Integrated bioinformatics analyses for GSDMB in carcinogenesis and progression of bladder cancer

Introduction

Globally, bladder cancer (BLCA) is a highly prevalent disease with high mortality (1), accounting for 573,278 new cases and 212,536 related deaths each year (2). Approximately 70% of BLCA cases are initially non-muscle-invasive bladder cancer (NMIBC) but can progress to muscle-invasive bladder cancer (MIBC), while 25% are diagnosed as MIBC. Targeted agents and immunotherapy are promising treatments, yet despite advancements, the cancer often progresses aggressively, resulting in poor survival rates (2). For recurrent or distant metastatic disease, especially MIBC, treatments are often inadequate, with a 50% 5-year survival rate for metastatic progression. Studying the genes and molecular pathways involved in BLCA carcinogenesis could uncover oncogenesis mechanisms and improve cancer diagnosis, treatment, and prevention. Apoptosis and senescence were once seen as key defenses against cancer in damaged or abnormal cells (3).

Typically, pyroptosis refers to lytic, programmed, and inflammatory cell death, triggered by inflammasomes and carried out by N-terminal gasdermin pores. It plays a vital part in swelling of cells, lysis of plasma membranes, fragmentation of chromatin, and release of proinflammatory molecules (4). Gasdermins (GSDMs) are a protein family that takes part in the process of programmed cell death (PCD), specifically in pyroptosis, leading to cell proliferation, tumorigenesis (5), secretion of interleukin (IL) family members and pyroptotic lytic cell death (6).

Gasdermin B (GSDMB) is a member of the GSDMs and participates in the regulation of cell pyroptosis (7). Generally, GSDMB exhibits pleiotropic roles in tumorigenesis and progression, impacting cancers differently in a tissue-specific manner or in varying pathological conditions (8). Overexpression of GSDMB is linked to increased risk and enhanced invasiveness in breast cancer (BC), gastric cancer, and non-small cell lung cancer (NSCLC) (9-11). A recent study conducted by Wang et al. found that GSDMB expression in tumors is much higher than that in normal tissue. However, a higher GSDMB level was associated with lower tumor grade and stage, and higher overall survival (OS) rate (12). Contrary to its pro-tumor effects, GSDMB can function as a tumor suppressor in specific circumstances. In the same study, Wang et al. also found that anlotinib can induce an increase in the secretion of inflammatory factors and tumor-killing factors in GSDMB-positive BLCA and improve the anti-tumor effect in mice. Other studies have also shown that GSDMB is associated with a better prognosis (13-15). Moreover, Zhou et al. proved that when the inhibitory checkpoint was blocked by antibody to programmed cell death 1 (PD-1), the overexpression of GSDMB in mouse cancer cells promoted the clearance of tumors by cytotoxic T lymphocytes, indicating that Gasdermin-mediated pyroptosis may enhance anti-tumor immunity (16).

Overall, the biological function of GSDMB is largely unknown due to its opposite role in promoting and suppressing tumor progression in different tissues and genetic backgrounds (17). In our previous study, it was proved that some of the pyroptosis-related genes (PRGs) were closely related to BLCA prognosis and clinicopathological features, but the exact function of GSDMB in BLCA is not well understood (). Currently, no research on the role of GSDMB in the onset of BLCA has been reported. Therefore, we believe that it is meaningful to further explore the role of GSDMB in the carcinogenesis, development, and treatment of BLCA. The current research indicated that GSDMB might be considered as an essential gene for the onset of BLCA, as well as contributing factor for prognosis, tumor microenvironment (TME), and immunotherapy. We present this article in accordance with the TRIPOD reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0085/rc).

Methods

Mendelian randomization causality analysis based on cis-eQTL for GSDMB

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Mendelian randomization (MR) analysis uses genetic variants which are robustly associated with exposure as instrumental variables (IVs) to explore causal associations between the exposure and outcome (18). Cis-expression quantitative trait loci (cis-eQTL, exposure) can alter mRNA expression, leading to changes in the level, timing, and localization of gene expression which can significantly cause variations in individual phenotypes. Significant cis-eQTLs for GSDMB were viewed as IVs representing exposure with a minor allele frequency >0.01, which was obtained from a publicly available dataset containing 17,142 single-nucleotide polymorphisms (SNPs) and 31,684 blood samples from European individuals issued by the eQTLGen consortium (https://www.eqtlgen.org/cis-eqtls.html) (19). All relevant SNPs selected as genetic instruments met a P<5e-8 threshold and were defined as within 100 kb of a probe in either direction. To prevent bias from weak IVs, the F-statistic for each SNP was calculated by the formula: F = R² (N – K − 1) / [(1 – R²) K], with values over 10 indicating a valid instrument (20). All IVs for the MR analysis were investigated in PubMed, LDlink, and the genome-wide association study (GWAS) Catalog to control for confounding factors. rs4580194 was identified as an IV for both GSDMB expression and body mass index (BMI). Given BMI’s established causal link to BLCA (21), rs4580194 may act as a confounder and was therefore excluded. Table S1 provides the valid IVs for GSDMB cis-eQTL. Linkage disequilibrium (LD) scores were calculated to retain independent SNPs (r²<0.3 within 100 kb). Three BLCA-related traits, malignant neoplasm of bladder, BLCA, and bladder surgery, were considered as the outcome factors. The IVs for outcome were drawn from publicly available genetic association studies conducted in European populations by different consortiums: UK Biobank (https://biobank.ctsu.ox.ac.uk/), FinnGen database (https://www.finngen.fi/en/access_results) (22) (release 10) and Medical Research Council (MRC) Integrative Epidemiology Unit (IEU) consortium (https://gwas.mrcieu.ac.uk/) (23) (GWAS ID: ieu-b-4874, ukb-b-18477). We also used genetic association study data for benign neoplasm of bladder in the FinnGen database (release 10) as negative controls to further confirm the association between GSDMB level and malignancy instead of benign tumors. Information regarding GWAS studies used in our study is detailed in Table 1.

