Abstract

   Bladder Pain Syndrome/Interstitial Cystitis encompasses Hunner lesion
   and non-Hunner lesion (NHIC) subtypes, characterized by chronic pelvic
   pain and urinary symptoms. The two subtypes show distinct clinical
   presentations, yet their underlying mechanisms remain poorly
   understood. Advancements in biological detection technologies allow for
   deeper insights into disease pathogenesis through integrated
   multi-omics approaches. This study aims to delineate the differences
   between NHIC and HIC using multi-omics analysis to uncover distinct
   microbiome, metabolome, and transcriptome profiles, thereby elucidating
   their unique pathogenic mechanisms. We conducted a comprehensive
   analysis involving bladder tissue samples and urine specimens from NHIC
   and HIC patients. Genomic, metabolomic, and transcriptomic data were
   obtained using high-throughput sequencing techniques. Differentially
   expressed genes, metabolites, and microbial communities were identified
   and subjected to pathway enrichment analysis to explore their
   associations with disease subtypes. Our findings revealed significant
   differences in the urinary microbiota, with NHIC showing a predominance
   of Lactobacillus and Enterococcus, while HIC exhibited higher levels of
   Pseudomonas and Gardnerella. Metabolomic analysis identified altered
   pathways, such as arginine and proline metabolism in NHIC and steroid
   hormone biosynthesis in HIC. Transcriptomic profiling highlighted
   upregulation of immune response genes in HIC, particularly those
   involved in mast cell degranulation and viral infection pathways. In
   contrast, NHIC was associated with increased expression of metabolic
   and energy-related pathways. This multi-omics analysis revealed
   distinct pathogenic signatures between NHIC and HIC, indicating that
   HIC may be driven by chronic immune dysregulation and past infections,
   while NHIC is more closely associated with metabolic disturbances.
   These insights suggest potential avenues for tailored therapeutic
   interventions based on the molecular characteristics of each subtype.
   Further research is warranted to confirm these findings and explore
   their clinical implications.

   Keywords: Bladder pain syndrome, Multi-omics analysis, Microbiome,
   Metabolomics, Transcriptomics, Disease mechanism

   Subject terms: Bladder, Immunology, Microbiology, Urology

Introduction

   Bladder Pain Syndrome/Interstitial Cystitis (BPS/IC), is a chronic
   inflammatory bladder condition characterized by pain or pelvic
   discomfort associated with a full bladder, along with lower urinary
   tract symptoms such as frequency or urgency^[32]1.

   In 1987, Fall et al. posited that Bladder Pain Syndrome/Interstitial
   Cystitis is divisible into Hunner lesion Interstitial Cystitis (HIC)
   and Non-Hunner lesion Interstitial Cystitis (NHIC), as they noted
   marked histological disparities between the two subtypes and advocated
   for their independent investigation^[33]2. For a majority of HIC
   patients, cauterization or excision of the Hunner ulcers yields a
   significant therapeutic benefit. Thus, patients afflicted with HIC who
   receive cystoscopy with hydrodistension exhibit superior symptom
   amelioration and protracted relief, as compared to those suffering from
   NHIC^[34]3. In view of the divergent clinical course of the two
   subtypes of BPS/IC discussed earlier, it is apparent that
   individualized treatment assumes primacy, underscoring the exigency to
   fathom the pathogenesis of both subtypes.

   In recent years, the relentless advancement of biological detection
   technologies, coupled with the continuous evolution of microbiome,
   transcriptomics, metabolomics, proteomics, and other research
   methodologies, have enabled experts to unravel the molecular
   underpinnings of numerous intractable diseases, providing a bedrock for
   the implementation of individualized clinical interventions.

   Isobel Walton et al. meticulously culled and statistically analyzed six
   studies that investigated urine microbiota profiles in healthy female
   controls and those afflicted with BPS/IC, yielded the following key
   insights: (1) Female BPS/IC patients displayed substantial
   inter-individual variations in urine microbiome profiles; (2)
   Currently, no persuasive evidence exists to support clinically
   significant differences in diversity or species composition between
   female BPS/IC and control groups; (3) Lactobacillus species emerged as
   the most prevalent dominant genus in both female BPS/IC and control
   samples; (4) Pain severity in female BPS/IC patients did not show any
   significant correlation with urine microbiota; (5) As of now, no
   promising candidate molecule or microbiome model has emerged for female
   BPS/IC^[35]4.

   In May 2021, J Curis Nickel spearheaded a pioneering investigation into
   the divergent distribution of urine microbiota in patients afflicted
   with HIC and NHIC, postulated that bacterial involvement in HIC
   pathology remained unsubstantiated, and there was no significant
   variation in species abundance and diversity between the two subtypes.
   Nevertheless, a few male samples exhibited significantly disparate
   abundances of Negativicoccus, Porphyromonas somalae, Mobiluncus
   curtisii, and Corynebacterium renale. Notably, the first two bacterial
   species are wound-associated anaerobic gram-negative bacteria, and the
   third is known to be associated with bacterial vaginosis in women. The
   study proposes that catheterized specimens may circumvent the
   limitations imposed by the interference of urethral fixed microbiota
   and potentially uncover subtle dissimilarities between the two
   subtypes^[36]5.

   A study of 42 women, comprising 20 patients with interstitial cystitis
   (IC) and 22 controls, implemented 16SrRNA sequencing and liquid
   chromatography coupled with mass spectrometry for urine microbiota and
   metabolite analysis. The findings indicated that the expression of four
   pathogenic bacterial species—Serratia, Brachybacter, Porphyromonas, and
   Citrate Acidobacteria—was significantly upregulated in the IC group.
   Additionally, alterations in IC metabolite characteristics appeared to
   be linked to sphingosine metabolism, amino acid metabolism, and fatty
   acid biosynthesis^[37]6.

   Japanese scholars have subsequently undertaken whole transcriptome
   sequencing on bladder biopsy specimens from non-Hunner type IC (NHIC)
   and Hunner type IC (HIC) patients. They discovered that HIC patients
   and NHIC patients displayed varying genomic and histological features,
   with a correlation to upregulated pathways that are involved in immune
   response and infection. Conversely, NHIC patients lacked these specific
   characteristics^[38]7.

   Despite the emergence of numerous studies on the mechanism of BPS/IC,
   there is still a deficiency in the exploration of multi-omics
   combinations such as microbiome, metabolome, transcriptome, and so
   forth. In our previous research on this topic, we compared the
   demographic and clinical characteristics of NHIC and HIC
   patients^[39]8. It can be observed that there were no statistically
   significant differences between the two groups in terms of age, disease
   duration, or voiding diaries, making it difficult to distinguish the
   two groups based solely on clinical symptoms. However, significant
   differences were observed in urodynamic and cystoscopic findings
   between the two groups. Therefore, this project aims to investigate the
   multi-omics distinctions of NHIC and HIC types, further explore the
   pathogenesis of the two types, and explore the correlation between
   clinical symptoms and molecular mechanisms of patients with different
   types. Ultimately, the goal is to provide a foundation for the creation
   of individualized treatment plans for BPS/IC patients, improve patient
   prognosis, and reduce patient burdens.

Results

Urine microbiome analysis

   An investigation was conducted involving 18 subjects with non-Hunner
   type interstitial cystitis (NHIC) and 12 subjects with Hunner type
   interstitial cystitis (HIC). Two NHIC and five HIC subjects provided
   both mid-morning urine samples and intraoperative catheterization
   samples, resulting in a total of 37 urine samples that met sequencing
   quality control standards. Subsequent analysis of the microbiome was
   performed on these samples.

