Dysbiosis of the gut microbiota in calcium oxalate nephrolithiasis is associated with impaired short-chain fatty acid production and systemic metabolomic disruptions Authors: Xi Chen, Fangxing Zhang, Lang Cheng, Decao Niu, Jianpei Hu, Shengzhu Huang, Fubo Wang, Guijian Pang, Caisheng Huang, Mingli Li, Chao Wang, Zengnan Mo Published: 2025 ABSTRACT BackgroundThe prevalence of calcium oxalate (CaOx) kidney stones is increasing, yet the underlying mechanisms remain incompletely understood. Emerging evidence suggests that gut microbiota—particularly short-chain fatty acid (SCFA)-producing bacteria—may modulate host metabolism and inflammation, thereby influencing stone formation. However, the mechanistic links between gut dysbiosis, metabolic disturbances, and CaOx stone pathophysiology remain to be fully elucidated. This study investigates gut microbiota composition, SCFA levels, and metabolomic alterations in CaOx stone formers (CSF), aiming to uncover potential pathophysiological mechanisms and therapeutic targets.ResultsAmong 59 CSF and 60 healthy controls (HC), CSF exhibited significantly reduced microbial richness, with marked depletion of SCFA-producing bacteria such as Faecalibacterium prausnitzii and Eubacterium rectale. This dysbiosis was associated with decreased fecal and plasma SCFA levels, reduced 24-h urinary citrate, and widespread metabolic disturbances, particularly in tryptophan metabolism and the citrate cycle. Plasma SCFA levels were positively correlated with urinary citrate excretion, suggesting a regulatory link within the gut-kidney axis. Mendelian randomization analysis suggested that Bacteroides thetaiotaomicron may be a potential microbial risk factor for stone formation (OR = 1.26, 95% CI: 1.03–1.54, p = 0.028). In a hyperoxaluria rat model, interventions with F. prausnitzii, E. rectale, or sodium butyrate reduced renal CaOx crystal deposition and kidney injury.ConclusionsOur findings highlight the central role of SCFA-producing bacteria and their metabolites in maintaining metabolic balance and protecting against CaOx stone formation. Gut dysbiosis and reduced SCFA levels appear to drive metabolic changes that contribute to stone development. B. thetaiotaomicron may increase stone risk, while F. prausnitzii, E. rectale, and sodium butyrate show therapeutic potential. These insights support further exploration of microbiome-based strategies for the prevention and personalized management of kidney stones. Video AbstractSupplementary InformationThe online version contains supplementary material available at 10.1186/s40168-025-02224-5. Background Calcium oxalate (CaOx) stones, the most common type of kidney stones [1], impose a substantial burden on global healthcare due to their high incidence and recurrence rates [2–4]. While multiple factors contribute to stone formation, emerging evidence emphasizes the crucial role of the gut microbiota in their pathogenesis. Studies have shown that children with kidney stones experience significantly higher antibiotic exposure within the first 3 years of life [5], which has been strongly associated with an increased risk of stone development [6]. Antibiotic use may promote conditions conducive to stone formation by disrupting the gut microbiota and impairing the intestinal barrier [7]. Additionally, numerous studies have reported marked alterations in gut microbial composition and function in kidney stone formers [8–13]; however, a clear mechanistic understanding of how gut dysbiosis contributes to stone formation remains lacking. Gut-derived short-chain fatty acids (SCFAs)—including acetate, butyrate, and propionate—not only maintain gut health but also play essential roles in modulating immune responses, inflammation, and systemic metabolic homeostasis [14, 15]. Oxidative stress and the associated inflammatory immune response are key drivers of stone formation [1]. SCFAs have been shown to inhibit CaOx stone development by modulating immune pathways and reducing urinary oxalate excretion [16, 17]. Notably, several studies have consistently reported a significant reduction in the relative abundance of certain SCFA-producing bacteria in the gut microbiota of kidney stone patients compared to healthy controls [5, 10, 12, 13, 18–20]. However, gaps remain regarding the trends in SCFA levels in CaOx stone formers (CSF) and how the depletion of SCFA-producing bacteria and SCFA deficiency contribute to stone formation. Kidney stone formation is a complex, multi-step process involving crystal nucleation, growth, aggregation, and retention [21], each potentially influenced by distinct metabolic factors [22]. The metabolome—representing the full array of small-molecule metabolites involved in intrinsic biochemical processes—is influenced by genetic variation, environmental exposure, and their interactions [23]. While previous metabolomic studies on kidney stones have primarily focused on individual sample types (e.g., fecal, blood or urine) [5, 20, 24], an integrative multi-sample strategy is essential to uncover the full landscape of metabolic perturbations involved in stone formation. Comprehensive profiling of fecal, plasma, and urinary metabolomes in parallel provides a more holistic view of host metabolism and may reveal key metabolic alterations that would otherwise remain undetected. Moreover, integrated analyses linking gut microbiota dysbiosis, metabolomic perturbations, and nephrolithiasis remain rare. Such integrative approaches are essential for elucidating the mechanistic pathways through which microbial imbalances contribute to kidney stone formation. In this study, we aimed to comprehensively investigate the relationship between the gut microbiota and CaOx stone formation, with a focus on SCFAs and associated metabolic disturbances. By integrating metagenomic, metabolomic, and clinical data (referred to as the'phenome'), we sought to uncover novel mechanistic insights and therapeutic targets for kidney stone prevention. The study workflow is shown in Additional file 2: Fig. S1. Methods Recruitment of human subjects and sample collection A total of 59 CaOx stone formers (CSF) and 60 healthy volunteers were recruited from the First People's Hospital of Yulin between April and September 2022. Eligibility criteria for CSF included: age 20–70 years; a diagnosis of upper urinary tract stones confirmed by imaging (X-ray, ultrasound, or CT); and post-surgical stone composition analysis confirming CaOx as the primary component using infrared spectroscopy. Patients with pure calcium oxalate monohydrate (COM), pure calcium oxalate dihydrate (COD), or mixed COM/COD composition were all classified as CSF. Healthy controls (HC), also aged 20–70 years, had no history of urolithiasis, as confirmed by urinary ultrasound. Exclusion criteria for all participants were: (1) antibiotic use within the past month; (2) probiotic use within the past two weeks; (3) urinary tract malformations, polycystic kidney disease, hyperthyroidism, hyperparathyroidism, primary hyperoxaluria, neurogenic bladder, organ transplantation, inflammatory bowel disease, or gastrointestinal surgery; (4) severe metabolic diseases, malignancies, autoimmune diseases, or chronic kidney disease. Recurrent stones were defined as multiple stone episodes of stone formation within a period (typically months or years) after complete clearance following treatment. Peripheral venous blood, stool, and the midstream urine sample from the second void of the morning were collected from all subjects in the morning while fasting. A 24-h urine sample was also collected using containers pre-filled with 30 ml of 6 mol/L hydrochloric acid. All samples were stored at 4℃ immediately after collection and transferred to −80℃ within 2 h. Metadata were collected through questionnaire surveys, clinical chemistry analyses, and imaging examinations (Additional file 1: Table S1). Detailed methods are provided in Additional file 3: Supplementary Methods. Ethical approval was granted by the Guangxi Medical University Medical Research Ethics Committee (No. 20220107). Written informed consent was obtained from all participants. Fecal DNA extraction, Metagenomic Sequencing, and Microbial Profiling Fecal DNA was extracted using a Magnetic Soil and Stool DNA Kit (TIANGEN, DP712) and subjected to sequencing on the MGI-tech DNBSEQ-T7 platform (150 bp paired-end). Quality control was performed using KneadData (version 0.12.0, http://huttenhower.sph.harvard.edu/kneaddata). Low-quality bases were trimmed from the 3’ end of the reads with Trimmomatic and reads shorter than 50 nucleotides were discarded. Host-derived sequences were removed by mapping reads to human genome (hg37 build) using Bowtie 2 (version 2.5.1). The processing scripts are available on GitHub (https://github.com/YongxinLiu/EasyMetagenome) [25, 26]. Taxonomic and functional annotations were carried out using MetaPhIAn (version 4.0.6) and HuMAnN3 (version 3.6). Total high-quality paired sequences were classified into 13 phyla, 155 classes, 173 orders, 207 families, 587 genera, and 1095 species. Metagenomic analysis, restricted to microorganisms present in ≥ 5% of samples [27], was performed in R (version 4.3.1). Species-level alpha diversity indices (richness, Chao1 and Shannon) were