All MR analyses were conducted using the “TwoSampleMR” package in R 4.2.1. Data were sourced from publicly available databases and organized with the “tidyverse” and “data.table” packages. The primary analysis utilized the inverse variance weighted (IVW) estimator. As a test for robust causal link, we performed weighted median estimator (WME), single-mode and weighted-mode MR. Sensitivity analyses were conducted to evaluate pleiotropy, heterogeneity, and the robustness of the results. Specifically, MR-PRESSO, which identifies and removes outlier SNPs, and MR-Egger intercept test were utilized to assess potential directional pleiotropy. Cochran’s Q statistic was employed to quantify heterogeneity across the IVs. A significant deviation (P<0.05) suggests the presence of pleiotropy or heterogeneity. Additionally, a leave-one-out sensitivity analysis was performed to identify potentially significant individual SNPs.

GSDMB expression profile and clinical characteristics

Transcriptome profiling data and clinical information for The Cancer Genome Atlas (TCGA)-BLCA were downloaded from TCGA on August 29, 2022. Expression data of BLCA samples were normalized for comparison. This study analyzed 406 bladder tumor tissues and 19 adjacent normal tissues from the TCGA dataset (https://portal.gdc.cancer.gov/). After excluding patients without clinical data, 406 patients’ data were eligible for analysis. Replicate values from the same patient were averaged and merged.

GSDMB mRNA expression differences between tumor and normal tissues across 33 cancer types were analyzed using the TIMER2 database (http://timer.cistrome.org/). Difference in the expression levels of GSDMB mRNA between BLCA and normal tissue was investigated using paired and unpaired tests. Additionally, the association between GSDMB expression and clinicopathological features (age at diagnosis, sex, pathological stage, T stage, N stage, and M stage) was assessed in the TCGA-BLCA cohort. The general clinical characteristics of these BLCA patients were summarized in Table 2.

Overall and disease-specific survival analyses

To assess the prognostic significance of GSDMB mRNA expression in BLCA patients, Kaplan-Meier survival analyses of OS and progression-free survival (PFS) were performed by dividing the patients into high and low groups according to their median GSDMB mRNA expression level with the “survival” R language package. Univariate and multivariate Cox regression analysis was conducted to estimate the prognostic value of GSDMB. Time-dependent receiver operating characteristic (ROC) curves and nomogram models for prognosis analysis were created to assess the predictive capacity of GSDMB expression and nomogram, using “timeROC” and “survival” packages in R, respectively. The calibration curve presents a linear relationship between concentration and response using a least squares method.

The risk score was calculated by adding the product of each gene expression value with the corresponding risk coefficient. The median risk score was selected as the cutoff value to divide the patients into a high-risk group and a low-risk group. A 3D principal component analysis (PCA) plot was created to assess the potential differences between the high- and low-risk groups with R package “scatterplot3d”.

Functional enrichment analyses of differentially expressed genes (DEGs)

The DEGs between high and low GSDMB expression subgroups in the TCGA-BLCA dataset were identified with cutoffs of adjusted P<0.05 and |log₂FC| >1.5. The results of the DEGs are shown in the heatmap and volcano plot created with the “pheatmap” and “ggplot2” R packages, respectively. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment (https://www.kegg.jp/) and Gene Ontology (GO) enrichment (https://geneontology.org/) analysis were performed using the “clusterProfiler” package in R. In addition, gene set enrichment analysis (GSEA) (https://www.gsea-msigdb.org/gsea/msigdb/) was conducted to investigate the correlation and relative contribution between gene expression and phenotype (24).

Immune cell infiltration analysis and tumor mutational burden analysis

TME has been recognized as playing an essential role in regulating tumor immune suppression, distant metastasis, local resistance, and the targeted therapy response (25,26). The exact regulation of pyroptosis and GSDMs on TME is not very clear; they have opposite impacts in different regions and evolving periods of TME (27). To investigate the correlation between the expression of GSDMB and TME characteristics, the “estimate” package in R was used to determine the ratio of the immune-stromal component of BLCA cases, which includes three scores: the immune score, stromal score, and ESTIMATE score. A higher score represents a larger ratio of the corresponding component (28). Besides, the correlation of GSDMB expression with infiltration of 22 immune cell types was analyzed using the “CIBERSORT” R package with Spearman correlation. A P value <0.05 was set as the screening condition of the Wilcoxon test to obtain the immune cell infiltrate matrix.

Tumor mutational burden (TMB) is reported as the total number of mutations per megabase (mut/Mb) in the tumor genome, which is a novel target for predicting the response to immunotherapy (29,30). This metric has been shown to correlate with patient response to both cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and PDCD1 inhibition in several tumor types, including urothelial cancer (31,32). It is hypothesized that highly mutated tumors are prone to stimulate the emergence of neoantigens and the production of highly immunogenic antibodies that trigger anti-tumor immunity. To investigate the correlation between TMB and GSDMB expression, the somatic mutation data obtained from TCGA-BLCA was analyzed with “maftools” R package. Then a survival analysis was performed to explore the TMB effect on BLCA prognosis.