Exploring α and β diversity analysis in the MID and CATH cohorts

   Initially, we performed Partial Least Squares Discriminant Analysis
   (PLS-DA) to classify the samples, designating the self-retained early
   morning mid-morning urine samples as “MID.” and the intraoperative
   catheterization samples as “CATH.” Fig. [40]1A visually reveals
   discernible distinctions between the two groups, but some samples still
   had close distances, with overlapping confidence ellipses, indicating
   no significant differences between the groups. Figure [41]1B–D
   illustrate the genus-level comparison between the two groups, examining
   the Chao1 index (P = 0.8983), the Shannon index (P = 0.3067), and the
   Simpson index (P = 0.4433).For β-diversity analysis, we employed
   principal coordinate analysis (PCoA) and non-metric multidimensional
   scaling analysis (NMDS) to investigate inter-sample dissimilarities. As
   depicted in Fig. [42]1E, [43]F, PCoA analysis did not yield a complete
   distinction between the two groups, with sample distances exhibiting no
   statistical significance (P > 0.05). Similarly, NMDS analysis
   positioned the CATH group within the inner circle and the MID group in
   the outer circle, again with no statistically significant difference
   (P > 0.05). This intriguingly suggests that the catheterization
   sampling method may mitigate the influence of certain miscellaneous
   bacteria.

Fig. 1.

   [44]Fig. 1
   [45]Open in a new tab

   Urine microbiome analysis (A) PLS-DA analysis at the genus level for
   the MID and CATH groups; (B–F) Analysis of α and β diversity in the MID
   and CATH groups. (B) Chao1 index; (C) Shannon index; (D) Simpson index;
   (E) PCoA analysis; (F) NMDS analysis. (G) Analysis of community
   composition in the MID and CATH groups.

   Despite observing some dissimilarities in α and β diversity at the
   genus level, the lack of statistical significance reaffirms the earlier
   findings from the PLS-DA analysis. Consequently, it appears that
   although some variations exist between the two sampling techniques, the
   overall species richness and evenness remain comparatively consistent.

Deciphering the community composition analysis of the MID and CATH cohort

   Figure [46]1G distinctly portrays the analogous community structure of
   the two groups, yet accentuates the considerable dissimilarity in the
   relative abundance of each bacterial genus. Noteworthy disparities
   manifest in the MID group, where the proportion of numerous bacterial
   genera notably surpasses that of the CATH group.

Unveiling the species composition analysis of the NHIC and HIC cohorts

   As depicted in Fig. [47]2A, the Venn map analysis of urine flora in the
   NHIC and HIC groups revealed that 852 species were shared at the genus
   level, with 303 species exclusive to the HIC group and 273 species
   unique to the NHIC group. Notably, the number of species observed in
   the NHIC group was slightly smaller than that in the HIC group.

Fig. 2.

   [48]Fig. 2
   [49]Open in a new tab

   Comprehensive Insights into Species Composition, α Diversity, and β
   Diversity in NHIC and HIC Cohorts (A) Venn diagram analysis of
   bacterial species in the NHIC and HIC microbial communities; (B) PLS-DA
   analysis at the genus level for the NHIC and HIC groups; (C)
   Distribution of dominant species in the NHIC and HIC microbial
   communities using a Circos plot; (D, E, F, G and H) Analysis of α and β
   diversity at the genus level in the NHIC and HIC groups;(D) Chao1
   index; (E) Shannon index; (F) Simpson index; (G) PCoA analysis; (H)
   NMDS analysis.) (I) showcases the LEfSe analysis for discriminating
   species differences between the NHIC and HIC groups.

   Upon juxtaposing the results in Fig. S1A and Fig. S1B, a conspicuous
   discrepancy arises, wherein the NHIC cohort exhibits a pronounced
   prevalence over the HIC cohort in the proportions of Lactobacillus,
   Klebsiella, Enterococcus, and Prevosia. Conversely, in comparison to
   the NHIC group, the HIC group demonstrates a statistically significant
   advantage in the prevalence of Pseudomonas, Gardnerella, and Chia.

   Figure [50]2C portrays a Circos diagram that intricately illustrates
   the relationship between samples and species. The small semicircle on
   the left depicts the species composition within each sample, with the
   outer ribbon color-coded to represent the corresponding group and the
   inner ribbon reflecting the species, with the length denoting its
   relative abundance. Conversely, the large semicircle on the right
   showcases the distribution of species among different samples at this
   taxonomic level. Here, the outer ribbon represents the species, the
   inner ribbon is color-coded to signify different groups, and the length
   reflects the proportion of each species in a sample. The discernible
   divergence in dominant urine microbiota between the two groups becomes
   readily apparent.

Exploring α and β diversity analysis in the NHIC and HIC cohorts

   An initial PLS-DA analysis was conducted to assess sample grouping, as
   presented in Fig. [51]2B. While some differences were observed between
   the two groups, certain samples remained in close proximity, with
   overlapping regions of the confidence ellipse indicating a lack of
   significant difference.

   The assessment of α diversity was conducted utilizing the Chao1 index,
   the Shannon index, and the Simpson index. As depicted in Fig. [52]2D, E
   and F, the box plot demonstrates that the NHIC group exhibited greater
   species richness and uniformity compared to the HIC group. However, at
   the genus level, there were no substantial variations between the two
   groups, as evidenced by the Chao1 index (P = 0.4779), Shannon index
   (P = 0.7), and Simpson index (P = 0.8552).

   The PCoA and NMDS analyses are presented in Fig. [53]2G, H,
   respectively. The confidence ellipse of the PCoA analysis between the
   two groups exhibited a cross-overlapping segment with a P value greater
   than 0.05, although some samples showed clear differences. Similarly,
   the confidence ellipses of the two groups in the NMDS analysis
   partially overlapped, with a P value greater than 0.05 and a stress
   value lower than 0.2, indicating that the grouping model has some
   significance. However, there were no significant differences in β
   diversity between the two groups at the genus level.

   Despite observing some variations in the α and β diversity at the genus
   level between the NHIC and HIC groups, the dissimilarities were not
   statistically significant, which was in line with the outcomes of the
   PLS-DA analysis. These findings indicate that the microbiota analysis
   cannot completely distinguish between NHIC and HIC subtypes, as both
   subtypes represent distinct presentations of BPS/IC.

Unraveling species disparities between NHIC and HIC cohorts

   The LDA discriminant histogram quantified the microbial taxa in both
   groups, utilizing LDA analysis (a form of linear regression analysis)
   to obtain LDA scores. Enrichment was considered significant if the taxa
   exhibited P values less than 0.05 and LDA scores greater than or equal
   to 2.

   Figure [54]2I illustrates the results of LEfSe analysis, which revealed
   significant differences in microbial taxa between the NHIC and HIC
   groups. Specifically, the Proteus genus was found to be significantly
   enriched in the HIC group, while the Digestive Coccus genus was
   significantly enriched in the NHIC group. Moreover, many common
   environmental bacteria, such as those found in sludge and flowers, were
   significantly enriched in the NHIC group.

Urine metabolomics analysis

   The metabolomics study included urine samples from 17 NHIC patients and
   12 HIC patients that met the quality control criteria. To eliminate the
   impact of urinary retention mode, all patients’ urine samples were
   obtained through catheterization and retained.

