The gut microbe *Bacteroides fragilis* ameliorates renal fibrosis in mice

Chronic kidney disease (CKD) affects approximately 10% of the population and typically progresses to renal fibrosis, characterized by myofibroblast proliferation, α-SMA expression, and excessive extracellular matrix deposition, ultimately reducing glomerular filtration and causing renal injury. Current management (antihypertensives, BP and glucose control, dialysis, transplantation) lacks effective anti-fibrotic therapies. Growing evidence implicates gut–kidney crosstalk: CKD alters gut microbiota, increasing uremic toxin-producing taxa and decreasing beneficial bacteria; dysbiosis and barrier dysfunction can exacerbate CKD progression. This study investigates whether the gut commensal Bacteroides fragilis, known for anti-inflammatory effects in other organs, confers protection against renal fibrosis and explores underlying mechanisms, focusing on LPS reduction, metabolite 1,5-anhydroglucitol (1,5-AG), and renal pathways involving SGLT2 and TGR5.

CKD is associated with gut dysbiosis that shifts microbial composition toward increased Enterobacteriaceae, Clostridiaceae, Pseudomonadaceae, and Bacteroidiaceae and reduced Lactobacillaceae, Bifidobacteriaceae, and Prevotellaceae, promoting uremic toxins (indoxyl sulfate, p-cresyl sulfate, TMAO) and chronic inflammation through barrier disruption and ammonium production. Protective microbiota-derived metabolites (e.g., indolepropionic acid, SCFAs, neurotransmitters) are reduced in CKD, accelerating progression. B. fragilis, an obligate anaerobe prevalent in the lower gut, produces polysaccharide A and SCFAs that induce IL-10+ Tregs and dampen inflammation in models of intra-abdominal abscess, IBD, autism spectrum disorder, and asthma. However, its role in CKD/renal fibrosis had not been explored prior to this work.

Human studies: Two fecal cohorts assessed B. fragilis abundance by qPCR: discovery (10 CKD, 10 matched controls; Renmin Hospital of Wuhan University) and validation (15 CKD, 15 controls; Putuo People’s Hospital). Correlations with blood urea nitrogen (BUN) and serum creatinine (Scr) were analyzed. Serum metabolomics included untargeted GC-MS (115 CKD, 113 controls; Affiliated Hospital of Nanjing University of Chinese Medicine) and targeted validations: GC-MS (110 CKD, 110 controls; Ningbo Hospital of Zhejiang University) and LC-MS (100 CKD, 100 controls; Putuo People’s Hospital) focusing on 1,5-AG. Immunohistochemistry assessed TGR5 in IgA nephropathy tissues. GEO datasets were mined for SLC5A2 (SGLT2) mRNA in kidney diseases. Animal models: Male ICR mice underwent UUO surgery or adenine-induced CKD. Interventions included oral live or heat-killed B. fragilis (NCTC 9343; 2×10^6 CFU/mL gavage), intraperitoneal 1,5-AG (100 mg/kg), madecassoside (Mad; 80 mg/kg p.o. or i.p.), madecassic acid (MA; 40–80 mg/kg p.o.), and an oral antibiotics cocktail to deplete microbiota. An empagliflozin group (10 mg/kg p.o.) tested SGLT2 dependence. Outcomes included renal index, serum biochemistry (BUN, Scr, TC, TG), fecal/serum LPS and cytokines (ELISA), histology (H&E, Masson), immunofluorescence (vimentin, E-cadherin), and Western blots (fibrosis markers; TGF-β/Smad; Nrf2/Keap1/12-LOX/Rac1; SGLT2; tight junction proteins; TGR5). Metabolomics/proteomics: GC-MS untargeted metabolomics profiled UUO vs sham vs UUO+BF serum, with OPLS-DA/PCA, VIP≥1.0 and FDR-adjusted p≤0.05 identifying differential metabolites; targeted GC-MS/LC-MS quantified 1,5-AG. Label-free quantitative proteomics of kidneys (MaxQuant, FDR<1%) identified differentially expressed proteins (fold change ≥5, adj p≤0.01), followed by GO/KEGG. Molecular docking and 100 ns MD simulations (MM-PBSA) evaluated 1,5-AG binding to TGR5 and SGLT2. Cell studies: HK-2 and HMC cells modeled fibrosis/inflammation; primary mouse renal tubular cells (PRTC) were stimulated with TGF-β1 (10 ng/mL) + high glucose (30 mM). 1,5-AG (50 µM) treatment assessed effects on fibrosis/inflammation markers and Nrf2/Keap1/HO-1. TGR5 function was probed via siRNA knockdown and the antagonist SBI-115 (10 µM). cAMP was measured in adenine mice. HEK293 cells stably expressing SGLT2 evaluated uptake of 1,5-AG-13C6 vs WT cells. In vitro screening of 14 natural products identified compounds modulating B. fragilis growth (OD600 over 0–72 h), highlighting madecassoside. Statistics: Group comparisons used Mann–Whitney U tests, Student’s t tests, and one-way ANOVA with Sidak’s post hoc tests; multivariate analyses via OPLS-DA/PCA; significance at adjusted p<0.05.