Immune response prediction

We next performed an analysis of Spearman correlations to assess the relationship between GSDMB and several widely discussed immune checkpoint genes’ expression levels, including lymphocyte activating 3 (LAG-3), B7-H1 (CD274), hepatitis A virus cellular receptor 2 (HAVCR2), CTLA-4, T-cell immunoglobulin and ITIM domain (TIGIT), Siglec 15, PDCD1, PDCD1 ligand 2 (PDCD1LG2). Meanwhile, based on the immunophenoscore (IPS) downloaded from The Cancer Immunome Atlas (TCIA) (https://tcia.at/home), we assessed its value in predicting immunotherapy response. IPS reliably predicts responses to anti-CTLA-4 and anti-PD-1 therapies and generally correlates positively with immune checkpoint blockade (ICB) response. Differences in IPS between the low and high GSDMB expression subgroups were analyzed by the Wilcoxon test. Besides, the IMvigor210, GSE78220, and GSE67501 cohorts from Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) were applied in our study to further verify the prognostic value of GSDMB expression for immunotherapy.

Drug sensitivity analysis

Utilizing data from the Genomics of Drug Sensitivity in Cancer (GDSC) database (https://www.cancerrxgene.org/), the study estimated the anti-cancer drug sensitivity of BLCA patients with varying levels of GSDMB expression. The results were presented in terms of half-maximal inhibitory concentration (IC₅₀) values of drugs, employing the “oncoPredict” R package. Statistical analysis was conducted utilizing Pearson’s test. Visualization of the results was achieved through the “ggpubr” and “ggplot2” packages in R.

Statistical analysis

For normally distributed data comparison, statistical significance was assessed by Student’s t-test or analysis of variance (ANOVA) as appropriate. For non-normally distributed data comparison, statistical significance was assessed by Wilcoxon test or Kruskal-Wallis test as appropriate. An estimation of the correlation between GSDMB expression and certain variables was performed using the Spearman correlation test. The R language was employed for downloading data, performing analysis and producing graphs (version 4.2.1). A P value <0.05 was considered statistically significant.

Results

MR analysis of GSDMB expression in blood and BLCA

We performed MR analyses using GSDMB cis-eQTLs as the exposure and BLCA-related IVs as the outcomes. Thirty-nine independent SNPs associated with circulating GSDMB identified by the latest GWAS were extracted. The results in the IVW analysis demonstrated a causal correlation between GSDMB expression in circulation and malignant neoplasm of bladder (UK Biobank, P=0.011; FinnGen database, P=3.98e-7), BLCA [MRC-IEU (ieu-b-4874), P=3.55e-4], bladder surgery (UK Biobank, ukb-b-18477, P=0.005), which suggested that a positive correlation exists between circulating GSDMB and risk of BLCA and bladder surgery. The causal relationships are enhanced by the consistency of results between different methods (Figure 1A). No significant risk association between GSDMB and benign neoplasm of bladder (P=0.860) was found. The forest plot and scatter plot of causal estimates of GSDMB expression on BLCA could be found in Figures S1,S2. The exposure and outcome variants were distributed symmetrically about the combined effect size in the funnel plot (Figure S3). Based on Cochran’s Q test and funnel plot, no significant heterogeneity across the IVs was observed for any trait (P>0.05, Figure 1A, Table S2). The MR-PRESSO and MR-Egger intercept test showed that there was no directional pleiotropy in all analyses (P>0.05, Figure 1A, Table S3), and the leave-one-out test did not identify any variants that had a strong effect on the overall results (Figure S4).

Figure 1 (A) Forest plots displaying the findings from MR analysis of GSDMB-eQTLs in blood with BLCA and bladder surgery risk. Data are represented as ORs with 95% CI. (B) Expression difference of GSDMB between cancer tissues and control samples in diverse cancer types. (C) Expression differences in GSDMB between tumor patients and normal samples in BLCA. *, P<0.05; **, P<0.01; ***, P<0.001. BLCA, bladder cancer; CI, confidence interval; eQTLs, expression quantitative trait loci; FE, fixed effects; IEU, Integrative Epidemiology Unit; IVW, inverse variance weighted; MR, Mendelian randomization; OR, odds ratio; TPM, transcripts per million; UKB, UK Biobank.

Differential expression pattern of GSDMB in pan-cancer analysis

To explore the role of GSDMB in the development of tumor, we analyzed the mRNA expression levels of GSDMB in 35 types of cancer. As shown in Figure 1B, GSDMB expression levels were significantly higher than those in adjacent normal tissues in 15 of 35 types of cancer, comprising BLCA, CESC, ESCA, HNSC, HNSC-HPV+, KIRC, KIRP, LIHC, LUAD, LUSC, PCPG, PRAD, STAD, THCA and UCEC. Subsequently, Figure 1C revealed that the mRNA expression level of GSDMB was significantly higher in TCGA-BLCA tissues than in corresponding adjacent normal tissues (P<0.01, P<0.05, respectively), indicating that GSDMB might play a vital role in the pathogenesis of BLCA.

