Bladder carcinoma (BLCA) is the prevalent malignancy of the urinary tract and the fourth most common malignancy in men, with a high mortality and morbidity rate (1, 2). In 2020, the number of new BLCA cases was predicted to be about 81,400, and the number of new deaths was calculated to be about 17,980 in the United States (3). Of these, non-muscle invasive bladder carcinoma (NMIBC), which is prone to recurrence, accounts for approximately 75%, and muscle invasive bladder carcinoma (MIBC) and metastatic disease, which are more aggressive, account for 25% (4, 5). Currently, cisplatin-based chemotherapy remains the first-line treatment option for advanced BLCA, but the response period to chemotherapy is limited and the prognosis of patients remains dismal, with a 5-year relative overall survival rate of only 15% for metastatic disease (5, 6). Moreover, studies have indicated that the wild-type TP53 gene expression profile was resistant to neoadjuvant chemotherapy in MIBC tumors (7). The highly heterogeneous and genomically unstable nature of BLCA have also been identified by an increasing number of studies (8, 9). The genetic prognostic biomarkers of BLCA have been emerging in recent years. An example was minichromosome maintenance protein 5 (MCM5), a biomarker of cell proliferation, which is expressed in normal urinary epithelium only in cells within the basal proliferative lumen and can be used to diagnose BLCA (10). A clinical trial demonstrated that the detection of MCM5 in urine sediment was a biomarker for predicting BLCA recurrence and prognosis. However, it was not widely used in clinical practice because its false-positive rate was higher than that of cystoscopy (11). Thus, searching for gene expression characteristics of BLCA patients and identifying newly effective prognostic markers are essential procedures for personalized treatment strategies to improve clinical outcomes.

Pyroptosis is a novel non-apoptotic mode of programmed cell death (PCD) mediated by the gasdermin family, characterized by pore formation on the cytoplasmic membrane, cell swelling, and rupture, and accompanied by the efflux of cell contents. Currently, the gasdermin family comprises six human homologs, gasdermin A–D (GSDMA–D), gasdermin E (GSDME, alternatively known as DFNA5), and Pejvakin (PJVK, alternatively known as DFNB59) (12). Pyroptosis can be induced by both the classical inflammasome pathway, which is activation of caspase-1 cutting the GSDMD protein, and the non-classical inflammasome pathway, which is bacterial lipopolysaccharide binding to caspase-4/5/11 cutting the GSDMD protein (13, 14). In addition, it was recently found that the serine protease granzyme A (GZMA) in cytotoxic lymphocytes could access target cells via perforin and evoke cancer cells to undergo pyroptosis by cleaving activated GSDMB protein, as well as activate immune response (15). Meanwhile, emerging lines of evidence also suggested that selective delivery of active gasdermin protein to cancer cells or targeted cleavage of GSDME by granzymes can inspire potent antitumor immunity (16, 17). Produced during pyroptosis, IL-1β and IL-18 can impair the recruitment of neutrophils, whereas GSDMD acted as an anti-inflammatory agent to stimulate neutrophil death (18). Moreover, GSDMD also increased cleavage in activated CD8⁺ T cells and enhanced their cytolytic ability (19). Indeed, immune infiltration of natural killer (NK), M1 macrophages, CD8⁺ T, and type 1 T helper (Th1) cells typically represented a good prognosis, while high-level infiltration of neutrophils, M2 macrophages, MDSCs, Treg cells, and Th2 cells was frequently correlated with a poor prognosis (20, 21). Interestingly, Hou J et al. demonstrated that GSDMC expression was correlated with breast cancer survival prognosis (22).

These studies have demonstrated that pyroptosis was inextricably bound to tumorigenesis, invasion, and tumor immune microenvironment, implicating that induction of tumor cell pyroptosis would be a potential strategy to strengthen the efficacy of anti-cancer therapy. Moreover, it has been determined that pyroptosis-related genes (PRGs) have been developed as markers of tumor prognosis in a variety of carcinomas, such as lung cancer (23), hepatocellular carcinoma (24), and breast cancer (25), by establishing prognostic models of pyroptosis genes. However, there was relatively less research on its specific functions in BLCA. Hence, we systematically researched and validated the expression levels of genes associated with pyroptosis in normal bladder and BLCA tissues to study the prognostic value of these genes. We also analyzed the relationship between PRGs and tumor immune microenvironment by the TIMER website.

