White-to-brown adipose switching promotes bladder cancer progression

Highlights

• •Perivesical adipose tissue invasion marks a key prognostic turning point in BCa.
• •Adipose-invasive tumors show higher PTHrP expression in BCa.
• •PTHrP-PTHR-PKA signaling drives browning and lipolysis in perivesical adipocytes.
• •Adipocyte-derived FFAs drive lipid metabolic reprogramming in bladder tumor cells.
• •Blocking perivesical adipose browning suppresses tumor growth and metastasis in vivo.

Abstract

The invasion of bladder cancer into perivesical adipose tissue represents a watershed event in tumor progression, accompanied by a dramatic deterioration in clinical outcomes. However, the underlying molecular mechanisms governing this cancer-adipose tissue crosstalk remain poorly elucidated. Here, we systematically characterize a bidirectional regulatory network between bladder cancer cells and perivesical adipocytes. Our findings demonstrate that bladder cancer cells secreted parathyroid-hormone-related protein (PTHrP), which induces browning of perivesical adipose tissue through activation of protein kinase A (PKA) signaling. Conversely, thermogenesis induced by browning perivesical adipose tissues leads to the release of excessive free fatty acids. These free fatty acids are subsequently taken up by bladder cancer cells, where they promote lipid metabolic reprogramming and thereby enhance cancer cell proliferation, invasion, and metastatic potential. In vivo experiments further validate that treatment with H89, a specific PKA inhibitor, effectively reverses perivesical adipose tissue browning and attenuates bladder cancer progression. Collectively, our data clarify that PTHrP secreted by bladder cancer cells drives perivesical adipose tissue browning to accelerate cancer progression, providing a novel potential therapeutic target for bladder cancer intervention.

Introduction

Bladder cancer is the ninth most common malignancy and the 13th most common cause of cancer death worldwide [1]. Bladder cancer is clinically divided into non-muscle-invasive (NMIBC) and muscle-invasive (MIBC) types. The latter exhibits markedly poorer clinical outcomes than NMIBC [2,3]. However, the most pronounced prognostic decline occurs between stages T2 and T3, when tumors breach the muscle layer and invade the perivesical adipose tissue. Previous studies have reported 5-year overall survival rates of 64–84% for pT2N0, 25–56% for pT3N0, and 19–44% for pT4N0 [4,5]. Despite this sharp transition in disease aggressiveness, the mechanisms by which adipose invasion promotes bladder cancer progression remain largely unknown.

In mammals, adipose tissue has two main types—white adipose tissue (WAT) and brown adipose tissue (BAT)—with distinct morphological, functional, and metabolic traits [6]. WAT stores energy in intracellular lipid droplets, while BAT dissipates energy as heat. Notably, brown-like adipocytes can also be induced in WAT depots, a process called WAT browning or beiging [7]. WAT browning is regulated by mechanisms including prolonged cold exposure, adrenergic activation, the myokine irisin, and the prostaglandin synthesis enzyme cyclooxygenase (COX)-2 [8]. Emerging evidences have documented the occurrence of WAT browning in the context of several cancers [[9], [10], [11], [12]]. Perivesical adipose tissue is a subtype of WAT found adjacent to the bladder fascia, which surrounds and supports the bladder [13]. However, it remains unclear whether bladder cancer is capable of inducing browning in adipose tissue, particularly in the perivesical adipose depots relevant to its progression.

WAT browning is classically characterized by two key features: the upregulation of uncoupling protein 1 (UCP1) expression and the enhancement of mitochondrial biogenesis [9,14]. Both processed collectively elevate adipocyte metabolic activity and drive alterations in the profile of metabolites they secrete. Adipocytes primarily function in the storage and release of free fatty acids (FFAs), a process that fulfills both local tissue and systemic metabolic demands [15]. A key factor in cancer progression is the role of cancer-associated adipocytes, which create a metabolic niche that promotes malignant cell survival and proliferation, primarily by providing FFAs as an essential energy source [[15], [16], [17]].

Here, we compared perivesical WAT samples from bladder cancer patients and identified definitive evidence of browning in this adipose depot. Based on this key observation, we further delineated the detailed bidirectional crosstalk between bladder cancer cells and perivesical adipose tissue: specifically, parathyroid hormone-related protein (PTHrP) secreted by bladder cancer cells induces browning of perivesical WAT. In turn, the FFAs released by these browned perivesical adipocytes are taken up by bladder cancer cells, which subsequently upregulate the malignant cells’ lipid metabolic activity. Collectively, these findings reveal that adjacent beige adipocytes in the bladder tumor microenvironment play a critical tumor-supportive role, establishing a functional metabolic loop that facilitates bladder cancer progression.

Materials and methods

Patients and tissue samples

This study was approved by the internal review and ethics boards of Sun Yat-sen Memorial Hospital (SYSMH) in accordance with the Declaration of Helsinki.

This study utilized MIBC tissues and matched adjacent normal tissues from 413 bladder cancer (BCa) patients undergoing radical cystectomy at SYSMH between 2014 and 2020. Patient clinical information is summarized in Supplementary Table 1. In addition, fresh tumor tissues and paired normal bladder mucosal tissues from an additional three patients with pathologically confirmed T3 or T4 disease were collected and used for experimental analyses. Detailed information for these samples is provided in Supplementary Table 2. Tumor stage and grade followed the WHO criteria and UICC TNM 7th edition (2010). All patients received regular follow-up. Overall survival (OS) and disease-free survival (DFS) were calculated from surgery to death, censored at last follow-up for survivors.

Cell culture

Human bladder cancer cell lines T24, 5637, UM-UC-3, and normal uroepithelial cell line (SV-HUC-1) were all purchased from Procell Life Science & Technology (Wuhan, China). The 5637 cell lines were cultured in RPMI-1640 (Gibco, USA), UM-UC-3, and T24 in DMEM (Gibco, USA). The SV-HUC-1 cells were cultured in Ham’s F-12 K medium (Gibco, USA). All media were supplemented with 10% fetal bovine serum (FBS) (Gibco, USA) and 1% penicillin–streptomycin solution (Gibco, USA). All the cell lines were cultured in a cell incubator at 37°C and 5% CO₂ and regular testing was conducted to verify that they were not contaminated with mycoplasma.