calculated using the vegan package (version 2.6.5) and compared between groups via the Wilcoxon rank-sum test (p < 0.05). Beta diversity was visualized using non-metric multidimensional scaling (NMDS) based on Bray–Curtis distances. Adonis and Analysis of Similarities (ANOSIM) tests were conducted using vegan package. Differentially abundant gut microbial taxa between CSF and HC were identified by Linear Discriminant Analysis Effect Size (LEfSe), with thresholds of p < 0.05 (Wilcoxon test) and Linear discriminant analysis (LDA) score > 2. Spearman correlation analysis was employed to examine associations between gut microbiota and clinical phenotypes. In addition, the envfit function in vegan package was used to quantify associations between clinical features and inter-sample similarity, weighted by Bray–Curtis distance. Additional details regarding the co-occurrence network construction and microbial functional annotation are provided in Additional file 3: Supplementary Methods. Metabolomic profiling Widely targeted metabolomics analysis was performed on fecal, plasma, and urinary samples using the ABSciex QTRAP ® LC–MS/MS platform (Metware Biotechnology Co., Ltd., Wuhan, China), identifying 1562 fecal, 786 plasma, and 1484 urinary annotated metabolites. Detailed protocols for sample preparation and metabolome profiling are provided in Additional file 3: Supplementary Methods. Principal component analysis (PCA) was used to visualize metabolomic differences between CSF and HC using the ade4 package in R. Permutation-based multivariate analysis of variance (PERMANOVA) was conducted to assess statistical significance between groups. Differential metabolites were screened using ten-fold cross-validated orthogonal partial least squares discriminant analysis (OPLS-DA) in combination with a Student’s t-test, selecting those with a variable importance in projection (VIP) > 1 and false discovery rate (FDR) < 0.05 as candidate biomarkers. Partial correlation analysis, adjusted for sex, age, and BMI, was performed to examine associations between differential metabolites and clinical variables, including stone size and recurrent stones. Pathway enrichment analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG), topological analysis, and visualization were conducted in MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/). The impact of metabolites in pathways was determined by betweenness centrality, with an impact value > 0 as the threshold for pathway selection. Clusters of co-abundant differential metabolites—termed differential metabolite modules (DMMs)—were identified using weighted gene co-expression network analysis (WGCNA) [28] with the dynamic hybrid tree-cutting. WGCNA parameters were as follows: for plasma metabolites, a minimum cluster size of 3, a deepSplit value of 4, and a merged cutHeight of 0.2; for urine metabolites, a minimum cluster size of 2, a deepSplit value of 4, and a merged cutHeight of 0.3; and for fecal metabolites, a minimum cluster size of 2, a deepSplit value of 4, and a merged cutHeight of 0.2. Quantification of SCFAs, oxalate, and citrate The concentrations of SCFAs in feces and plasma were quantified by gas chromatography-mass spectrometry (GC–MS/MS), while 24-h urinary oxalate and citrate levels were determined using liquid chromatography-tandem mass spectrometry (LC–MS/MS). Detailed protocols for sample preparation and detection are provided in Additional file 3: Supplementary Methods. Mendelian randomization (MR) Genetic summary statistics for urolithiasis were sourced from a genome-wide association study (GWAS) dataset (ebi-a-GCST90018715) in BioBank Japan, comprising 11,699 cases and 167,027 controls, adjusted for age, age2, sex, age × sex and the top 20 principal components [29]. Previous studies, especially those based on the international MiBioGen consortium, have explored causal links between gut microbiota (from phylum to genus levels) and urolithiasis. In the present study, we further searched for GWAS datasets of gut bacterial species that were significantly different between CSF and HC using the IEU Open GWAS project (https://gwas.mrcieu.ac.uk/). However, such GWAS datasets were only available for a subset of these species (see Additional file 1: Table S8). Given the limited number of loci reaching genome-wide significance in gut microbiome GWAS, we adopted a suggestive threshold (p < 1 × 10⁻5), as recommended by previous studies, to select SNPs as instrumental variables (IVs) for each species [30, 31]. To ensure the independence of IVs, we applied a linkage disequilibrium threshold of R2 < 0.001 and a clumping distance of 10,000 kb. For reverse MR analysis, SNPs associated with urolithiasis were selected using a more stringent threshold of p < 1 × 10–8. MR analyses were performed using the TwoSampleMR package in R (version 4.3.1), with inverse variance weighting (IVW) as the primary method to estimate causal effect across SNPs. Complementary methods, including MR-Egger, weighted median, weighted mode, and simple mode, were used to check for pleiotropy and robustness. Cochran’s Q test was used to test the heterogeneity across SNPs. Sensitivity analyses, including MR-Egger to examine horizontal pleiotropy and a"leave-one-SNP-out"approach, were applied to evaluate the influence of individual SNPs on overall MR estimates and address potential pleiotropic effects. Animal experiment Five-week-old male Sprague–Dawley (SD) rats, weighing approximately 151–175 g, were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Foshan, Guangdong, China) and housed in a controlled environment with 12-h light/dark cycle at 26 ± 1 °C. All animal procedures were approved by Guangxi Medical University Medical Research Ethics Committee (Protocol No. 20220107). Thirty rats were acclimatized for one week with a standard diet and sterile saline, then randomly assigned to five groups: control, model, Eubacterium rectale (E. rectale), Faecalibacterium prausnitzii (F. prausnitzii) and butyrate (Fig. 10A). Details on the cultivation and preparation of E. rectale and F. prausnitzii are provided in Additional file 3: Supplementary Methods. The control group received sterile saline and a standard diet, while the remaining groups were fed a 3% glycolate diet for four weeks to induce hyperoxaluria. The control and model groups received 1 mL/day ddH2O by oral gavage, while the E. rectale and F. prausnitzii groups were given daily oral gavage of E. rectale (1 × 109 CFU/mL, 1 mL/day) and F. prausnitzii (1 × 109 CFU/mL, 1 mL/day) suspensions, respectively. The butyrate group received 1000 mg/kg/day of sodium butyrate by oral gavage. Rat body weights were recorded every six days. After four weeks, 24-h urine samples were collected from all rats using metabolic cages, with 1 mL of 6 mol/L hydrochloric acid added per 10 mL of urine to prevent oxalate degradation. Urinary citrate and oxalate concentrations were measured using LC–MS/MS, as previously described. Rats were anesthetized by intraperitoneal injection of 4% (w/v) chloral hydrate, and blood was collected from the inferior vena cava. Blood samples were centrifuged at 3000 rpm for 10 min at 4℃ to obtain serum, which was stored at −80℃ until further analysis. Kidneys were harvested and stored at −80℃, while portions of renal tissue were fixed in 10% formaldehyde and embedded in paraffin. Details of histological analysis and real-time quantitative PCR (qRT-PCR) are provided in Additional file 3: Supplementary Methods. Statistical analysis The sample size of the case–control study was not predetermined using statistical methods. Data were analyzed using SPSS (version 20), GraphPad Prism (version 8.0.2), and R (version 4.3.1). Categorical variables were presented as counts and percentages, while continuous variables were expressed as mean ± standard deviation or median (interquartile range), depending on distribution. Group comparisons for categorical variables were performed using the chi-square test. For continuous variables, an unpaired two-sided t-test was used for normally distributed data, while the Wilcoxon rank-sum test was applied to non-normally distributed data. Outliers, defined as values more than three standard deviations from the mean, were excluded. Spearman’s correlation p-values were adjusted using the Benjamini–Hochberg method. A univariate unconditional logistic regression model was used to calculate the odds ratio (OR) and 95% confidence intervals (CIs). Variables significantly associated with CaOx nephrolithiasis in the univariate analysis (p < 0.05), along with sex, age and BMI, were included in the multivariable logistic regression model. Full statistical details, including sample size and p-values, are provided in the figure legends. All statistical tests were two-sided, with a significance threshold of p < 0.05. Results Characteristics of the study population Following a rigorous screening process (detailed in the “ Methods” section), we recruited 59 CaOx stone formers (CSF) and 60 healthy controls (HC). The clinical characteristics of the study population are summarized in Table 1. Gender and age were comparable between the two groups. However, the CSF group exhibited significantly higher BMI, waist-to-hip ratio, prevalence of central obesity and