Bioinformatics analysis and clinical validation identify FBXW9 as a biomarker in bladder urothelial carcinoma

Introduction

Bladder urothelial carcinoma (BLCA) is the second most common malignancy of the urinary system worldwide, accounting for over 540,000 new cases and approximately 200,000 deaths annually. Globally, it ranks as the ninth most frequently diagnosed cancer and the thirteenth leading cause of cancer-related mortality (1). The current standard of care for BLCA involves surgical resection followed by adjuvant chemotherapy (2); however, the selection of chemotherapeutic regimens continues to rely heavily on empirical clinical judgment (3). In recent years, novel systemic therapies, including immune checkpoint inhibitors (e.g., pembrolizumab) and antibody-drug conjugates (e.g., enfortumab vedotin), have been increasingly utilized in patients with advanced disease who progress after platinum-based chemotherapy or are cisplatin-ineligible, offering new options for those who may not benefit from traditional treatments. Nevertheless, the efficacy of these novel therapies is highly heterogeneous. Clinical factors such as bone metastases and performance status are closely associated with treatment outcomes. This highlights the importance of precise patient selection (4-6). Although multiple prognostic classification systems, such as the Uromol-2016, Van-Kessel, and Seiler classifications have been proposed (7,8), the marked heterogeneity of BLCA presents considerable obstacles to consistent molecular subtyping. Thus, a deeper understanding of BLCA pathogenesis is essential. Robust prognostic biomarkers and novel therapeutic targets are also urgently needed. These will help improve diagnostic accuracy and treatment outcomes.

Ubiquitination is a crucial post-translational modification process that regulates protein homeostasis, carried out through the coordinated actions of E1 activating enzymes, E2 conjugating enzymes, and E3 ligases. Among the numerous E3 ligases, F-box proteins represent one of the most extensively studied classes (9). These proteins typically recognize phosphorylated substrates and mediate either proteasome-dependent ubiquitination (e.g., via lysine 48 or lysine 11 linkages) or non-proteasome-dependent ubiquitination (e.g., via lysine 63 linkages) (10). Based on structural features, F-box proteins are classified into three major subfamilies: FBXL, FBXO, and FBXW (11). Members of the FBXW subfamily all contain WD-40 domains, which play a key role in substrate recognition (12). Recent studies have shown that certain FBXW members can function as oncogenes or tumor suppressors depending on the cellular context. For example, FBXW1 plays a role in preserving the properties of cancer stem cells in glioblastoma through the targeting of GLI2 (13); in acute myeloid leukemia, FBXW4 is significantly expressed, and its expression level is linked to unfavorable patient prognosis (14); and FBXW7 regulates peritoneal metastasis in gastric cancer by mediating the ubiquitination of BGN protein (15).

Among the FBXW family members, FBXW9 has recently emerged as a particularly intriguing candidate. A comprehensive pan-cancer analysis revealed that FBXW9 expression was associated with prognostic alterations in 14 out of 41 cancer subtypes, ranking first among all FBXW family members, followed by FBXW5 (10/41) and FBXW4 (9/41) (16). The downregulation of FBXW9 in breast cancer cells results in the upregulation of p21, a target gene of TP53, and FBXW9-related genes are significantly enriched in functions related to the MYC signaling pathway (17). However, current research on the function of FBXW9 in BLCA remains unclear, and its potential role in the initiation and progression of the disease has not been defined.

This study aims to integrate bioinformatics analysis with immunohistochemical (IHC) experimental validation to systematically elucidate the expression pattern, prognostic value, and biological functions of FBXW9 in BLCA, as well as the related signaling pathways involved in tumorigenesis and progression. Furthermore, we investigate the profile of gene mutations and the status of DNA methylation in FBXW9, while also exploring its regulatory functions within the tumor immune microenvironment. This research contributes to establishing a novel theoretical basis and potential strategies for the clinical diagnosis and targeted treatment of BLCA. We present this article in accordance with the REMARK reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-1-1002/rc).

Methods

Data processing

RNA sequencing data from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) databases were integrated. All data were uniformly processed using the Toil pipeline, through which raw count data were converted into transcripts per million (TPM) format. The ComBat algorithm was applied to correct for batch effects between the TCGA and GTEx datasets (18). The effectiveness of batch effect removal was verified by principal component analysis (PCA). The RNA sequencing (RNA-seq) data and corresponding clinical information were downloaded via the GDC Data Portal on August, 2022. This retrospective analysis utilized the TCGA-BLCA cohort, which included patients diagnosed between 1999 and 2013. This dataset comprises 412 tumor samples associated with BLCA and 19 adjacent tissue samples (Table 1). To enhance the robustness of differential expression analysis, 9 normal bladder tissue samples from the GTEx database were also included, resulting in a combined analysis of 412 tumor samples and 28 normal samples (19 adjacent +9 GTEx). Messenger RNA (mRNA) expression datasets GSE13507 and GSE37815 were retrieved from the GEO database through the National Center for Biotechnology Information (NCBI) portal (https://www.ncbi.nlm.nih.gov/), and robust multi-array average (RMA) normalization was applied to these microarray datasets. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Survival analysis