OPLS-DA analysis of NHIC and HIC groups

   OPLS-DA offers a means to assess model classification efficacy by
   evaluating R2X, R2Y, Q2, and OPLS-DA score maps. For this analysis, the
   R ropls package was employed, and the VIP (Variable Importance for the
   Projection) value served as a metric to gauge the significance of a
   variable (i.e., characteristic peak) in explaining both the X dataset
   and its associated Y dataset. Variables with a VIP value greater than 1
   were considered important and utilized as a screening criterion for
   potential biomarkers.

   In Fig. [55]3A, the NHIC and HIC samples are clearly differentiated in
   the positive ion mode, forming distinct clusters represented by blue
   and red circles, respectively. Each individual sample is depicted as a
   scattered small square. Notably, the ellipses of the two groups exhibit
   discernible differences, with the NHIC group appearing as a
   predominantly horizontal ellipse, and the HIC group appearing as a
   predominantly longitudinal ellipse with partial overlap in the middle.
   The R2X value is 0.408cum, while the R2Y value is 0.589cum, with a Q2
   value of − 0.416cum, indicating the model’s performance. Furthermore,
   the PR2Y value is 0.705, and PQ2 is 0.88.

Fig. 3.

   [56]Fig. 3
   [57]Open in a new tab

   Integrated Metabolomic Profiling and Pathway Enrichment Analysis
   Reveals Contrasts in NHIC and HIC Groups. (A) displays the score plot
   of OPLS-DA analysis in positive ion mode for the NHIC and HIC
   metabolomic profiles; (B) presents the score plot of OPLS-DA analysis
   in negative ion mode for the NHIC and HIC metabolomic profiles; (C) VIP
   (Variable Importance in the Projection) plot for the positive ion mode
   analysis of the NHIC and HIC metabolomic profiles; (D) VIP (Variable
   Importance in the Projection) plot for the negative ion mode analysis
   of the NHIC and HIC metabolomic profiles; (F) and (E) Bubble plot of
   significantly enriched KEGG pathways for differential metabolites in
   the NHIC and HIC groups.(F) Positive ion mode; (E) Negative ion mode).

   As demonstrated in Fig. [58]3B, in the negative ion mode, the NHIC and
   HIC groups form two longitudinal ellipses with only a small overlap.
   The R2X value is 0.3cum, the R2Y value is 0.747cum, and the Q2 value is
   − 0.318cum, further assessing the model’s performance. Additionally,
   the PR2Y value is 0.27, and PQ2 is 0.7.

Unveiling metabolic disparities: differential analysis of NHIC and HIC
metabolomes

   In the positive ion mode, a comprehensive differential analysis heatmap
   of urine metabolites between the NHIC and HIC groups is depicted in
   Fig. S1C. Among the extensive array of metabolites analyzed, 17
   displayed statistically significant differences between the two groups.
   Remarkably, 16 of these metabolites exhibited down-regulated expression
   in the HIC group relative to the NHIC group. Singularly, one
   short-chain keto acid derivative displayed up-regulation in the HIC
   group. Figure [59]3C depicts the differential analysis VIP diagram,
   offering insight into the specific contribution of each metabolite. The
   abscissa represents the VIP value, while the ordinate displays the top
   15 differential metabolites. We derived the mean value of metabolite
   abundance for each sample in each group and conducted a z-score
   analysis accordingly. The color on the right indicates the level of
   metabolite abundance in different groups, where red denotes
   up-regulation, and blue represents down-regulation. Notably,
   metabolites with VIP > 1 were deemed to exhibit significant
   differences.

   In the negative ion mode, Fig. S1D showcases the differential
   metabolites, encompassing a total of 16 metabolites that exhibited
   remarkable differences between the NHIC and HIC groups. The VIP diagram
   in the negative ion mode is portrayed in Fig. [60]3D, shedding light on
   the specific contributions of the discernible metabolites.

DEMs enrichment analysis: KEGG analysis of NHIC and HIC metabolomes

   In the positive ion mode, KEGG enrichment analysis was carried out on
   the differential metabolites between the NHIC and HIC groups. Notably,
   six metabolic pathways exhibited statistical significance (P < 0.05)
   between the NHIC and HIC groups, namely: arginine and proline
   metabolism, glycine, serine, and threonine metabolism, neuroactive
   ligand receptor interactions, nicotine addiction, actin cytoskeleton
   regulation, and steroid hormone biosynthesis, as depicted in
   Fig. [61]3E. These findings suggest the potential involvement of these
   metabolic pathways in the development and progression of hypertension
   and signify promising targets for future treatment and prevention
   strategies.

   Similarly, in the negative ion mode, KEGG enrichment analysis was
   performed on the differential metabolites between the NHIC and HIC
   groups. Notably, primary bile acid biosynthesis, bile secretion,
   synaptic vesicle cycle, and vitamin B1 metabolism exhibited the most
   significant pathways with statistical differences (P < 0.05), as shown
   in Fig. [62]3F.

Transcriptomic analysis of bladder tissue

   In this study, 28 HIC specimens, and 20 NHIC specimens were included in
   the analysis.

Principal component analysis (PCA) of transcriptomics data

   To explore the global transcriptomic differences between the NHIC and
   HIC groups, we performed PCA on the transcriptomics dataset
   (Fig. [63]4A). The PCA plot reveals a clear separation between the two
   groups along the first principal component (PC1), which accounts for
   75.4% of the total variance. The second principal component (PC2)
   explains an additional 7.6% of the variance. Samples from the NHIC
   group (orange) and the HIC group (blue) form distinct clusters,
   indicating significant transcriptional differences between the groups.
   Additionally, the tight clustering within each group suggests good
   reproducibility and minimal intra-group variation.

Fig. 4.

   [64]Fig. 4
   [65]Open in a new tab

   Transcriptomic analysis of bladder tissue (A) PCA plot showing distinct
   clustering between the two groups, indicating transcriptional
   differences and good reproducibility within each group. (B) Volcano
   plot of differential gene expression analysis between the NHIC and HIC
   transcriptomic profiles; (C) KEGG enrichment analysis of the NHIC and
   HIC transcriptomic profiles. (D) GO (Gene Ontology) enrichment analysis
   of the NHIC and HIC transcriptomic profiles;

Decoding transcriptional disparities: differential analysis of NHIC and HIC
bladder tissue transcriptomes

   Transcriptome sequencing was executed on 28 HIC specimens and 20 NHIC
   specimens, followed by the generation of a volcano map analysis based
   on significantly differentially expressed genes from the two groups. In
   earlier studies and our previous basic studies, some BPS/IC-related
   potential pathogenic genes were identified. We examined the expression
   of these key focus genes and found that NGFR, BDNF, IL6R, IL1B, TGFBR1,
   ADCYAP1, CCL2, and CCL3 genes were significantly upregulated in the HIC
   group, whereas IL17D was downregulated significantly in the HIC group
   (Fig. [66]4B).

DEGs enrichment analysis: KEGG Analysis of NHIC and HIC bladder tissue
transcriptomes

   Figure [67]4C presents the enriched KEGG pathways in a visually
   informative manner. Notably, the figure illustrates 25 significantly
   enriched KEGG pathways, shedding light on the molecular processes
   underlying NHIC and HIC.

   The results of this analysis demonstrate a close relationship between
   these significantly enriched pathways and various aspects of bacterial
   and viral infections, immune responses, and autoimmune diseases. These
   findings align with the results from the GO enrichment analysis and
   further emphasize the crucial role of infection and immunity in the
   pathogenesis of both NHIC and HIC groups.

