3.1 The landscape of gasdermin family genes in pan-cancer

The differentially expressions of Gasdermin family genes in cancers were presented in Fig. 1A. GSDMA, GSDMB, GSDMC, GSDMD, GSDME and PJVK were significantly upregulated in BLCA, BRCA, COAD, ESCA, GBM, KICH, KIRC, KIRP, LIHC, LUAD, PRAD, and READ. GSDME was significantly down-regulated in HNSC and THCA. CNV affected the mRNA expression of Gasdermin family genes in cancers. GSDMD expression was positively associated with CNV in most cancers (Fig. 1B). GSDMB, GSDME, GSDMC, and PJVK showed positive association with CNV in KIRP, CESC, LUSC, ESCA, LUAD, HNSC, BRCA, and OV. GSDMA was only positively associated with GSDMA in KIRP and KIRC. The methylation levels significantly affected the mRNA expression of Gasdermin family. As shown in Fig. 1C, five Gasdermin family genes were negatively associated with methylation levels in cancers (Fig. 1C). All five genes showed high single nucleotide variants (SNV) in UCEC, SKCM, COAD, STAD, ESCA, LUAD, BLCA, and LUSC (Fig. 1D). The Fig. 1E and F showed the variant classification and type and gene alteration frequency. The missense mutation was main variant classification followed by nonsense, frame shirt del and splice site. T > G and T > A were primary variant type of SNP. All gene alteration frequencies were 33% for GSDMC, 24% for GSDME, 20% for GSDMA, 17% for PJVK, 15% for GSDMB, and 14% for GSDMD.

Fig. 1

Expression levels and correlations of Gasdermin family genes in pan-cancer. A Elevated or decreased expression levels of Gasdermin family genes in whole cancers. B Heatmap plots shows the correlations of CNV with mRNA expression of Gasdermin family genes in specific cancers. C Correlations between methylation and mRNA expression of inputted genes in the specific cancers. D Profile of SNV of Gasdermin family genes in specific cancers. E Variant classification and type of Gasdermin family genes in cancers. F Gene alteration levels of Gasdermin family genes in cancers

3.2 Prognostic value of gasdermin family genes in pan-cancer

Subsequently, the prognosis-predictive value of the Gasdermin family genes was further explored through the analysis of multiple databases. As indicated by K-M survival curves, the Gasdermin family gene expression is linked to the prognosis of many cancer types (Supplementary material 1). The macroscopic analysis revealed that in many cases, a lower expression level of GSDM genes was linked to a higher survival rate, such as individuals with KIRC, KICH, and Uveal Melanoma (UVM) having better overall survival. However, the opposite was seen in patients with BLCA and Skin Cutaneous Melanoma (SKCM), where enhanced expression of GSDM genes was linked to better OS. More specifically, for distinct types of malignancies, the expression of GSDMB genes had variable impacts on patient survival. For example, the higher expression level of the GSDMB gene was linked to lower OS in individuals with KIRC (N = 530, P < 0.001). Contrarily, higher GSDMB expression was favorable for the OS in individuals with BLCA (P < 0.001), SKCM (P = 0.011), and Uterine Carcinosarcoma (UCS) (P = 0.026). The expression of GSDMC has a positive impact on the overall survival of COAD (P = 0.031) and LGG (P = 0.007) but a negative impact on the OS in individuals with BRCA (P = 0.014), KICH (N = 64, P = 0.031), KIRC (P = 0.007), Pancreatic adenocarcinoma (PAAD) (P = 0.046), and Uveal Melanoma (UVM) (P = 0.007). The expression of GSDMD had a positive impact on the OS in BLCA individuals (P = 0.007), SKCM (P = 0.030), and UCEC (P = 0.047) but a negative impact on the OS in UVM individuals (P = 0.007) and Brain Lower Grade Glioma (LGG) (P < 0.001). The expression of GSDME had a positive impact on the OS in Adrenocortical carcinoma (ACC) individuals (P = 0.005), but a negative impact on the OS in KIRC individuals (P = 0.002), LIHC (P = 0.001), and STAD (N = 350, P = 0.040). Higher expression of PVJK had a positive impact on the OS in Acute Myeloid Leukemia (LAML) individuals (P = 0.002), Mesothelioma (MESO) (P = 0.010), PAAD (P = 0.037) and Sarcoma (SARC) (P = 0.005) but a negative impact on the OS in KIRC individuals (P < 0.001).

Additionally, a predictive evaluation was conducted as per the Hazard Ratio in pan-cancer using COX regression analysis. The outcomes indicated that GSDMA had a unfavored impact on OS of KIRC, but a favor impact in KP, LAML, MESO, and SARC (Fig. 2A). GSDMB had a negative impact on OS of HNSC, KICH, KIRC, LIHC, and UCEC, but favor impact in ACC and KIRP (Fig. 2B). GSDMC had a negative impact on LIHC, PAAD, SKCM, and THCA (Fig. 2C). Moreover, GSDMD acted as a high-risk factor in ACC, KIRC, LGG, and UVM and as a low-risk factor in KIRP and SKCM (Fig. 2D). GSDME played a role as an adverse factor in KIRC and LIHC (Fig. 2E). Ultimately, PVJK served as an unsatisfying factor in KICH and SCKM (Fig. 2F).