Histological evaluation and immunohistochemistry

For histological examination, tissues of bladder cancer patients were fixed with 4% paraformaldehyde in PBS, embedded into paraffin blocks, sectioned (5 μm), and then stained with Haematoxylin and Eosin (H&E). For immunohistochemistry (IHC) analysis, deparaffinized sections were incubated with primary anti-UCP1 antibody (1:1000, 83870-1-RR, Proteintech, China), Anti-Ki67 antibody (1:1000, GB111499-100, Servicebio, China) and Anti-PTHrP antibody (1:500, ab197358, Abcam, UK) at 4°C overnight, and then incubated with species-appropriate secondary antibodies (RGAR011, Proteintech, China), then stained with peroxidase and 3,3′-diaminobenzidine tetrahydrochloride and nuclei were counterstained with hematoxylin. The sections were captured with × 400 high-power fields by an ECLIPSE Ni-E/Ni-U microscope (Nikon, Japan) and analyzed by manual counting. For adipocyte size quantification, ImageJ software was used for analysis following the previous protocol [17].

Multiplex Immunofluorescence (MIF)

MIF staining was performed using a PANO 4-plex MIF Kit (0079100020, Panovue, China) following the manufacturer’s instructions. Briefly, sections were deparaffinized and hydrated. A microwave treatment was applied for antigen retrieval. After a 10-minute blocking step, sections were incubated with the primary antibodies at 4°C overnight. Next, the slices were incubated with the corresponding secondary antibodies (RGAR011, Proteintech, China) at 37°C for 10 min. Finally, one of the two tyramide signal amplification fluorophores (Opal 520 and 620) was added to the tissue sections for 10 min. This sequence of steps (except deparaffinisation and hydration) was repeated, starting with blocking and ending with the microwave treatment. The slides were sequentially counterstained with DAPI (C1006, Beyotime, China) and scanned with a Vectra Polaris Automated Quantitative Pathology Imaging System (Akoya Biosciences, USA). The primary antibodies used for the MIF staining were as follows: anti-UCP1 (1:3000, 83870-1-RR, Proteintech, China) and anti-Perilipin1 (1:2000, 27716-1-AP, Proteintech, China). The signals were evaluated and quantified using Phenochart and inForm Tissue Analysis software (Akoya Biosciences, USA).

Adipogenic differentiation

Perivesical white adipose tissue (WAT) derived stem cells (ADSCs) were routinely isolated from patients who underwent surgical resection in Sun Yat-sen Memorial Hospital, cultured and differentiated into mature adipocytes following the previously published method [18]. Briefly, the perivesical WATs were obtained, and then dissected, washed twice with 1 × PBS, minced into small pieces and digested for 40 min at 37°C in an isolation buffer containing antibiotics (2.5 mg/mL fungizone, 50 mg/mL gentamicin, 1% penicillin/streptomycin), 1 mg/mL type I collagenase (#HYE70005A, MedChemExpress, Shanghai, China) and 1 mg/mL type II collagenase (#HYE70005B, MedChemExpress, Shanghai, China). Digested tissues were passed through a sterile 100-mm nylon mesh (BD Falcon, USA) to remove undigested tissues and the remaining suspension was centrifuged at 500 × g for 5 min to generate ADSCs in the pellet. The pellet was resuspended in DMEM/F12 (Gibco, USA) medium containing 1% pen/strep and 10% FBS, passed through a 40-mm nylon mesh, centrifuged, resuspended, and plated on collagen-coated culture dishes (Corning, USA). ADSCs were induced to differentiate into adipocytes using a differentiation kit (HyCyte, ADHX-D102) in accordance with the manufacturer’s protocols. Briefly, ADSCs were grown in DMEM containing 10% FBS until 90% confluency. Then cells were cultured with the differentiation induction complete medium for 72 h, followed by culturing in the differentiation maintenance complete medium for 24 h. A total of three cycles of culture medium change were performed. At the end of the differentiation protocol, ADSC-derived adipocytes containing abundant lipid droplets were obtained and used for subsequent experiments.

Western blotting

Total protein from cells or tissues was harvested with ice-cold RIPA buffer (P0013B, Beyotime, China) supplemented with protease inhibitor cocktail (78440, Thermo Fisher Scientific, USA). The lysates were centrifuged for 20 min at 15,000 × g at 4°C to remove debris. Protein concentration was measured using the detergent-insensitive Pierce BCA protein assay kit (23227, Thermo Fisher Scientific, USA). NuPAGETM LDS-Sample buffer (NP0008, Thermo Fisher Scientific, USA) was added and samples were denatured at 95°C for 5 min. Equal amounts (30 μg) of protein mixed with 1 × loading buffer were loaded onto SDS-polyacrylamide gel electrophoresis (PAGE). After electrophoresis, proteins were transferred to polyvinylidene difluoride (PVDF) membranes (WJ001, EpiZyme, China) and blocked with 5% bovine serum albumin (BSA) in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h at room temperature. The membranes were then incubated with primary antibodies including anti-UCP1 (1:3000, 83870-1-RR, Proteintech, China), anti-PTHLH (1:2000, A3183, ABclonal, China), anti-PTHR (1:1000, 29115-1-AP, Proteintech, China), anti-Phospho-PKA Substrate (RRXS*/T*) (1:1000, 9624, CST, USA), anti-pHSL (Ser660) (1:1000, AF8026, Affinity, China), anti-HSL (1:1000, AF6403, Affinity, China), anti-pCREB (Ser133) (1:2000, 28792-1-AP, Proteintech, China), anti-CREB(1:5000, 12208-1-AP, Proteintech, China), anti-GAPDH (1:50000, HRP-60004, Proteintech, China) and anti-ɑTubulin (1:10000, HRP-80762, Proteintech, China) overnight at 4°C. The membranes were further incubated with the corresponding HRP-conjugated secondary antibodies (1:5000, RGAR001, Proteintech, China) for 1.5 h at room temperature. Finally, protein bands were imaged using an enhanced chemiluminescence detection kit (Pierce ECL Plus, Thermo Scientific) and analyzed with a MiniChemi 610 Chemiluminescent/Fluorescent Imaging and Analysis System (Beijing Sage Creation Science Co., China).

RNA isolation and qRT-PCR

Total RNA was extracted from frozen tissues or cultured cells using a reagent kit (400-100, GOONIE, China) and then underwent reverse transcription and analyzed by qRT-PCR using a reagent kit (R323-01, Vazyme Biotech, China) according to the manufacturer’s instructions. Reactions were conducted in the 96-well format with a LightCycler 480 II Instrument (Roche Diagnostics, Germany), and relative mRNA abundance was normalized to GAPDH using the 2^(-ΔΔCt) method. Primer sequences used are listed in Supplementary Table 3.