Survival analysis was conducted using the TCGA_BLCA datasets, employing the Kaplan-Meier method. RNA-seq expression data and corresponding clinical information for BLCA patients were downloaded from the TCGA database. Inclusion criteria were: (I) pathologically confirmed BLCA; (II) available FBXW9 expression data; (III) available survival data [overall survival (OS) or disease-specific survival (DSS)]. Exclusion criteria were: (I) survival time of zero or missing; (II) incomplete key clinicopathological information [e.g., tumor-node-metastasis (TNM) stage, pathological grade]. A total of 412 BLCA patients were included in the survival analysis. For clinical variables with a missing rate of less than 5% (e.g., pathological grade), multiple imputation was applied. For variables with higher missing rates, no imputation was performed, and these cases were excluded from the corresponding analyses to avoid potential bias. The clinical follow-up information was updated until the data download date, resulting in a median follow-up time of 26.4 months. This study utilized OS and DSS as clinical endpoints. OS was defined as the duration from the date of initial pathological diagnosis to death from any cause, while DSS was defined as the duration from the date of initial pathological diagnosis to death specifically due to BLCA. In the DSS analysis, patients who died from causes other than BLCA or who remained alive at the last follow-up were treated as censored data. Univariate Cox proportional hazards regression was performed to assess the prognostic impact of FBXW9 expression (as a dichotomous variable) along with other clinicopathological variables. Variables yielding a P<0.10 in univariate analysis were subsequently incorporated into a multivariable Cox model to identify independent prognostic factors. The results of the regression analyses are presented as forest plots (generated with the ggplot2 package).

Differentially expressed gene analysis

Samples were divided into high and low FBXW9 expression groups based on the median FBXW9 expression in the TCGA BLCA dataset. Differential expression gene (DEG) analysis was performed using the “DESeq2” package in R, with the filtering threshold set as an adjusted P<0.05 and |log₂ fold change| >2. Spearman correlation analysis was employed to investigate the relationship between FBXW9 expression and the top five upregulated and downregulated DEGs. A volcano plot was used to visualize the differentially expressed genes (DEGs).

Functional enrichment analysis

The “clusterProfiler” package was utilized to perform analyses on the DEGs, specifically conducting Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) assessments. For the GSEA, the gene sets utilized included c2.cp.all.v2022.1.Hs.symbols.gmt and c5.all.v2022.1.Hs.symbols.gmt (19), A significance threshold was established with an adjusted P value of less than 0.05 and a false discovery rate (FDR) not exceeding 0.25. Visualization of the results was carried out using ggplot2.

Protein-protein interaction (PPI) network analysis

PPI analysis was performed via the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database (confidence score >0.7), constructing a PPI network containing FBXW9 and its interacting proteins (20).

Gene mutations and DNA methylation

The cBioPortal platform (21) was utilized to examine gene mutations and copy number variations (CNVs) of FBXW9, aiming to investigate the link between these genetic changes and the prognosis of BLCA. Additionally, the University of ALabama at Birmingham CANcer data analysis Portal (UALCAN) database (22), was employed to assess the methylation status of the FBXW9 promoter region, while the MethSurv tool (23) was used to evaluate the prognostic significance of the methylation levels of FBXW9.

Immune infiltration analysis

The GSVA utilized the single-sample GSEA (ssGSEA) algorithm to determine enrichment scores for 24 different immune cell types. Tumor purity was estimated using the ESTIMATE algorithm. Correlation analysis was conducted to investigate the relationship between FBXW9 expression and the levels of immune cell infiltration. Additionally, the Wilcoxon rank-sum test was employed to compare the differences in immune cell infiltration between groups with high and low FBXW9 expression levels (24).

Immunohistochemistry

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by institutional review board of TNMShanghai Zhuoli Biotechnology Co., Ltd. (Approval No. SHLLS-BA-22101102) ofand informed consent was taken from all the patients. The tissue microarray ZL-BlaU961 was purchased from Shanghai Zhuoli Biotechnology Co., Ltd., (Shanghai, China), containing 96 cores (1.5 mm in diameter) from 48 cases of BLCA and their matched adjacent non-tumor tissues. Detailed clinicopathological data, including TNM stage, were available for all cases. Tissue samples were stored at room temperature and protected from light after paraffin embedding. Hematoxylin-eosin (HE) staining and IHC analysis were performed to compare FBXW9 expression between tumor and adjacent non-tumor tissues. For IHC, a rabbit anti-human FBXW9 polyclonal antibody (catalog No. NBP2-97608, Novus Bio, Littleton, CO, USA) was used at a dilution of 1:100. IHC scoring was conducted using digital image analysis with the APTIME image analysis system (SODA Data Technology Inc., Shanghai, China). The system incorporates a tissue region recognition module and employs artificial intelligence (AI) deep learning algorithms. Through parameter configuration, it accurately identifies the cytoplasm and/or nucleus of each positively stained cell, precisely localizing target regions including the nucleus, cell membrane, and cytoplasm. Based on predefined positive staining thresholds, staining patterns in different tissue regions are categorized into four grades: negative, 1+, 2+, and 3+. The system then performs precise cell counting, positive area measurement, average staining intensity calculation, and H-Score computation. The H-Score was calculated as follows: H-SCORE = ∑ (pi × i) = (percentage of weak intensity × 1) + (percentage of moderate intensity × 2) + (percentage of strong intensity × 3), where “pi” represents the proportion of positive signal area or positive cells, and “i” represents the staining intensity grade. The H-Score ranges from 0 to 300, with higher scores indicating stronger overall staining intensity.

