (n = 3 independent experiments). C Transepithelial resistance (TER) changes following TLR3 intervention in TCDCA-pretreated HUCs (n = 3 independent experiments). D, E Expression of ZO-1 and TNF-α after pentosan polysulfate sodium (PPS) administration in TCDCA-pretreated HUCs (n = 3 independent experiments). F TER changes following PPS intervention in TCDCA-pretreated HUCs (n = 3 independent experiments). G Experimental workflow showing TLR3 inhibitor and PPS administration to rats following intravesical TCDCA instillation (n = 5 per group). Created in BioRender. Chen, J. (2025) https://BioRender.com/3lccwo1. H Assessment of voiding function (one measurement per rat) and pain threshold after TLR3 inhibition (n = 5 per group). I Representative histological images of bladder tissues post-TLR3 inhibition (n = 5 per group; one section and field per rat). For panels (B, C, E, F, H): data are presented as median (IQR). Statistical analysis was performed using the Kruskal–Wallis test. ns not significant, IHC immunohistochemistry, Red arrow: urothelial thinning, detachment, and exposure, Blue line: LP lamina propria, U urothelium, M muscularis. Source data are provided as a Source Data file.