4DEGs enrichment analysis: GO analysis of NHIC and HIC bladder tissue
transcriptomes

   In our GO enrichment analysis (Fig. [68]4D), we made significant and
   noteworthy discoveries regarding the differentially expressed genes
   between the NHIC and HIC groups. Notably, we found a substantial
   enrichment of 2449 entries (P < 0.05) in the GO biological process
   category, encompassing numerous terms primarily associated with
   immunity. These terms included immune response, immune system
   processes, adaptive immune response, regulation of immune system
   processes, cell activation, leukocyte activation, immune response
   regulation, response to stimuli, inflammatory response, and others,
   highlighting the pivotal role of immune-related processes in the
   pathogenesis of both groups.

   Furthermore, the GO molecular function analysis revealed the
   significant enrichment of 356 entries (P < 0.05), which were closely
   associated with receptors, ligands, cytokines, and various other
   molecules involved in immune activity, cell signal transduction, and
   virus-host interactions. Terms such as receptor binding, antigen
   binding, cytokine binding, G protein-coupled receptor binding, and more
   were among the enriched molecular functions.

   Additionally, the GO cell components analysis yielded a remarkable
   enrichment of 189 entries (P < 0.05), mainly focusing on cellular
   membranes, extracellular regions, immunoglobulin complexes, vesicles,
   and secretory granules. These cellular components play essential roles
   in immune response, cytokine secretion, and intercellular signal
   transduction processes.

   The findings from the GO enrichment analysis strongly suggest that the
   differentially expressed genes between the NHIC and HIC groups are
   critical players in immune response and regulation. These genes likely
   play pivotal roles in the underlying mechanisms contributing to the
   pathogenesis of both groups, shedding light on the intricate and
   complex nature of bladder disorders like BPS/IC. The insights gained
   from this analysis may pave the way for further investigations and
   potential therapeutic strategies targeting the immune-related pathways
   implicated in these conditions.

GSEA analysis of NHIC and HIC bladder tissue

   Our Gene Set Enrichment Analysis (GSEA) of KEGG Pathways yielded
   insightful results, with a total of 326 pathways identified, out of
   which 177 exhibited significant enrichment. Among these pathways, our
   focus was primarily on those associated with infectious immunity,
   considering their potential relevance to the pathogenesis of the
   conditions.

   Notably, in the HIC group, several pathways related to viral infections
   were significantly upregulated, including Epstein-Barr virus infection
   and human cytomegalovirus infection. Additionally, pathways linked to
   bacterial infections such as pathogenic Escherichia coli infection and
   Staphylococcus aureus infection showed significant upregulation.
   Furthermore, key immune response-related pathways, such as
   cytokine-cytokine receptor interactions, antigen processing and
   presentation, natural killer cell-mediated cytotoxicity, T cell
   receptor signaling pathway, and B-cell receptor signaling pathway, were
   also found to be upregulated in the HIC group.

   In contrast, several pathways related to metabolism exhibited
   significant downregulation in the HIC group. These pathways included
   cytochrome P450 metabolism of heterogeneous substances, oxidative
   phosphorylation, retinol metabolism, butyrate metabolism, glycine,
   serine, and threonine metabolism, steroid hormone biosynthesis, valine,
   leucine, and isoleucine degradation, pentose and glucuronic acid
   interconversion, glutathione metabolism, PPAR signaling pathway, fatty
   acid metabolism, phenylalanine metabolism, peroxisome, tyrosine
   metabolism, fatty acid degradation, glyoxylic acid and dicarboxylic
   acid metabolism, carbon metabolism, porphyrin metabolism, propionate
   metabolism, arachidonic acid metabolism, cysteine and methionine
   metabolism, fat digestion and absorption, bile secretion, and more.

   To better understand the pathways relevant to the pathogenesis, we
   visualized the significant pathways in Fig. [69]5, providing a
   comprehensive overview of the molecular processes that may contribute
   to the development and progression of HIC.

Fig. 5.

   [70]Fig. 5
   [71]Open in a new tab

   GSEA (Gene Set Enrichment Analysis) analysis of the NHIC and HIC
   transcriptomic profiles.

Multi-omics integration analysis

Integration of microbiome and metabolomics: a comprehensive analysis

   In our study, we integrated metabolomics analysis and microbiota
   analysis by using the metabolite expression levels with significant
   differences between the NHIC and HIC groups as environmental factor
   data, and correlating them with the corresponding patient microbiota.
   This allowed us to construct a heat map illustrating the correlation
   between microbiota and metabolomics data. Specifically, Fig. [72]6A
   displays the correlation heat map for differential metabolites in
   positive ion mode, along with the corresponding patient microbiota,
   while Fig. [73]6C presents the correlation heat map for differential
   metabolites in negative ion mode, after matching the corresponding
   patient flora.

Fig. 6.

   [74]Fig. 6
   [75]Open in a new tab

   Multi-omics Integration Analysis (A) Heatmap depicting the correlation
   between differential metabolites in the positive ion mode and microbial
   community in the NHIC and HIC groups; (B) Venn diagram depicting the
   overlapping KEGG enrichment pathways between the metabolomic and
   transcriptomic datasets; (C) Heatmap illustrating the correlation
   between differential metabolites in the negative ion mode and the
   microbial community in the NHIC and HIC groups.

Integration of metabolomics and transcriptomics: unraveling molecular
insights

   Figure [76]6B portrays an analysis of the distinct and overlapping
   portions of the data between each group. In total, there are 19
   pathways, with KEGG-enriched metabolic pathways in negative ion mode
   intersecting with those enriched by transcriptome KEGG, summing up to
   23 pathways. Additionally, after intersecting the metabolic pathways
   and transcriptomes in positive and negative ion modes simultaneously,
   six pathways were identified, namely steroid hormone biosynthesis,
   glycine, serine, and threonine metabolism, neuroactive ligand receptor
   interaction, synaptic vesicle cycle, bile secretion, and metabolic
   pathway. Further, the positive ion mode and transcriptome intersection
   after enrichment revealed 13 unique metabolic pathways, including actin
   cytoskeleton regulation, arginine and proline metabolism, nicotine
   addiction, pancreatic secretion, cholinergic synapse, gastric acid
   secretion, sphingomyelin metabolism, insulin secretion, saliva
   secretion, cAMP signaling pathway, taste conduction, tryptophan
   metabolism, and glycerophospholipid metabolism. Likewise, the negative
   ion mode and transcriptome intersection and enrichment after enrichment
   yielded 17 unique metabolic pathways, namely thiamine metabolism,
   porphyrin metabolism, fatty acid biosynthesis, pantothenate and CoA
   biosynthesis, mineral absorption, glutathione metabolism, glyoxylic
   acid, and dicarboxylic acid metabolism, β-alanine metabolism, aminoacyl
   tRNA biosynthesis, lysine degradation, carbon metabolism, protein
   digestion and absorption, central carbon metabolism of cancer,
   pyrimidine metabolism, purine metabolism, amino acid biosynthesis, ABC
   transport, and other metabolic pathways.