Value of GSDMB expression in prognosis prediction of BLCA patients

The univariate Cox regression (Figure 2A) showed that GSDMB expression level (HR: 0.662, P<0.001) was linked to a more favorable OS. Age [hazard ratio (HR): 1.036, P<0.001] and clinical stage (HR: 1.699, P<0.001) are strongly associated with poor outcomes. A similar conclusion also emerged from the multivariate Cox regression analyses (Figure 2B), which indicated that GSDMB expression level (HR: 0.701, P<0.001), age (HR: 1.027, P<0.001), and clinical stage (HR: 1.602, P<0.001) are independent prognostic factors for BLCA patients. Time-dependent ROC curves indicated that GSDMB expression could effectively distinguish between tumor and normal tissues (Figure 2C), with AUC values of 0.660, 0.669, and 0.682 for 1, 3, and 5 years, respectively. Kaplan-Meier plots and log-rank analyses (Figure 2D) showed that higher GSDMB expression was significantly linked to better OS (HR: 0.37, P<0.001) and PFS (HR: 0.36, P=0.002) in BLCA patients from the TCGA dataset.

Figure 2 (A,B) Forest plot for the univariate and multivariate Cox proportional hazard regression model. GSDMB could act as an independent predictor of poor survival rate in BLCA patients. (C) Time-dependent ROC analysis of GSDMB expression predicts the 1-, 3-, and 5-year survival rates in BLCA patients. (D) Kaplan-Meier survival curves of OS and PFS for BLCA patients between high- and low-GSDMB expression groups. (E) A nomogram model integrating key clinicopathologic factors and GSDMB expression levels to predict survival probability at 1, 3, and 5 years. (F) Time-dependent ROC of the nomogram to predict the 1-, 3-, and 5-year survival rates in BLCA patients. (G) Calibration plot of the nomogram for predicting the OS likelihood. (H) The distribution of risk scores based on GSDMB expression and results of the PCA between low-risk and high-risk groups. ***, P<0.001. AUC, area under the curve; BLCA, bladder cancer; CI, confidence interval; OS, overall survival; PCA, principal component analysis; PFS, progression-free survival; ROC, receiver operating characteristic.

Then we constructed an OS nomogram model based on the multivariate Cox regression analyses (Figure 2E), integrating GSDMB and clinicopathological factors (including age, clinical stage and sex) to predict the survival probabilities. ROC curves demonstrated AUC values of 0.741, 0.734, and 0.758 for 1, 3, and 5 years, respectively (Figure 2F), indicating GSDMB’s effective predictive power for OS in BLCA patients. The calibration curve validated the alignment between the nomogram’s predicted OS and actual OS proportions (Figure 2G). Figure 2H illustrates the risk score distribution and survival status. The PCA analysis confirmed that the expression levels of OS prognostic GSDMB included in model construction were significant for distinguishing high- and low-risk patients.

DEG identification between high and low GSDMB expression subgroups

Subsequently, the DEGs between low and high GSDMB expression subgroups were identified by fold change method, following the criteria of false discovery rate (FDR) <0.005 and |log₂FC| ≥1.5. The heatmap of the top 30 DEGs based on order of logFC (Figure 3A) and the volcano plot (Figure 3B) demonstrated DEGs that have a strong association with GSDMB. Table S4 lists the top 10 significantly upregulated and downregulated DEGs. In Figure 3C, the chord diagram displays the gene pairs with the top 5 highest and lowest correlation coefficients linked to GSDMB. We identified 5 DEGs significantly related to GSDMB: BICDL2, ELF3, ORMDL3, PPFIBP2, and TTLL3, all positively correlated with GSDMB (Figure 5D-5H, P<0.001, correlation coefficient >0.6).

Figure 3 (A,B) Heatmap of the top 30 significant differentially expressed genes and Volcano plot of differentially expressed genes between the high- and low-GSDMB expression groups. (C) Gene co-expression profile of GSDMB. (D,E) GO enrichment analysis showed the BP, CC, and MF of co-expressed genes with GSDMB. (F) Significantly enriched KEGG terms obtained from KEGG enrichment analysis of co-expressed genes with GSDMB. (G) A combined enrichment map of GSEA results between high and low GSDMB expression phenotypes. BP, biological process; CC, cellular component; FC, fold change; GO, Gene Ontology; GSEA, gene set enrichment analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

Figure 5 (A,B) Lollipop plot and boxplots to visualize the correlation between GSDMB and sensitivity to anti-cancer drugs in BLCA. ***, P<0.001. BLCA, bladder cancer.

Functional analyses of DEGs in the subgroup with different GSDMB expression

Functional enrichment analyses of DEGs were conducted using GO and KEGG to identify biological process differences between subgroups with varying GSDMB expression. Figure 3D highlights enriched functions in cellular components, molecular functions, and biological processes, while Figure 3E displays key GO functions and gene distributions. Significant signaling pathways were also explored in the KEGG database. Figure 3F displays the top 15 pathways significantly enriched by DEGs. GSEA of DEGs from the MSigDB collection showed notable differences in several pathways between high and low GSDMB expression phenotypes, with the top 5 pathways shown in Figure 3G.