Fig. 2

Forest plots presented hazard ratios of Gasdermin family genes for different cancers. A GSDMA. B GSDMB. C GSDMC. D GSDMD. E GSDME. F PJVK

Likewise, the link between GSDM family gene expression and the prognosis of cancers was elucidated in the PrognoScan database using the GEO dataset. Specific results are showcased in Table 1. Specifically, GSDMB served a negative predictive role in Breast cancer (DFS, DMFS & RFS) and soft tissue cancer (DMFS) but a positive prognosis-predictive role in Colorectal cancer (DFS and OS), Lung cancer (OS & RFS), Brain cancer (OS), Blood cancer (OS) and Skin cancer (OS). GSDMC played a negative prognosis-predictive role in Breast cancer (DFS, DMFS & RFS) and Lung cancer (OS). GSDMD served as a good prognostic factor in Breast cancer (DMFS & RFS) and Ovarian cancer (DFS & OS) but a poor prognostic factor in Eye cancer (DMFS), Brain cancer (OS), Blood cancer (OS) and Lung cancer (RFS).

Table 1 Prognostic value of GADMB, GSDMC and GSDMD in some cancers from GEO dataset3.3 Correlations of GSDM genes with tumor microenvironment, stemness score, and immune status

In order to find out the correlation between the expression of GSDM genes and immune subtypes, correlation analysis was carried out [26]. As a milieu comprising immune components, tumor vessels, extracellular matrix, malignant cells, and signaling molecules, tumor microenvironment (TME) heterogeneity is crucial for the advancement, metastasis, and clinical prognosis of tumors [27]. Thus, further immune infiltration analysis was carried out to observe the link between the expression of GSDM family genes and TME in pan-cancer using the ESTIMATE algorithm. The outcomes of this research exposed that GSDM family genes have a significant association (positive/negative) with stromal and immune scores in pan-cancer. On the other hand, GSDM gene expression correlates significantly with RNAs and DNAs in pan-cancer (Fig. 3). Figure 4 showed the correlations of GSVA scores of GSDM family genes with immune cell infiltration levels in cancers. We found that NK cells, Th1, CD8 T cells, cytotoxic and exhausted cells showed positive associations with GSVA score in most cancers. CD8 naïve, Neutrophil, Th17, central memory CD4 naïve, and B cells showed negative associations with GSVA score in most cancers.

Fig. 3

Correlations of Gasdermin family genes with tumor microenvironment, tumor purity, and tumor stem cells characteristics (The larger dots indicate smaller P values, while blank grids signify that the P value is greater than 0.05)

Fig. 4

Correlation of Gasdermin family genes with immune cell infiltrations levels in different cancers

3.4 Drug susceptibility of GSDM genes in pan-cancer

For the purpose of determining the possible association between the expression of GSDM family genes and drug susceptibility in various cancers from the CellMiner™ database, correlation analysis was made using the R software. The results demonstrated that GSDMA expression had a positive link to Dexrazoxane drug susceptibility, while GSDMB had a positive link to Nelarabine, Fluphenazine, and Perifosine drug susceptibility. GSDMC expression was negatively linked to Ixazomib citrate, Midostaurin, Bortezomib, Pralatrexate, AT-13387, Vismodegib, and Vincristine and was positively linked to Gefitnib and Lificguat. GSDMD demonstrated a positive link to the susceptibility to Fludarabine, Cladribine, and 5-fluoro deoxy uridine 10. All the results are illustrated in Fig. 5A and 5B.

Fig. 5

Drug sensitivity analysis of Gasdermin family genes in whole cancers. A GDSC drug sensitivity and mRNA expression. B CTRP drug sensitivity and mRNA expression

3.5 High GSDMB is correlated with clinical features in KIRC

The KIRC tissues were compared to adjacent healthy tissues in order to investigate the expression of GSDMB in KIRC patients. The outcomes highlighted that GSDMB overexpressed in KIRC tissues (p = 2.835*e−4, Fig. 6A). The receiver operating curve (ROC) was generated to inspect the effectiveness of the differential expression of GSDMB in distinguishing normal and KIRC tissues. The area under the curve (AUC) of GSDMB was 0.813 (Fig. 6B), which indicated significant diagnostic value and thus a potential biomarker for KIRC therapy.

Fig. 6

Differential expression and correlation of GSDMB with clinical characteristic. A GSDMB is elevated in KIRC tissue. B Receiver operating characteristics of GSDMB for diagnosing KIRC. C–F The expression of GSDMB is associated with clinical grade and stage

Furthermore, the GSDMB expression and its clinical characteristics in individuals with KIRC were assessed. The clinical and GSDMB expression data of 520 individuals were collected from the TCGA database. The individuals with KIRC were categorized into high- and low-GSDMB expression groups according to the mean expression (Table 2). The outcomes revealed that high GSDMB expression was linked to histologic grade (p = 0.0071, Fig. 6C), pathologic stage (p = 0.00016, Fig. 6D), T stage (T1-2 vs T3-4, p = 0.00027, Fig. 6E), and M stage (p = 0.031, Fig. 6F). The results suggested that high GSDMB expression correlated with T (p = 0.028), M (p = 0.008), and pharmaceutical (p = 0.002) (Table 2).