Statistical analysis

Data analysis was performed using IBM SPSS Statistics 26 and R software (version 4.2.1). Continuous variables were presented as median or mean ± standard deviation, while categorical variables were expressed as frequency (percentage). The significance of differences in FBXW9 expression between unpaired tissues was assessed using the Wilcoxon rank-sum test, whereas comparisons within paired tissues were performed using paired sample t-tests. In the analysis of associations between clinical characteristics and FBXW9 expression (dichotomized as high/low), continuous variables were evaluated with the Wilcoxon rank-sum test, and categorical variables were assessed using the Chi-squared test or Fisher’s exact test. Correlation analysis was conducted using Spearman’s rank correlation. To identify independent prognostic factors, a multivariate Cox proportional hazards regression model was constructed. Model construction details variables with P<0.10 in univariate analysis were included in the initial multivariate model. Backward stepwise regression was employed for variable selection, retaining variables with P<0.05 in the final model. Hypothesis testing: the proportional hazards assumption was verified by testing Schoenfeld residuals. GSEA was conducted on MSigDB hallmark and KEGG gene sets using GSEA software. An FDR <0.25 was considered statistically significant. Unless otherwise specified, all tests used a two-sided P<0.05 as the threshold for statistical significance.

Results

FBXW9 is Highly Expressed in BLCA

Pan-cancer analysis based on TCGA database demonstrated that FBXW9 was significantly overexpressed in several cancer types, including BLCA, breast cancer, cholangiocarcinoma, esophageal carcinoma, and renal cell carcinoma (Figure 1A). Compared to normal tissues, FBXW9 expression was markedly upregulated in tissues from patients with BLCA (P<0.001) (Figure 1B). Analysis of paired BLCA tissues and their corresponding adjacent non-tumor tissues showed significant overexpression of FBXW9 in the tumor tissues (P<0.001) (Figure 1C). This finding was further validated using the Gene Expression Omnibus (GEO) database, where both the GSE13507 and GSE37815 datasets consistently demonstrated significantly elevated FBXW9 expression in BLCA tissues (P<0.001) (Figure 1D,1E). High FBXW9 expression was significantly correlated with more advanced TNM stage, pathological stage, and histological grade in BLCA (P<0.05) (Figure 1F-1H, Figure S1A,S1B). Univariate logistic regression analysis indicated a significant association between FBXW9 expression and race [odds ratio (OR) =0.532, 95% confidence interval (CI): 0.308–0.916, P=0.02] (Table 2).

Figure 1 Expression patterns and clinicopathological associations of FBXW9 in BLCA. (A) Pan-cancer analysis of FBXW9 expression in tumor tissues compared with normal tissues across various cancer types. (B) FBXW9 expression in BLCA tissues versus unpaired normal bladder tissues from TCGA. (C) Paired analysis of FBXW9 expression in BLCA tissues and adjacent normal tissues from TCGA. (D,E) Validation of FBXW9 upregulation in BLCA using independent GEO datasets GSE13507 and GSE37815. (F-H) Association between FBXW9 expression levels and clinicopathological parameters including pathologic TNM stage. Statistical significance is indicated as ns, not significant; ***, P<0.001. BLCA, bladder urothelial carcinoma; GTEx, Genotype-Tissue Expression; TCGA, The Cancer Genome Atlas; TNM, tumor-node-metastasis; TPM, transcripts per million.

High FBXW9 expression is associated with poor prognosis in patients with BLCA

The Kaplan-Meier technique was employed to assess the relationship between FBXW9 expression levels and prognosis in BLCA. Patients were categorized into groups with either high or low expression of FBXW9 by utilizing the “surv_cutpoint” algorithm. The findings indicated that high levels of FBXW9 expression correlated significantly with diminished OS [hazard ratio (HR) =1.52, 95% CI: 1.11–2.07, P=0.008] and worsened DSS (HR =1.52, 95% CI: 1.06–2.20, P=0.02) (Figure 2A,2B). A deeper exploration into the differences in OS across various clinical subgroups indicated that individuals exhibiting high FBXW9 expression experienced poorer outcomes in several categories, such as females, whites, patients with a body mass index (BMI) of 25 kg/m² or lower, those presenting a high histological grade, instances of lymphovascular invasion, non-smokers, and those at pathological T3 stage (all P<0.05) (Figure S2A-S2G).

Figure 2 Predictive ability of FBXW9 for BLCA. Kaplan-Meier survival curves for OS (A) and DSS (B) in TCGA. (C,D,E) FBXW9 diagnostic ROC curves in the TCGA dataset, GSE13507, GSE37815. (F) The capacity of FBXW9 to predict OS at one, three, and five years was assessed using time-dependent ROC analysis of TCGA data. (G) A forest plot derived from univariate Cox analysis illustrating OS. AUC, area under the curve; BLCA, bladder urothelial carcinoma; BMI, body mass index; CI, confidence interval; CI, confidence interval; DSS, disease-specific survival; FPR, false positive rate; HR, hazard ratio; M, metastasis; N, node; OS, overall survival; PD, progressive disease; PR, partial response; ROC, receiver operating characteristic; SD, stable disease; TCGA, The Cancer Genome Atlas; TPR, true positive rate.

Diagnostic and prognostic value of FBXW9 in BLCA

The diagnostic efficacy of FBXW9 for BLCA was evaluated through receiver operating characteristic (ROC) analysis using the TCGA database. The results indicated that FBXW9 exhibited good predictive performance, with an area under the curve (AUC) value of 0.875 in the TCGA dataset (Figure 2C). Subsequent validation in the GSE13507 and GSE37815 datasets yielded AUC values of 0.717 and 0.972, respectively, confirming the diagnostic efficacy of FBXW9 (Figure 2D,2E). Further analysis of clinicopathological subgroups also confirmed the robust predictive ability of FBXW9 (Figure S3A-S3F). Additionally, time-dependent ROC curves derived from the TCGA database further validated FBXW9’s predictive utility for survival rates at 1, 3, and 5 years (Figure 2F).