Correlation analysis of GSDMB expression and immune cell infiltration in the TCGA-BLCA dataset

To explore the immune characteristics of GSDMB, CIBERSORT analysis was subsequently performed in the TCGA-BLCA dataset. In the TCGA-BLCA dataset, high GSDMB expression was linked to increased infiltration of plasma cells, CD8 T cells, follicular helper T cells, monocytes, and activated dendritic cells. Conversely, low GSDMB expression was associated with higher levels of resting memory CD4 T cells, M0 macrophages, and M2 macrophages (P<0.05, Figure 4A). The correlation coefficient and P value of each cell are shown in Figure 4B.

Figure 4 (A,B) Differences in infiltrating immune cells in each step of the cancer immune cycle and their correlations with GSDMB. (C) Comparison of stromal score, immune score, and ESTIMATE score between low-and high-GSDMB groups. (D) GSDMB was positively correlated with tumor mutational burden, which indicated better survival. (E) Immune-checkpoint-relevant genes expressed in high- and low-GSDMB groups. (F) Relationship between GSDMB and immune-checkpoint-relevant genes or immune-activation-relevant genes. (G) IPS analysis of PD-1 and CTLA-4 between low- and high-GSDMB groups. (H) Comparison of GSDMB expression level between responders and non-responders to anti-PD-L1 treatment in 3 immunotherapy cohorts. *, P<0.05; **, P<0.01; ***, P<0.001. IPS, immunophenoscore; PD-1, programmed cell death protein 1; PD-L1, programmed death-ligand 1; TME, tumor microenvironment.

Furthermore, we investigated the relationship between GSDMB expression and TME. As a result, the TME score revealed that patients with high GSDMB expression exhibited a significantly lower stromal score (P<0.0001), immune score (P=0.001), and ESTIMATE score (P<0.0001, Figure 4C). MIBC, high-grade, and later-stage BLCA had significantly higher immune and stromal scores than NMIBC, low-grade, and earlier-stage groups. Higher scores were closely linked to poorer cancer-specific survival (CSS). We found that GSDMB expression correlated with a gradual increase in mutation load (Figure 4D). Kaplan-Meier curves showed that a high mutation load was significantly associated with better OS in BLCA (P<0.05). Therefore, all of the above results indicated that GSDMB plays a protective role in tumor invasion and metastasis, and may enhance the efficacy of immunotherapies.

Value of GSDMB in predicting immunotherapy response

Because responses to immune checkpoint–targeted therapies depend on checkpoint expression, checkpoint expression levels may indicate treatment efficacy. We analyzed the difference in immune checkpoint expression between patients with high and low GSDMB levels. Patients with low GSDMB had higher levels of LAG-3, CD274, HAVCR2, CTLA-4, TIGIT, PDCD1, and PDCD1LG2, but lower levels of SIGLEC15 (Figure 4E). A heatmap was created to illustrate the correlation between immune-related genes and GSDMB-related immune checkpoints (Figure 4F).

Additionally, the results of IPS of BLCA patients were downloaded from the TCIA website. Given their great significance among the immune checkpoints, PD-1 and CTLA-4 were included in the IPS analysis to assess GSDMB’s immune properties. It was revealed that the immunogenicity for CTLA-4 and PD-1 immunotherapy between low and high GSDMB groups was different. The average IPS of the high GSDMB expression group was significantly higher than that of the low GSDMB expression group in 3 subgroups: “CTLA4⁻/PD1⁻”, “CTLA4⁻/PD1⁺”, and “CTLA4⁺/PD-1⁻”, while there was no statistically significant difference in IPS between high and low GSDMB expression in the “CTLA4⁺/PD1⁺” subgroup (Figure 4G).

Biomarkers that can effectively predict the efficacy of immunotherapy drugs are currently lacking; such biomarkers would further improve the precision of immunotherapy. In this sense, we examined whether GSDMB expression can predict responses to anti-PD-L1 therapy using three cohorts: IMvigor210, GSE78220, and GSE67501. As a result, patients who responded to anti-PD-L1 treatment showed higher GSDMB expression in the GSE67501 cohort than the non-response subgroup, yet in the other two comparisons there were no significant differences (Figure 4H).

Drug sensitivity analysis of GSDMB

For a better clinical therapeutic strategy for BLCA, we examined the correlation between GSDMB expression and the sensitivity to anti-tumor drugs (IC₅₀), across different GSDMB expression levels using the GDSC database. Among all identified drugs, 112 drugs were significantly associated with the expression of GSDMB (Table S5). In Figure 5A, the top 38 drugs are presented and sorted according to the P values of correlation. We measured IC₅₀ levels of several commonly used anti-cancer drugs (Figure 5B). Results showed that higher GSDMB expression was associated with reduced sensitivity to drugs such as gemcitabine, sorafenib, ipatasertib, leflunomide and wee1 inhibitor, but increased sensitivity to dasatinib, talazoparib, luminespib and obatoclax mesylate. These findings underscore the pivotal role of GSDMB in modulating drug responsiveness.

Briefly, high GSDMB expression was associated with increased infiltration of anti-tumor immune cells (e.g., CD8⁺ T cells and activated dendritic cells) and decreased infiltration of immunosuppressive cells (e.g., M2 macrophages), suggesting a more active immune status. It was also correlated with lower stromal, immune, and ESTIMATE scores, as well as higher mutation load, which was linked to better OS. In addition, GSDMB expression showed significant associations with sensitivity to multiple anti-tumor drugs, indicating its potential value in guiding therapeutic strategies.