Online data analysis

The scRNA-seq data for eight primary tumors of bladder cancer (two low-grade bladder urothelial tumors and six high-grade bladder urothelial tumors) along with 3 adjacent normal mucosae, were retrieved from https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA662018. Raw FASTQ files were processed using Cell Ranger (version 2.2.0, 10x Genomics) to demultiplex cellular barcodes, align reads to the human reference transcriptome, and generate gene–cell count matrices. The resulting matrices were imported into Scanpy (version 1.9.3, Python) as AnnData objects for downstream analyses.

Cells with fewer than 1,000 unique molecular identifiers (UMIs), greater than 20% mitochondrial-derived UMIs, or >6,000 detected genes were excluded to remove low-quality cells and potential doublets. Genes expressed in fewer than three cells were also filtered out. After quality control, a total of 80,016 high-quality single cells were retained. Gene expression matrices were normalized and log-transformed using the normalize total and log1p functions in Scanpy. The top 2,000 highly variable genes were identified, followed by principal component analysis (PCA) for dimensionality reduction. To correct for batch effects across datasets, the Harmony algorithm was applied, chosen for its efficiency and ability to preserve cell-type integrity. After batch correction, the neighborhood graph was constructed with the neighbors function, and cells were clustered using the Leiden algorithm. Low-dimensional visualization was performed with UMAP, enabling identification of distinct cell populations based on canonical marker gene expression. Epithelial cells derived from normal tissues were annotated as normal epithelial cells, while epithelial cells derived from tumor samples were defined as tumor cells. Differential expression of multiple target genes between epithelial cells from adjacent normal mucosa and bladder tumor tissues was assessed using non-parametric Mann–Whitney U tests. Differences with P < 0.05 were considered statistically significant.

Transcriptome data and survival data of bladder cancer patients in different disease stages were retrieved from the TCGA-BLCA datasets from UCSC XENA (http://xena.ucsc.edu/) in this study. Differentially expressed genes between disease stages were identified, and gene set enrichment analysis (GSEA) was then performed on the ranked gene list using gene sets from the MSigDB database to explore stage-associated biological pathways.

Enzyme-linked immunosorbent assay (ELISA)

The concentrations of PTHrP in culture supernatants of normal and bladder cancer cells were measured by ELISA Kits (E-EL-H1478, Elabscience, China) according to the manufacturer's instructions.

Measurements of lipolysis in vitro

Lipolytic activities were assessed by measuring FFA release as previously described [19]. In brief, adipocytes with or without PTHR knockout were cultured in a 6-well plate in serum-free DMEM, with or without H89 (20 mM) for 3 h. Then, the medium was replaced by fresh DMEM and incubated for an additional 3 h at 37°C, and the medium in each well was collected for lipolytic measurement. FFA levels were assessed by Free fatty acid assay kit (BC0590, Solarbio, China) according to the manufacturer’s instructions.

Oil red O staining

Adipocytes or tissue slices of bladder cancer patients were stained with Oil Red O solution (G1015, Servicebio, China) according to the manufacturer's instructions. Briefly, samples were washed three times with PBS and fixed in 4% formalin. After fixation, Oil red O was diluted in isopropanol with water (3:2), filtered the solution through a 0.45-µm filter, and incubated the fixed samples for 1 h at room temperature. Lipid droplets were observed by using an Olympus IX73 microscope.

Assessment of lipid transfer between cells

To assess lipid transfer between cells, a co-culture experiment was performed as previously described [15]. Briefly, ADSCs (1 × 10⁴ cells per well) were seeded in the upper chamber of a 6-well Transwell system with a 0.4-μm membrane (3412, Corning, USA) and induced to differentiate into adipocytes. Adipocytes were subsequently stained with BODIPY stain (1 μg/mL in PBS, D3922, Invitrogen, USA) for 4 h at 37°C to label intracellular lipids. Subsequently, adipocytes were washed three times with 1 × PBS containing 0.2% fatty acid-free BSA to remove the uncombined dye. T24 cells were cultured in 6-well plates (5 × 10⁴ cells per well) and cultured overnight. The 6-well 0.4-μm transwells containing BODIPY-labeled adipocytes were then placed onto the wells containing T24 cells in 6-well plates and co-cultured for 48 h. T24 cells were harvested and washed with PBS. T24 cells positive for BODIPY were detected by flow cytometry or immunofluorescence staining.

Flow cytometry

For lipid content detection, T24 cells co-cultured with BODIPY-labeled adipocytes as described above were analyzed using a flow cytometer (Beckman CytoFLEX S), and FlowJo software was used to process the data.

Seahorse XFe24 measurements

The Seahorse Bioscience XFe24 Flux Analyzer (Agilent, USA) was used to measure the oxygen consumption rate (OCR) of tumor cells according to the manufacturer’s protocol [20]. Briefly, wild type or PTHR-knockout adipocytes were co-cultured with T24 cells for 48 h in the presence or absence with an anti-PTHrP neutralizing antibody (10 µg/mL, RHC93101, Antibodysystem), or an equivalent concentration of Mouse IgG1 isotype control. The resulting adipocytes (T24-educated Adipo) were washed three times with PBS and their conditioned medium (T24-educated Adipo-CM) were collected and subjected to centrifugation at 2,000 × g for 5 min to eliminate cellular debris. Bladder cancer cells were seeded in culture plates and treated with various T24-educated Adipo-CM (50% Adipo-CM + 50% medium) for 24 h. Cells were incubated in XF assay medium (containing 10 mM glucose and 2 mM l-glutamine) for 1 h without CO₂ before oxygen consumption analysis. The mitochondrial stress test utilizes sequential injections of oligomycin (4.5 mM), Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 2 mM), and rotenone/antimycin A in combination (1 mM). The parameters of basal respiration and maximal respiration were automatically calculated by the WAVE software (Agilent). All parameters were normalized to the total protein amount in individual wells using the BCA protein assay (Thermo Fisher Scientific, USA).

Cell proliferation assay

The Cell Counting Kit-8 (CCK-8) and Colony formation were used to assess the proliferation of bladder cancer cells as previously described [21]. Briefly, bladder cancer cells were treated with CM from T24-educated Adipocytes for 5 days. Cultured medium was replaced with 100 μL fresh DMEM, and 10 μL CCK-8 (CCK8-K1018-5, APExBIO, USA) was added into each well (time point set as 0 h). Then, the plate was incubated at 37°C for 60 min, after which the absorbance at 450 nm was read using a TECAN Spark microplate reader. Cell proliferation assays were also performed at time points of 24, 48, 72, 96, and 120 h.

