Skip to main content
NCBI home page
Discover Oncology logoLink to Discover Oncology
. 2026 Mar 23;17:674. doi: 10.1007/s12672-026-04851-2

Construction and evaluation of a bladder cancer prognosis model based on super-enhancer-associated genes

Lieyu Xu 1,2,#, Zhenhao Zeng 2,#, Xinchang Zou 1,#, Zunwei Zhu 2, Tao Zeng 1,3,✉
PMCID: PMC13133299  PMID: 41870815

Abstract

Introduction

Bladder cancer (BLCA) is a malignant tumour that occurs on the mucosa of the bladder. It accounts for the first place in the incidence of genitourinary tumours in China. BLCA is characterized by high recurrence rate and poor survival rate. There is still a research gap regarding super-enhancer-related genes (SERGs) in BLCA.

Methods

The The Cancer Genome Atlas Bladder Urothelial Carcinoma (TCGA-BLCA) and GSE31684 were subjected into this study. In addition, the Super-Enhancer Archive database was used to identify SERGs. Differential expression analysis was used to analyse the differentially expressed genes (DEGs) between the BLCA and control groups. The DEGs were overlapped with SERGs to get candidate genes in TCGA-BLCA, which were analyzed for Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG). Univariate Cox, Least Absolute Shrinkage and Selection Operator (Lasso) regression and multivariate Cox regression analyses were used to build the risk model for BLCA. Survival analyses and validation of the model were performed by Kaplan-Meier (K-M) curves and Receiver Operating Characteristic (ROC) curve, respectively. In addition, using the estimating relative subsets of RNA transcripts (CIBERSORT) algorithm, 22 immune cell proportions were calculated. The drug sensitivity was also analyzed in this study.

Results

First of all, based on the TCGA-BLCA, 70 DE-SERGs were yielded. A prognosis model based on MXRA7, PLEKHG4B and ATP2B4 was finally constructed. ROC curves revealed that the prognosis model was a good predictor of BLCA outcomes. Immune infiltration analysis revealed that risk score was positively associated with T cells CD4 memory resting, Mast cells resting and Macrophages M2 and negatively associated with Dendritic cells activated and T cells CD8. Besides, AZD8186, BMS-754,807, JQ1, KRAS (G12C) Inhibitor and NU7441 were the top five sensitivity drugs for BLCA.

Conclusion

Three genes (MXRA7, PLEKHG4B and ATP2B4) were identified to construct a SERG-related model in BLCA, which provides a basis for understanding BLCA pathogenesis and new insights into BLCA treatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-026-04851-2.

Keywords: Bladder cancer, Super-enhancer-related genes, Prognosis model, GEO, TCGA

Introduction

Bladder cancer is one of the common malignant tumors of the urinary system, causing morbidity and mortality that have been of significant concern worldwide [1]. Global cancer statistics for 2020 indicate that bladder cancer was the 10th most commonly diagnosed malignancy [2]. American Cancer Society statistics show that 82,290 new cases of BC are expected to occur in 2023, causing approximately 16,710 new deaths [3]. Bladder cancer can be divided into two types, including non-muscle invasive and muscle invasive bladder cancer [4]. The primary treatment for non-muscle invasive bladder cancer (NMIBC) is transurethral resection of bladder tumor (TURBT), but tumor control results are unsatisfactory, with recurrence rates of 15–61% at 1 year and 31–78% at 5 years [5]. Muscle-invasive bladder cancer (MIBC) is significantly more malignant and has an unfavorable prognosis [6], with a 5-year survival rate of only 5% in patients with distant metastasis [7]. Thus, the search for new biomarkers is necessary to increase survival and improve the prognosis of bladder cancer patients.

Super-enhancers, first discovered in mouse embryonic stem cells (ESCs) in 2013, are clusters of enhancers located near genes that regulate gene expression. Subsequent research has shown that super-enhancers play a crucial role in human cells (health and disease) [8, 9]. Recently, super-enhancers have been identified as playing an important role in cancer initiation and progression [10]. Super-enhancers can regulate tumor progression through a variety of modulations, such as activation of oncogenic signaling pathways, modulation of chromatin regulators, and regulation of transcription [11, 12]. Wu et al. [13] established a breast cancer prognostic model based on six super-enhancer-related genes (ZIC2, NFE2, FOXJ1, KLF15, POU3F2 and SPIB) and verified that downregulation of the ZIC2 gene inhibited breast cancer cell proliferation and migration. In acute myeloid leukemia (AML), the super-enhancer-associated gene CAPG may promote AML disease progression by regulating the NF-κB signaling pathway [14]. The super-enhancer-related gene TMSB10 was found to be involved in epithelial-mesenchymal transition in lung adenocarcinoma [15]. These studies show that super-enhancer-related genes are expected to be potential biomarkers for tumor diagnosis and therapy. Therefore, we constructed a bladder cancer prognostic model based on super-enhancer-related genes to provide a theoretical basis for the clinical treatment of bladder cancer.