Pyroptosis is a lysis form of programmed death triggered by inflammasome and can rapidly release cellular contents such as pro-inflammatory molecules (IL-1β and IL-18) and antigens into the tumor microenvironment to activate the immune response of the body (27). A recent study has identified GSDM-mediated pyroptosis as a mechanism of cytotoxic lymphocyte killing to potentiate antitumor immunity (15). It was also confirmed that GSDME-mediated pyroptosis could promote T-cell recruitment through the release of relevant cytokines and trigger the activation of T cells to stimulate T-cell infiltration into tumor tissues (28). However, GSDME was often silenced in tumor tissues with promoter methylation, thus avoiding the occurrence of chemotherapeutic drug-induced pyroptosis and resisting antitumor effects (27, 29, 30). Hence, it is worthwhile to clarify the expression of PRGs in tumor tissues, and inducing pyroptosis in tumor cells would be possible and deemed as a robust method to promote the effect of oncotherapy.

With the development of bioinformatics and high-throughput sequencing technologies, it has become a reality to predict the prognosis of BLCA by assessing their risk level. For example, the risk models constructed based on m⁶A-related non-coding RNAs (31), pyroptosis-related non-coding RNAs (32), and ferroptosis-related genes (33), respectively, have evaluated the prognosis of BLCA patients or the response to immunotherapy. These models were constructed from different mechanisms to predict the prognosis of BLCA, yet unfortunately, the prognostic relationship and immune infiltration characteristics of pyroptosis-related mRNAs with BLCA patients were not considered.

In our study, we first explored the DEPRGs in normal and BLCA tissues, in which 31 genes were differentially expressed compared to normal tissue. These data suggested that differentially expressed genes may exert an underestimated role in the tumorigenesis and progression of BLCA. Then, we highlighted the function of DEPRGs using GO and KEGG pathway enrichment analysis. Predictably, we discovered that these genes were primarily relevant to the NLR signaling pathway, necroptosis, and regulation of IL-1β production. NLRs played an important regulatory role in inflammation-associated tumorigenesis. NLRs can specifically and sensitively recognize pathogen-associated molecular patterns, and activate caspase-1 or caspase-11, enhancing the production and release of IL-1β and IL-18 to recruit immune cells and promote the pyroptosis process (34, 35). Moreover, a study indicated that activation of the NF-κB/NLRP3/Caspase-1 pyroptosis signaling pathway exerted antitumor effects by increasing GSDMD levels and inhibiting cell proliferation and epithelial–mesenchymal transition progression (36). Therefore, DEPRGs may be potential targets for therapeutic and prognostic markers of BLCA.

Next, the expression of four key genes (CRTAC1, GSDMB, AIM2, and FOXO3) has been confirmed to be consistent with the bioinformatics analysis by IHC. However, the absence of difference in FOXO3 expression in cancerous and non-cancerous tissue may be related to both small sample size and the possibility that the FOXO3 gene was not transcribed in some adjacent non-cancerous bladder tissue. After that, they were screened to be associated with prognostic survival in BLCA by univariate and multifactorial Cox regression analysis. In addition, patient risk scores were calculated from mRNA expression levels and risk factors of each patient to obtain risk curves and K-M survival curves. Single- and multi-factor independent prognostic and clinical relevance analyses were conducted using the pyroptosis prognostic model to identify the relationship between gene expression levels and clinical prognosis. It was indicated that these genes were correlated with age, sex, grade, stage, and TNM subtype in the clinicopathological features (p < 0.05) and could be treated as a predictor of BLCA prognosis.

Cartilage acidic protein 1 (CRTAC1) is a glycosylated extracellular matrix molecule of human articular cartilage secreted by chondrocytes (37). They used gene expression data analysis from public databases to find that CRTAC1 expression was lower in lung adenocarcinoma (LUAD) than in normal tissues, demonstrating that the CRTAC1 gene was a protective gene in LUAD, and the high expression of CRTAC1 gene had a better prognosis (38). Equally, our study indicated that high expression of CRTAC1 was associated with earlier T (T1–2), N (N0) subtype, earlier stage (I and II), and low grades in BLCA patients. It suggested that the CRTAC1 gene could also be regarded as a protective gene in BLCA. Presently, there were no studies of CRTAC1 in BLCA or the relationship between CRTAC1 and pyroptosis, but our studies may provide new findings on prognostic markers in BLCA. The mechanism of CRTAC1 also needs further validation in BLCA.