For colony formation, T24 cells (1 × 10³ cells per well) were inoculated in six-well plates, and groups were assigned as described above, followed by timely replacement of the medium according to cell morphology and growth conditions. The plates were cultured for 10–14 days until visible colonies appeared at the bottom of the wells. After washing twice with PBS, cells were fixed with 4% paraformaldehyde for 15 min and subsequently stained with 1% crystal violet for 30 min. The number of colonies was then counted and recorded.

Cell migration

Bladder cancer cell migration was determined in a scratch wound assay as previously described [22]. Bladder cancer cells (8 × 10⁴ cells/well) were seeded and cultured to 100% confluence in the lower chamber in a complete medium supplemented with 10 ng/ml mitomycin-C for 2 h to inhibit cell proliferation. Cells were scratched vertically with a 200-μL pipette tip at a constant speed, washed twice with PBS to remove detached cells, and photographed under a microscope, and the scratch area at 0 h was recorded. After the serum-free medium was added for another 24 h, the scratch area was photographed again to record. The relative wound density (the ratio of the occupied area to the total area of the initial scratched region) was measured using ImageJ software.

Transwell

Complete medium was added to the lower chamber of the 24-well plate (Corning, USA), and serum-free medium was added to the upper chamber (Corning, USA). Before cell seeding, the upper chamber was pre-coated with LDEV-Free Cultrex™ Matrix (40183ES, Yeasen Biotech, China) according to the manufacturer’s instructions. 5 × 10⁴ cells were inoculated in the upper chamber, while the 24-well plate was shaken and cultured in the incubator for 24-36 h. Then 4% paraformaldehyde was added to the lower chamber to fix for 20 min. After cleaning again with PBS, 1% crystal violet was taken and stained for 30 min. Under the microscope, 3-6 fields were randomly selected for observation and counting [21].

Animal experiments

All mouse procedures were approved by the SYSMH Institutional Animal Care and Use Committee and followed approved guidelines. For the subcutaneous xenograft tumor model, 100 μL of a cell suspension containing 1 × 10⁶ T24 cells alone or 1 × 10⁶ T24 cells mixed with 4 × 10⁶ adipocytes was injected subcutaneously into the right axilla of female BALB/c nude mice (aged 4 to 6 weeks, GemPharmatech). Tumor diameter was recorded once every seven days from the beginning of subcutaneous transplantation, and tumor mass was recorded after tumor dissection.

For the lung metastasis model, 1 × 10⁶ T24 cells pretreated with control CM or T24-educated adipocyte CM were injected into the tail vein of BALB/c nude mice. The progression of metastases was monitored 4 weeks later by bioluminescent (BL) imaging using the IVIS Spectrum system (PerkinElmer) 10 min after intraperitoneal injection of 100 μL of d-luciferin (15 mg/mL). Mice were euthanized, and lungs were harvested and subjected to 4% paraformaldehyde for the subsequent hematoxylin-eosin staining. For lung nodules analysis, a minimum of three sections were excised at 100 mm intervals, and metastases in each section were counted manually on an ECLIPSE Ni-E/Ni-U microscope (Nikon, Japan).

Results

Bladder cancer induces adjacent adipose browning

Clinically, patients with MIBC exhibit substantially poorer survival than those with NMIBC, and disease outcomes deteriorate progressively with higher T stage [23,24]. The most striking decline occurs between T2 and T3, when tumors breach the muscularis propria and infiltrate the perivesical adipose tissue [[2], [4], [5], [25]]. Consistent with this clinical pattern, analysis of the SYSMH cohort revealed a similar trend (Fig. 1A). Differential expression analysis of TCGA-BLCA T2 versus T3 tumors, followed by GSEA, demonstrated that T3 tumors were enriched for pathways associated with adipocyte browning and epithelial–mesenchymal transition, accompanied by a concomitant downregulation of apoptosis-related pathways (Fig. 1B). To further substantiate these transcriptomic findings at the tissue level, we examined the histological features of perivesical adipose tissue in patients with bladder cancer. Adipose samples were collected from individuals with bladder cancer and stratified into two categories according to anatomical location relative to the tumor: adjacent adipose tissue (directly adherent to the tumor), distal adipose tissue (obtained from sites >5 cm away from the tumor margin). Hematoxylin and eosin (H&E) staining was performed on these specimens, and morphological analysis revealed that adipocytes in adjacent adipose tissue exhibited a smaller size compared with those in distal adipose tissue from the same individuals (Fig. 1C). Furthermore, immunohistochemistry staining was conducted to evaluate protein expression patterns, and the results indicated that uncoupling protein 1 (UCP1) expression was significantly higher in adipocytes from tumor-adjacent adipose tissue compared with those located from distal adipose tissue (Fig. 1D). We further observed that the increased UCP1 fluorescence signal in tumor-adjacent adipocytes showed a high degree of colocalization with Perilipin1, a specific marker for adipocytes (Fig. 1E). Additionally, western blotting analysis verified that the protein level of UCP1 was significantly higher in adjacent adipocytes than in distal adipocytes (Fig. 1F). Moreover, adipose tissue adjacent to bladder cancer displayed elevated expression of genes associated with thermogenesis, including UCP1, PGC1α, and DIO2 (Fig. 1G). To investigate whether factors derived from bladder cancer cells can induce thermogenic gene expression in adipocytes, we conducted functional experiments using cancer cell-conditioned medium (CM). Adipocytes treated with bladder cancer cell-CM showed increased UCP1 fluorescent intensity compared with those treated with control CM (Fig. 1H). Consistently, western blotting analysis demonstrated that adipocytes exposed to bladder cancer cell CM exhibited significantly higher UCP1 protein levels (Fig. 1I). Furthermore, treatment with bladder cancer cell CM effectively stimulated the expression of UCP1, PGC1α, and DIO2 in adipocytes (Fig. 1J). Collectively, these data suggested that adipose tissue adjacent to the bladder cancer underwent a browning process.

Fig. 1.