In this study, we screened DE-SERGs in bladder cancer by using multiple databases and constructed a prognostic model of DE-SERGs in bladder cancer. It was further analyzed in relation to the immune microenvironment and drug sensitivity. Our study may help provide new insights into the prognostic assessment and molecular mechanisms of bladder cancer.

Materials and methods

Data source

The TCGA-BLCA dataset was extracted from the Cancer Genome Atlas database (https://www.cancer.gov/ccg/), which contained 405 BLCA and 19 control samples. The GSE31684 dataset had been extracted from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), including 53 BLCA samples with complete overall survival (OS). Besides, we collected 695 SERGs from Super-Enhancer Archive (SEA) database (http://sea.edbc.org).

Identification of differentially expressed super enhancer related genes (DE-SERGs) in BLCA

Firstly, we employed the “DESeq2” in R to DEGs in the TCGA-BLCA dataset, with the screening criteria: P-adj < 0.05 and |log2FC| > 2 [16]. Then, DE-SERGs were obtained by overlaping DEGs and SERGs by the ggVennDiagram [17]. We performed enrichment analysis by the “clusterProfiler” in R in order to identify related functions and pathways for DE-SERGs in TCGA-BLCA based on GO and the KEGG datebase [18].

Construction risk model in BLCA

The TCGA-BLCA sample were used as a training set (N = 424) to build a model to predict survival in BLCA patients, and GSE84433 (N = 53) was used as an external test set. The risk model of BLCA was established by univariate Cox, Least absolute shrinkage and selection operator (Lasso) regression and multivariate cox regression analyses in TCGA-BLCA dataset [18]. To plot the expression of prognostic genes in the TCGA-BLCA dataset, the “ggplot2” package in R was used [19]. The performance of prognostic risk model was assessed by Kaplan-Meier (K-M) survival curves [20], ROC and risk curves in TCGA-BLCA [21]. In the same way, the K-M, ROC and risk curves were used in testing set to verify the risk model.

Independent prognostic analysis and nomogram construction

The screening of independent prognostic factors through univariate and multivariate Cox analysis of clinical pathological factors (age, grade, and TNM stage) and risk score. Moreover, a nomogram was created using the “rms” in R [22], and which clinical utility was evaluateed by Calibration curves, ROC curves, and decision curve analysis (DCA) curve [23].

Correlation analysis in clinical traits and risk score

Firstly, we performed the rank sum test to examine differences in clinical trait groups and risk scores [24]. Then, survival analyses were performed using “survival” in R for different clinical traits [25]. In addition, the expression heatmap of prognostic genes in different risk groups and different clinical features was mapped using “pheatmap” in R [26]. Besides, the sample KEGG pathway scores were calculated in the training set using the “GSVA” in R package [27].

Analysis of immune microenvironment in TCGA-BLCA

The high and low risk group ESTIMATE analyses were performed by the ESTIMATE algorithms [28], and the variability between two risk groups was assessed using the rank-sum test [29]. Besides, the correlation between these scores and risk scores was also analyzed in this study. Then, a total of 22 immune cell scores were calculated for all BLCA samples by applying CIBERSORT algorithm [30], and which the difference was compared through t-test between different risk group. The differences between risk groups in the expression of 42 immune checkpoints were assessed. Finally, the TIDE score was calculated to analyse the response to immunotherapy in different risk groups [31].

Drug sensitivity analysis

The drug sensitivity in the TCGA-BLCA sample was investigated. The oncePredict” and T-test were used for the calculation of the half maximal inhibitory concentration (IC50) of the drugs and for the analysis of differences between risk groups, respectively [32].

Quantitative real-time PCR

TRIzol reagent (CWBIO, China) was used for extraction of total RNA. The FastReal qPCR ProMix kit (TIANGEN, FP217, China) was used f to perform qRT-PCR. β-actin as a reference gene and 2−∆∆CT method was used to calculate the relative expression. The sequences of the primers are listed in Supplementary Table 1.

Results

Identification of DE-SERGs in BLCA

The volcano plot (Fig. 1A) and the heatmap (Fig. 1B) demonstrated that 1,563 DEGs were obtained in the TCGA-BLCA dataset, of which 787 were up-regulated and 776 were down-regulated. There were 70 DE-SERGs obtained by intersecting DEGs between BLCA and control samples with SERGs (Fig. 1C). Figure 1D showed the GO enrichment results for DE-SERGs. In terms of biological function, genes were significantly enriched in ‘regulation of angiogenesis’, ‘regulation of vasculature development’, ‘muscle organ development’, etc. In terms of cell composition, genes were significantly enriched in items such as ‘collagen containing extracellular matrix’, ‘myofiber’, ‘contracting fiber’, and ‘sarcomete’. In terms of molecular function, the genes were found to be important in ‘glycosaminoglycan binding’, ‘actin binding’, ‘heparin binding’ and ‘extracellular matrix structural constituent’ pathways. Besides, the KEGG enrichment results showed that DE-SERGs were significantly enriched in the pathways such as ‘aldosterone synthesis and secretion’, ‘proteoglycans in cancer’, ‘cortisol synthesis and secretion’ and pathways related to ‘Cushing’s syndrome’ (Fig. 1E).