hypertension, as well as a greater proportion of individuals with a family history of urolithiasis. Additionally, CSF showed elevated triglyceride (TG) and reduced high-density lipoprotein (HDL) levels, and distinct 24-h urine metabolic profiles—most notably lower levels of citrate, potassium, and phosphorus—whereas urinary oxalate levels were comparable between the two groups (Table 1). The HC group reported higher levels of physical activity and greater dietary diversity than the CSF group (Additional file 1: Table S1). Multivariate logistic regression analysis further identified a family history of urolithiasis (OR = 13.07, p = 0.001) as a significant risk factor for CaOx stone formation. In contrast, higher HDL levels (OR = 0.05, p = 0.001) and urinary citrate levels (OR = 0.993, p = 0.001) were identified as protective factors (Additional file 1: Table S2). Table 1Clinical characteristics of study participantsClinical CharacteristicsCSF(n = 59)HC(n = 60)P ValueDemographic Female, n (%)31 (52.5%)35 (58.3%)0.525 Age, years50.0 (45.0,55.0)47.5 (39.5,54.5)0.077 BMI, kg/m224.45 ± 3.8823.17 ± 2.870.042 WHR0.92 ± 0.070.87 ± 0.06< 0.001 Central obesity, n (%)45 (76.3%)30 (50.0%)0.003 SBP, mmHg132.0 (124.0,147.0)121.0 (112.5,133.0)< 0.001 DBP, mmHg87.83 ± 12.7379.55 ± 11.02< 0.001 Hypertension, n (%)37 (62.7%)15 (25.0%)< 0.001 Family history of urolithiasis, n (%)32 (54.2%)9 (15.0%)< 0.001 Diabetes, n (%)5 (8.5%)1 (1.7%)0.09Clinical Index FPG, mmol/L4.82 (4.47,5.24)5.14 (4.90,5.43)0.001 ALT, U/L15.0 (11.0,21.0)17.5 (13.0,22.5)0.112 AST, U/L17.0 (15.0,21.0)21.5 (18.5,25.5)< 0.001 Serum Cr, µmol/L89.0 (72.3,109.0)74.0 (62.0,81.0)< 0.001 Urea, µmol/L4.82 (3.98,5.94)4.62 (3.58,5.25)0.056 UA, µmol/L343.69 ± 103.37341.58 ± 87.130.904 TG, mmol/L1.51 (1.16,2.47)1.29 (0.95,1.92)0.014 TC, mmol/L4.83 (4.23,5.38)4.64 (4.20,5.31)0.510 HDL, mmol/L1.00 ± 0.301.32 ± 0.30< 0.001 LDL, mmol/L2.84 ± 0.792.99 ± 0.890.338 Urine pH6.0 (5.5,6.5)6 (5.5,6.0)0.063 USG1.010 ± 0.0051.015 ± 0.007 < 0.00124-h Urine Chemistries Urine volume, ml1800 (1300,2600)1600 (1240,2000)0.097 Sodium, mmol/24 h123.20 (88.00,165.20)139.06 (92.80,177.25)0.649 Magnesium, mmol/24 h1.90 (1.40,3.20)2.09 (1.72,2.84)0.591 Calcium, mmol/24 h2.31 (1.22,4.56)3.28 (2.16,4.65)0.063 Phosphorus, mmol/24 h14.60 ± 6.2117.28 ± 5.490.015 Potassium, mmol/24 h24.04 (18.58,36.92)33.56 (25.92,39.07)0.004 Oxalate, mg/24 h29.63 (20.02,40.11)32.61 (24.70,45.19)0.173 Citrate, mg/24 h94.89 (53.19,175.42)313.80 (190.32,477.16)< 0.001 Cystine, mg/24 h17.36 (11.06,26.89)17.815 (13.52,27.98)0.485Abbreviations: CaOx Calcium oxalate, CSF CaOx stone former, HC Healthy controls, BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, WHR Waist-to-hip ratio, FPG Fasting plasma glucose, HDL High-density lipoprotein, LDL Low-density lipoprotein, ALT Alanine aminotransferase, AST Aspartate aminotransferase, UA Uric acid, TG Triglyceride, TC Total cholesterol, USG Urine specific gravity, Cr Creatinine Table 1Clinical characteristics of study participantsClinical CharacteristicsCSF(n = 59)HC(n = 60)P ValueDemographic Female, n (%)31 (52.5%)35 (58.3%)0.525 Age, years50.0 (45.0,55.0)47.5 (39.5,54.5)0.077 BMI, kg/m224.45 ± 3.8823.17 ± 2.870.042 WHR0.92 ± 0.070.87 ± 0.06< 0.001 Central obesity, n (%)45 (76.3%)30 (50.0%)0.003 SBP, mmHg132.0 (124.0,147.0)121.0 (112.5,133.0)< 0.001 DBP, mmHg87.83 ± 12.7379.55 ± 11.02< 0.001 Hypertension, n (%)37 (62.7%)15 (25.0%)< 0.001 Family history of urolithiasis, n (%)32 (54.2%)9 (15.0%)< 0.001 Diabetes, n (%)5 (8.5%)1 (1.7%)0.09Clinical Index FPG, mmol/L4.82 (4.47,5.24)5.14 (4.90,5.43)0.001 ALT, U/L15.0 (11.0,21.0)17.5 (13.0,22.5)0.112 AST, U/L17.0 (15.0,21.0)21.5 (18.5,25.5)< 0.001 Serum Cr, µmol/L89.0 (72.3,109.0)74.0 (62.0,81.0)< 0.001 Urea, µmol/L4.82 (3.98,5.94)4.62 (3.58,5.25)0.056 UA, µmol/L343.69 ± 103.37341.58 ± 87.130.904 TG, mmol/L1.51 (1.16,2.47)1.29 (0.95,1.92)0.014 TC, mmol/L4.83 (4.23,5.38)4.64 (4.20,5.31)0.510 HDL, mmol/L1.00 ± 0.301.32 ± 0.30< 0.001 LDL, mmol/L2.84 ± 0.792.99 ± 0.890.338 Urine pH6.0 (5.5,6.5)6 (5.5,6.0)0.063 USG1.010 ± 0.0051.015 ± 0.007 < 0.00124-h Urine Chemistries Urine volume, ml1800 (1300,2600)1600 (1240,2000)0.097 Sodium, mmol/24 h123.20 (88.00,165.20)139.06 (92.80,177.25)0.649 Magnesium, mmol/24 h1.90 (1.40,3.20)2.09 (1.72,2.84)0.591 Calcium, mmol/24 h2.31 (1.22,4.56)3.28 (2.16,4.65)0.063 Phosphorus, mmol/24 h14.60 ± 6.2117.28 ± 5.490.015 Potassium, mmol/24 h24.04 (18.58,36.92)33.56 (25.92,39.07)0.004 Oxalate, mg/24 h29.63 (20.02,40.11)32.61 (24.70,45.19)0.173 Citrate, mg/24 h94.89 (53.19,175.42)313.80 (190.32,477.16)< 0.001 Cystine, mg/24 h17.36 (11.06,26.89)17.815 (13.52,27.98)0.485Abbreviations: CaOx Calcium oxalate, CSF CaOx stone former, HC Healthy controls, BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, WHR Waist-to-hip ratio, FPG Fasting plasma glucose, HDL High-density lipoprotein, LDL Low-density lipoprotein, ALT Alanine aminotransferase, AST Aspartate aminotransferase, UA Uric acid, TG Triglyceride, TC Total cholesterol, USG Urine specific gravity, Cr Creatinine Clinical characteristics of study participants Abbreviations: CaOx Calcium oxalate, CSF CaOx stone former, HC Healthy controls, BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, WHR Waist-to-hip ratio, FPG Fasting plasma glucose, HDL High-density lipoprotein, LDL Low-density lipoprotein, ALT Alanine aminotransferase, AST Aspartate aminotransferase, UA Uric acid, TG Triglyceride, TC Total cholesterol, USG Urine specific gravity, Cr Creatinine Alterations in gut microbiome composition in CaOx stone formers Shotgun metagenomic sequencing of 119 fecal samples (59 CSF and 60 HC) was performed using MetaPhIAn 4, identifying 11 phyla, 59 classes, 72 orders, 96 families, 237 genera, and 464 species (Additional file 1: Table S3 and Additional file 2: Fig. S2). Microbial richness, as assessed by species count and the Chao1 index, was significantly lower in CSF than in HC. In contrast, microbial diversity and evenness, as reflected by the Shannon index, did not differ significantly between the two groups (Fig. 1A–C). β-diversity analysis, conducted using NMDS based on Bray–Curtis distances, showed no significant differences in overall microbial community structure between CSF and HC (ANOSIM test: R = 0.013, p = 0.09, permutations = 999, Fig. 1D).Fig. 1Characteristics of the gut microbiome in CaOx stone formers (CSF). A-C Comparisons of α-diversity metrics between the CSF and healthy controls (HC), including species richness (A), Chao1 index (B), and Shannon index (C). D Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distance indicating similarities in gut microbiome composition between the CSF and HC, with statistical significance assessed by ANOSIM. E Linear discriminant analysis effect size (LEfSe) identified 39 bacterial species with significantly different relative abundances (LDA > 2). F SCFA-producing bacterial species with significantly different relative abundances between groups, including several major butyrate producers. On the x-axis, species names in red indicate higher relative abundance in CSF, while those in blue indicate higher relative abundance in HC. G Correlation between differential SCFA-producing bacterial species and key clinical indicators. * P < 0.05. Cr = creatinine; HDL = high-density lipoprotein; Oxalate/24 h = 24-h urinary oxalate; Citrate/24 h = 24-h urinary citrate Characteristics of the gut microbiome in CaOx stone formers (CSF). A-C Comparisons of α-diversity metrics between the CSF and healthy controls (HC), including species richness (A), Chao1 index (B), and Shannon index (C). D Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distance indicating similarities in gut microbiome composition between the CSF and HC, with statistical significance assessed by ANOSIM. E Linear discriminant analysis effect size (LEfSe) identified 39 bacterial species with significantly different relative abundances (LDA > 2). F SCFA-producing bacterial species with significantly different relative abundances between groups, including several major butyrate producers. On the x-axis, species names in red indicate higher relative abundance in CSF, while those in blue indicate higher relative abundance in HC. G Correlation between differential SCFA-producing bacterial species and key clinical indicators. * P < 0.05. Cr = creatinine; HDL = high-density lipoprotein; Oxalate/24 h = 24-h urinary oxalate; Citrate/24 h = 24-h urinary citrate At the species level, LEfSe analysis identified 39 differentially abundant species between the two groups (p < 0.05, LDA > 2.0, Fig. 1E). Of these, 20 species were annotated as SCFA producers based on prior literature [15, 32–34], including E. rectale and F. prausnitzii, two dominant butyrate-producing bacteria in the human colon known to play key roles in gut and systemic health [35]. Fourteen of these SCFA-producing species were significantly depleted in CSF compared to HC (Fig. 1F). Notably, the relative abundance of these bacteria was positively correlated with 24-h urinary citrate excretion, a known inhibitor of kidney stone formation (Fig. 1G). Hyperoxaluria is a well-established risk factor for kidney stone formation, and Oxalobacter formigenes (O. formigenes) is the most effective oxalate-degrading bacterium in the