Bladder cancer cells induced browning of adjacent adipocytes. (A) Kaplan–Meier analysis of overall survival according to T stage (T2–T4) in MIBC patients from the Sun Yat-sen Memorial Hospital (SYSMH) cohort (T2, n = 284; T3, n = 84; and T4, n = 45). (B) Gene Set Enrichment Analysis (GSEA) enrichment analysis of differential expression between T2 and T3 bladder cancer in TCGA-BLCA cohort. (C) Representative H&E staining images (left) and quantification of the average area of adipocytes in sections of distal and adjacent adipose tissue of patients with bladder cancer. Scale bar, 50 μm. (D) Representative immunohistochemistry staining for UCP1 (left) and quantification of the percentage of UCP1 positive cells in sections of distal and tumor-adjacent adipose tissue of individuals with bladder cancer. Scale bar, 50 μm. (E) Representative immunofluorescent staining for UCP1 and Perilipin1 (left) and quantification of the intensity of UCP1 in sections of distal and tumor-adjacent adipose tissue of individuals with bladder cancer. Scale bar, 50 μm. (F) Representative immunoblots (left) and quantification (right) of UCP1 in distal and tumor-adjacent adipose tissue of individuals with bladder cancer. (G) The mRNA levels of UCP1, PGC1ɑ, and DIO2 in adipocytes harvested from distal and tumor-adjacent adipose tissue of individuals with bladder cancer were evaluated by qRT-PCR. (H) Representative immunofluorescent staining for UCP1 and Perilipin1 (left) and quantification of the intensity of UCP1 in adipocytes treated with the CM from bladder cancer cells or control CM for 48 h. (I) Representative immunoblots (left) and quantification (right) of UCP1 in adipocytes treated with the CM from bladder cancer cells or control CM. (J) The mRNA levels of UCP1, PGC1ɑ, and DIO2 in adipocytes treated with the CM from bladder cancer cells or control CM. Data were expressed as means ± SEM (C-J). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by log-rank test (A), or unpaired two-tailed Student’s t-test (C-J).

Bladder cancer-derived PTHrP causes adipose browning

Previous studies have shown that several tumor-secreted cytokines might activate white adipocytes browning, such as parathyroid hormone-related protein (PTHrP), zinc-alpha-2-glycoprotein-1 (AZGP1), growth/differentiation factor 15 (GDF15), and interleukin-6 (IL6) [10,[26], [27], [28]]. To determine whether a similar mechanism exists in bladder cancer, we compared the expression of these cytokines between bladder tumor cells and normal urothelial epithelial cells using publicly available single-cell RNA-sequencing data (BioProject PRJNA662018, NCBI SRA). Among all candidates, only PTHLH, which encodes PTHrP, showed a significant increase in bladder cancer cells relative to normal epithelial cells (Fig. 2A, Figures S1A-C). To evaluate the association between PTHLH expression and pathological stage, we interrogated TCGA-BLCA and stratified tumors by adipose invasion into T1/T2 (no perivesical fat invasion) versus T3/T4 (perivesical fat invasion). PTHLH expression was markedly higher in the T3/T4 group than in T1/T2 (p = 0.0049) (Fig. 2B). Moreover, elevated PTHLH expression was significantly associated with poorer overall survival (Fig. 2C) and reduced disease-free survival (Fig. 2D). Therefore, we investigated the level of PTHrP of bladder cancer at the tissue level. Fresh bladder urothelial carcinoma tissues from patients with pathologically confirmed T3 or T4 stage, along with paired fresh normal bladder mucosal tissues, were collected for analysis. Western blotting showed substantially higher PTHrP protein abundance in bladder cancer tissues compared with paired normal mucosae (Fig. 2E). Furthermore, qRT-PCR showed that the mRNA expression of PTHLH was markedly upregulated in tumor samples (Fig. 2F). Consistently, Enzyme-linked Immunosorbent Assay (ELISA) demonstrated that bladder cancer cell-derived CM contained significantly higher levels of secreted PTHrP than control CM (Fig. 2G). To validate the functional significance of tumor-derived PTHrP, we treated adipocytes with recombinant PTHrP, which was sufficient to induce adipocyte browning, as indicated by increased UCP1 protein expression (Fig. 2H). Evaluation of several bladder cancer cell lines further revealed that T24 cells secreted the highest amount of PTHrP among the tested lines, which was substantially higher than that detected in UM-UC-3 cells, the second highest among the tested lines (Fig. S1D). Western blot and qRT-PCR analyses indicated that UM-UC-3 induced only a modest increase in UCP1 expression in adipocytes (Figure S1E and S1F). Accordingly, the subsequent analyses focused primarily on the T24 cell line. Next, neutralizing PTHrP in bladder cancer cell CM with a specific anti-PTHrP antibody eliminated the bladder cancer cell CM-driven upregulation of UCP1 in adipocytes (Fig. 2I). Likewise, knockout of the PTHrP cell surface receptor (PTHR) in adipocytes efficiently attenuated the bladder cancer cell CM-induced UCP1 upregulation (Figs. 2J and 2K). These data suggest that bladder cancer-secreted PTHrP plays a central role in promoting browning of adjacent adipose tissue.

Fig. 2.

Bladder cancer cell-secreted PTHrP induced perivesical adipose tissue browning. (A) Expression of PTHLH in epithelial cells from adjacent normal bladder mucosa and bladder tumor tissues based on a publicly available single-cell RNA-seq dataset (PRJNA662018). (B) The mRNA levels of PTHLH in individuals with bladder cancer of T1&T2 or T3&T4 according to TCGA database. (C and D) The association of PTHLH mRNA levels with overall survival (C) and disease-free survival rates (D) of individuals with bladder cancer, based on TCGA database(n = 397). The median expression was used as the cutoff value. (E) Representative immunoblots (left) and quantification (right) of PTHrP in normal mucosa and bladder tumor tissues, n = 3. (F) The mRNA levels of PTHLH in normal mucosa and bladder tumor tissues were evaluated by qRT-PCR, n = 3. (G) Levels of PTHrP on CM from normal mucosa and bladder tumor tissues were evaluated using ELISA, n = 3. (H) Representative immunoblots (left) and quantification (right) of UCP1 in adipocytes treated with PTHrP or DMSO (vehicle control) for 48 h. (I) Adipocytes were treated with control CM or CM from bladder cancer cells in the presence or absence of neutralizing anti-PTHrP or control IgG antibody. Representative immunoblots (left) and quantification (right) of UCP1 in indicated adipocytes. (J and K) Adipocytes were transduced with or without control (sgCtrl) or PTHR sgRNA (sgPTHR). (J) Representative immunoblots of PTHR in indicated adipocytes. (K) Indicated adipocytes were treated with CM from bladder cancer cells or control CM for 48 h. Representative immunoblot (left) and quantification (right) of UCP1 in adipocytes. Data were expressed as means ± SEM (A, E-J, and K) and as the median with interquartile range (B). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by Mann-Whitney U-test (A, B), log-rank test and Cox proportional hazards regression (C, D), unpaired two-tailed Student’s t-test (E-H), or one-way ANOVA (I and K).