Fig. 1.

Fig. 1

Open in a new tab

Screening of DE-SERGs and its GO and KEGG enrichment analysis in bladder cancer. A Volcano map analysis of differentially expressed genes in bladder cancer. B Heatmap analysis of differentially expressed genes in bladder cancer. C DE-SERGS were screened by taking the intersection of differentially expressed genes and super-enhancer related genes. D GO enrichment analysis of DE-SERGs. E KEGG enrichment analysis of DE-SERGs. GO: Gene Ontology, DE-SERGs: differentially expressed super enhancer related genes, KEGG: Kyoto Encyclopedia of Genes and Genomes

A model was constructed for the prognosis of patients with BLCA

There were 22 genes that were identified as having prognostic values for BLCA by univariate Cox regression analysis (Fig. 2A). When lambda = 0.034, error rate was the lowest, and five genes (MXRA7, HSPG2, PLEKHG4B, ATP2B4, and BDKRB2) were yielded through LASSO regression (Fig. 2B and C), and the genes were used for multivariate Cox regression analysis. Then, 3 genes were identified to construct prognostic risk model, namely MXRA7, PLEKHG4B and ATP2B4 (Fig. 2D). The formula for calculating the risk score is: riskScore = 0.321*MXRA7 + 0.269*PLEKHG4B + 0.143*ATP2B4. Patients were categorised into high and low risk groups based on the median value of the risk score in the TCGA-BLCA dataset. As shown in Fig. 2E, the expression of the ATP2B4 and MXRA7 genes was significantly lower in the BLCA group than in the control group, while the opposite was true for the PLEKHG4B gene, with p-values less than 0.05.

Fig. 2.

Fig. 2

Open in a new tab

Regression analysis and construction of prognostic models. A univariate Cox regression analysis screened 22 DE-SERGs with prognostic value. B-C Lasso regression analysis of DE-SERGs. D Multivariate Cox regression analysis screened for genes used to construct prognostic risk models. E The expression analysis of MXRA7, PLEKHG4B, and ATP2B4 in bladder cancer group tissues and normal tissues. DE-SERGs: differentially expressed super enhancer related genes

Evaluation and validation of prognostic models

Significant differences in survival were seen between the risk groups (p < 0.05), with those in the higher risk group having a worse outcome and those in the lower risk having a better outcome (Fig. 3A). The predictive model was effective as the AUC at 1, 3 and 5 years on the ROC curve was all greater than 0.6 (Fig. 3B). The risk values of patients increased sequentially from left to right in the risk curve (Fig. 3C). Prognostic model validation was performed using GSE31684. A significant difference in survival between different risk groups was shown by the K-M survival curves (P < 0.05) (Fig. 3D), and the AUC values were all greater than 0.6 (Fig. 3E). Similarly, Fig. 3F showed the risk curve of the prognostic model in GSE31684. These results were consistent with the training set.

Fig. 3.

Fig. 3

Open in a new tab

The role of the DE-SERGs risk model for prognostic assessment in bladder cancer. A Kaplan–Meier curve analysis comparing overall survival between the high-risk and low-risk groups in training. B Time-independent ROC curve of this DE-SERGs risk model in training. C Risk curve analysis for high and low risk groups in training. D Kaplan–Meier curve analysis comparing overall survival between the high-risk and low-risk groups in validation. E Time-independent ROC curve of this DE-SERGs risk model in validation. F Risk curve analysis for high and low risk groups in validation. DE-SERGs: differentially expressed super enhancer related genes, ROC: receiver operating characteristic

Construction and validation of nomogram related to clinical prognosis

As can be seen from Fig. 4A and B, two indicators were identified, pathologic_N and riskScore, and their P-values were less than 0.05 based on univariate and multivariate Cox analysis. The survival rate of 1/3/5-year were predicted according to the total score in the nomogram (Fig. 4C), and a higher score indicated a lower survival rate. The slopes of the calibration curves were 0.7317 (1 year), 0.3594 (3 years) and 0.2059 (5 years), respectively. They indicated that the best prediction effect was at 1 year (Fig. 4D). It can be seen that the AUC values were all greater than 0.6 (Fig. 4E). Finally, the decision curve suggested that this nomogram had good clinical utility (Fig. 4F and H).

Fig. 4.

Fig. 4

Open in a new tab

Prognostic predictive effect analysis based on the DE-SERGs risk model. A Univariate Cox regression methods to analyze the prognostic significance of clinicopathologic factors and risk groups. B Multivariate Cox regression analysis was performed to determine independent prognostic indicators. C Construction of a nomogram based on multivariate Cox analysis to predict the probability of OS in patients with bladder cancer at 1,3 and 5 years. D The calibration plots of 1, 3 and 5-years OS for nomogram. E ROC curve for 1, 3 and 5-years OS based on nomogram. F-H Decision curves for 1-, 3- and 5-years OS of the nomogram. DE-SERGs: differentially expressed super enhancer related genes, OS: overall survival

Application of risk scores to clinical features

Patients with high grade had significantly higher risk scores than those with low grade, and the risk scores were different among different tumour N-stages (Figs. 5A − 5B). The results showed that the risk model constructed in this study can be applied to clinicopathological features in Fig. 5C and D. In addition, based on the signalling pathway score, the top 10 pathways with significant KEGG enrichment can be seen in Fig. 5E.