Discussion

In the present study, genetic analysis with eQTLs was performed first and uncovered the causal relationship between higher expression of GSDMB in circulating blood and increased risk of BLCA and bladder surgery. GSDMB mRNA were highly expressed in 27 tumors types by pan-cancer analysis. Subsequent investigation uncovered a correlation between GSDMB and BLCA. Bioinformatics analysis using TCGA high-throughput RNA sequencing data demonstrated that GSDMB expression was significantly higher in BLCA tissues than in adjacent non-tumor tissues. However, high-level GSDMB mRNA expression was found to be connected with a lower TME score, lower sensitivity to most anti-cancer drugs, higher TMB and better prognosis. The Kaplan-Meier survival analysis showed that the high GSDMB expression group had a better OS prognosis. The univariate and multivariate analysis showed that high GSDMB mRNA expression was a protective independent prognostic factor for BLCA. High GSDMB expression may contribute to improved prognosis by enhancing pyroptosis-mediated anti-tumor immunity, leading to increased activation and infiltration of cytotoxic immune cells. In addition, GSDMB may serve as a downstream effector of immune cell-derived signals, amplifying immune-mediated tumor cell death and reinforcing its role as an independent protective prognostic factor (33).

These findings suggest that GSDMB may exert context-dependent and potentially dual biological effects in BLCA progression. On the one hand, elevated GSDMB expression may promote tumor initiation and progression through its association with increased TMB and potential involvement in inflammatory signaling pathways, which could facilitate tumor development at an early stage. On the other hand, high GSDMB expression is associated with enhanced infiltration of cytotoxic immune cells (e.g., CD8⁺ T cells and activated dendritic cells) and reduced immunosuppressive components in the TME, suggesting a more immunologically active state that may contribute to improved prognosis and better response to immunotherapy.

Immune response induced by pyroptosis activation is a double-edged sword that affects all stages of tumorigenesis. In the case of promoting tumorigenesis, GSDMs have been found within amplicons, genomic regions that are amplified during the course of cancer development (34). A study on BLCA revealed that GSDMB enhances STAT3 phosphorylation, alters glucose metabolism, and promotes tumor growth (). Lutkowska et al. discovered that the rs8067378 A>G variant increases GSDMB expression, contributing to cervical squamous cell carcinoma (). Additionally, GSDMB overexpression has shown pro-tumor effects in vitro and in vivo, with prognostic relevance in BC (8).

Pyroptosis may enhance immune evasion, thereby conferring protection against tumorigenesis. This cell death process begins with inflammasome assembly, leading to gasdermin D (GSDMD) cleavage and the release of pro-inflammatory cytokines such as IL-1β and IL-18, especially in chronic inflammation (4). They may assist tumor cells in evading immune surveillance and host cell immune responses, thereby promoting local immunosuppression and undermining the development of an antigen-specific immune response (35). Studies have demonstrated that activated caspase-1 can trigger pyroptosis and the release of pro-inflammatory cytokines, promoting tumorigenesis in human hepatocellular carcinoma and esophageal cancer (36,37). In addition to that, it was found that depletion of GSDMD inhibited NSCLC. Molecules located upstream of GSDMD in the pyroptotic pathway have the potential to trigger apoptosis. GSDMD-deficient macrophages exhibited increased activity of apoptotic caspases compared with caspase-1 deficient macrophages (38). These might be some potential reasons why elevated pyroptosis levels may influence the risk of BLCA.

Conversely, GSDMB exhibits anti-tumor properties through the activation of its pore-forming pyroptotic function within malignant cells. The pore-forming function of GSDMB can be activated via targeted nanotherapy or cleavage by GZMA, leading to pyroptosis-mediated apoptosis in cancer cells. Analysis at the molecular level has shown that CD8+ T cells and natural killer (NK) cells secrete GZMA+ and GSDMB+ cells. Reduced levels of GZMA+ and GSDMB+ immune cells in melanoma specimens suggest a weakened anti-tumor immune response, resulting in decreased pyroptosis capability and potentially influencing the anti-melanoma characteristics (39).

Triggering pyroptosis could convert immune “cold” tumors to “hot” tumors, which may alter the TME and the influx of tumor-infiltrating lymphocytes (TILs) (27). The inflammatory state of the TME has implications for the response to immune checkpoint inhibitor (ICI) therapy and is closely associated with the prognosis of cancer patients (40,41). We found that lower GSDMB was correlated with higher expression of some classical immune checkpoints. For BLCA patients with high expression of GSDMB, immune checkpoint combination therapy may achieve better efficacy. Anlotinib’s effectiveness in inhibiting growth of BLCA with higher GSDMB may be partly explained by this. Zhou et al. proved that when the inhibitory checkpoint is blocked by antibody to PDCD1, the expression of GSDMB in mouse colon cancer cells promoted the clearance of tumors by cytotoxic T lymphocytes, thereby elevating the anti-tumor efficacy (16). Specifically, exogenous GSDMB overexpression did not affect tumor progression, but PDCD1 ICIs activated GSDMB pyroptotic activity in tumor cells through NK and CD4⁺ T cells. Studies in mouse models have shown that GSDMB expression enhances BC development specifically in the presence of HER2-driven tumors (42). Experiments also showed that full-length GSDMB isoform 2 was equally expressed in both HER2/NEU oncogene and polyoma middle-T antigen models, suggesting that differences in tumor development were due to biological context, not GSDMB overexpression levels. Similar findings were noted in gastric carcinomas (42). Hence, it is plausible to hypothesize that GSDMB promotes anti-drug functions depending on the biological context.