Univariate Cox regression analysis was conducted to investigate the association between FBXW9 expression and the prognosis of BLCA. The results of the univariate analysis indicated that high FBXW9 expression (high vs. low, HR =1.347, P=0.047), lymphovascular invasion (yes vs. no, HR =2.247, P<0.001), and pathological M stage (M1 vs. M0, HR =3.112, P=0.002) were significantly correlated with poorer OS in patients with BLCA (Figure 2G). However, the multivariable Cox regression analysis revealed that FBXW9 was not an independent prognostic factor (Table S1).

Identification of DEGs and functional enrichment analysis in BLCA

A total of 99 DEGs were identified between the high and low FBXW9 expression groups, including 28 upregulated and 71 downregulated genes (adjusted P<0.05 and |log₂ fold change| >2) (Figure 3A). Based on adjusted P values, the top five most significantly upregulated and downregulated DEGs (PRSS56, SOX14, FGF3, H3Y1, ATOH1, PAGE2, LINC02512, PNMA5, SST, GPC3-AS1) were selected for further analysis, and their correlation with FBXW9 expression was visualized using a heatmap (Figure 3B). Functional annotation of the DEGs through GO and KEGG enrichment analyses revealed that the primary biological processes involved cellular calcium ion homeostasis, humoral immune response, and monocyte chemotaxis. The main cellular components were associated with the external side of the plasma membrane, intrinsic components of the postsynaptic membrane, and integral components of the postsynaptic membrane. Key molecular functions included signaling receptor activator activity, receptor ligand activity, and C-C chemokine receptor (CCR) chemokine receptor binding. KEGG analysis indicated that these DEGs were predominantly enriched in pathways such as neuroactive ligand-receptor interaction, cytokine-cytokine receptor interaction, and viral protein interaction with cytokine and cytokine receptor (Figure 3C,3D).

Figure 3 Examination of DEGs and functional enrichment related to FBXW9 in BLCA. (A) Volcano plot illustrating the DEGs between groups with high and low FBXW9 expression. (B) Heatmap depicting the top five genes that are either upregulated or downregulated in relation to FBXW9 levels. (C) Bubble plot for GO and KEGG enrichment assessment. (D) Circular diagram representing the GO and KEGG terms associated with the identified DEGs. (E) Results of GSEA gene set enrichment in the high FBXW9 expression group. (F) Results of GSEA gene set enrichment in the low FBXW9 expression group. *, P<0.05; **, P<0.01. BLCA, bladder urothelial carcinoma; BP, biological process; CC, cellular component; DEGs, differentially expressed genes; FC, fold change; GO, Gene Ontology; GSEA, Gene Set Enrichment Analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; ns, not significant; TPM, transcripts per million.

We performed GSEA to explore the association between FBXW9 expression and various biological processes. The results indicated significant correlations with processes such as polyamine metabolism, G2/M phase progression, proteasome degradation, DNA synthesis, and glycolysis (Figure 3E, Figure S4A). Conversely, low FBXW9 expression was significantly associated with processes and pathways including the core matrisome, collagen degradation, extracellular matrix glycoproteins, complement cascade, and coagulation cascade (Figure 3F, Figure S4B).

PPI network analysis and ferroptosis-related gene analysis

A network of interactions between proteins related to genes associated with FBXW9 was developed utilizing the STRING database. The analysis identified CAND1, CCT3, CCT4, CCT5, CCT7, CCT8, CUL1, FBXL5, FBXW2, FBXW4, FBXW5, FBXW10, NEDD8, RBX1, and SKP1 as the genes most significantly associated with FBXW9 (Figure 4A,4B). Comparative expression analysis of these genes in BLCA revealed that in tumor tissues, the expression of CCT3, CCT4, CCT5, CCT7, CCT8, NEDD8, and RBX1 was significantly higher than in adjacent non-tumor tissues, while the expression of genes such as FBXL5, FBXW4, and SKP1 was significantly lower (Figure 4C). Correlation analysis further demonstrated the relationships among these FBXW9 related genes (Figure 4D). Subsequent prognostic analysis indicated that high expression of most of these genes was associated with poor prognosis in patients with BLCA (Figure S5A-S5O). Further analysis using GO and KEGG pathways revealed that FBXW9-related genes are primarily involved in biological functions and pathways such as protein localization regulation, ATP hydrolysis, unfolded protein binding, protein folding, TGF-β signaling pathway, circadian rhythm, and cell cycle (Figure 4E).

Figure 4 FBXW9-related genes and their functional analysis. (A) PPI network of FBXW9‐related genes. (B) Annotation and correlation coefficients of 15 FBXW9-related genes. (C) Expression of FBXW9-related genes in BLCA. (D) Correlation between FBXW9 and related genes. (E) GO/KEGG functional enrichment analysis of FBXW9-related genes. ns: P≥0.05; *, P<0.05; **, P<0.01; ***, P<0.001. BLCA, bladder urothelial carcinoma; BLCA, bladder urothelial carcinoma; BP, biological process; CC, cellular component; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; ns, not significant; PPI, protein-protein interaction.