PTHrP induces adipose browning and lipolysis via the PKA signaling pathway

PTHR is a G protein-coupled receptor that activates the cyclic adenosine monophosphate (cAMP)-dependent protein kinase A (PKA) signaling pathway [29]. The treatment of adipocytes with PTHrP promoted the phosphorylation of the PKA substrates CREB and hormone-sensitive lipase (HSL) (Fig. 3A). When PKA activity was blocked with H89, a selective PKA inhibitor, the transcriptional regulation induced by PTHrP was also completely inhibited (Fig. 3A). In line with this, knockout of PTHR in adipocytes also prevented the phosphorylation events induced by cancer cell–derived CM. (Fig. 3B). Notably, treatment of adipocytes with H89 and knockout of PTHR in adipocytes also abolished the PTHrP-driven upregulation of UCP1 (Figs. 3C and 3D). HSL is a key lipolytic enzyme in adipocytes that catalyzes triglyceride hydrolysis to generate free fatty acids (FFAs) and glycerol [30,31]. We therefore examined FFA release levels in adipocytes. We found that PTHrP treatment significantly increased FFA release levels in adipocytes, and this effect was reversed by H89 treatment (Fig. 3E). Similarly, adipocytes treated with bladder cancer cell-derived CM exhibited a marked elevation in FFA release, which was inhibited by PTHR knockout in adipocytes (Fig. 3F). Consistently, assessment of lipid content by Oil Red O staining revealed that PTHrP treatment reduced lipid accumulation in adipocytes, and this decrease was reversed by H89 treatment or knockout of PTHR in adipocytes (Fig. 3G and 3H). Of note, CM prepared from UM-UC-3 cells triggered a mild but significant increase in FFA release from adipocytes and a concomitant reduction in intracellular lipid content (Figures S2A and S2B). This observation is consistent with the relatively low level of PTHrP secreted by UM-UC-3 cells, and these phenotypic changes were abrogated by H89 treatment (Figures S2A and S2B). Taken together, these results strongly suggested that bladder cancer cell-derived PTHrP mediated its effects on UCP1 transcription and subsequent promotion of adipocyte lipolysis in a PKA pathway-dependent manner.

Fig. 3.

PTHrP induces adipose browning through the PKA signaling pathway. (A-H) Adipocytes were subjected to two experimental treatments: (A, C, E, and G) adipocytes were treated with 100 ng/mL PTHrP or DMSO (vehicle control) in the presence or absence of H89 (50 µM), (B, D, F, and H) adipocytes with or without PTHR knockout were treated with bladder cancer cell-derived CM or control CM. (A and B) Representative immunoblots (left) of phosphorylated PKA substrates, p-HSL (Ser660), total HSL, p-CREB (Ser133), and total CREB in adipocytes with indicated treatments, with quantification (right) of the pHSL/total HSL and p-CREB/total CREB ratios. (C and D) Representative immunoblots (left) and quantification (right) of UCP1 in indicated adipocytes. (E and F) FFA release levels of indicated adipocytes were measured using a colorimetric assay. (G and H) Representative Oil Red O staining (left) and quantification (right) of neutral lipids in indicated adipocytes. Scale bars, 50 μm. Data were expressed as means ± SEM (A-H). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA (A-H).

FFA released by adipocytes was taken up by bladder cancer cells and upregulated their lipid metabolism

Adipocytes are known to promote tumor progression by supplying energy-rich lipids, such as FFAs, thereby enhancing lipid metabolic activity within tumor cells [[15], [16], [17],32]. Based on this, we hypothesized that FFAs released from adipocytes following PTHrP-induced lipolysis may be taken up by bladder cancer cells to support metabolic reprogramming. Oil-red O staining revealed that bladder cancer cells that were adjacent to adipocytes exhibited higher lipid contents compared with those distant to adipocytes (Fig. 4A). Next, we labeled adipocytes with fluorescence lipid dye BODIPY and co-cultured them with bladder cancer cells in a Transwell system (adipocytes in the upper chamber and cancer cells in the lower chamber). We found that bladder cancer cells co-cultured with these adipocytes exhibited significantly increased BODIPY signals compared with those without co-culture (Fig. 4B). Notably, treating bladder cancer cells with neutralizing anti-PTHrP antibody or knocking out PTHR in adipocytes abrogated this effect (Figs. 4B and 4C). Additionally, flow cytometry confirmed that a notable increase in BODIPY intensity in bladder cancer cells when co-cultured with adipocytes, relative to those cultured alone. This elevated BODIPY intensity was suppressed, however, by anti-PTHrP treatment of bladder cancer cells or PTHR knockout in adipocytes (Figs. 4D and 4E).

Fig. 4.

FFA released from beige adipocytes was taken up by bladder cancer cells and upregulated lipid metabolism. (A) Representative Oil Red O staining image of frozen sections from individuals with bladder cancer. Regions of interest are labeled as follows: A, tumor area distant from adipocytes; B, tumor area adjacent to adipocytes; C, adipocyte-rich area. Yellow dashed lines, the tumor–adipocyte interface; white dashed lines, tumor regions; white square, related magnifications. Scale bar, 100 μm. (B-E) Bladder cancer cells were co-cultured with or without BODIPY-labeled adipocytes, either (B, D) with/without neutralizing anti-PTHrP antibody treatment or (C, E) with/without PTHR knockout in adipocytes. Relative BODIPY fluorescence intensity in bladder cancer cells was quantified via immunofluorescence staining (B, C) or flow cytometry (D, E), with representative images or histograms (left) and quantification (right) of BODIPY staining presented. Scale bar, 20 μm. (F and G) In the presence or absence of neutralizing anti-PTHrP antibody (F), bladder cancer cells were co-cultured with or without PTHR-knockout or wild-type adipocytes (G). The mRNA levels of CD36, CPT1A, PPARɑ, and PPARγ in indicated bladder cancer cells were evaluated by qRT-PCR. (H and I) In the presence or absence of anti-PTHrP antibody (H), bladder cancer cells were co-cultured for 48 h with or without PTHR-knockout or wild-type adipocytes (I). CM from T24-educated adipocytes were then collected to treat T24 cells for 24 h. Representative oxygen consumption curves (left) and quantification of basal OCR (middle) and maximal OCR (right) of indicated bladder cancer cells. Oligo, oligomycin; FCCP, fluorocarbonyl cyanide phenylhydrazone; Rot/AA, rotenone and antimycin A. Data were expressed as means ± SEM (B-I). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA (B-I).