The differential isoform expression of GSDMB could lead to distinct functional consequences in normal and pathological contexts. NK cell attacks led to pyroptosis in GSDMB3-expressing cells, mixed pyroptosis and apoptosis in GSDMB4-expressing cells, and only apoptosis in GSDMB1/2-expressing cells, indicating GSDMB3 as the most functional form. In BLCA, elevated GSDMB3 levels are correlated with better survival, whereas the influence of other isoforms is negligible (43). Manipulating GSDMB splicing to elevate cytotoxic isoforms and diminish noncytotoxic variants may augment anti-tumor immunity and improve immunotherapy.

Nevertheless, there exist a few limitations in our study. First, the study was mainly based on TCGA datasets, while histological validation was not conducted. Next, the accuracy of the database and the choice of statistical methods may affect the interpretation of the research results. At last, although our hypothesis had been verified with bioinformatics methods, it still lacks experimental proof to verify the molecular mechanism and human experiments to further understand the functional role of GSDMB. A tissue-specific dataset for MR analysis is also warranted to confirm the findings. These issues represent some direction for future research.

Conclusions

In summary, by combining genetic association analyses and TCGA data-based analysis, we discovered that GSDMB has been linked to both anti- and pro-tumor functions in BLCA, which indicates that pyroptosis acts as a “double-edged sword” in tumorigenesis and progression. This study provides new insights for elucidating the pathogenesis and molecular targets of BLCA, which suggests that different measures should be taken to inhibit or promote pyroptosis in the prevention or treatment of BLCA, depending on the biological context. However, the specific pathogenesis and molecular mechanisms still require further exploration.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0085/rc

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0085/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-1-0085/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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