Discussion

   Previous research has documented that non-pregnant, healthy female
   vaginal microbiota are dominated by certain bacterial genera such as
   lactobacillus, actinomycetes, Prevosia, Weroncoccus, Streptococcus,
   Proteus, Bacteroides, and Burkholderia^[77]9. Among these,
   lactobacillus is the most abundant genus, thriving in anaerobic
   conditions within the vaginal environment. It is capable of producing
   lactic acid, hydrogen peroxide, bacteriocins, and other antimicrobial
   agents, which help to prevent the invasion of foreign pathogens and to
   maintain a healthy microbiome within the vagina^[78]9. In contrast,
   Gardnerella is the most common genus associated with bacterial
   vaginosis in women, and Trichomonas vaginalis, a species of
   Gardnerella, has virulence factors that enable it to adhere to
   epithelial cells, forming biofilm communities that give it a
   competitive advantage over Lactobacillus^[79]10. Klebsiella, a pathogen
   found commonly in healthy individuals within the nose, throat, skin,
   and intestines, can cause a range of infections such as pneumonia,
   surgical wound infections, urinary tract infections, and sepsis^[80]11.
   Prevotella is a common dominant bacterium found in the oral cavity and
   is linked to various diseases such as inflammatory autoimmune
   disorders, opportunistic infections, bacterial vaginosis, and oral
   biofilm formation. As a common intestinal flora, it is also involved in
   the regulation of the intestinal inflammatory response^[81]12. Cardos
   allogeneis is a rare cause of urinary tract infection, and its
   association with the disease is unclear^[82]13. The Staphylococcus
   genus is found on human skin and mucous membranes, with Staphylococcus
   aureus and Staphylococcus epidermidis being common pathogens^[83]14.
   Vengoldia macrofungalis is a conditionally pathogenic bacterium
   associated with nosocomial device infections^[84]15. Streptococcus
   species are found as physiological flora in the human mouth and
   intestines, and some of these streptococci can cause opportunistic
   infections when human immunity is compromised^[85]16. Enterococci are
   common microbiota in the human gut microbiota and are also common
   pathogens causing urinary tract infections, soft tissue infections, and
   medical device infections^[86]17. The Paracoccus genus is taxonomically
   diverse, with more than 80 newly isolated species from various
   primitive and contaminated environments. These spherical gram-negative
   bacteria have diverse metabolic properties, which are useful for
   bioplastic synthesis and biodegradation^[87]18. Bacterium aquaticus is
   commonly found in soil and freshwater environments^[88]19,[89]20.

   Our analysis of the microflora present in catheterized urine samples
   and self-retained mid-morning urine samples revealed no significant
   differences in the α diversity and β diversity indices, indicating
   similar richness and evenness of bacterial species in both groups.
   Furthermore, our comparison of the NMDS analysis between the two groups
   indicated that the MID group encompassed the CATH group, suggesting
   that the midstream urine collected by the patient contained a broader
   range of bacterial species and also implying that catheterization could
   exclude some bacterial species from the sample.

   We further analyzed the dominant microbial communities in both groups
   and found that the MID group had a predominance of beneficial vaginal
   resident bacteria such as Lactobacillus, as well as common gut bacteria
   like Prevotella, Streptococcus, and Proteus. Additionally, there were
   various pathogenic bacteria present, including Staphylococcus,
   Klebsiella, and Cardos allogeneis, which are commonly associated with
   infections. In the MID group, there were also common environmental
   bacteria such as Streptococcus and Pseudomonas, which were typically
   found in water and soil. This suggested that midstream urine samples
   collected by patients were more susceptible to contamination from
   vaginal microbiota, gut microbiota, and environmental microbiota. The
   CATH group, on the other hand, had a higher proportion of Pseudomonas,
   which was a common pathogenic bacterium in hospital-acquired
   infections. This highlighted the importance of strict aseptic
   techniques during catheterization to prevent sample contamination from
   these bacteria.

   NHIC and HIC, being two subtypes of a disease, share a plethora of
   similarities. In terms of urine microbiota analysis, although there
   exist minor distinctions in the α and β diversity, they are not
   statistically significant, thereby rendering urine flora alone
   insufficient to distinguish between the two groups. A prior research
   study, encompassing 21 women with HIC and 22 women with NHIC, also
   attests to the lack of marked differences in species abundance between
   HIC and NHIC patients. This observation displayed the current dearth of
   unique pathogenic microbiota, enabling to differentiate NHIC from
   HIC^[90]5.

   Our investigation has also established that while there is no
   significant variation in species abundance and evenness in the urine
   microbiota between NHIC and HIC patients, their dominant flora types
   differ considerably. The NHIC group presented with a clear
   preponderance of Lactobacillus, Klebsiella, Enterococcus,
   Staphylococcus, and Prevosia, whereas the HIC group had a marked
   predominance of Pseudomonas and Gardnerella. Using the Lefse
   multi-level species discrimination analysis, we discovered that the
   Proteus genus was significantly enriched in the HIC group, while the
   Digestive Coccus genus was significantly enriched in the NHIC group,
   thus displaying significant distinctions between the two groups.
   Proteus mirabilis and Proteus vulgaris of the Proteus genus are common
   opportunistic pathogenic bacteria that induce urinary tract infections
   by utilizing urease to decompose urea nitrogen, increase urine pH, and
   by invading and colonizing the host urinary tract through various
   virulence factors such as urease, stone formation, fimbriae and other
   adhesions, iron and zinc acquisition, protease and toxin, biofilm
   formation, and pathological mechanism regulation. Literature has also
   demonstrated a correlation between the increase of Proteus and
   inflammatory bowel disease. Digestive coccus is a normal flora present
   widely in the mouth, intestines, female reproductive tract, and skin,
   and is a gram-positive obligate anaerobic bacterium that can result in
   bacteremia and various purulent infections in the human body under
   pathogenic circumstances. The dissimilarities in dominant flora
   composition and the differences in enriched flora between the two
   groups imply dissimilarities in pathogenesis.

   In our transcriptome differential gene analysis, we observed a
   significant upregulation of IL1β, IL6R, CCL3, TGFβR1, NGFR, and other
   genes in the HIC group, which are involved in the signal transduction
   mechanism of mast cell degranulation^[91]21. These findings led us to
   postulate that the inflammatory response associated with the HIC group
   may be closely linked to mast cells. Moreover, we found that pathways
   related to pathogen infections such as cytomegalovirus and Epstein-Barr
   virus, as well as bacterial infections including Escherichia coli and
   Staphylococcus aureus, were significantly upregulated in the HIC group.
   Similarly, pathways related to rheumatoid immune diseases like
   rheumatoid arthritis, systemic lupus erythematosus, autoimmune thyroid
   disease, and inflammatory bowel disease were also upregulated in this
   group, suggesting that HIC patients may have a history of recurrent
   bacterial and viral infections that caused abnormal autoimmune
   mechanisms, resulting in the development and progression of the
   disease. Previous literature has also reported multiple/frequent
   rheumatic immune diseases such as systemic lupus erythematosus,
   rheumatoid arthritis, Sjogren’s syndrome, and inflammatory bowel
   disease^[92]22–[93]25, indicating that BPS/IC may be immune-related
   inflammation caused by autoimmune dysfunction, and this autoimmune
   mechanism link is more prominent in patients with HIC in our study.
   Conversely, in the NHIC group, we observed significant upregulation of
   various metabolic pathways such as amino acids, fatty acids, carbon
   metabolism, and oxidative phosphorylation, which suggests that NHIC
   patients may have nutritional metabolism disorders that contribute to
   the development and progression of the disease.