• B. fragilis is reduced in CKD: qPCR showed significantly decreased fecal B. fragilis in CKD vs controls in two cohorts (n=10 per group, p=0.0005; n=15 per group, p<0.0001). B. fragilis negatively correlated with BUN (r=-0.6018, p=0.0050) and Scr (r=-0.5219, p=0.0183).
• Live B. fragilis attenuates UUO-induced renal fibrosis: Oral live, but not heat-killed, B. fragilis improved renal morphology, reduced renal index, and improved BUN, Scr, TC, and TG in UUO mice. Fibrosis markers (collagen I, fibronectin, α-SMA) decreased; E-cadherin increased and vimentin decreased; histological injury and fibrosis scores were ameliorated (multiple comparisons significant as reported in Fig. 1E–L).
• LPS and inflammation reduced: CKD patients and UUO mice had elevated fecal and serum LPS; live B. fragilis significantly lowered LPS and pro-inflammatory cytokines (IL-1β, IL-6, TNF-α) in feces and serum, and reduced their renal mRNA expression.
• Key pathways modulated: B. fragilis inhibited TGF-β/Smad signaling (reduced TGF-β, Smad2/3) and oxidative stress (reduced Keap1, 12-LOX, Rac1; increased Nrf2) in UUO kidneys.
• Metabolomics identified 1,5-AG: UUO decreased serum 1,5-AG; B. fragilis restored it (targeted GC-MS). In human cohorts, 1,5-AG was markedly lower in CKD: GC-MS external validation 1: 15.15±5.21 µg/mL (controls) vs 5.56±3.11 µg/mL (CKD); LC-MS external validation 2: 18.24±6.64 µg/mL (controls) vs 4.24±4.43 µg/mL (CKD); both p<0.0001.
• 1,5-AG is renoprotective in vivo: 1,5-AG administration improved renal morphology, reduced renal index, decreased fibrosis markers and improved BUN, Scr, TC, TG, lowered LPS and inflammation, suppressed TGF-β/Smad, and activated Nrf2/Keap1 in UUO mice.
• 1,5-AG acts via TGR5: TGR5 was reduced in human IgA nephropathy and in UUO/adenine models; 1,5-AG restored TGR5 levels in vivo and in PRTC. 1,5-AG increased cAMP in adenine mice. Docking/MD supported 1,5-AG binding to TGR5 (five H-bonds; MM-PBSA ΔG ≈ -4.48 kcal/mol). TGR5 knockdown or antagonism (SBI-115) abolished 1,5-AG’s anti-fibrotic effects and Nrf2 activation in vitro.
• SGLT2 transports 1,5-AG and is upregulated by B. fragilis: Proteomics showed SGLT2 downregulated in UUO; Western blot/qPCR confirmed. B. fragilis restored renal SGLT2 expression in UUO and adenine models. Docking/MD supported 1,5-AG–SGLT2 binding (five H-bonds; MM-PBSA ΔG ≈ -19.73 kcal/mol). Empagliflozin reduced serum 1,5-AG (p=0.0003); HEK293-SGLT2 cells showed 1.7-fold higher 1,5-AG-13C6 uptake vs WT (p=0.0022), indicating 1,5-AG is an SGLT2 substrate.
• Madecassoside (Mad) promotes B. fragilis and reduces fibrosis in a microbiota-dependent manner: Of 14 screened natural products, only Mad enhanced B. fragilis growth in vitro. Oral (but not i.p.) Mad improved renal morphology, histology, function, and decreased fibrosis markers; effects were abrogated by broad-spectrum antibiotics. Mad decreased oxidative stress and TGF-β/Smad activation. Madecassic acid (MA), the hydrolysis product, was ineffective. In vivo, Mad restored reduced B. fragilis abundance in UUO mice.
• Validation in adenine model: B. fragilis, 1,5-AG, and Mad each improved histology, reduced BUN/Scr, and decreased collagen I, fibronectin, and α-SMA; MA was ineffective. B. fragilis increased SGLT2 in adenine kidneys; B. fragilis and 1,5-AG reduced fecal/serum LPS and improved intestinal tight junction proteins (occludin, ZO-1).

The study demonstrates that the commensal B. fragilis mitigates renal fibrosis in two mouse models, aligning with its known anti-inflammatory roles in other organs. The findings mechanistically link gut microbiota modulation to renal protection through two complementary axes: (1) reduction of systemic and fecal LPS, preserving intestinal barrier integrity and dampening inflammatory signaling; and (2) restoration of serum 1,5-AG via upregulation of renal SGLT2, with 1,5-AG acting as a TGR5 agonist to activate cAMP and the Nrf2/Keap1 pathway while suppressing TGF-β/Smad-mediated fibrogenesis. The human data corroborate decreased B. fragilis abundance and markedly reduced 1,5-AG levels in CKD, with negative correlations to kidney injury markers. Pharmacological and genetic interventions (empagliflozin, HEK293-SGLT2 uptake, TGR5 knockdown/antagonist) substantiate SGLT2’s role in 1,5-AG reabsorption and TGR5’s necessity for anti-fibrotic signaling. Identification of madecassoside as a B. fragilis growth promoter highlights a feasible, microbiota-dependent strategy to boost endogenous protective pathways. Collectively, the results support targeting the gut–kidney axis—lowering LPS, elevating 1,5-AG, and engaging TGR5/Nrf2—to counter renal fibrosis.

Oral administration of live B. fragilis attenuates renal fibrosis in UUO and adenine mouse models by lowering LPS-driven inflammation and restoring serum 1,5-AG through upregulation of renal SGLT2. 1,5-AG acts as a TGR5 agonist to inhibit oxidative stress and TGF-β/Smad signaling, reducing fibrotic remodeling. Madecassoside, identified as a B. fragilis growth modulator, effectively rescues B. fragilis abundance and mitigates fibrosis in a gut microbiota-dependent manner. These findings position microbiome modulation—via probiotics or microbiota-active small molecules—as a promising therapeutic strategy for progressive renal fibrosis. Future work should translate these findings to clinical interventions, define strain-specific safety/efficacy, and explore dosing, durability, and interactions with SGLT2 inhibitors.