gut, known to reduce oxalate absorption into the bloodstream. Animal studies have demonstrated that increasing the abundance of O. formigenes can significantly inhibit kidney stone formation [36]. Moreover, decreased colonization of O. formigenes has been identified as a potential risk factor for kidney stone development in previous studies [37–39]. However, in our study, no significant differences in the abundance of O. formigenes or the expression of key oxalate-degrading enzymes were observed between CSF and HC (Additional file 2: Fig. S3), a result consistent with certain earlier reports [12, 18, 19]. To further investigate inter-species relationships within the gut microbiota of CSF and HC, co-abundance networks were constructed using FlashWeave [40]. The Stone Network comprised 88 species (nodes) and 96 edges (Additional file 2: Fig. S4), while the Control Network included 89 species (nodes) and 116 edges (Additional file 2: Fig. S5). Significant differences were observed in degree, average shortest path length, and closeness centrality between the two networks (Additional file 2: Fig. S6A, B, D), indicating tighter interactions among bacterial species in the Control Network. Notably, SCFA-producing bacterial species in the Control Network exhibited shorter average shortest path length and higher closeness centrality (Additional file 2: Fig. S6F, H), suggesting stronger regulatory interactions within the healthy gut microbiome. Impaired SCFA-related and broader metabolic functions in the gut microbiota of CaOx stone formers To investigate the functional characteristics of the gut microbiome in CSF, HUMAnN3 was used to identify 3768 KEGG orthologues (KOs), among which 306 showed significant differences in abundance between CSF and HC (Wilcoxon rank-sum test, FDR < 0.05). Notably, only five KOs were enriched in CSF compared to HC (Additional file 1: Table S4). However, NMDS analysis of KO profiles based on the Bray–Curtis distances revealed no significant differences in overall microbial functional composition between the groups (ANOSIM test: R = 0.014, p = 0.083, permutations = 999) (Fig. 2A). KEGG pathway enrichment analysis of the differential KOs highlighted a depletion of SCFA synthesis-related metabolic pathways, including pyruvate metabolism, propanoate metabolism, and butanoate metabolism. All KOs corresponding to the top 20 enriched pathways were downregulated in CSF, suggesting a reduced metabolic capacity in these pathways (Fig. 2B).Fig. 2Alterations in metabolic functions of the gut microbiota in CaOx stone formers (CSF). A Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distance shows no significant separation in gut microbial functions between CSF and healthy controls (HC), with ANOSIM analysis evaluating β-diversity differences. B Pathway enrichment analysis of significantly different KEGG orthologs (KOs) (Top 20 enriched pathways). C-E Comparison of the relative abundance of SCFA synthesis-related metabolic function modules between CSF and HC, with MF0087 and MF0093 annotations derived from GMMs, and M00307 annotations derived from the KEGG functional module reference framework. *, **, *** indicate FDR < 0.05, FDR < 0.01, and FDR < 0.0001, respectively Alterations in metabolic functions of the gut microbiota in CaOx stone formers (CSF). A Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distance shows no significant separation in gut microbial functions between CSF and healthy controls (HC), with ANOSIM analysis evaluating β-diversity differences. B Pathway enrichment analysis of significantly different KEGG orthologs (KOs) (Top 20 enriched pathways). C-E Comparison of the relative abundance of SCFA synthesis-related metabolic function modules between CSF and HC, with MF0087 and MF0093 annotations derived from GMMs, and M00307 annotations derived from the KEGG functional module reference framework. *, **, *** indicate FDR < 0.05, FDR < 0.01, and FDR < 0.0001, respectively Further analysis revealed that SCFA synthesis-related metabolic modules—such as Pyruvate oxidation (M00307), acetyl-CoA to crotonyl-CoA (MF0087), and propionate production (MF0093)—were significantly less abundant in CSF (Fig. 2C–E), indicating a diminished SCFA synthesis capacity. To enhance statistical rigor, we reanalyzed these modules using MaAsLin2 with adjustment for age, sex, and BMI, which confirmed the same trends observed the initial Wilcoxon rank-sum test (Additional file 2: Fig. S7). In addition, several other modules related to the citrate cycle, amino acids synthesis, and the biosynthesis of vitamins and cofactors were also depleted in CSF (Additional file 2: Fig. S8), suggesting broader metabolic dysfunction that may contribute to metabolic imbalances in these individuals. Decreased SCFA levels in CaOx stone formers and their association with bacterial species and SCFA synthesis-related metabolic modules Plasma levels of acetate and butyrate were significantly lower in CSF compared to HC (Fig. 3A). Although fecal levels of acetate, butyrate and propionate exhibited a downward trend in CSF, these differences were not statistically significant (Fig. 3B). Notably, participants without a family history of urolithiasis had significantly higher levels of fecal acetate, propionate, total SCFAs, and plasma acetate than those with a family history. (Additional file 2: Fig. S9). Correlation analysis revealed that most bacterial species enriched in HC were positively associated with fecal and plasma levels of acetate and butyrate (Fig. 3C). Similarly, SCFA synthesis-related metabolic modules—including pyruvate oxidation (M00307), acetyl-CoA to crotonyl-CoA (MF0087), and propionate production (MF0093)—were positively correlated with acetate and butyrate levels in both feces and plasma (Fig. 3D). Furthermore, plasma acetate and butyrate levels were positively correlated with 24-h urinary citrate excretion (Fig. 3E). Plasma butyrate was also positively associated with HDL levels (Fig. 3F), suggesting a potential link between SCFA levels and kidney stone risk.Fig. 3Alterations in plasma and fecal levels of major SCFAs in CaOx stone formers (CSF). A-B Comparison of plasma (A) and fecal (B) levels of major SCFAs between CSF and healthy controls (HC). C Heatmap illustrating the correlations between plasma and fecal SCFA levels and bacterial species with significantly different abundances between CSF and HC. D Heatmap showing the correlations between plasma and fecal SCFA levels and SCFA synthesis-related function modules that are depleted in CSF. E Plasma levels of acetate and butyrate positively correlate with 24-h urinary citrate. F Plasma butyrate levels are positively correlated with plasma HDL levels. AA = acetic acid; PA = propionic acid; BA = butyric acid Alterations in plasma and fecal levels of major SCFAs in CaOx stone formers (CSF). A-B Comparison of plasma (A) and fecal (B) levels of major SCFAs between CSF and healthy controls (HC). C Heatmap illustrating the correlations between plasma and fecal SCFA levels and bacterial species with significantly different abundances between CSF and HC. D Heatmap showing the correlations between plasma and fecal SCFA levels and SCFA synthesis-related function modules that are depleted in CSF. E Plasma levels of acetate and butyrate positively correlate with 24-h urinary citrate. F Plasma butyrate levels are positively correlated with plasma HDL levels. AA = acetic acid; PA = propionic acid; BA = butyric acid The impact of clinical phenotypes and SCFA-producing bacteria on the gut microbiome The gut microbiome is shaped by various internal and external factors. Our analysis revealed significant associations between fecal SCFAs (acetic acid, propionic acid, butyric acid, isobutyric acid and valeric acid), dietary intake of vegetable and cereal, and variations in gut microbiota composition (adonis p < 0.05, Fig. 4A). Notably, fecal SCFAs collectively explained 12.1% of the variance in microbial composition (Fig. 4A) and, together with vegetable intake, steered the gut microbiome composition toward a healthy state (Fig. 4C), underscoring their pivotal role in shaping gut microbial structure. SCFAs have been shown to influence both host metabolism and microbial functions [41]. In our study, fecal propionic acid, butyric acid, and valeric acid significantly contributed to microbial functional variation, accounting for 11.6% of the total variance (Fig. 4B), and were associated with a shift in microbial function toward patterns characteristic of the HC group (Fig. 4D).Fig. 4Influence of clinical phenotypes and SCFA-producing bacteria on gut microbiota. A Effect sizes of phenotype indices that significantly contribute to the variance (R2) of gut microbiota composition (adonis p < 0.05). B Effect sizes of phenotype indices that significantly contribute to the variance (R2) of gut microbiota metabolic function (adonis p < 0.05). C Non-metric multidimensional scaling (NMDS) plot showing the linear fit between gut microbiota species composition and clinical phenotypes. D NMDS plot illustrating the linear fit between gut microbiota metabolic function and clinical phenotypes. E Effect sizes of SCFA-producing bacterial species that significantly contribute to the variance (R2) of gut microbiota composition (adonis p < 0.05). F Effect sizes of SCFA-producing bacterial species that significantly contribute to the variance of gut microbiota metabolic function (adonis p < 0.05). AA = acetic acid; PA = propionic acid; BA = butyric acid; IBA = isobutyric acid; VA = valeric acid; TG = triglycerides Influence of clinical phenotypes and SCFA-producing bacteria on gut microbiota. A Effect sizes of phenotype indices that significantly contribute to the variance (R2) of gut microbiota composition (adonis p < 0.05). B Effect sizes of phenotype indices that significantly contribute to the variance (R2) of gut microbiota metabolic function (adonis p < 0.05). C Non-metric multidimensional scaling (NMDS) plot showing the linear fit between gut microbiota species composition and clinical phenotypes. D NMDS plot illustrating the linear fit between gut microbiota metabolic function and clinical phenotypes. E Effect sizes of SCFA-producing bacterial species that significantly contribute to the variance (R2) of gut microbiota composition (adonis p < 0.05). F Effect sizes of SCFA-producing bacterial species that significantly contribute to the variance of gut microbiota metabolic function (adonis p < 0.05). AA = acetic acid; PA = propionic acid; BA = butyric acid; IBA = isobutyric acid; VA = valeric acid; TG = triglycerides Additionally, 20 differential SCFA-producing bacterial species together explained 24.1% of the variance in microbial composition. Among them, Bacteroides thetaiotaomicron (B. thetaiotaomicron) accounted for the largest proportion of explained variance (Fig. 4E). Furthermore, B. thetaiotaomicron, Dialister hominis, Clostridium sp., and Lachnospiraceae bacterium significantly influenced the functional composition of the gut microbiota, with B. thetaiotaomicron alone explained 12.1% of the variance in metabolic functions, highlighting its critical role in modulating gut microbial metabolism (Fig. 4F). Global metabolic changes in CaOx stone formers Metabolomic profiling of fecal, plasma, and urinary samples revealed substantial differences in the metabolic landscapes between CSF and HC (Fig. 5A–C). A total of 306 fecal metabolites, 185 plasma metabolites, and 446 urinary metabolites were differentially abundant between the two groups (Additional file 1: Table S5–S7; Additional file 2: Fig. S10). Partial Spearman correlation analysis identified several differential metabolites across all three sample types that were significantly associated with stone volume or recurrent stones, with the majority of associations observed in urine (Additional file 2: Fig. S11A–C). These findings suggest that fluctuations in the urinary metabolome are closely linked to stone formation and progression. Notably, plasma levels of succinate—a key intermediate in the tricarboxylic acid (TCA) cycle—were significantly reduced in CSF and inversely correlated with stone volume (Additional file 2: Fig. S11B, D), highlighting its potential as a therapeutic target. Procrustes analysis demonstrated significant cooperativity among fecal, plasma and fecal metabolome profiles, with the strongest association observed between fecal and plasma profiles (Additional file 2: Fig. S12).Fig. 5Distinct fecal, plasma, and urinary metabolic features associated with CaOx stone formers (CSF). A–C Principal component analysis (PCA)showing differences in metabolite profiles of feces (A), plasma (B), and urine (C) between CSF and healthy controls (HC). D Effect sizes of phenotype indices that significantly contribute to the variance (R2) in fecal metabolite profiles (adonis p < 0.05), based on all subjects (59 CSF and 60 HC). E Effect sizes of phenotype indices that significantly contribute to the variance (R2) in plasma metabolite profiles (adonis p < 0.05). F Effect sizes of phenotype indices which significantly contribute to the variance(R2) in urinary metabolite profiles (adonis p < 0.05). G Mantel test results showing the associations between SCFA levels and global metabolic changes. AA = acetic acid; PA = propionic acid; BA = butyric acid; IBA = isobutyric acid; VA = valeric acid; IVA = isovaleric acid; CA = caproic acid; GPAQ = Global physical activity questionnaire; BMI = body mass index; WHR = waist-to-hip ratio; Cr = creatinine; Cl = chloride; Mg = magnesium; Ca = calcium; HDL = high-density lipoprotein; TC = total cholesterol; TG = triglycerides; AST = aspartate aminotransferase; UA = uric acid; Na/24 h = 24-h urinary sodium; USG = urine specific gravity Distinct fecal, plasma, and urinary metabolic features associated with CaOx stone formers (CSF). A–C Principal component analysis (PCA)showing differences in metabolite profiles of feces (A), plasma (B), and urine (C) between CSF and healthy controls (HC). D Effect sizes of phenotype indices that significantly contribute to the variance (R2) in fecal metabolite profiles (adonis p < 0.05), based on all subjects (59 CSF and 60 HC). E Effect sizes of phenotype indices that significantly contribute to the variance (R2) in plasma metabolite profiles (adonis p < 0.05). F Effect sizes of phenotype indices which significantly contribute to the variance(R2) in urinary metabolite profiles (adonis p < 0.05). G Mantel test results showing the associations between SCFA levels and global metabolic changes. AA = acetic acid; PA = propionic acid; BA = butyric acid; IBA = isobutyric acid; VA = valeric acid; IVA = isovaleric acid; CA = caproic acid; GPAQ = Global physical activity questionnaire; BMI = body mass index; WHR = waist-to-hip ratio; Cr = creatinine; Cl = chloride; Mg = magnesium; Ca = calcium; HDL = high-density lipoprotein; TC = total cholesterol; TG = triglycerides; AST = aspartate aminotransferase; UA = uric acid; Na/24 h = 24-h urinary sodium; USG = urine specific gravity Fecal SCFAs accounted for 15.7%, 11.1% and 7.1% of the variance in the fecal, plasma and urinary metabolomes, respectively, and serving as key contributors to the global metabolic differences between CSF and HC (Fig. 5D–F). Gender exerted a modest yet statistically significant effect on metabolomic profiles across fecal (R2 = 1.3%), plasma (R2 = 3.2%), and urine (R2 = 3.7%) samples of CSF (all p < 0.05; Fig. 5D–F), suggesting possible sex-related metabolic differences in stone formation. Importantly, stone status independently explained the largest variance in the urinary metabolome (R2 = 5.1%) (Fig. 5F). Mantel test results further revealed significant correlations between fecal SCFA levels and global metabolic alterations across fecal, plasma, and urinary profiles, whereas no such association was observed for plasma SCFAs (Fig. 5G), suggesting that fecal SCFAs may play a broader role in shaping host metabolism. Disruption in the kynurenine pathway of tryptophan metabolism in CaOx stone formers To investigate metabolic pathways associated with kidney stone formation, we performed KEGG pathway enrichment analysis on differential metabolites identified in fecal, plasma, and urinary samples (Additional file 2: Fig S13–S15). Upregulated metabolites in CSF were collectively enriched in seven pathways, including tryptophan (Trp) metabolism, steroid hormone biosynthesis, and vitamin B6 metabolism (Additional file 2: Fig. S16A), while downregulated metabolites were enriched in pathways, such as Trp metabolism, the citrate cycle, and primary bile acid biosynthesis (Additional file 2: Fig. S16B). An imbalance in Trp metabolism appears to be a common feature in CSF. Over 95% of Trp is degraded via the kynurenine (Kyn) pathway, and dysregulation of this pathway has been implicated in various diseases. The Kyn/Trp ratio serves as a marker of pathway activity and has been correlated with clinical features and adverse outcomes in various conditions [42]. Increased Trp catabolism is recognized as a systemic phenomenon in inflammation and may affect impacting host immunity [43]. Our analysis showed significantly elevated levels of Kyn and an increased Kyn/Trp ratio in both plasma and urine samples from CSF (Fig. 6A, B). Furthermore, the Kyn/Trp ratio was associated with inflammation, dyslipidemia, impaired kidney function, and reduced 24-h urinary citrate excretion (Fig. 6C). Notably, plasma SCFA levels were inversely associated with the Kyn/Trp ratio in both plasma and urine (Fig. 6D–G). These findings suggest that the disruption of the Kyn pathway may contribute to CaOx stones formation and may represent a promising therapeutic target.Fig. 6Disruption of the kynurenine pathway of tryptophan metabolism in CaOx stone formers (CSF). A Alterations of kynurenine pathway-related metabolites in CSF. Each filled circle represents a metabolite identified in fecal, plasma, and urine samples. The circle is divided into three segments: the first segment represents the fecal metabolome, the middle segment represents the plasma metabolome, and the last segment represents the urinary metabolome. The color of each segment indicates the log2-transformed fold change (log2FC) in metabolite abundance between CSF and healthy controls (HC), with yellow indicating upregulation and blue indicating downregulation. B Comparison of kynurenine/tryptophan (Kyn/Trp) ratios in plasma and urine between CSF and HC. C Heatmap illustrating correlations between key clinical indices (e.g., BMI, creatinine, citrate, and