   Previous studies investigating BPS/IC, when combined with normal
   control group urine flora and metabolome analysis, have found that the
   expression of Serratia, Brachybacter, Porphyromonas, and
   Citherobacterium was upregulated in the BPS/IC group. The metabolic
   pathways associated with sphingosine metabolism, fatty acid
   biosynthesis, and amino acid metabolism in the BPS/IC group were also
   significantly different from those in the control group. These studies
   have suggested that an increase in plant sphingosine and S1p may be
   related to BPS/IC symptoms mediated by bladder mast cell
   activation^[94]6. However, to date, no studies have been found that
   investigate the microbiome combined with metabolomic analysis
   associated with the two subtypes of HIC.

   These findings suggest that there may be a relationship between the
   altered microbiome and metabolic pathways in HIC and NHIC. The negative
   correlations between specific bacteria and metabolites may indicate
   potential mechanisms underlying the pathogenesis of the disease. For
   example, the negative correlation between Prevobacterium and histidine
   suggests that the metabolism of this amino acid may be affected in NHIC
   patients. Similarly, the negative correlation between Pseudomonas and
   uracil and 5′-acetyladenosine suggests that this bacterium may affect
   nucleotide metabolism in HIC patients. These findings provide a
   starting point for further research into the metabolic pathways
   involved in the pathogenesis of HIC and NHIC.

   The downregulation of actin cytoskeleton-related pathways in the HIC
   group may indeed contribute to bladder contractile dysfunction in
   patients with ICH. Actin cytoskeleton plays an essential role in
   regulating the contraction and relaxation of smooth muscle cells in the
   bladder wall. The contraction of smooth muscle cells is primarily
   driven by the interaction between actin and myosin filaments. The actin
   cytoskeleton provides a structural framework for the formation of
   stress fibers, which generate tension to drive bladder contraction. The
   downregulation of actin cytoskeleton-related pathways may result in the
   destabilization of stress fibers and impair the contraction of smooth
   muscle cells, leading to bladder dysfunction. Additionally, the
   downregulation of insulin secretion and pyrimidine and purine
   metabolism pathways in the HIC group may also contribute to bladder
   dysfunction, as these pathways are involved in regulating smooth muscle
   cell contraction and relaxation. Insulin regulates bladder smooth
   muscle cell contractility by modulating the activity of potassium
   channels, which are crucial for maintaining the membrane potential of
   smooth muscle cells. Pyrimidine and purine metabolism pathways are also
   involved in regulating smooth muscle cell contraction by modulating the
   availability of intracellular calcium, which is a key mediator of
   smooth muscle cell contraction.

   We undertook a comprehensive review of domestic and foreign literature
   on differential metabolites and enriched metabolic pathways to further
   elucidate the intricate interplay between the functions of differential
   microbiota, metabolites, and metabolic pathways enriched by
   differential genes and diseases. Histidine, an essential amino acid, is
   synthesized via the action of histamine decarboxylase. Mast cells store
   histamine in their granules and release histamine primarily through
   IgE-dependent degranulation.

   Histamine mainly induces allergic reactions and inflammation via H1 and
   H4 receptors and participates in and regulates the immune response
   primarily through H1, H2, and H4 receptors^[95]26. Previous studies
   have confirmed that mast cell degranulation releases histamine in the
   inflammatory response of interstitial cystitis^[96]27. Through
   multi-omics joint analysis, we can infer that changes in histidine
   levels in NHIC patient urine can regulate mast cell degranulation and
   histamine release, potentially contributing to the different
   pathogenesis of NHIC and HIC phenotypes.

   Benedikte Richter^[97]28 found that YKL-40 and mast cells are
   associated with detrusor fibrosis and can be used to assess bladder
   fibrosis. HIC patients exhibit lower bladder capacity and more severe
   bladder fibrosis, indirectly suggesting that mast cells may be involved
   in the progression of HIC. This is consistent with our research
   findings.

   However, in recent years, there has been an increasing focus on
   immune-related hypotheses, suggesting that mast cells may not play a
   central role in the pathogenesis of HIC. Yoshiyuki Akiyama and
   colleagues^[98]29 conducted transcriptomic analyses of Hunner lesions
   and non-lesion areas, finding no significant differences in
   lymphoplasmacytic and mast cell infiltration or the CD4/CD8 T
   lymphocyte ratio. Robert M. Moldwin and colleagues^[99]30 confirmed
   through immunohistochemistry that the activation of B cells and plasma
   cells is associated with the disease state of HIC. However, their study
   design could not determine whether the presence of these cells is a
   cause or a result of the disease. Kwang Jin Ko and colleagues^[100]31
   proposed that, compared to non-Hunner lesions, Hunner lesions exhibit
   reduced expression of anti-inflammatory and immunosuppressive factors,
   along with changes in immune cell populations. They suggested that
   macrophage polarization might be involved in the progression from
   non-Hunner lesions to Hunner lesions.

   Based on a comprehensive review of various studies, the mechanisms
   underlying HIC remain complex and diverse. Mast cells and immune cells
   may play distinct roles at different stages of the disease or in
   different patient subgroups. Furthermore, the pathological
   characteristics of HIC may vary among individuals, which could also
   explain the discrepancies in research findings.

   β-Hydroxy-β-methylbutyrate, a metabolite of leucine in essential amino
   acids, can stimulate protein synthesis by upregulating the mTOR pathway
   and anabolic signaling pathways such as growth hormone. This compound
   can also reduce proteolysis by downregulating catabolic signaling
   pathways such as ubiquitin proteasome and autophagic lysosomal system.
   This has significant implications for muscle damage repair, as
   β-hydroxy-β-methylbutyrate has been shown to be effective in treating
   various cachexia forms of muscle wasting in in vitro
   experiments^[101]32. NHIC patients had a higher abundance of
   lactobacilli than HIC patients, and lactobacillus was significantly
   negatively correlated with β-hydroxy-β-methylbutyrate. This observation
   suggests that protein synthesis was reduced, and the muscle injury
   repair pathway was not active in NHIC patients.

   Uracil is a base specific to RNA. Recently, it has been found that the
   misbinding of cytosine deamination or DNA replication deoxyuridine
   5′-triphosphate nucleotide (dUTP) leads to uracil in DNA. This compound
   has a barrier against infectious pathogens and is both a factor in
   antibody diversification in adaptive immunity and an effective
   antiviral drug^[102]33. Uridine, an endogenous small molecule
   metabolite composed of uracil and ribose, is involved in the synthesis
   of biofilms and RNA, which is of great significance for cell growth and
   function. In the environment of ischemia and hypoxia, uridine can
   increase lipid and glucose metabolism and prevent cell
   necrosis^[103]34. Our study revealed that Pseudomonas in HIC patients
   were the dominant bacteria and were significantly negatively correlated
   with uracil. We speculate that Pseudomonas may be involved in
   regulating pyrimidine metabolism, thereby mediating the necrosis of
   bladder urothelial cells and the occurrence of bladder wall ulcers.

   Our study has certain limitations. Due to the challenges in sample
   collection, the final sample size included in the study was relatively
   small. In future research, we plan to improve the collection process
   and expand the sample size for further analysis. Additionally, this
   study only included a disease group and lacked a healthy control group.
   For urine collection, we used catheterized urine sampling, which is
   more precise for microbiome and metabolomics analysis compared to
   midstream morning urine. However, transcriptomic analysis also required
   bladder biopsy tissues, which are invasive procedures and not feasible
   for healthy individuals. As a result, we did not include a healthy
   control group in this study, making it impossible to determine whether
   the observed differences are disease-specific. In future studies, our
   research team will continue to refine the methodology and further
   explore the differences between disease and healthy control groups.