Fig. 5.

Fig. 5

Open in a new tab

The relationship between DE-SERGs risk scores and clinical characteristics. A The risk for grade 1 (high grade) was significantly higher than the risk for grade 0 (low grade). B N0 had a significantly lower risk than N1 and N2. C Kaplan–Meier curves were analyzed for survival in the high-risk group versus the low-risk group of patients in grade 1. D Kaplan–Meier curve analysis of survival in patients with N0 in the high-risk group versus the low-risk group. E KEGG pathway analysis in high-risk and low-risk groups. DE-SERGs: differentially expressed super enhancer related genes, KEGG: Kyoto Encyclopedia of Genes and Genomes. *** P < 0.001

The role of prognosis models in the BLCA microenvironment

The results of the rank sum test showed that the ESTIMATED, Immune and Stromal scores were significantly different between the high and low risk groups (Fig. 6A). The risk score and these three scores were significantly positively correlated (Fig. 6B and D). The samples in TCGA-BLCA were subjected to immune infiltration analysis in Fig. 6E. Among them, there were 5 types of immune cells with P < 0.05 by t-test (Fig. 6F). Immune infiltration analysis revealed that risk score was positively associated with T cells CD4 memory resting, Mast cells resting and Macrophages M2 and negatively associated with Dendritic cells activated and T cells CD8 (Fig. 7A and E). The high - risk group showed a better response to immunotherapy via the TIDE score (Fig. 7F). Expression of immune checkpoints was in most cases significantly different in risk groups (Fig. 7G). Figure 8 shows the top 6 drugs with significant differences, which are AZD8186, BMS-754,807, JQ1, KRAS (G12C) inhibitor, and NU7441 respectively.

Fig. 6.

Fig. 6

Open in a new tab

The association of DE-SERGs risk scores with the immune microenvironment. A Comparison of ESTIMATEScore, ImmuneScore, and StromalScore in high-risk and low-risk groups. B-D DE-SERGs risk score correlated with immune score, stromal score, and ESTIMATE score. E Heatmap analysis of 22 immune cell abundances. F compare the differences of immune cell infiltration in the high-risk and low-risk groups. DE-SERGs: differentially expressed super enhancer related genes. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 7.

Fig. 7

Open in a new tab

DE-SERGs risk score in association with immune infiltration and immune checkpoints. A-E 5 immune cells significantly associated with DE-SERGs risk score. F Predicting the response of high and low risk groups to ICB by TIDE algorithm. G Comparison of immune checkpoint expression in high-risk and low-risk groups. DE-SERGs: differentially expressed super enhancer related genes. ICB: immune checkpoint blockade, TIDE: tumor immune dysfunction and exclusion. * P < 0.05, ** P < 0.01, *** P < 0.001

Fig. 8.

Fig. 8

Open in a new tab

Drugs with significantly different IC50 in the high- and low-risk groups

Validation of SERGs gene expression

The SERGs of bladder cancer expression level were detected by qRT-PCR. We found that ATP2B4 and MXRA7 were significantly higher expressed in human normal uroepithelial cells (SV-HCU-1) than in bladder cancer cell lines (T24, 5637, EJ, UMUC3). In contrast, PLEKHG4B was significantly more highly expressed than SV-HCU-1 in the bladder cancer cell lines (T24, 5637, EJ, J82, UMUC3) (Fig. 9). Additionally, these findings were validated in 22 cases of patient tissues we collected (Fig. 10).

Fig. 9.

Fig. 9

Open in a new tab

The expression levels of ATP2B4, MXR7A and PLEKHG4B in bladder cancer cell lines were detected by qRT-PCR. ns not significant, *p < 0.05, **p < 0.01, ***p < 0.001

Fig. 10.

Fig. 10

Open in a new tab

Expression levels of ATP2B4, MXR7A, and PLEKHG4B detected by qRT-PCR in 22 cases of bladder cancer tissues, ns not significant, *p < 0.05, **p < 0.01, ***p < 0.001

Discussion

Super-enhancers are defined as large clusters of enhancers in genomic DNA. Compared to typical enhancers, they show a higher degree of mediation, binding to a greater number of transcription factors associated with transcriptional activity, stronger regulation of gene expression levels, and show differential binding to tissue-specific transcription factors, among other [12, 33]. An increasing number of studies have found that super-enhancers are closely associated with cancer and that super-enhancers are involved in cancer development through various pathways [11, 12]. The construction of prognostic models based on super-enhancer-related genes has been reported in various cancers. Qing Wu et al. [13] found that a risk score model based on 6 SERGs (ZIC2, NFE2, FOXJ1, KLF15, POU3F2 and SPIB) was effective in predicting the prognosis of breast cancer patients and was associated with immune infiltration. A prognostic model constructed based on 10 super-enhancer-related genes (S100A11, LZTS2, CYP2S1, ZNF552, PSMG1, GJC1, NXN, and DCBLD2) was found to be effective in predicting the survival rate of colon cancer patients [34]. Xueyan Wei et al. found that models constructed with four super-enhancer-related genes (RTKN2, HS3ST5, SQSTM1, ETV4) could accurately predict the prognosis of hepatocellular carcinoma [35]. However, no relevant studies have been reported in bladder cancer.