other metabolic markers) and the Kyn/Trp ratio in plasma and urine. D Scatter plot showing a significant negative correlation between plasma acetic acid levels and the plasma Kyn/Trp ratio in CSF. E–G Negative correlations between plasma acetic acid, butyric acid, and total SCFA levels and the urinary Kyn/Trp ratio. *, **, *** indicate P < 0.05, P < 0.01, and P < 0.0001, respectively. AA = acetic acid; BA = butyric acid; Cr = creatinine; HDL = high-density lipoprotein; TC = total cholesterol; TG = triglycerides; UA = uric acid; USG = urine specific gravity; WBC = White blood cell. NEUT (%) = Neutrophil percentage; LYM (%) = lymphocyte percentage; NLR = Neutrophil-to-lymphocyte ratio Disruption of the kynurenine pathway of tryptophan metabolism in CaOx stone formers (CSF). A Alterations of kynurenine pathway-related metabolites in CSF. Each filled circle represents a metabolite identified in fecal, plasma, and urine samples. The circle is divided into three segments: the first segment represents the fecal metabolome, the middle segment represents the plasma metabolome, and the last segment represents the urinary metabolome. The color of each segment indicates the log2-transformed fold change (log2FC) in metabolite abundance between CSF and healthy controls (HC), with yellow indicating upregulation and blue indicating downregulation. B Comparison of kynurenine/tryptophan (Kyn/Trp) ratios in plasma and urine between CSF and HC. C Heatmap illustrating correlations between key clinical indices (e.g., BMI, creatinine, citrate, and other metabolic markers) and the Kyn/Trp ratio in plasma and urine. D Scatter plot showing a significant negative correlation between plasma acetic acid levels and the plasma Kyn/Trp ratio in CSF. E–G Negative correlations between plasma acetic acid, butyric acid, and total SCFA levels and the urinary Kyn/Trp ratio. *, **, *** indicate P < 0.05, P < 0.01, and P < 0.0001, respectively. AA = acetic acid; BA = butyric acid; Cr = creatinine; HDL = high-density lipoprotein; TC = total cholesterol; TG = triglycerides; UA = uric acid; USG = urine specific gravity; WBC = White blood cell. NEUT (%) = Neutrophil percentage; LYM (%) = lymphocyte percentage; NLR = Neutrophil-to-lymphocyte ratio Microbiome alterations mediate the metabolome changes in CaOx stone formers and link to clinical phenotypes Given the significant impact of fecal SCFAs on the metabolic alterations in CSF and the depletion of SCFA-producing bacteria in their gut, we investigated the connection between the gut microbiome and metabolomic profiles. Our analysis revealed that the gut microbiome accounted for 49.1%, 27.6% and 27.6% of the variance in fecal, plasma, and urinary metabolomes of CSF, respectively (Fig. 7A). Notably, its influence on the metabolome exceeded that of host phenotypes in both CSF and HC (Fig. 7A, B), underscoring its critical role in shaping the host metabolic landscape. Dysbiosis of SCFA-producing bacterial species contributed to metabolic alterations across all three sample types, with B. thetaiotaomicron exerting particularly notable effects on both fecal and plasma profiles (Fig. 8B–D).Fig. 7The gut microbiome modulates host global metabolomes in CaOx stone formers (CSF). A The proportions of total variation in fecal, plasma, and urine metabolomes explained by the gut microbiome and host phenome in CSF and healthy controls (HC). B Effect sizes of SCFA-producing bacterial species significantly contributing to variance(R2) in fecal metabolite profiles. C Effect sizes of SCFA-producing bacterial species significantly contributing to variance (R2) in plasma metabolite profiles. D Effect sizes of SCFA-producing bacterial species significantly contributing to variance (R.2) in urinary metabolite profiles. + indicates P < 0.05Fig. 8Associations of B. thetaiotaomicron and F. prausnitzii with microbial functions, host metabolome, and stone-related phenotypes. Correlations between the abundance of these bacterial species and microbial functional modules, as well as between microbial functional modules and metabolite modules, were analyzed using MaAsLin2, with adjustments for sex, age, and BMI as confounding covariates. Associations between metabolite modules and stone-related phenotypes were evaluated using partial correlation analysis, also adjusted for sex, age, and BMI. Grey edges indicate negative correlations, whereas orange edges indicate positive correlations. The metabolite module number is followed by the Class II substance categories corresponding to the core metabolites in the metabolite module, along with the names of the core metabolites The gut microbiome modulates host global metabolomes in CaOx stone formers (CSF). A The proportions of total variation in fecal, plasma, and urine metabolomes explained by the gut microbiome and host phenome in CSF and healthy controls (HC). B Effect sizes of SCFA-producing bacterial species significantly contributing to variance(R2) in fecal metabolite profiles. C Effect sizes of SCFA-producing bacterial species significantly contributing to variance (R2) in plasma metabolite profiles. D Effect sizes of SCFA-producing bacterial species significantly contributing to variance (R.2) in urinary metabolite profiles. + indicates P < 0.05 Associations of B. thetaiotaomicron and F. prausnitzii with microbial functions, host metabolome, and stone-related phenotypes. Correlations between the abundance of these bacterial species and microbial functional modules, as well as between microbial functional modules and metabolite modules, were analyzed using MaAsLin2, with adjustments for sex, age, and BMI as confounding covariates. Associations between metabolite modules and stone-related phenotypes were evaluated using partial correlation analysis, also adjusted for sex, age, and BMI. Grey edges indicate negative correlations, whereas orange edges indicate positive correlations. The metabolite module number is followed by the Class II substance categories corresponding to the core metabolites in the metabolite module, along with the names of the core metabolites Despite distinct global metabolomic differences between CSF and HC, interpreting the physiological relevance of the hundreds of differential metabolites remains challenging. To address this, we applied WGCNA to cluster co-varying differential metabolites into distinct modules and identify key metabolites (Additional file 1: Table S5-7). A total of 36, 33 and 36 differential metabolite modules (DMMs) were identified in fecal, plasma, and urinary samples, respectively (Additional file 2: Fig. S17–S19). Dysbiosis of SCFA-producing bacterial species may contribute to the CaOx stones formation by altering host metabolic profile. To further explore this mechanism, we constructed a network linking “SCFA-producing bacteria → microbial functional modules → host metabolite modules → stone-related phenotypes” (Additional file 2: Fig. S20). Our findings suggest that F. prausnitzii might inhibit stone formation by enhancing microbial metabolic functions such as the citrate cycle and amino acid synthesis. These functions positively regulate DMMs related to oxidized lipids, organic acid, sphingolipids, and phenolic acids, contributing to increased serum HDL and elevated 24-h urinary citrate excretion (Fig. 8). In contrast, B. thetaiotaomicron appeared to promote stone formation by modulating tryptophan and pantothenate biosynthesis, upregulating Tyr-Ala (a small peptide), and downregulating organic acid and its derivatives. These alterations were associated with decreased HDL and reduced urinary citrate levels (Fig. 8). Exploratory causal inference between the gut microbiome and urolithiasis using Mendelian randomization (MR) Randomized controlled trials (RCTs) are considered the gold standard for establishing causal relationships; however, they are often impractical for investigating the link between gut microbiota and urolithiasis. As an alternative, MR leverages single nucleotide polymorphisms (SNPs) as instrumental variables (IVs) to infer causality while minimizing biases from confounding and limited sample sizes [44]. In this study, we conducted a two-sample MR analysis to explore the potential causal relationships between species-level gut microbiota and urolithiasis using publicly available GWAS data. Using the inverse variance weighted (IVW) method as the primary MR approach, our analysis suggested that B. thetaiotaomicron might be a potential microbial risk factor for urolithiasis (OR = 1.26, 95% CI: 1.03–1.54, p = 0.028). In contrast, other species, including F. prausnitzii and E. rectale, showed no significant associations (Additional file 1: Table S9). Reverse MR analysis suggested a potential causal effect of urolithiasis on the decreased abundance of E. rectale (OR = 0.91, 95% CI: 0.82–0.99, p = 0.042; Additional file 1: Table S10). These relationships are visualized in scatter plots (Fig. 9 A, B). No significant heterogeneity or directional horizontal pleiotropy was detected based on Cochran’s Q test and the MR-Egger regression intercepts (all p > 0.05; Additional file 1: Table S11, S12). Visual inspections, including funnel plots and leave-one-out analyses, further supported the robustness of these findings (Additional file 