We next evaluated whether adipocyte-derived FFAs altered lipid metabolic programming in bladder cancer cells. Compared to bladder cancer cells cultured alone, those co-cultured with adipocytes showed increased expression of molecules linked to FFA uptake and metabolism, including CD36 (a scavenger receptor that mediates FFA uptake), CPT1A (the rate-limiting enzyme in long-chain FFA oxidation), peroxisome proliferator activated receptor alpha (PPARα), and PPARγ—with this upregulated expression being inhibited by either treating bladder cancer cells with neutralizing anti-PTHrP antibody or knocking out PTHR in adipocytes (Figs. 4F and 4G). To further validate the bidirectional communication between bladder cancer cells and adipocytes, we co-cultured T24 cells with wild-type or PTHR-knockout adipocytes, in the presence or absence of either neutralizing anti-PTHrP or control IgG antibody. CM from T24-educated adipocytes was then collected to stimulate T24 cells. Seahorse assays showed that, compared with control CM, CM from T24-educated adipocytes markedly increased both basal and maximal oxygen consumption rates (OCR) in T24 cells, whereas this increase was reversed either by anti-PTHrP treatment during co-culture or by PTHR knockout in adipocytes (Figs. 4H and 4I).

Collectively, these findings suggested that FFAs released from adipocytes in response to bladder cancer cell-derived PTHrP are subsequently taken up by tumor cells, and further promote lipid metabolism in cancer cells.

Inhibiting adipocyte browning suppressed bladder cancer progression

To further investigate the contribution of adipocytes to bladder cancer progression, we first stimulated T24 cells with CM from T24-educated adipocytes. CCK-8 assays revealed a pronounced increase in the proliferation of T24 cells exposed to CM from T24-educated adipocytes. Notably, this promotive effect was almost completely abrogated either by adding anti-PTHrP antibodies during co-culture or by PTHR knockout in adipocytes (Figs. 5A and 5B). Consistent with these findings, colony formation assays demonstrated that CM from T24-educated adipocytes substantially enhanced the clonogenic capacity of T24 cells; however, both neutralization of PTHrP and knockout of PTHR in adipocytes effectively reversed this enhancement (Figs. 5C and 5D). In addition to proliferative changes, functional assays further supported an oncogenic role for adipocyte-derived FFAs. Specifically, transwell and scratch assays showed that CM from T24-educated adipocytes markedly promoted the migration and invasion of T24 cells, whereas these pro-migratory and pro-invasive effects were almost completely abolished by either PTHrP neutralization or PTHR knockout in adipocytes (Figs. 5E and 5H). In contrast, supernatants from UM-UC-3-educated adipocytes induced only modest increases in tumor cell proliferation, migration, and invasion, and most of these effects did not reach statistical significance (Figs. S3A-S3C). To determine whether FFAs mediate the pro-tumorigenic effects observed above, we treated bladder cancer cells with increasing concentrations of FFAs (25, 50, and 100 μM). The results showed that exposure to FFAs at concentrations ranging from 25 to 100 μM significantly promoted the proliferation, invasion, and migration of T24 cells. Among these concentrations, FFAs at 50 μM exerted the most pronounced promotive effects on T24 cell malignant behaviors (Figures S3D-S3H).

Fig. 5.

Inhibition of adipocyte browning suppressed bladder cancer progression. (A-H) In the presence or absence of anti-PTHrP antibody (A, C, E, and G), bladder cancer cells were co-cultured with or without PTHR-knockout or wild-type adipocytes (B, D, F, and H). CM from T24-educated adipocytes were then collected to treat T24 cells. (A and B) The proliferation rates of T24 cells were evaluated by CCK8 assay. (C and D) The proliferation ability of T24 cells was detected by colony formation assay. Representative images and quantification of bladder cancer cells were shown. (E and F) Representative images and quantification of the number of invading T24 cells were clarified by Transwell assay. (G and H) Representative images and quantification of T24 cell migration ability were determined via scratch assay. (I and J) Nude mice were subcutaneously injected with T24 cells alone or co-injected with adipocytes, with or without anti-PTHrP antibody (I) or H89 (J). (I and J) Representative tumor images (left), tumor volume quantification (middle), and tumor weight quantification (right). (K and L) T24 cells were treated with CM from T24-educated adipocytes with or without anti-PTHrP antibody (K) or H89 (L) for two weeks. Nude mice were intravenously injected with indicated bladder cancer cells. Representative H&E-stained images of lung metastases and quantification of lung metastasis lesions are shown. Scale bar, 2.5 mm. Data were expressed as means ± SEM (A-L). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA (A-L).

To validate these observations in vivo, we subcutaneously co-injected T24 cells with adipocytes at a 4:1 ratio into BALB/c nude mice, using mice receiving T24 cells alone as controls [11]. Co-injection with adipocytes resulted in significantly larger tumors, while both anti-PTHrP administration and pharmacologic inhibition of PKA signaling by H89 markedly suppressed this tumor-promoting effect (Figs. 5I and 5J). In contrast, in mice injected with T24 cells alone, treatment with H89 or anti-PTHrP alone did not significantly alter tumor progression (Figures S4A and S4B). Supporting these results, immunohistochemistry revealed increased Ki-67 expression in tumors from co-injected mice, an effect again reversed by either anti-PTHrP or H89 treatment (Figures S4C and S4D). Moreover, Western blot analysis demonstrated reduced E-cadherin and elevated N-cadherin and vimentin in tumors from the co-injection group, indicative of an EMT-like shift, which was counteracted by PTHrP blockade (Figure S4E). We next sought to determine whether adipocyte-derived FFAs also influence metastatic dissemination. To model prolonged exposure to adipocyte signals, we pre-treated bladder cancer cells with CM from T24-educated adipocytes to stimulate T24 cells for 2 weeks. Following this 2-week pre-treatment, we injected the indicated T24 cells into BALB/c nude mice via the tail vein, with untreated T24 cells serving as controls. Five weeks post-injection, H&E staining revealed a striking increase in lung metastatic lesions in mice receiving CM-pretreated T24 cells, whereas both PTHrP neutralization and H89 administration markedly reduced metastatic burden (Figs. 5K and 5L). Collectively, these data suggest that suppressing adipocyte browning can restrain bladder cancer progression.