In this study, we screened 1563 differentially expressed genes from the TCGA-BLCA database, of which 70 genes overlapped with SERGs from the SEA database. By GO and KEGG analysis, we found that these 70 genes are closely related to muscle system process, regulation of angiogenesis, proteoglycans in cancer, etc. Angiogenesis has been shown to play an important role in the development of bladder cancer [36], and Asiliki Papadaki et al. found that two secreted proteoglycans (OMD and PRELP) can inhibit bladder cancer progression by suppressing EMT [37]. Using univariate, multivariate and lasso regression analysis, we screened three SERGs (MXRA7, PLEKHG4B, and ATP2B4) for the construction of a bladder cancer prognostic model. Previous studies have shown that these three genes are associated with malignant tumors. MXRA7 was found to play a critical role in blocking differentiation in human acute promyelocytic leukemia cells [38]. Yu Zhang et.al [39] A risk prognostic model constructed based on multiple RNAs (PLEKHG4B as one of the mRNAs) proved to be important for prognostic assessment of ovarian cancer patients. Yuanming Pan et al. [40] found that ATP2B4 inhibits gastric cancer progression by regulating mitochondrial metabolism and apoptosis.

Based on these three genes, we construct risk assessment models. According to this model, we categorized bladder cancer patients into high-risk and low-risk groups, and validated the model by KM curves and ROC curves in the training set and validation set. Then, a prognostic nomogram was constructed using the risk model with pathological N-staging, in which we verified that it had a positive predictive effect on OS in bladder cancer patients. Further analysis by time-dependent ROC curves and DAC curves confirmed promising predictive ability with prognostic nomogram. In addition, we noted differences in immune scores between the high-risk and low-risk groups, with ESTIMATE Score, ImmuneScore, and StromalScore significantly higher in the high-risk group than in the low-risk group, and further analyses revealed that risk scores were significantly positively correlated with ESTIMATEScore, ImmuneScore, and StromalScore. We also noted that Macrophages M2, Mast cells resting, T cells CD4 memory resting, Dendritic cells activated and T cells CD8 infiltration were significantly different in the high and low risk groups. Risk score was significantly positively related to Macrophages M2, Mast cells resting, and T cells CD4 memory resting cell infiltration program, while it was significantly negatively related to Dendritic cells activated and T cells CD8 cell infiltration.

The tumor microenvironment has an important impact on the progression and treatment of bladder cancer, especially the role of the immune microenvironment in bladder cancer [41]. Cytotoxic CD4 + T cells have been found to play an important role in antitumour immunity by directly killing tumor cells in bladder cancer, and two CD4 + T cell subpopulations of the expressing granzyme K (GZMK) and granzyme B (GZMB), respectively, have been identified [42]. CD8 + T cells in bladder cancer act as cytotoxic immune cells that can kill tumor cells by recognizing MHCI [43]. The role of M2 macrophages in bladder cancer is uncertain, with some studies suggesting that they may be associated with a poor prognosis and, conversely, with a favorable prognosis in bladder cancer [44]. Therefore, the role of M2 macrophages in bladder cancer needs to be further investigated [44]. Immunological status plays an important role in the surveillance and treatment of bladder cancer [45]. Our risk model based on SERGs was strongly associated with immune infiltration. A higher degree of T cell CD4 memory resting infiltration in a high-risk population means fewer activated CD4 + T cells and less toxicity to tumor cells. CD8 + T was negatively correlated with the degree of risk, suggesting that patients in the high-risk group had less CD8 + T cell infiltration and less toxic effects on tumor cells. These results suggest that the risk model we have constructed may provide guidance for immunotherapy in bladder cancer patients. Further analysis revealed significant differences in mostly immune checkpoints in the high-risk and low-risk groups. Tumor immune dysfunction and exclusion (TIDE) can be used to predict patient responsiveness to immune checkpoint blockade (ICB). High TIDE scores are associated with poor ICB therapy efficacy and short survival after receiving ICB therapy [46].

Our in silico analyses revealed two distinct but exploratory translational insights. First, elevated TIDE scores in the high-risk group suggest a potential association between our risk signature and an immunosuppressive tumor microenvironment, which may correlate with reduced responsiveness to immune checkpoint blockade—a hypothesis that warrants direct experimental and clinical validation. Second, the differential predicted IC50 values for certain targeted agents (e.g., AZD8186, BMS-754807) between risk groups highlight candidate drugs that may exhibit differential efficacy. It is critical to emphasize that these computational associations are strictly hypothesis-generating; they nominate specific agents and immunological phenotypes for future investigation but do not constitute predictive biomarkers for clinical use.