GSDMB, as a member of the GSDM family, has also been observed to resemble the pyroptosis feature (39). However, the mechanism of GSDMB-induced pyroptosis was not quite clear and consistent. It was found that GSDMB-induced pyroptosis may be achieved by caspase-1 cleavage of GSDMB protein in 293T cells (40, 41). In contrast, Chen et al. demonstrated that GSDMB induced a conformational change in caspase-4 protein, which, in turn, increased caspase-4 enzyme activity and accelerated the cleavage of GSDMD, consequently causing non-canonical pyroptosis (42). In addition, the association between the high or low expression of GSDMB in normal and tumor tissues and prognosis was also controversial. Based on several studies (39, 43–45), hyperexpression of GSDMB in hepatocellular carcinoma, gastric cancer, and cervical cancer was considered as an oncogene that would appear to be involved in cancer invasion and metastasis. However, Shao Feng’s team indicated that CTLs could mediate target cell pyroptosis through GZMA cleavage of GSDMB in target cells. Respectively, they engrafted CT26 cells and B16-F10 cells with recombinant human GSDMB into mice subcutaneously, and found that GZMA/GSDMB pathway-mediated pyroptosis strengthened tumor clearance in the mouse model by enhancing the therapeutic efficacy of antitumor immunity. Meanwhile, they revealed that high GSDMB expression was relevant with OS benefit in BLCA patients (15). This is consistent with our findings. Our research also indicated that GSDMB was highly expressed in BLCA and hyper-expression of GSDMB was associated with low risk, low-grade infiltration, early stage (I and II), no lymph node infiltration (N0), and no distant metastasis (M0) of BLCA patients, all p < 0.05. Thus, GSDMB could serve as a protective prognostic biomarker for BLCA.

The forkhead box O class (FOXO) family is a widely expressed transcription factor, mainly including FOXO1, FOXO3, FOXO4, and FOXO6. The FOXO family is involved in several biological processes that were related to cell metabolism, cell proliferation, and apoptosis (46). Among them, FOXO3, also known as FOXO3a, plays a significant role in the regulation of cellular processes by targeting the expression and activity of effector genes. Meanwhile, the FOXO3 activity can be inhibited by an imbalance between kinases and phosphatases, significantly affecting cellular processes and even inducing carcinogenesis (47). The expression of FOXO3 in various tumors exhibited variable pro-carcinogenic and anti-carcinogenic effects. Qian et al. demonstrated that high expression of FOXO3 was relevant with the progression of glioblastoma, suggesting a poor prognosis for patients with glioblastoma (48), the same result for triple-negative breast cancer (49) and hepatocellular carcinoma (50). However, downregulation of FOXO3 promoted DNMT3B overexpression, leading to tumor growth in lung cancer, while clinical findings also showed that there was a high DNMT3B, low FOXO3a, and high MDM2 expression associated with low survival rates in lung cancer patients (51). Moreover, transcriptional inhibition of the FOXO3 gene directly and downregulation of FOXO3 protein phosphorylation promoted proliferation and migration of uroepithelial cancer cells in vivo and in vitro (52). Overexpression of circular RNA FOXO3 suppressed the proliferation, migration, and invasion of BLCA cell lines as well as promoted their apoptosis (53, 54). In contrast to our findings, FOXO3 was an indicator of poor prognosis in BLCA because it was enriched in high-risk groups, and the mechanism remains to be investigated. The study showed that downregulation of FOXO3 could inhibit NLRP3 inflammasome-mediated endothelial cell pyroptosis (55). However, the relationship between FOXO3 and pyroptosis in tumors has been rarely reported. Therefore, it is essential to further explore the regulation of FOXO3 on tumor cell pyroptosis.