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. Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33. [Crossref] [PubMed]
2. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
3. Kulbay M, Paimboeuf A, Ozdemir D, et al. Review of cancer cell resistance mechanisms to apoptosis and actual targeted therapies. J Cell Biochem 2022;123:1736-61. [Crossref] [PubMed]
4. Rao Z, Zhu Y, Yang P, et al. Pyroptosis in inflammatory diseases and cancer. Theranostics 2022;12:4310-29. [Crossref] [PubMed]
5. Zou J, Zheng Y, Huang Y, et al. The Versatile Gasdermin Family: Their Function and Roles in Diseases. Front Immunol 2021;12:751533. [Crossref] [PubMed]
6. Ivanov AI, Rana N, Privitera G, et al. The enigmatic roles of epithelial gasdermin B: Recent discoveries and controversies. Trends Cell Biol 2023;33:48-59. [Crossref] [PubMed]
7. Li L, Li Y, Bai Y. Role of GSDMB in Pyroptosis and Cancer. Cancer Manag Res 2020;12:3033-43. [Crossref] [PubMed]
8. Sarrio D, Rojo-Sebastián A, Teijo A, et al. Gasdermin-B Pro-Tumor Function in Novel Knock-in Mouse Models Depends on the in vivo Biological Context. Front Cell Dev Biol 2022;10:813929. [Crossref] [PubMed]
9. Zhang X, Liu R. Pyroptosis-related genes GSDMB, GSDMC, and AIM2 polymorphisms are associated with risk of non-small cell lung cancer in a Chinese Han population. Front Genet 2023;14:1212465. [Crossref] [PubMed]
10. Hergueta-Redondo M, Sarrio D, Molina-Crespo Á, et al. Gasdermin B expression predicts poor clinical outcome in HER2-positive breast cancer. Oncotarget 2016;7:56295-308. [Crossref] [PubMed]
11. Molina-Crespo Á, Cadete A, Sarrio D, et al. Intracellular Delivery of an Antibody Targeting Gasdermin-B Reduces HER2 Breast Cancer Aggressiveness. Clin Cancer Res 2019;25:4846-58. [Crossref] [PubMed]
12. Wang C, Cao Q, Zhang S, et al. Anlotinib Enhances the Therapeutic Effect of Bladder Cancer with GSDMB Expression: Analyzed from TCGA Bladder Cancer Database & Mouse Bladder Cancer Cell Line. Pharmgenomics Pers Med 2023;16:219-28. [Crossref] [PubMed]
13. Wang L, Wang Y, Wang J, et al. Identification of a Prognosis-Related Risk Signature for Bladder Cancer to Predict Survival and Immune Landscapes. J Immunol Res 2021;2021:3236384. [Crossref] [PubMed]
14. Nie S, Huili Y, He Y, et al. Identification of Bladder Cancer Subtypes Based on Necroptosis-Related Genes, Construction of a Prognostic Model. Front Surg 2022;9:860857. [Crossref] [PubMed]
15. Gu L, Chen Y, Li X, et al. Integrated Analysis and Identification of Critical RNA-Binding Proteins in Bladder Cancer. Cancers (Basel) 2022;14:3739. [Crossref] [PubMed]
16. Zhou Z, He H, Wang K, et al. Granzyme A from cytotoxic lymphocytes cleaves GSDMB to trigger pyroptosis in target cells. Science 2020;368:eaaz7548. [Crossref] [PubMed]
17. Sarrió D, Martínez-Val J, Molina-Crespo Á, et al. The multifaceted roles of gasdermins in cancer biology and oncologic therapies. Biochim Biophys Acta Rev Cancer 2021;1876:188635. [Crossref] [PubMed]
18. Lawlor DA, Harbord RM, Sterne JA, et al. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med 2008;27:1133-63. [Crossref] [PubMed]
19. Võsa U, Claringbould A, Westra HJ, et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat Genet 2021;53:1300-10. [Crossref] [PubMed]
20. Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol 2011;40:755-64. [Crossref] [PubMed]
21. Qiu Y, Jiang Z, Zhang J. Causal effects of BMI, waist circumference, and body fat percentage on the risk of bladder cancer: A Mendelian randomization study. Medicine (Baltimore) 2024;103:e38231. [Crossref] [PubMed]
22. Kurki MI, Karjalainen J, Palta P, et al. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature 2023;613:508-18. [Crossref] [PubMed]
23. Elsworth B, Mitchell R, Raistrick C, et al. MRC IEU UK Biobank GWAS pipeline, version 2. 2019.
24. Mubeen S, Tom Kodamullil A, Hofmann-Apitius M, et al. On the influence of several factors on pathway enrichment analysis. Brief Bioinform 2022;23:bbac143. [Crossref] [PubMed]
25. Sherman MH, Beatty GL. Tumor Microenvironment in Pancreatic Cancer Pathogenesis and Therapeutic Resistance. Annu Rev Pathol 2023;18:123-48. [Crossref] [PubMed]
26. Wang J, Qin D, Tao Z, et al. Identification of cuproptosis-related subtypes, construction of a prognosis model, and tumor microenvironment landscape in gastric cancer. Front Immunol 2022;13:1056932. [Crossref] [PubMed]
27. Du T, Gao J, Li P, et al. Pyroptosis, metabolism, and tumor immune microenvironment. Clin Transl Med 2021;11:e492. [Crossref] [PubMed]
28. Chen ZA, Tian H, Yao DM, et al. Identification of a Ferroptosis-Related Signature Model Including mRNAs and lncRNAs for Predicting Prognosis and Immune Activity in Hepatocellular Carcinoma. Front Oncol 2021;11:738477. [Crossref] [PubMed]
29. Zhao Q, Gao S, Chen X, et al. POC1A, prognostic biomarker of immunosuppressive microenvironment in cancer. Aging (Albany NY) 2022;14:5195-210. [Crossref] [PubMed]
30. Liu L, Bai X, Wang J, et al. Combination of TMB and CNA Stratifies Prognostic and Predictive Responses to Immunotherapy Across Metastatic Cancer. Clin Cancer Res 2019;25:7413-23. [Crossref] [PubMed]
31. Cristescu R, Mogg R, Ayers M, et al. Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 2018;362:eaar3593. [Crossref] [PubMed]
32. Le DT, Durham JN, Smith KN, et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science 2017;357:409-13. [Crossref] [PubMed]
33. Yang W, Hu X, Li X. Unveiling the role of gasdermin B in cancer and inflammatory disease: from molecular mechanisms to therapeutic strategies. PeerJ 2025;13:e19392. [Crossref] [PubMed]
34. Rühl S, Broz P. Regulation of Lytic and Non-Lytic Functions of Gasdermin Pores. J Mol Biol 2022;434:167246. [Crossref] [PubMed]
35. Li L, Jiang M, Qi L, et al. Pyroptosis, a new bridge to tumor immunity. Cancer Sci 2021;112:3979-94. [Crossref] [PubMed]
36. Barber G, Anand A. Characterizing caspase-1 involvement during esophageal disease progression. Cancer Immunol Immunother 2020;69:2635-49. [Crossref] [PubMed]
37. Luan J, Ju D. Inflammasome: A Double-Edged Sword in Liver Diseases. Front Immunol 2018;9:2201. [Crossref] [PubMed]
38. He WT, Wan H, Hu L, et al. Gasdermin D is an executor of pyroptosis and required for interleukin-1beta secretion. Cell Res 2015;25:1285-98. [Crossref] [PubMed]
39. Zhang Y, Bai Y, Ma XX, et al. Clinical-mediated discovery of pyroptosis in CD8(+) T cell and NK cell reveals melanoma heterogeneity by single-cell and bulk sequence. Cell Death Dis 2023;14:553. [Crossref] [PubMed]
40. Zhang Z, Chen L, Chen H, et al. Pan-cancer landscape of T-cell exhaustion heterogeneity within the tumor microenvironment revealed a progressive roadmap of hierarchical dysfunction associated with prognosis and therapeutic efficacy. EBioMedicine 2022;83:104207. [Crossref] [PubMed]
41. Jackstadt R, van Hooff SR, Leach JD, et al. Epithelial NOTCH Signaling Rewires the Tumor Microenvironment of Colorectal Cancer to Drive Poor-Prognosis Subtypes and Metastasis. Cancer Cell 2019;36:319-336.e7. [Crossref] [PubMed]
42. Gámez-Chiachio M, Molina-Crespo Á, Ramos-Nebot C, et al. Gasdermin B over-expression modulates HER2-targeted therapy resistance by inducing protective autophagy through Rab7 activation. J Exp Clin Cancer Res 2022;41:285. [Crossref] [PubMed]
43. Kong Q, Xia S, Pan X, et al. Alternative splicing of GSDMB modulates killer lymphocyte-triggered pyroptosis. Sci Immunol 2023;8:eadg3196. [Crossref] [PubMed]