2: Fig. S21 and S22). However, given the lenient GWAS significance thresholds applied for instrument selection and the species-level resolution of the microbiome data—which may introduce bias and ambiguity—the causal inferences drawn from these analyses should be interpreted with caution.Fig. 9Scatter plots illustrating causal relationships between gut microbiota and urolithiasis through Mendelian randomization (MR) analysis. A Forward MR analysis suggests Bacteroides thetaiotaomicron as a potential risk factor for urolithiasis. B Reverse MR analysis indicates a possible link between urolithiasis and reduced abundance of Eubacterium rectale Scatter plots illustrating causal relationships between gut microbiota and urolithiasis through Mendelian randomization (MR) analysis. A Forward MR analysis suggests Bacteroides thetaiotaomicron as a potential risk factor for urolithiasis. B Reverse MR analysis indicates a possible link between urolithiasis and reduced abundance of Eubacterium rectale Effects of SCFA-producing bacteria and sodium butyrate on kidney stone formation in a hyperoxaluria rat model Given the depletion of key butyrate-producing bacteria (F. prausnitzii and E. rectale) in CSF, we conducted a 4-week intervention using these bacteria and sodium butyrate in a glycolate-induced hyperoxaluria rat model to evaluate their effects on CaOx stone formation (Fig. 10A). No significant differences in body weight were observed among the groups during the experiment (Fig. 10B). Hematoxylin–eosin (HE) and Pizzolato staining revealed extensive crystal deposition and tubular dilation in model group, which were notably reduced in the E. rectale, F. prausnitzii, and sodium butyrate groups (Fig. 10D, E). The kidney index (ratio of kidney weight to body weight), along with serum urea and creatinine levels, were significantly elevated in the model group compared to controls but were markedly decreased following sodium butyrate intervention (Fig. 10C, F, G). Although reductions were also observed in the E. rectale and F. prausnitzii groups, these did not reach statistical significance. Urinary oxalate levels were elevated in both the model and intervention groups; however, the F. prausnitzii group showed a significant reduction compared to the model group (Fig. 10H). Urinary citrate excretion was significantly decreased in the model, E. rectale and F. prausnitzii groups, while the sodium butyrate group showed levels comparable to those in the control group (Fig. 10). Additionally, glycolate administration significantly upregulated renal expression of osteopontin (OPN) and interleukin-6 (IL-6), both of which were effectively suppressed at the mRNA level by E. rectale, F. prausnitzii, and sodium butyrate (Fig. 10J, K). These findings suggested that E. rectale, F. prausnitzii, and especially sodium butyrate effectively reduce CaOx crystal deposition and renal inflammation, with sodium butyrate exhibiting the greatest efficacy in alleviating kidney damage.Fig. 10Animal experiments validating the role of E. rectale, F. prausnitzii, and sodium butyrate in kidney stone formation. A Schematic diagram of the animal study design. Twenty-four Sprague–Dawley (SD) rats were treated with glycolate for 4 weeks. One group (n = 6) received E. rectale treatment, a second group (n = 6) received F. prausnitzii treatment, a third group (n = 6) received sodium butyrate treatment, and a fourth group (n = 6) received vehicle treatment. Six SD rats served as controls, receiving vehicle without glycolate. B Body weight changes of rats throughout the study. C Kidney index (kidney weight/body weight ratio) across the five groups. D Hematoxylin–eosin (HE) staining and Pizzolato staining of kidney tissue. E Proportion of Pizzolato staining area in kidney tissues across groups. F, G Serum urea (F) and Serum creatinine (G) concentrations in the five groups. H, I Levels of 24-h urinary oxalate(H) and citrate(I) in the five groups. J, K Relative mRNA expression of OPN (J) and IL-6 (K) among groups. Data are presented as the mean ± SD and analyzed using Student's t- test between two groups. *P < 0.05, ** P < 0.01, *** P < 0.001 Animal experiments validating the role of E. rectale, F. prausnitzii, and sodium butyrate in kidney stone formation. A Schematic diagram of the animal study design. Twenty-four Sprague–Dawley (SD) rats were treated with glycolate for 4 weeks. One group (n = 6) received E. rectale treatment, a second group (n = 6) received F. prausnitzii treatment, a third group (n = 6) received sodium butyrate treatment, and a fourth group (n = 6) received vehicle treatment. Six SD rats served as controls, receiving vehicle without glycolate. B Body weight changes of rats throughout the study. C Kidney index (kidney weight/body weight ratio) across the five groups. D Hematoxylin–eosin (HE) staining and Pizzolato staining of kidney tissue. E Proportion of Pizzolato staining area in kidney tissues across groups. F, G Serum urea (F) and Serum creatinine (G) concentrations in the five groups. H, I Levels of 24-h urinary oxalate(H) and citrate(I) in the five groups. J, K Relative mRNA expression of OPN (J) and IL-6 (K) among groups. Data are presented as the mean ± SD and analyzed using Student's t- test between two groups. *P < 0.05, ** P < 0.01, *** P < 0.001 Discussion With the increasing prevalence of metabolic syndrome, the incidence of kidney stones has risen correspondingly [45, 46]. Emerging evidence highlights the pivotal role of gut microbiota in various health conditions, including hypertension, diabetes, obesity, and cardiovascular diseases [47–52]. Our findings expand on this evidence, demonstrating that gut dysbiosis—particularly the depletion of SCFA-producing bacteria—likely plays a critical role in CaOx kidney stone formation. Such dysbiosis disrupts essential metabolites production, impairs nutrient absorption, and potentially alters inflammatory responses, thereby fostering a metabolic environment conducive to stone formation. The significant shifts observed in both gut microbiota composition and host metabolomic profiles in our study emphasize the intricate relationship between microbial functions and metabolic pathways linked to urolithiasis. A notable finding in our study is the significant reduction in gut microbial richness in CSF, consistent with findings from previous reports [12, 53]. In particular, the depletion of key SCFA-producing bacteria, such as E. rectale and F. prausnitzii, is consistent with the observations reported by Denburg et al. [5]. The decline of these SCFA-producing bacteria has been implicated in the pathogenesis of several chronic diseases [54], and their reduction in CSF may contribute to the decreased SCFA levels observed in our study. SCFAs play critical roles in immunomodulation, inflammation reduction, and antioxidative protection [14, 55]. Their reduction may impair renal immune responses and disrupt the regulation of inflammation, potentially contributing to kidney stone formation. Urinary citrate, a well-known inhibitor of kidney stones, plays a critical role by binding calcium and preventing crystal aggregation [56]. Reduced urinary citrate levels are strongly associated with a higher risk of kidney stone formation [57]. Our results demonstrated that the abundance of SCFA-producing bacteria was significantly positively correlated with the higher plasma acetate and butyrate levels, which, in turn, showed a significant positive correlation with urinary citrate excretion. This observation suggests the existence of a potential gut microbiota-urinary citrate axis influencing stone pathogenesis, underscoring the gut-kidney axis as a critical factor in the pathophysiology of stone formation. Interactions among gut microbiota, such as acetate cross-feeding, are crucial for maintaining microbial homeostasis and host metabolic balance [58]. We found that fecal SCFAs explained significant variance in gut microbial composition and function, promoting a shift toward a healthier microbial state. Furthermore, SCFA-producing bacteria, particularly B. thetaiotaomicron, played a significant role in shaping microbial composition and function. These findings highlight the pivotal role of SCFAs and their producers in regulating gut homeostasis and maintaining systemic metabolic stability. Our metabolomic analyses revealed profound differences in the metabolic landscapes of CSF and healthy controls, with multiple disrupted pathways implicated in CaOx stone formation. Key metabolites of the tricarboxylic acid (TCA) cycle were notably downregulated across fecal, plasma, and urinary samples from CSF, consistent with previous studies [59]. Succinic acid, a critical intermediate of the TCA cycle, was inversely correlated with stone volume in our study, which may exert protective effects by modulating inflammatory responses, inhibiting cellular adhesion, and preventing osteogenic differentiation [60]. These metabolic disturbances suggest that impairments in the TCA cycle may contribute to early stone pathogenesis. Disruptions in tryptophan metabolism also emerged as a significant finding, as evidenced by the elevated kynurenine/tryptophan (Kyn/Trp) ratio in CSF. This alteration was associated with heightened inflammation, dyslipidemia, aggravated