High expression of UCP1 predicts poor prognosis in BCa patients

To evaluate the clinical relevance of PTHrP and UCP1 in human BCa specimens from patients with T3–T4 stage. IHC staining of 129 cases from the SYSMH cohort revealed a positive correlation between PTHrP expression in tumor cells and UCP1 expression in adjacent adipocytes (Fig. 6A). Kaplan–Meier analysis further showed that high UCP1 expression was significantly associated with poorer overall survival (OS) and disease-free survival (DFS) in patients with BCa (Figs. 6B and 6C). Collectively, these findings indicate that adipocyte browning driven by the PTHrP–PTHR-PKA signaling axis facilitates bladder cancer progression and highlight this signaling pathway as a potential therapeutic target.

Fig. 6.

High expression of UCP1 predicts the poor prognosis in BCa patients. (A) Correlation analysis of the tumor PTHrP positivity rate and the adipocyte UCP1 positivity rate among 129 patients with T3–T4 bladder cancer (BCa) from the SYSMH cohort. (B and C) The association of UCP1 levels with overall survival (B) and disease-free survival rates (C) of individuals with bladder cancer, based on SYSMH-cohort (n = 129). (D) Schematic model illustrating the proposed PTHrP–PTHR–PKA signaling axis driving BCa progression. Kaplan–Meier survival curves with log-rank test; HRs were derived from Cox proportional hazards regression (B and C).

Discussion

Bladder cancer progression involving perivesical fat infiltration is closely correlated with an inferior patient survival outcome. However, the precise molecular and metabolic mechanisms underlying this clinically relevant association remain poorly defined. In the present study, we identified a key regulatory pathway: bladder cancer cells secrete PTHrP, which drives the browning of adjacent perivesical adipose tissue. This browning process, in turn, leads to the release of FFAs into the tumor microenvironment. Collectively, this cascade constitutes a novel putative mechanism that fuels enhanced bladder cancer growth, providing a functional explanation for the adverse prognostic significance of perivesical fat infiltration (Fig. 6D).

Recent studies have demonstrated that WAT browning induced by lung carcinoma contributes to cancer cachexia, manifesting as a wasting disorder of adipose tissue [10,33]. A latest study revealed that beige adipocytes in perirenal adipose tissues secrete lactate, which enhances tumor growth, invasion, and metastasis in clear cell renal cell carcinoma (ccRCC) [11]. In the present study, we identified a distinct mechanism in bladder cancer: the malignancy induces browning of perivesical adipose tissue, and the resulting browned adipocytes release FFAs that directly support bladder cancer growth. This aligns with prior observations that cancer-associated adipocytes promote tumor progression in ovarian cancer, myeloma, and breast cancer via FFA secretion [[15], [16], [17],34]. The difference between our findings and the ccRCC study may lie in metabolic preference: ccRCC preferentially converts FFAs into neutral triglycerides rather than relying on fatty acid oxidation (FAO) as its primary energy source [35,36]. Notably, many tumors utilize adipocyte-derived FFAs and are highly dependent on FAO for ATP production [[16], [32], [37], [38]]. In our study, the uptake of adipocyte-derived FFA led to marked upregulation of key lipid metabolic regulators in bladder cancer cells, such as CD36, CPT1A, PPARα, and PPARγ. More importantly, Seahorse extracellular flux analysis confirmed that the uptake of FFAs significantly enhances both basal and maximal mitochondrial FAO activity in these malignant cells.

Recent studies have shown that tumor-derived PTHrP drives adipose tissue browning and cachexia in tumor-bearing mice [[10], [11], [12]]. Consistent with these findings, our study revealed that PTHrP expression is significantly higher in bladder cancer tissues compared to normal bladder tissues. Moreover, we obtained evidence that bladder cancer-derived PTHrP induces browning of perivesical adipose tissue and further promotes lipolysis in adipocytes. Importantly, while similar effects were also observed using UM-UC-3 cells, the magnitude of adipocyte browning and lipolysis induced by UM-UC-3 was markedly weaker, consistent with its lower level of PTHrP secretion. Notably, PTHrP has been shown to stimulate UCP1 expression and enhance cellular respiration via a mechanism involving the protein kinase A (PKA) signaling cascade [[10], [12], [29]]. HSL, the key enzyme mediating lipolysis, is a well-characterized substrate of the PKA signaling pathway [[39], [40], [41]]. In line with this regulatory axis, we found that bladder cancer-derived PTHrP significantly promotes lipolysis in perivesical adipocytes, and the FFAs released from these adipocytes are subsequently taken up by bladder cancer cells.

In summary, our study reveals a critical regulatory loop: bladder cancer-derived PTHrP induces browning of perivesical adipose tissue and further promotes lipolysis in these adipocytes. The FFAs released through this process are subsequently taken up by bladder cancer cells, which in turn upregulates lipid metabolic activity within the malignant cells. This cascade of events collectively fuels bladder cancer progression.

Ethics approval and consent to participate

The Ethics Committee of Sun Yat-sen Memorial Hospital (SYSMH) approved this study (No. SYSKY-2025-963-01). As for experiments involving animals, the protocol was conducted with the regulations of the Animal Protection and Use Committee of SYSMH and approved by its Animal Experiment Ethics Committee (No. AP20250290).

Availability of supporting data

In this study, the scRNA-Seq data were deposited in the SRA database, and the Bioproject number is PRJNA662018. Transcriptome data and survival data of bladder cancer patients in different disease stages were retrieved from the TCGA-BLCA datasets from UCSC XENA (http://xena.ucsc.edu/) in this study.

Funding support

This study was funded by the National Key Research and Development Program of China (Grant No 2023YFC2507003, W.H.); the Noncommunicable Chronic Diseases-National Science and Technology Major Project (Grant No 2024ZD0525700, W.H.); the National Natural Science Foundation of China (Grant No 82573227, W.H.).

CRediT authorship contribution statement

Mingchao Gao: Writing – review & editing, Writing – original draft, Resources, Methodology, Investigation, Data curation. Chunni Li: Methodology, Investigation. Wenjie Li: Software, Methodology, Data curation. Junyi Xie: Software, Methodology. Juntian Long: Validation, Investigation, Formal analysis, Data curation. Mingli Luo: Methodology, Investigation. Jintao Hu: Investigation, Data curation. Cong Lai: Methodology. Tianhang Lan: Methodology. Dongxi Zhu: Methodology. Wenlong Zhong: Resources, Methodology. Wang He: Writing – review & editing, Supervision, Resources, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.