Ferroptosis is increasingly recognized as a critical mechanism in BLCA progression, and emerging evidence has identified multiple F-box family proteins—including FBXW5, FBXW7, FBXO10, and FBXO39—as direct regulators of ferroptosis signaling pathways in various cancers (25-28). As FBXW9 belongs to the same protein family, we explored its potential association with ferroptosis. Based on the TCGA database, the association between FBXW9 and its ferroptosis-related genes was invested. The results showed that FBXW9 expression was significantly positively correlated with several key ferroptosis-related genes, including ATP5MC3, CARS1, CS, EMC2, FANCD2, FDFT1, GPX4, HSPA5, LPCAT3, MT1G, RPL8, SLC1A5, and TFRC (Figure 5A,5B). It should be noted that the above correlation analysis between FBXW9 and ferroptosis-related genes is exploratory in nature, intended to provide preliminary insights into the potential biological functions of FBXW9 rather than to establish a direct regulatory mechanism.

Figure 5 Relationship of FBXW9 with genes associated with the ferroptosis pathway. (A) Heatmap illustrating the correlation between FBXW9 levels and genes related to ferroptosis. (B) Levels of ferroptosis-related genes classified by high and low FBXW9 expression groups. *, P<0.05; **, P<0.01; ***, P<0.001.

FBXW9 gene mutation and methylation analysis

Based on two BLCA datasets from the cBioPortal platform (BGI, Nat Genet 2013; Cancer Discov 2014; MSKCC, JCO Precis Oncol 2024; TCGA, Firehose Legacy), we analyzed FBXW9 gene mutations and CNVs. The results showed that the FBXW9 gene exhibited alterations such as missense mutations, amplifications, and splice site mutations, with a frequency of 0.5% (Figure 6A). Further Kaplan-Meier survival analysis indicated that FBXW9 gene mutations were not correlated with OS (P=0.81) or disease-free survival (P=0.77) in patients (Figure 6B,6C). Additionally, multiple studies have confirmed the important role of abnormal DNA methylation in the early development of BLCA. An extensive examination of the DNA methylation levels of FBXW9 and the prognostic implications of CpG islands was performed utilizing the MethSurv database. The results indicated that the majority of CpG sites displayed a state of hypomethylation (Figure 6D). An examination of the UALCAN database indicated that the levels of DNA methylation in the promoter region of FBXW9 were reduced in BLCA tissues in comparison to normal tissues (P=0.002) (Figure 6E). Among them, low methylation levels at five CpG sites cg12110584, cg21564527, cg19721801, cg05868316 and cg02753619 were correlated with prognosis in BLCA (Figure 6F-6I). To investigate the potential epigenetic mechanism underlying FBXW9 upregulation, we analyzed the correlation between FBXW9 CpG methylation and its mRNA expression in the TCGA-BLCA cohort. A significant negative correlation was observed (r=−0.131, P=0.007), suggesting that CpG hypomethylation may contribute to the transcriptional upregulation of FBXW9 in BLCA (Figure S6).

Figure 6 The impact of mutations and DNA methylation levels of FBXW9 on the prognosis of BLCA. (A) Levels of mutation for FBXW9 as presented in the cBioPortal OncoPrint. (B) The relationship between mutations in the FBXW9 gene and OS in BLCA patients. (C) The connection between FBXW9 gene mutations and DFS in BLCA cases. (D) FBXW9 methylation levels in BLCA as reported by the UALCAN database. (E) The relationship between FBXW9 mRNA expression levels and methylation levels according to the MethSurv database. (F) Analysis of the correlation between FBXW9 methylation levels and prognosis in BLCA. The Kaplan-Meier survival curve for FBXW9 in (G) cg12110584, (H) cg21564527, (I) cg19721801. BGI, Beijing Genomics Institute; BLCA, bladder urothelial carcinoma; DFCI/MSK, Dana-Farber Cancer Institute/Memorial Sloan Kettering Cancer Center; DFS, disease-free survival; HR, hazard ratio; mRNA, messenger RNA; MSKCC, Memorial Sloan Kettering Cancer Center; OS, overall survival; TCGA, The Cancer Genome Atlas; UALCAN, University of ALabama at Birmingham CANcer data analysis Portal.

FBXW9 expression is significantly associated with immune infiltration

Earlier research has highlighted the significance of immune cells within tumors in human malignancies (29). In this investigation, we examined the relationship between FBXW9 and the presence of 24 different types of immune cells infiltrating the tumor microenvironment by employing the ssGSEA algorithm. Comparison of the enrichment scores of the 24 immune cell types between the high and low FBXW9 expression groups revealed that the high-expression group had higher enrichment scores for Th2 cells, while the low-expression group showed higher enrichment scores for activated dendritic cells (aDCs), B cells, cytotoxic cells, DCs, immature dendritic cells (iDCs), mast cells, plasmacytoid dendritic cells (pDCs), T cells, and TFH cells (Figure 7A,7B). FBXW9 expression showed a significant but modest correlation with the infiltration levels of Th2 cells (r=0.248, P<0.001) and γδT cells (r=0.131, P=0.008) (Figure 7C,7D), while it was negatively correlated with pDCs (r=−0.209, P<0.001), iDCs (r=−0.155, P=0.002), DCs (r=−0.133, P=0.007), T cells (r=−0.130, P=0.008), mast cells (r=−0.111, P=0.02), and aDCS (r=−0.111, P=0.03) (Figure 7A,7E-7J).

Figure 7 Relationship between FBXW9 expression and immune cell infiltration. (A) A bubble plot illustrating the connection between FBXW9 and 24 types of immune cells. (B) Levels of immune infiltration among various immune cell types associated with high and low FBXW9 expression. Scatter plots showing the relationship between FBXW9 expression levels and (C) Th2 cells, (D) Tgd cells, (E) pDC cells, (F) iDC cells, (G) DC cells, (H) T cells, (I) Mast cells, and (J) aDC cells. ns: P≥0.05; *, P<0.05; **, P<0.01; ***, P<0.001. ns, not significant; TPM, transcripts per million.