renal dysfunction, and reduced urinary citrate excretion. Interestingly, higher plasma SCFA levels were associated with a lower Kyn/Trp ratio, indicating that SCFAs may play critical roles in regulating tryptophan metabolism and mitigate inflammation. Dysregulation of tryptophan metabolism has been linked to various diseases, and its kynurenine pathway is significantly altered in inflammation-driven chronic conditions [42, 61]. Indoleamine 2,3-dioxygenase (IDO), a key rate-limiting enzyme, catalyzes the conversion of tryptophan to kynurenine, a process upregulated by pro-inflammatory cytokines [42, 62]. Inhibition of IDO has shown potential benefits, including improved insulin sensitivity, regulating lipid metabolism, and enhancing intestinal barrier function [63]. Thus, IDO may represent a potential therapeutic target for metabolic diseases and CaOx stone formation, warranting further investigation. The reduction of butyrate-producing bacteria is associated with metabolic disturbances that may act as upstream determinants of CaOx stone formation [5]. While plasma metabolite levels are influenced by genetics, diet, and gut microbiota, gut microbiota plays a dominant role in explaining inter-individual differences in plasma metabolomes [64]. In our study, fecal SCFA levels significantly influenced the global metabolome, accounting for notable variances in fecal, plasma, and urinary metabolomes. Furthermore, specific SCFA-producing bacterial species, such as B. thetaiotaomicron, played a pivotal role in these variations. This highlights the importance of SCFA-producing bacteria in shaping host metabolic profiles. By constructing a network linking"SCFA-producing bacteria → microbial functional modules → host metabolite modules → stone-related phenotypes,"we propose a potential mechanistic link through which gut dysbiosis may induce metabolic disruptions and contribute to CaOx stone formation. Of particular interest, B. thetaiotaomicron was enriched in CSF, and Mendelian randomization (MR) analysis suggested a potential causal association between this species and an increased risk of urolithiasis. While B. thetaiotaomicron is recognized for its role in modulating immune responses and maintaining gut integrity [65–67], its enrichment in CSF suggests it may exert diverse and complex effects in different biological contexts. This bacterium’s potential involvement in kidney stone pathogenesis warrants further investigation. Our animal experiments further supported the protective role of key SCFA-producing bacteria in CaOx stone formation. Treatment with E. rectale, F. prausnitzii, or sodium butyrate significantly reduced renal crystal deposition in rats with glycolate-induced hyperoxaluria, as evidenced by a significant reduction in crystal areas observed in histological analyses (p < 0.05). However, these bacterial interventions did not lead to statistically significant improvements across all renal function indicators, although a trend toward amelioration was observed. Several factors may have limited the observed therapeutic effects. First, the sample size in each treatment group was limited (n = 6), which may have reduced the statistical power to detect subtle yet biologically meaningful effects. Second, the intervention period of four weeks may not have been sufficient for these obligate anaerobes to achieve stable colonization or exert their full physiological impact. Third, the glycolate-induced hyperoxaluria model we employed results in pronounced renal injury and extensive CaOx crystal deposition, which could have overwhelmed the more subtle anti-inflammatory and anti-crystallization effects of the bacterial interventions. Notably, F. prausnitzii markedly reduced urinary oxalate excretion compared to the model group, while sodium butyrate reversed citrate depletion and alleviated kidney injury. These results suggest that while the bacterial treatments may offer renal protective benefits, optimization of dosage, treatment duration, or model severity may be required to achieve more pronounced therapeutic effects. Collectively, these findings underscore the therapeutic potential of SCFA-producing bacteria and their metabolites as promising candidates for preventing and managing CaOx kidney stones. F. prausnitzii and E. rectale are vital indicators of gut health, with F. prausnitzii particularly noted for maintaining gut barrier integrity and exhibiting strong anti-inflammatory and antioxidative properties [68–70], while E. rectale serves as a primary producer of butyrate and valerate [71]. Reduced levels of these SCFA-producing bacteria have been widely associated with chronic kidney disease, coronary heart disease, ulcerative colitis, Crohn’s disease, and diabetes [72–77]. In our study, the depletion of F. prausnitzii and E. rectale likely led to decreased SCFA production, thereby potentially weakening the anti-inflammatory effects critical for gut and kidney health. F. prausnitzii has been shown to alleviate renal dysfunction and reduce inflammation and fibrosis, as well as lower serum uremic toxins, primarily through butyrate production and activation of GPR43 receptors [72]. Similarly, E. rectale modulates immune responses by suppressing the TLR4/MyD88/NF-kB signaling pathway, thereby attenuating inflammation [71]. Given their crucial roles in maintaining host health, further studies are warranted to elucidate the precise mechanisms by which F. prausnitzii and E. rectale influence CaOx kidney stone formation. Despite these insights, our study has several limitations that warrant consideration. First, the use of relative quantification in metagenomics and metabolomics restricted our ability to assess absolute microbial or metabolite abundances. Second, the identification of differential metabolites partially relied on the supervised OPLS-DA model, which, while enhancing sensitivity, may be prone to overfitting and exaggerating group separation, especially when biological variation is subtle. To mitigate this potential bias, we incorporated unsupervised PCA and confirmed group-level differences using Adonis tests (p < 0.001). Nevertheless, the risk of overestimation cannot be entirely ruled out. Third, while our integrative multi-omics analysis revealed potentially meaningful associations between gut microbial features, metabolic perturbations, and clinical phenotypes, these findings remain correlative. Further mechanistic studies or causal inference strategies, such as intervention trials or integrated host-genome and microbiome analyses, are needed to validate their biological relevance. In addition, although we adjusted for key covariates such as age, sex, and BMI in most analyses, residual confounding from factors such as diet composition, antibiotic use, and physical activity cannot be excluded. Due to sample size constraints and collinearity among variables, these factors were not included in multivariate models. As these residual confounders may influence both gut microbiota composition and metabolite profiles, potentially biasing observed associations, future studies should employ larger cohorts and consider sensitivity analyses—such as excluding individuals with recent antibiotic exposure, probiotic intake, extreme dietary habits, significant weight changes, or chronic comorbidities—to improve the robustness of microbiome-metabolome associations and strengthen causal inferences. Regarding Mendelian randomization, the GWAS dataset of urolithiasis used in this study lacked stratification by stone type, potentially limiting the generalizability of our findings. Moreover, due to the small sample size of the gut microbiome GWAS dataset, we adopted a lenient p-value threshold of < 1.0 × 10−5 to select instrumental variables for the gut microbiota, which is higher than the conventional genome-wide significance level (p < 5 × 10−8), potentially affecting the reliability of the results. Finally, although all enrolled patients were confirmed to have CaOx stones via infrared spectroscopy, we did not stratify them by subtype (e.g., COM, COD, or mixed forms), despite the biological heterogeneity among these categories. Future studies incorporating stone subtype classification and stratified analyses may help elucidate subtype-specific mechanisms underlying nephrolithiasis formation. Conclusions By integrating multi-omics analyses, our study reveals significant dysbiosis in SCFA-producing bacteria and widespread metabolic perturbations in CSF. The depletion of key SCFA-producing species, accompanied by reduced SCFA levels, is strongly associated with host metabolic imbalances, potentially driving CaOx stone formation and progression. Moreover, our findings suggest that B. thetaiotaomicron may act as a potential microbial risk factor, while highlighting the therapeutic promise of F. prausnitzii, E. rectale, and sodium butyrate in mitigating renal CaOx deposition and injury. Collectively, these insights enhance our understanding of gut–kidney interactions in nephrolithiasis and provide a rationale for microbiota-targeted strategies in the prevention and personalized management of kidney stones. Supplementary Information Additional file 1.Additional file 2.Additional file 3. Additional file 1. Additional file 2. Additional file 3.