Conclusion

   The pathogenesis of NHIC and HIC is still enigmatic, and further
   fundamental experiments are necessary to confirm the findings. However,
   the information gleaned from bioinformatics analysis provides a crucial
   stepping stone towards comprehending the underlying mechanisms of the
   two phenotypes.

   Through the application of multi-omics analysis, we have established
   that NHIC and HIC exhibit notable differences in their microbiome,
   metabolomics, and transcriptomics, thereby indicating that the
   pathogenesis of these two types is distinct. Specifically, our
   examination revealed significant variances in the Proteus and peptic
   genera in the NHIC group, although further inquiry is warranted to
   ascertain whether these genera can be utilized as markers for
   distinguishing between the two types. Our transcriptome difference
   analysis unveiled that the upregulation of inflammation-related genes
   in the HIC group is closely linked to the mast cell degranulation
   transduction mechanism. Additionally, bacterial viral infection and
   rheumatic immune disease-related pathways were considerably upregulated
   in the HIC group, thereby signifying that individuals with HIC may have
   suffered from repeated bacterial viral infections in the past, causing
   an aberration in their autoimmune mechanism, which led to the
   development and manifestation of the disease. With regards to metabolic
   pathways, several nutrient and energy consumption-related processes
   were significantly upregulated in the NHIC group, thereby suggesting
   that these patients may have nutrient and energy metabolism disorders
   that contribute to the development of the disease. Our metabolome
   combined transcriptome analysis also indicated that bladder systolic
   dysfunction in HIC patients could be related to the downregulation of
   actin cytoskeleton-related pathways. Through our multi-omics joint
   analysis, we have inferred that Prévobacterium species, which is
   advantageous in the urine of NHIC patients, may regulate the
   inflammatory response by altering histidine changes, which could affect
   mast cell degranulation and histamine release. Furthermore, the
   dominant bacterium Lactobacillus may negatively regulate
   β-hydroxy-β-methylbutyrate, which is implicated in the damage and
   repair mechanism of muscle. In contrast, Pseudomonas, which is
   significantly dominant in the urine of HIC patients, could be involved
   in regulating pyrimidine metabolism, lipid, and glucose metabolism,
   thereby mediating the necrosis of bladder urothelial cells and the
   formation of bladder wall ulcers. It is important to note that these
   are only preliminary inferences, as the occurrence and development of a
   disease is a complex interplay between genetics, environment, and other
   factors.

Methods

Ethical declaration and informed consent

   The proposal of this research project has been approved by the Ethics
   Committee of Peking University People’s Hospital and is in line with
   the provisions of the Declaration of Helsinki. The ethical approval
   number is 2022PHB400-001. All patients signed informed consent.

Specimen acquisition

   The specimen repository compiled bladder tissue samples from two
   cohorts of NHIC and HIC patients who underwent cystoscopy with
   hydrodilation surgery. HIC is a type of pancystitis^[104]35, therefore,
   during the biopsy procedure, we collected two samples from the Hunner
   lesion and one sample each from the surrounding mucosa (upper, lower,
   left, and right), for a total of six points. After the surgeon
   extracted the bladder tissue samples, the laboratory personnel
   immediately performed liquid nitrogen flash freezing. The samples were
   then transported to a liquid nitrogen tank and, once a sufficient
   number of samples were collected, they were shipped on dry ice for
   further processing.

   Urine samples were collected via two different retention methods:
   surgeon-retained urine before surgery and early mid-morning urine
   specimens retained by the patients before surgery. The patients’
   self-retained mid-morning urine sample was categorized as the MID
   group, while the intraoperative catheterization sample was designated
   as the CATH group.

   The urine samples collected from patients were transported to the
   laboratory, where they were centrifuged at 3000 rpm for 10 min at 4 °C.
   The clear supernatant was then aliquoted into centrifuge tubes, with
   1.5 ml per tube. Four tubes were prepared for each sample: two for
   urine microbiome analysis and two for metabolomics analysis. After
   proper labeling, the samples were flash-frozen in liquid nitrogen for
   15 min and stored in a − 80 °C freezer. Once a sufficient number of
   samples were collected, they were shipped on dry ice for further
   processing.

   DNA extraction and PCR amplification: Total microbial genomic DNA was
   extracted from urine samples using FastDNA® Spin Kit for Soil (MP
   Biomedicals, U.S.) according to manufacturer’s instructions. The
   quality and concentration of DNA were determined by 1.0% agarose gel
   electrophoresis and a NanoDrop2000 spectrophotometer (Thermo
   Scientific, United States) and kept at − 80 ℃ prior to further use. The
   hypervariable region V3-V4 of the bacterial 16S rRNA gene were
   amplified with primer pairs 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and
   806R(5′-GGACTACHVGGGTWTCTAAT-3′)^[105]36 byT100 Thermal Cycler PCR
   thermocycler (BIO-RAD, USA). The PCR reaction mixture including 4 μL
   5 × Fast Pfu buffer, 2 μL 2.5 mM dNTPs, 0.8 μL each primer (5 μM),
   0.4 μL Fast Pfu polymerase, 0.2uL BSA, 10 ng of template DNA, and ddH2O
   to a final volume of 20 µL. PCR amplification cycling conditions were
   as follows: initial denaturation at 95 ℃ for 3 min, followed by 29
   cycles of denaturing at 95 ℃ for 30 s, annealing at 53 ℃ for 30 s and
   extension at 72 ℃for 45 s, and single extension at 72 ℃ for 10 min, and
   end at 10 ℃. The PCR product was extracted from 2% agarose gel and
   purified using the PCR Clean-Up Kit (YuHua, Shanghai, China) according
   to manufacturer’s instructions and quantified using Qubit 4.0 (Thermo
   Fisher Scientific, USA) .

   Extraction of urine metabolites was performed using the
   methanol/acetonitrile extraction method. After the samples were slowly
   thawed at 4 °C, 100 µL of each sample was taken and mixed with 400 µL
   of pre-cooled methanol/acetonitrile solution (1:1, v/v). The mixture
   was vortexed and left to stand at − 20 °C for 30 min, followed by
   centrifugation at 14,000g for 20 min at 4 °C. The supernatant was
   collected and vacuum-dried. For mass spectrometry analysis, the dried
   metabolites were reconstituted in 100 µL of acetonitrile–water solution
   (acetonitrile:water, 1:1, v/v), vortexed, and centrifuged again at
   14,000g for 15 min at 4 °C. The resulting supernatant was used for
   sample injection and analysis.

Approaches for investigating the urinary microbiome

   Genomic DNA is first extracted from the sample, and the conserved
   region of rDNA is amplified using specific primers with barcodes. The
   PCR amplification products are then quantified and purified before
   being combined in equal amounts to construct a sequencing library.
   Illumina PE250 sequencing is then performed on the library, and the
   resulting raw reads are filtered to eliminate low-quality sequences.
   The remaining high-quality reads are assembled and stitched into tags,
   which are then filtered, resulting in clean tags. The clean tags are
   used for clustering based on UPARSE algorithm, and the resulting
   effective tags are analyzed for OTU abundance statistics. The processed
   data is then subjected to species annotation, α diversity analysis, β
   diversity analysis, community function prediction, and other analyses.
   If valid groupings exist, between-group comparisons and statistical
   tests are performed. Finally, more advanced analyses such as CCA are
   conducted in conjunction with other factors, such as environmental
   factors or metabolomics results, to answer questions regarding the
   relationship between microbes and the environment.