Immunohistochemical (IHC) validation of high FBXW9 expression in BLCA

To validate the reliability of the bioinformatics analysis results, this study collected BLCA tissues and adjacent normal tissue samples, and detected the expression of FBXW9 at the protein level through IHC analysis (Figure 8A). The IHC staining scoring results showed that, compared with adjacent normal tissues, the expression of FBXW9 was significantly upregulated in BLCA tumor tissues (Figure 8B,8C). These experimental results are consistent with the conclusions from the earlier bioinformatics analysis, further confirming the high expression characteristics of FBXW9 in BLCA at the protein level.

Figure 8 Validation of FBXW9 expression in BLCA patients. (A) Representative IHC images of FBXW9 expression in BLCA and adjacent tissues. (B,C) IHC score of immunohistochemical staining of FBXW9 in BLCA and adjacent liver tissues. ***, P<0.001. Bar indicates 500 μm. BLCA, bladder urothelial carcinoma; IHC, immunohistochemical.

Discussion

BLCA has a high recurrence rate, a high risk of progression, and marked heterogeneity. Its prevention and treatment remain a major challenge in urological oncology. According to the Cancer Statistics Report 2024, the overall incidence of BLCA in males is gradually declining. However, the proportion of advanced cases remains high. In contrast, the incidence and disease-specific mortality in female patients have steadily increased over the past decade (1). Clinically, most patients are diagnosed at muscle-invasive or locally advanced stages. This severely limits treatment options and impairs long-term survival (2). The current lack of highly sensitive and specific early screening tools largely contributes to this dilemma. Therefore, developing novel biomarker systems for early warning, stage stratification, and prognosis assessment is crucial for improving clinical management of BLCA.

Integrated analysis of multi-omics data from TCGA and GEO databases revealed that FBXW9 is significantly overexpressed in multiple cancer types. These include BLCA, breast cancer, cholangiocarcinoma, esophageal carcinoma, and renal cell carcinoma. In independent BLCA cohorts, FBXW9 expression was consistently upregulated in tumor tissues. This was observed in both paired and unpaired comparisons. FBXW9 expression positively correlated with clinicopathological parameters such as tumor grade and lymph node metastasis. Patients with high FBXW9 expression had significantly shorter OS. Notably, ROC curve analysis indicated that FBXW9 has excellent diagnostic performance for BLCA (AUC >0.85). These findings suggest that FBXW9 may serve as a novel molecular marker for this malignancy, providing a potential target for early diagnosis and prognostic assessment.

We further explored the biological functions and related pathways of FBXW9. GO and KEGG enrichment analyses showed that FBXW9 is primarily involved in several biological processes. These include cellular calcium ion homeostasis, humoral immune response, and monocyte chemotaxis. Related genes were enriched in cellular components such as the external side of the plasma membrane and the postsynaptic membrane. They were also associated with molecular functions including signaling receptor activation, receptor ligand binding, and CCR chemokine receptor binding. KEGG pathway analysis revealed that DEGs are significantly enriched in signaling networks. These include neuroactive ligand-receptor interaction and cytokine-cytokine receptor interaction.

PPI network analysis suggested that FBXW9 may contribute to BLCA progression. It may regulate biological processes such as protein localization, ATP hydrolysis, unfolded protein binding, protein folding, TGF-β signaling, circadian rhythm, and the cell cycle. Further immune correlation analysis revealed that FBXW9 expression is significantly positively correlated with the infiltration of Th2 cells and γδT cells. In contrast, it was negatively correlated with the infiltration of pDCs and iDCs. The observed correlations between FBXW9 expression and immune cell infiltration were modest in magnitude (r values ranging from −0.209 to 0.248). These findings are exploratory and hypothesis-generating rather than mechanistic. Future functional studies are warranted to elucidate the underlying mechanisms.

Immune checkpoint inhibitors have significantly transformed the treatment landscape for advanced urothelial carcinoma. However, accurately identifying patients most likely to benefit remains a major clinical challenge. A multicenter real-world study based on the ARON-2 database validated an immunotherapy prognostic score. This score incorporates sex, performance status, and liver metastases. It demonstrated the ability to effectively stratify patient outcomes (30). In the present study, we observed that high FBXW9 expression was significantly correlated with the infiltration levels of multiple immune cell types in BLCA tissues. This was particularly evident for Th2 cells and γδT cells. In light of the aforementioned real-world evidence, we hypothesize that FBXW9 expression level may serve as a potential predictive biomarker for immunotherapy response. This may be especially relevant for patients with bone metastases or poor performance status. Future prospective cohort studies are warranted to further explore the association between FBXW9 expression and clinical outcomes following immunotherapy.

The management of immune-related adverse events represents another critical aspect of clinical practice. A meta-analysis of 16 studies (including 4,658 patients) systematically evaluated the association between sex and the risk of immune-related adverse events (31). The overall analysis did not reveal a statistically significant difference in the incidence of these events between males and females. However, the study suggested that sex differences may manifest in the type and severity of adverse events. This highlights the necessity of implementing sex-based individualized monitoring strategies in immunotherapy management. This finding further reinforces the importance of personalized treatment from a safety perspective. In this study, we further investigated the sex-specific role of FBXW9 expression in BLCA. Among female patients, the high FBXW9 expression group exhibited significantly worse prognosis (P=0.009). No such association was observed in male patients. This finding suggests that the prognostic value of FBXW9 may be sex-dependent. Its biological functions may be mediated through distinct signaling pathways in male and female patients.