This study has several important limitations that must be acknowledged. Primarily, the model is derived from and validated within retrospective public cohorts. Its true prognostic and predictive utility requires rigorous validation in prospective, multi-center studies with standardized clinical annotations. Second, while the core genes (MXRA7, PLEKHG4B, ATP2B4) show compelling RNA-level associations, their protein-level expression and biological function in bladder cancer pathogenesis remain to be fully characterized. Our expanded validation with 22 additional paired samples confirmed tumor overexpression for ATP2B4 and PLEKHG4B, but the trend for MXRA7 was not statistically significant, underscoring the complexity of translating transcriptomic signatures.

Looking forward, the translational journey of this SERG-based model necessitates a multi-pronged research agenda focused on validation, mechanism, and clinical integration. The immediate priority is to transform the bioinformatic signature into a robust, clinical-grade assay and to prospectively validate its prognostic power in independent, multi-center cohorts. Concurrently, the biological hypotheses generated herein—particularly the signature’s association with immune evasion and differential drug sensitivity—must be causally investigated through functional genomics in preclinical models. The ultimate test of its clinical value will be its integration into prospective therapeutic trials, evaluating its utility in stratifying patients for tailored interventions, including immunotherapy or molecularly targeted agents.

Conclusion

In conclusion, we constructed a bladder cancer prognostic model based on3-SERGs. The SERGs model we constructed can effectively predict the prognosis of bladder cancer patients, which is correlated with the immune infiltration of bladder cancer, and has a certain guiding role in the immunotherapy of bladder cancer patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (16.5KB, doc)

Author contributions

LYX designed this experiment, collected and organized data, and wrote the article. ZHZ, LYX and XCZ collected the data and revised the article. ZWZ and XCZ participated in data compilation. TZ was responsible for capital acquisition, supervision and resources. All authors read and approved the final manuscript.

Funding

This work was supported by Science and Technology Program Project of Jiangxi Provincial Health Commission (202130051), Jiangxi Provincial Health Commission project (202610361) and Science and Technology Program Project of Jiangxi Provincial Administration of Traditional Chinese Medicine (2020A0384).

Data availability

The datasets analyzed during this study are available in the following public repositories: The Super-Enhancer Archive [http://www.licpathway.net/sedb/], The Cancer Genome Atlas Bladder Urothelial Carcinoma (TCGA-BLCA) dataset [https://portal.gdc.cancer.gov/projects/TCGA-BLCA], and Gene Expression Omnibus (GEO) datasets GSE31684 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE31684] and GSE84433 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE84433]. All other data generated or analyzed during this study are included in this published article and its supplementary information files, or are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study and its protocols were approved by the Biomedical Research Ethics Committee of the Second Affiliated Hospital of Nanchang University and the Ethics Committee of Jiangxi Provincial People’s Hospital. The study utilized human bladder cancer tissue samples collected from patients undergoing radical cystectomy at these two medical centers. Written informed consent was obtained from all participating patients after they were fully informed about the study’s purpose, procedures, potential risks, and benefits. All patient data and samples were anonymized and handled in strict accordance with the ethical standards of the institutional review boards and with the principles of the Declaration of Helsinki. All patients were informed and agreed to participate in the study. All experiments were conducted in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

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

Lieyu Xu, Zhenhao Zeng and Xinchang Zou have contributed equally to this work.