Approaches for urinary metabolomics analysis

   Following sample collection, specimens undergo rigorous quality
   control, screening, and subsequent formal experimentation and
   bioinformatics analysis. During experimental procedures, a combination
   of positive and negative ion modes (POS and NEG, respectively) is
   employed, allowing for enhanced coverage and more robust detection
   results. In the subsequent data analysis phase, separate analyses of
   positive and negative ion modes are performed.

Qualitative and quantitative analysis of metabolites

   Non-targeted metabolomics utilizes high-resolution mass spectrometry
   (HRMS) detection technology to detect as many molecular characteristic
   peaks in the sample as possible. The technique combines public
   databases such as Mass Bank, Metlin, MoNA, and self-built secondary
   mass spectrometry databases for substance annotation.

Differential metabolite identification and analysis

   We integrated multivariate statistical analysis of the VIP value of
   OPLS-DA and univariate statistical analysis using T-test P value to
   screen for differential metabolites across various comparison groups.
   The set threshold for differential significance was: VIP ≥ 1 and T-test
   P < 0.05 in the OPLS-DA model. The VIP plot of OPLS-DA is employed to
   showcase the significance of variables and their contribution towards
   sample differentiation.

Metabolic pathway analysis

   Following the identification of differential metabolites, we performed
   KEGG enrichment analysis to determine the key biochemical metabolic
   pathways and signal transduction pathways that were impacted by these
   metabolites across groups.

Liquid chromatography-mass spectrometry analysis

   Liquid chromatography-mass spectrometry (LC–MS/MS) analysis was
   performed using an Agilent 1290 Infinity LC system coupled to a SCIEX
   TripleTOF 6600 mass spectrometer. Chromatographic separation was
   achieved on a Waters ACQUITY UPLC BEH Amide column (1.7 μm,
   2.1 × 100 mm) maintained at 25 °C. The mobile phase consisted of (A)
   25 mM ammonium acetate/25 mM ammonium hydroxide in water and (B)
   acetonitrile, delivered at 0.5 mL/min with the following gradient: 95%
   B (0–0.5 min), linear reduction to 65% B (0.5–7 min), further reduction
   to 40% B (7–8 min), isocratic hold (8–9 min), and re-equilibration at
   95% B (9.1–12 min). Electrospray ionization operated at ± 5500 V with
   source gases set to 60 psi (Gas1 and Gas2) and 30 psi curtain gas, at
   600 °C source temperature. Full-scan MS data (m/z 60–1000) and
   data-dependent MS/MS scans (m/z 25–1000) were acquired in
   high-sensitivity mode using information-dependent acquisition (IDA)
   with collision energy of 35 ± 15 eV and declustering potential
   of ± 60 V. Quality control samples (pooled from all specimens) were
   injected every 10 samples to monitor system stability (retention time
   RSD < 0.1 min; intensity RSD < 15%). Raw data were converted to .mzML
   format (ProteoWizard v3.0.6428), processed using XCMS Online (v3.7.1)
   with centWave peak detection (m/z tolerance = 10 ppm; peak
   width = 10–60 s), and normalized to total peak area after filtering
   features with > 50% missingness within experimental groups.

Transcriptomic research methods of bladder biopsy

   After collecting tissue samples, RNA extraction was conducted using the
   Trizol reagent kit (Invitrogen, Carlsbad, CA, USA), according to the
   manufacturer’s instructions. Agilent 2100 Bioanalyzer (Agilent
   Technologies, Palo Alto, CA, USA) was used to assess RNA quality. The
   total RNA extraction was followed by the enrichment of eukaryotic mRNA
   using Oligo(dT). Subsequently, the enriched mRNA underwent
   fragmentation using a fragmentation buffer, followed by reverse
   transcription into cDNA utilizing the NEBNext Ultra RNA Library Prep
   Kit for Illumina (NEB#7530, New England Biolabs, Ipswich, MA, USA). The
   resulting library was subjected to sequencing on an Illumina
   Novaseq6000 platform by Gene Denovo Biotechnology Co. (Guangzhou,
   China).

   After ribosomal RNA was aligned, the genome alignment was performed
   using HISAT2^[106]37, and quantification was conducted based on
   StringTie^[107]38. The FPKM values were calculated using RSEM (v1.3.3,
   URL: [108]https://github.com/deweylab/RSEM/archive/v1.3.3.tar.gz)
   software. Differential analysis between groups was performed using
   DESeq2(Fold change ≥ 2 and FDR < 0.05). Geneontology (GO) and Kyoto
   Encyclopedia of Genes and Genomes^[109]39–[110]41 (KEGG) enrichment
   analyses of DEmRNAs were performed using hypergeometric Tests.

   Additionally, gene set enrichment analysis (GSEA) was used for the
   elucidation of the function of differentially expressed genes between
   groups.

Integrated analysis

   Statistical analysis and visualization in this study were conducted
   using R software (version 3.6.3).

Integration of metabolomics and microbiome analysis

   Significant differences in metabolite expression between the NHIC and
   HIC groups in metabolomics analysis were used as environmental factor
   data and matched with the corresponding patient’s microbiome analysis.
   This allowed for the generation of a correlation heat map for the
   integration of microbiome and metabolomics analysis.

Integration of metabolomics and transcriptomics analysis

   We analyzed the unique and shared portions of the data between each
   group and visualized the results using the ggplot2 and VennDiagram
   packages. The intersection of metabolite pathways enriched by KEGG
   analysis in positive and negative ion modes with pathways enriched by
   transcriptomics analysis yielded the shared pathways.

Supplementary Information

   [111]Supplementary Information 1.^ (1.3MB, pdf)
   [112]Supplementary Information 2.^ (1.3MB, pdf)

Author contributions

   Lin Zhu: Methodology, Investigation, Formal Analysis, Data Curation,
   Visualization, Writing—Original Draft. Hanwei Ke: Investigation,
   Visualization, Formal Analysis, Writing—Original Draft, Writing—Review
   & Editing. Qi Wang: Supervision, Validation. Kexin Xu:
   Conceptualization, Project Administration, Supervision, Funding
   Acquisition. Xiaofang Chen: Conceptualization, Supervision, Funding
   Acquisition.

Funding

   This work was supported by the National Natural Science Foundation of
   China (81970660) and Beijing Natural Science Foundation (7242153).

Data availability

   The raw sequence data reported in this paper have been deposited in the
   Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in
   National Genomics Data Center (Nucleic Acids Res 2022), China National
   Center for Bioinformation/Beijing Institute of Genomics, Chinese
   Academy of Sciences (GSA-Human: HRA012042) that are publicly accessible
   at [113]https://ngdc.cncb.ac.cn/gsa-human.

Declarations

Competing interests

   The authors declare no competing interests.

Ethics approval and consent to participate

   This study was approved by the Medical Ethics Committee of Peking
   University People’s Hospital (approval number: 2022PHB400-001). All
   procedures followed applicable guidelines and regulations. Informed
   written consent was also obtained from the patients in this study.

Consent for publication

   All authors consent to the publication of this manuscript and declare
   no conficts of interest.

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

   Hanwei Ke and Lin Zhu equally contributed to this work.

Contributor Information

   Kexin Xu, Email: cavinx@yeah.net.

   Xiaofang Chen, Email: chenxiaofang2023@yeah.net.

Supplementary Information

   The online version contains supplementary material available at
   10.1038/s41598-025-12010-w.

References