Considering the aforementioned meta-analysis evidence demonstrating sex differences in immune-related adverse events, we speculate that FBXW9 expression levels may not only correlate with tumor progression. They may also potentially influence the tolerability and response to immunotherapy in a sex-specific manner. However, the relationship between FBXW9 expression and clinical outcomes of immunotherapy remains unclear. Its association with the risk of adverse events also requires further investigation. Future studies are warranted to further explore the sex-specific biological functions of FBXW9. They should also evaluate its potential as a predictive biomarker for immunotherapy efficacy or safety. This could provide novel strategies for personalized precision medicine in BLCA.

Although this study systematically evaluated the diagnostic and prognostic value of FBXW9 in BLCA, several limitations should be acknowledged. The analysis integrated multiple databases with IHC validation. First, this study was based on retrospective analyses of publicly available databases. These may introduce inherent selection bias and sample heterogeneity. This could affect the generalizability of the findings. Second, univariate analysis revealed a significant association between FBXW9 expression and patient prognosis. However, this association did not achieve statistical significance as an independent prognostic factor in multivariable Cox regression analysis. This may be due to collinearity between FBXW9 expression and established clinicopathological factors such as tumor stage and grade. Its prognostic information may be partially captured by these well-recognized indicators. Additionally, the limited sample size of the TCGA cohort may have constrained the statistical power of the multivariable analysis. Therefore, the prognostic value of FBXW9 should be interpreted with caution. Its independent predictive ability warrants further validation in larger, multicenter prospective cohorts. Third, the ComBat algorithm was applied to correct for batch effects when integrating the TCGA and GTEx datasets. However, potential batch differences across platforms and research centers may not have been completely eliminated. This could still have influenced the results. Fourth, current analyses of FBXW9 expression and function rely primarily on validation from online databases and clinical samples. They lack corroboration from in vitro cellular experiments and in vivo animal models. Future investigations should strengthen relevant functional experiments to verify the biological significance of FBXW9 in BLCA. Fifth, the analyses of immune infiltration and ferroptosis-related genes were largely correlative in nature. They lacked functional validation. Thus, these findings should be considered hypothesis-generating rather than mechanistic evidence.

Furthermore, high FBXW9 expression is associated with poor prognosis. This may be attributed to its biological function as an E3 ubiquitin ligase. FBXW9 may influence the progression of BLCA by promoting the degradation of certain tumor suppressors. It may also stabilize specific oncogenic proteins. Future functional experiments are warranted to validate the specific downstream substrates of FBXW9. They should also elucidate its precise mechanisms of action in BLCA.

Conclusions

In summary, the findings of this study provide preliminary evidence for FBXW9 as a potential biomarker in BLCA. However, its clinical utility warrants further validation. This includes larger, prospective, multicenter cohorts and in-depth mechanistic investigations.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the REMARK reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-1-1002/rc

Data Sharing Statement: Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-1-1002/dss