Absent in melanoma 2 (AIM2), a typical receptor for cytoplasmic DNA contained a positively charged HIN-200 structural domain that recognized the sugar-phosphate backbone of host ectopic negatively charged double-stranded DNA (dsDNA) forming an electrostatic attraction (56). Host ectopic dsDNA activated AIM2 inflammasomes that drove the secretion of proinflammatory cytokines IL-18 and IL-1β, inducing cellular pyroptosis (57). AIM2 exhibited both growth inhibitory effects in colon, hepatoma, nasopharyngeal, and lung cancers, and proliferation-promoting effects in cutaneous squamous cell carcinomas (SCCs) and endometrial cancers (57–60). Compared with normal tissues, components of NLRP3 and AIM2 inflammasomes were highly expressed in nasopharyngeal carcinoma tissues, which correlated with an increased chance of survival in nasopharyngeal carcinoma (58). On the other hand, the absence of AIM2 inhibited the growth of cutaneous SCC (59). Therefore, the role of AIM2 in tumorigenesis was bipartite, and its outcome depended on the type of cancer. According to our findings, AIM2 appeared to be an oncogenic gene, as it was highly expressed in BLCA tissues and significantly upregulated in high-grade BLCA.

Briefly, all four genes (CRTAC1, GSDMB, FOXO3, and AIM2) were prognostic indicators of BLCA pyroptosis, with CRTAC1 and GSDMB suggesting a better prognosis, and FOXO3 and AIM2 suggesting a poor prognosis.

PRGs worked through cell membrane pore formation to activate pyroptosis, enhance antitumor immunity, and inhibit tumor growth. Expression of pyroptosis genes strengthened phagocytosis of tumor cells by tumor-associated macrophages (TAM), as well as the count and function of tumor-infiltrating NK cells (17). However, TAM infiltration in BLCA disrupted the adaptive immune response and promoted tumor vascularization, which, in turn, led to tumor progression (61). Meanwhile, one study described that BLCA can be stratified as high or low risk on the basis of immune cells status (62). It was evident that the infiltration of immune cells was of great importance to the survival prognosis of BLCA. It was shown that BLCA recruited more T cells than normal bladder cells, and that infiltrating T cells increased the proliferation and invasion of BLCA cells by modulating the ERβ/c-MET or ERβ/IL-1/c-MET signaling pathways, promoting metastasis (63). Neutrophils may promote tumor aggressiveness and progression in BLCA owing to their stimulation of excessive inflammation, release of growth factors to malignant cells, and secretion of human neutrophil polypeptides (64). Within the present research, we indicated a remarkable relevance between the expression of PRGs (CRTAC1, GSDMB, FOXO3, and AIM2) and the infiltration of six types of immune cells, including CD4⁺ T cells, CD8⁺ T cells, B cells, macrophages, dendritic cells, and neutrophils, implying that these genes not only could serve as prognostic indicators, but also may reflect the immune status.

Chen et al. (65) have previously constructed a pyroptosis-related prognostic model to predict survival in BLCA patients, although their study contained only 53 pyroptosis genes compared with 80 pyroptosis genes in our study. Moreover, there was a lack of experimental validation in their study. Unfortunately, our study also has some limitations. It is unclear how these DEPRGs in the prognostic model we constructed modulate the underlying mechanisms of the BLCA process. A further exploration of their biological functions is required by elaborate experiments. Furthermore, this was a retrospective study, and we would like to conduct further well-designed prospective studies to confirm our findings.

In summary, our study has constructed a prognostic model based on pyroptosis in BLCA and identified four key prognostic genes (CRTAC1, GSDMB, FOXO3, and AIM2), which also reflected immune status. We expect that our findings will provide new insights into emerging immunotherapeutic targets, identify biomarkers to more precisely predict survival in BLCA patients, and provide data to support clinical formulation of personalized treatment plans.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .

The studies involving human participants were reviewed and approved by the Clinical Research Ethics Committee of Taizhou People’s Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Conception: CY, QG, and XY. Equally contributing to the paper, study design, data analysis, experiment execution, and drafting: JY and HL. Experiment execution: YW. Study supervision: PF and YZ. Final approval: all authors. All authors contributed to the article and approved the submitted version.

This research was supported by the Taizhou Science and Technology Support Plan (Social Development) project (TS201903), the China Postdoctoral Science Foundation (2019M663505), and the Postdoctoral Interdisciplinary Innovation Foundation, Sichuan University (No. 0040204153243).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.