References

  • 1.Liu X, Jiang J, Yu C, Wang Y, Sun Y, Tang J, et al. Secular trends in incidence and mortality of bladder cancer in China, 1990–2017: A joinpoint and age-period-cohort analysis. Cancer Epidemiol. 2019;61:95–103. [DOI] [PubMed] [Google Scholar]
  • 2.Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49. [DOI] [PubMed] [Google Scholar]
  • 3.Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17–48. [DOI] [PubMed] [Google Scholar]
  • 4.Dyrskjot L, Hansel DE, Efstathiou JA, Knowles MA, Galsky MD, Teoh J, et al. Bladder cancer. Nat Rev Dis Primers. 2023;9(1):58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Teoh JY, Kamat AM, Black PC, Grivas P, Shariat SF, Babjuk M. Recurrence mechanisms of non-muscle-invasive bladder cancer - a clinical perspective. Nat Rev Urol. 2022;19(5):280–94. [DOI] [PubMed] [Google Scholar]
  • 6.Kamoun A, de Reynies A, Allory Y, Sjodahl G, Robertson AG, Seiler R, et al. A Consensus Molecular Classification of Muscle-invasive Bladder Cancer. Eur Urol. 2020;77(4):420–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Dai H, Yu Z, Zhao Y et al. Integrating machine learning models with multi-omics analysis to decipher the prognostic significance of mitotic catastrophe heterogeneity in bladder cancer. Biol Direct. 2025;20(1):56. Published 2025 Apr 21. 10.1186/s13062-025-00650-x [DOI] [PMC free article] [PubMed]
  • 8.Ma C, Wang Z, Zhu X et al. Super enhancer-driven LINC01013 mediates hypoxia-induced mitochondrial dysfunction by HSPA9 to determine pulmonary arterial smooth muscle cell fate. Cell Mol Life Sci. 2026;83(1):57. Published 2026 Jan 6. 10.1007/s00018-025-06071-3 [DOI] [PMC free article] [PubMed]
  • 9.Sabari BR, Dall’Agnese A, Boija A, et al. Coactivator condensation at super-enhancers links phase separation and gene control. Science. 2018;361(6400):eaar3958. 10.1126/science.aar3958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Bacabac M, Xu W. Oncogenic super-enhancers in cancer: mechanisms and therapeutic targets. Cancer Metastasis Rev. 2023;42(2):471–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Tang F, Yang Z, Tan Y, Li Y. Super-enhancer function and its application in cancer targeted therapy. NPJ Precis Oncol. 2020;4:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Dai H, Zhang X, Yin L et al. Integrating machine learning and multi-omics analysis to explore Treg-associated programmed cell death features in clear cell renal cell carcinoma. Cancer Cell Int. 2026;26(1):15. Published 2026 Jan 9. 10.1186/s12935-025-04133-x [DOI] [PMC free article] [PubMed]
  • 13.Wu Q, Tao X, Luo Y, Zheng S, Lin N, Xie X. A novel super-enhancer-related gene signature predicts prognosis and immune microenvironment for breast cancer. BMC Cancer. 2023;23(1):776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ma Q, Zhao M, Long B, Li H. Super-enhancer-associated gene CAPG promotes AML progression. Commun Biol. 2023;6(1):622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Liu Y. Identification and comprehensive analysis of super-enhancer related genes involved in epithelial-to-mesenchymal transition in lung adenocarcinoma. PLoS ONE. 2023;18(9):e0291088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Liu S, Wang Z, Zhu R, Wang F, Cheng Y, Liu Y. Three Differential Expression Analysis Methods for RNA Sequencing: limma, EdgeR, DESeq2. J Vis Exp. 2021(175). [DOI] [PubMed]
  • 17.Li X, Yu W, Liang C, Xu Y, Zhang M, Ding X, et al. INHBA is a prognostic predictor for patients with colon adenocarcinoma. BMC Cancer. 2020;20(1):305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Dai H, Wang X, Wu Q, et al. Integrating genetic crosstalk between atherosclerosis and lung adenocarcinoma to advance precision diagnosis and treatment. Cell Div. 2025;20(1):27. 10.1186/s13008-025-00172-y. Published 2025 Nov 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wu X, Sui Z, Zhang H, Wang Y, Yu Z. Integrated Analysis of lncRNA-Mediated ceRNA Network in Lung Adenocarcinoma. Front Oncol. 2020;10:554759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Chen L, Lu D, Sun K, Xu Y, Hu P, Li X, et al. Identification of biomarkers associated with diagnosis and prognosis of colorectal cancer patients based on integrated bioinformatics analysis. Gene. 2019;692:119–25. [DOI] [PubMed] [Google Scholar]
  • 21.Dai K, Liu C, Guan G, Cai J, Wu L. Identification of immune infiltration-related genes as prognostic indicators for hepatocellular carcinoma. BMC Cancer. 2022;22(1):496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Iasonos A, Schrag D, Raj GV, Panageas KS. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol. 2008;26(8):1364–70. [DOI] [PubMed] [Google Scholar]
  • 23.Wang W, Liu J, Liu L. Development and Validation of a Prognostic Model for Predicting Overall Survival in Patients With Bladder Cancer: A SEER-Based Study. Front Oncol. 2021;11:692728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Dutta S, Datta S. A rank-sum test for clustered data when the number of subjects in a group within a cluster is informative. Biometrics. 2016;72(2):432–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Wu W, Dong J, Lv Y, Chang D. Cuproptosis-Related genes in the prognosis of colorectal cancer and their correlation with the tumor microenvironment. Front Genet. 2022;13:984158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Hu K. Become Competent in Generating RNA-Seq Heat Maps in One Day for Novices Without Prior R Experience. Methods Mol Biol. 2021;2239:269–303. [DOI] [PubMed] [Google Scholar]