Table 2 Association between GSDMB expression and clinical characteristics3.6 Prognosis and clinical value of GSDMB in KIRC

In the survival analysis of KIRC from the TCGA dataset, the results demonstrated that high GSDMB expression had a significant link to poor OS (p < 0.001, Fig. 7A); poor PFI (p = 0.016, Fig. 7B); unsatisfying DFI (p < 0.022, Fig. 7C) and poor DSS (p < 0.001, Fig. 7D). Subsequently, univariate cox regression analysis indicated high GSDMB expression was remarkably linked to poor OS (hazard ratio [HR] = 1.339, 95% CI 1.234–1.452, Fig. 7E). Multivariate cox regression analysis highlighted that high GSDMB expression served as an independent prognostic factor for OS in individuals with KIRC (HR = 1.202, 95% CI 1.099–1.315, Fig. 7F).

Fig. 7

GSDMB is an independent prognosis factor in KIRC. A–D Highly expressed GSDMB is associated with OS, PFI DFI and DSS. E Univariate cox regression of GSDMB for OS. F Multivariate cox regression of GSDMB for OS

Furthermore, to predict the 1-, 3-, and 5-year OS in KIRC patients from the TCGA dataset, the nomogram was developed using age, grade, stage, radiation, pharmaceutical, and GSDMB expression (Fig. 8A). The ROC analysis verified that the AUC at 3 years was 0.923 (Fig. 8B). Overall, the nomogram model provided a terrific calibration method for predicting 3-year OS (Fig. 8C).

Fig. 8

Establishment and validation of nomogram model based on prognostic signature genes. A Nomogram model established in KIRC patients. B AUC of GSDMB for predicting 3-year overall survival. C The 3-year calibration curves in the KIRC

3.7 Function and pathway analysis of GSDMB in KIRC

To investigate the potential interaction network of the differential expression genes in high- and low-GSDMB expression groups, PPI analysis was carried out. The results showed that many genes were correlated with GSDMB in KIRC (Fig. 9A). To further explore the specific molecular function and signaling pathways of the genes correlated with GSDMB, GO and KEGG enrichment analyses were carried out. The results of GO enrichment analysis showed genes correlated with GSDMB were mainly enriched in RNA splicing (Fig. 9B). KEGG enrichment also indicated a remarkable involvement of GSDMB in spliceosomes (Fig. 9C).

Fig. 9

Function analysis of GSDMB in KIRC. A PPI of genes correlated with GSDMB. B Functional enrichment of genes correlated with GSDMB. C KEGG pathway analysis of genes correlated with GSDMB

3.8 Correlations of GSDM genes with immune status and stemness score in KIRC

A correlation analysis demonstrated the possible correlation in KIRC between the GSDM family gene expression and immune types, stemness scores, or TME. The results highlighted that GSDMA, GSDMC, GSDMD, and GSDME were linked to immune subtypes in KIRC (Fig. 10A). These results were supported by further analysis indicating that GSDMA and GSDME were highly expressed in C6 but lowly expressed in C5. GSDMD exhibited a higher expression in C2 and a lower expression in C5. Different from other GSDM genes, GSDMC uniquely exhibited a higher expression in C5. Further results of correlation analysis in KIRC (Fig. 10B) indicated that GSDMB (R = −0.15, p = 0.0087) and PVJK (R = -0.12, p = 0.029) were negatively linked to RNAs, and contrarily, GSDME (R = 0.22, p = 8.4*e-5) was positively linked to RNAs. GSDME (R =−0.19, p = 0.0069) was also negatively linked to DNAs. Regarding TME, GSDMA and GSDME had a positive link to stromal, immune, and ESTIMATE scores (p < 0.05). GSDMB was positively linked to immune score (R = 0.22, p = 7.1*e-5) but negatively linked to stromal score (R = −0.12, p = 0.03). GSDMC had a negative link to stromal score (R = 0.14, p = 0.015), immune score (R = 0.097, p = 0.088), and ESTIMATE score (R = 0.14, p = 0.014). GSDMD correlated with stromal score negatively (R = −0.2, p = 0.00044) but immune score (R = 0.28, p = 5.1*e-7) and ESTIMATE score (R = −0.085, p = 0.14) positively. PVJK was negatively associated with stromal score (R = −0.24, p = 2.2*e-5) and ESTIMATE score (R = −0.16, p = 0.0046).

Fig. 10

Correlation of Gasdermin family genes with immune subtype, tumor microenvironment and stemness score in KIRC. A Differential expression of Gasdermin family genes among different immune subtype in KIRC. B correlation of Gasdermin family genes with stromal score, immune score, RNA stemness score and DNA stemness score in KIRC