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-1-1002/prf

Funding: This study was supported by Administration of Basic and Applied Basic Research Foundation of Guangdong Province (No. 2022A1515220194), Guangdong Province Science and Technology Innovation Strategy Special Fund (No. pdjh2023b1059), Guangdong University Innovation Team Project (Natural Science 2024KCXTD058), Medical Research Fund of Guangdong Province (Nos. A2023307, B2023280, and A2024543), the Scientific Research Fund of The First People’s Hospital of Zhaoqing (Nos. YJJ-2023-02-04 and YJJ-2025-01-008), and Zhaoqing Medical College Fund for Young Talent (Nos. Zqyq22-005 and Zqyq22-007).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-1-1002/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All analyses of human samples were carried out in compliance with the relevant ethical regulations and approved by Shanghai Zhuoli Biotech Co., Ltd (No. SHLLS-BA-22101102) and informed consent was taken from all the patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Antoni S, Ferlay J, Soerjomataram I, et al. Bladder Cancer Incidence and Mortality: A Global Overview and Recent Trends. Eur Urol 2017;71:96-108. [Crossref] [PubMed]
2. Compérat E, Amin MB, Cathomas R, et al. Current best practice for bladder cancer: a narrative review of diagnostics and treatments. Lancet 2022;400:1712-21. [Crossref] [PubMed]
3. Alifrangis C, McGovern U, Freeman A, et al. Molecular and histopathology directed therapy for advanced bladder cancer. Nat Rev Urol 2019;16:465-83. [Crossref] [PubMed]
4. Rizzo A, Myint ZW, Ciccimarra F, et al. Bone metastases in advanced urothelial carcinoma patients receiving enfortumab vedotin: A Real-World study from the ARON-2(EV) project. Clin Exp Med 2025;26:75. [Crossref] [PubMed]
5. Massari F, Santoni M, Takeshita H, et al. Global real-world experiences with pembrolizumab in advanced urothelial carcinoma after platinum-based chemotherapy: the ARON-2 study. Cancer Immunol Immunother 2024;73:106. [Crossref] [PubMed]
6. Santoni M, Myint ZW, Büttner T, et al. Real-world effectiveness of pembrolizumab as first-line therapy for cisplatin-ineligible patients with advanced urothelial carcinoma: the ARON-2 study. Cancer Immunol Immunother 2023;72:2961-70. [Crossref] [PubMed]
7. Seiler R, Gibb EA, Wang NQ, et al. Divergent Biological Response to Neoadjuvant Chemotherapy in Muscle-invasive Bladder Cancer. Clin Cancer Res 2019;25:5082-93. [Crossref] [PubMed]
8. van Kessel KEM, van der Keur KA, Dyrskjøt L, et al. Molecular Markers Increase Precision of the European Association of Urology Non-Muscle-Invasive Bladder Cancer Progression Risk Groups. Clin Cancer Res 2018;24:1586-93. [Crossref] [PubMed]
9. Deshaies RJ. SCF and Cullin/Ring H2-based ubiquitin ligases. Annu Rev Cell Dev Biol 1999;15:435-67. [Crossref] [PubMed]
10. Frescas D, Pagano M. Deregulated proteolysis by the F-box proteins SKP2 and beta-TrCP: tipping the scales of cancer. Nat Rev Cancer 2008;8:438-49. [Crossref] [PubMed]
11. Tekcham DS, Chen D, Liu Y, et al. F-box proteins and cancer: an update from functional and regulatory mechanism to therapeutic clinical prospects. Theranostics 2020;10:4150-67. [Crossref] [PubMed]
12. Zhang Q, Mady ASA, Ma Y, et al. The WD40 domain of FBXW7 is a poly(ADP-ribose)-binding domain that mediates the early DNA damage response. Nucleic Acids Res 2019;47:4039-53. [Crossref] [PubMed]
13. Wei P, Jiang J, Xiao M, et al. The transcript ENST00000444125 of lncRNA LINC01503 promotes cancer stem cell properties of glioblastoma cells via reducing FBXW1 mediated GLI2 degradation. Exp Cell Res 2022;412:113009. [Crossref] [PubMed]
14. Han Q, Zhang Q, Song H, et al. FBXW4 Is Highly Expressed and Associated With Poor Survival in Acute Myeloid Leukemia. Front Oncol 2020;10:149. [Crossref] [PubMed]
15. Huang G, Xiang Z, Wu H, et al. The lncRNA SEMA3B-AS1/HMGB1/FBXW7 Axis Mediates the Peritoneal Metastasis of Gastric Cancer by Regulating BGN Protein Ubiquitination. Oxid Med Cell Longev 2022;2022:5055684. [Crossref] [PubMed]
16. Yu S, Liang Z, Fan Z, et al. A Comprehensive Analysis Revealing FBXW9 as a Potential Prognostic and Immunological Biomarker in Breast Cancer. Int J Mol Sci 2023;24:5262. [Crossref] [PubMed]
17. Zhao H, Feng K, Lei J, et al. Identification of somatic mutation-driven enhancers and their clinical utility in breast cancer. iScience 2024;27:108780. [Crossref] [PubMed]
18. Vivian J, Rao AA, Nothaft FA, et al. Toil enables reproducible, open source, big biomedical data analyses. Nat Biotechnol 2017;35:314-6. [Crossref] [PubMed]
19. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
20. Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 2019;47:D607-13. [Crossref] [PubMed]
21. Gao J, Aksoy BA, Dogrusoz U, et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 2013;6:pl1. [Crossref] [PubMed]
22. Chandrashekar DS, Bashel B, Balasubramanya SAH, et al. UALCAN: A Portal for Facilitating Tumor Subgroup Gene Expression and Survival Analyses. Neoplasia 2017;19:649-58. [Crossref] [PubMed]
23. Modhukur V, Iljasenko T, Metsalu T, et al. MethSurv: a web tool to perform multivariable survival analysis using DNA methylation data. Epigenomics 2018;10:277-88. [Crossref] [PubMed]
24. Bindea G, Mlecnik B, Tosolini M, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 2013;39:782-95. [Crossref] [PubMed]
25. Yang L, Wang D, Yu N, et al. FBXW7 promotes osteoarthritis injury by regulating SLC7A11 ubiquitination degradation and chondrocyte ferroptosis. Pathol Res Pract 2026;277:156297. [Crossref] [PubMed]
26. Li T, Wang K, Chen Y, et al. FBXO39 knockdown promotes spermatogenesis impairment by inducing mitochondrial dysfunction and ferroptosis via inhibiting KDM5A ubiquitination and regulating H3K4me3 demethylation. Cell Biol Toxicol 2026;42:50. [Crossref] [PubMed]
27. Chen C, Li H, Zhang Z, et al. F-box/WD Repeat-Containing Protein 5 Promotes Breast Cancer Progression by Regulating Ferroptosis via Enhancing Krüppel-like Factor 13 Ubiquitination Through Phosphoinositide 3-Kinase/Serine/Threonine Protein Kinase Pathway. Rejuvenation Res 2025;28:226-38. [Crossref] [PubMed]
28. Hu J, Ma F, Chen X, et al. F-box only protein 10 protects against kidney tubulointerstitial fibrosis by inhibiting ACSL4-mediated lipid peroxidation and ferroptosis. Cell Signal 2025;132:111841. [Crossref] [PubMed]
29. Espinosa-Carrasco G, Chiu E, Scrivo A, et al. Intratumoral immune triads are required for immunotherapy-mediated elimination of solid tumors. Cancer Cell 2024;42:1202-1216.e8. [Crossref] [PubMed]
30. Rizzo A, Zayed A, Vitale E, et al. Prognostic immunotherapy score (PIS) in patients with advanced urothelial carcinoma treated with pembrolizumab: real-world data and validation from ARON-2 dataset. Clin Exp Metastasis 2025;42:52. [Crossref] [PubMed]
31. Vitale E, Rizzo A, Maistrello L, et al. Sex differences in adverse events among cancer patients receiving immune checkpoint inhibitors: the MOUSEION-07 systematic review and meta-analysis. Sci Rep 2024;14:28309. [Crossref] [PubMed]