  • 27.Lin W, Wang Y, Chen Y, Wang Q, Gu Z, Zhu Y. Role of Calcium Signaling Pathway-Related Gene Regulatory Networks in Ischemic Stroke Based on Multiple WGCNA and Single-Cell Analysis. Oxid Med Cell Longev. 2021;2021:8060477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Dai H, Zhao K, Zhao Y et al. Machine learning model in multi-omics perspective demystifies the prognostic significance of crotonylation heterogeneity in clear cell renal cell carcinoma. BMC Urol. 2025;25(1):229. Published 2025 Sep 15. 10.1186/s12894-025-01914-4 [DOI] [PMC free article] [PubMed]
  • 29.Arghittu A, Deiana G, Dettori M, Dempsey E, Masia MD, Plamieri A, et al. Web-based analysis on the role of Digital Media in Health Communication: the experience of VaccinarSinSardegna Website. Acta Biomed. 2021;92(S6):e2021456. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Liu J, Zhang D, Cao Y, Zhang H, Li J, Xu J, et al. Screening of crosstalk and pyroptosis-related genes linking periodontitis and osteoporosis based on bioinformatics and machine learning. Front Immunol. 2022;13:955441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Wang T, Dai L, Shen S, Yang Y, Yang M, Yang X, et al. Comprehensive Molecular Analyses of a Macrophage-Related Gene Signature With Regard to Prognosis, Immune Features, and Biomarkers for Immunotherapy in Hepatocellular Carcinoma Based on WGCNA and the LASSO Algorithm. Front Immunol. 2022;13:843408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Dai H, Zhang X, Zhao Y et al. ADME gene-driven prognostic model for bladder cancer: a breakthrough in predicting survival and personalized treatment. Hereditas. 2025;162(1):42. Published 2025 Mar 19. 10.1186/s41065-025-00409-4 [DOI] [PMC free article] [PubMed]
  • 33.Sengupta S, George RE. Super-Enhancer-Driven Transcriptional Dependencies in Cancer. Trends Cancer. 2017;3(4):269–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Tang Y, Sang S, Gao S, Xu W, Zhou H, Xia X. Mechanistic insights into super-enhancer-related genes as prognostic signatures in colon cancer. Aging. 2024;16(11):9918–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Wei X, Zhou Z, Long M, Lin Q, Qiu M, Chen P, et al. A novel signature constructed by super-enhancer-related genes for the prediction of prognosis in hepatocellular carcinoma and associated with immune infiltration. Front Oncol. 2023;13:1043203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Elayat G, Punev I, Selim A. An Overview of Angiogenesis in Bladder Cancer. Curr Oncol Rep. 2023;25(7):709–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Papadaki V, Asada K, Watson JK, Tamura T, Leung A, Hopkins J et al. Two Secreted Proteoglycans, Activators of Urothelial Cell-Cell Adhesion, Negatively Contribute to Bladder Cancer Initiation and Progression. Cancers (Basel). 2020;12(11). [DOI] [PMC free article] [PubMed]
  • 38.Sun Z, Lin D, Shen Y, Ma K, Wang B, Liu H, et al. Critical role of MXRA7 in differentiation blockade in human acute promyelocytic leukemia cells. Exp Hematol. 2023;125–126:45–54. [DOI] [PubMed] [Google Scholar]
  • 39.Zhang Y, Ye Q, He J, Chen P, Wan J, Li J, et al. Recurrence-Associated Multi-RNA Signature to Predict Disease-Free Survival for Ovarian Cancer Patients. Biomed Res Int. 2020;2020:1618527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Pan Y, Wang X, He Y, Lin S, Zhu M, Li Y, et al. Tumor suppressor ATP4B serve as a promising biomarker for worsening of gastric atrophy and poor differentiation. Gastric Cancer. 2021;24(2):314–26. [DOI] [PubMed] [Google Scholar]
  • 41.Patwardhan MV, Mahendran R. The Bladder Tumor Microenvironment Components That Modulate the Tumor and Impact Therapy. Int J Mol Sci. 2023;24(15). [DOI] [PMC free article] [PubMed]
  • 42.Oh DY, Kwek SS, Raju SS, Li T, McCarthy E, Chow E, et al. Intratumoral CD4(+) T Cells Mediate Anti-tumor Cytotoxicity in Human Bladder Cancer. Cell. 2020;181(7):1612–e2513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Philip M, Schietinger A. CD8(+) T cell differentiation and dysfunction in cancer. Nat Rev Immunol. 2022;22(4):209–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Sharifi L, Nowroozi MR, Amini E, Arami MK, Ayati M, Mohsenzadegan M. A review on the role of M2 macrophages in bladder cancer; pathophysiology and targeting. Int Immunopharmacol. 2019;76:105880. [DOI] [PubMed] [Google Scholar]
  • 45.Pan S, Li S, Zhan Y, Chen X, Sun M, Liu X, et al. Immune status for monitoring and treatment of bladder cancer. Front Immunol. 2022;13:963877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Zheng K, Hai Y, Chen H, Zhang Y, Hu X, Ni K. Tumor immune dysfunction and exclusion subtypes in bladder cancer and pan-cancer: a novel molecular subtyping strategy and immunotherapeutic prediction model. J Transl Med. 2024;22(1):365. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (16.5KB, doc)

Data Availability Statement

The datasets analyzed during this study are available in the following public repositories: The Super-Enhancer Archive [http://www.licpathway.net/sedb/], The Cancer Genome Atlas Bladder Urothelial Carcinoma (TCGA-BLCA) dataset [https://portal.gdc.cancer.gov/projects/TCGA-BLCA], and Gene Expression Omnibus (GEO) datasets GSE31684 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE31684] and GSE84433 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE84433]. All other data generated or analyzed during this study are included in this published article and its supplementary information files, or are available from the corresponding author upon reasonable request.

Articles from Discover Oncology are provided here courtesy of Springer

RESOURCES