Pancancer analysis reveals the role of disulfidptosis in predicting prognosis, immune infiltration and immunotherapy response in tumors

Abstract

Disulfidptosis has been reported as a novel cell death process, suggesting a therapeutic strategy for cancer treatment. Herein, we constructed a multiomics data analysis to reveal the effects of disulfidptosis in tumors. Data for 33 kinds of tumors were downloaded from UCSC Xene, and disulfidptosis-related genes (DRGs) were selected from a previous study. After finishing processing data by the R packages, the expression and coexpression of DRGs in different tumors were assessed as well as copy number variations. The interaction network was drawn by STRING, and the activity of disulfidptosis was compared to the single-sample gene set enrichment analysis algorithm. Subsequently, the differences in DRGs for prognosis and clinicopathological features were evaluated, and the tumor immune microenvironment was assessed by the TIMER and TISCH databases. Tumor mutation burden, stem cell features and microsatellite instability were applied to predict drug resistance, and the expression of checkpoints was identified for the prediction of immunotherapy. Moreover, the TCIA, CellMiner and Enrichr databases were also utilized for selecting potential agents. Ten DRGs were differentially expressed in tumors, and the plots of coexpression and interaction revealed their correlation. Survival analysis suggested SLC7A11 as the most prognosis-related DRG with the most significant results. Additionally, the comparison also reflected the differences in DRGs in the status of pathologic lymph node metastasis for 5 types of tumors. The tumor immune microenvironment showed commonality among tumors based on immune infiltration and single-cell sequencing, and the analysis of tumor mutation burden, stemness and microsatellite instability showed a mostly positive correlation with DRGs. Moreover, referring to the prediction about clinical treatment, most DRGs can enhance sensitivity to chemotherapeutic agents but decrease the response to immune inhibitors with increasing expression. In this study, a primarily synthetic landscape of disulfidptosis in tumors was established and provided guidance for further exploration and investigation.

1. Introduction

Malignant tumors appear to be a global health problem threatening human life and leading to death.^[1,2] For patients with solid tumors, surgical resection combined with pre- and/or postoperative individual comprehensive treatment is considered the major choice; nevertheless, for advanced tumors, which always exhibit poor prognosis, surgical treatment may not promote satisfactory results and negative incisal margins.^[3,4] In addition, for nonsolid cancers (e.g., hematologic tumors), precise chemotherapy is considered the preferred treatment under the recommendation guidelines.^[5] However, existing therapeutic plans have several limitations, and patients may suffer from tumor recurrence, lymph node metastasis and even distant metastases after systematic treatment.^[6–8] Therefore, more precise and effective medical technology is still necessary and essential for further detailed and precise treatment. During the past decades, with the development of life science, immunotherapy has achieved beneficial results in several tumors (e.g., melanoma).^[9,10] Specifically, immune checkpoint inhibitor (ICI) therapy has become the major immunotherapeutic plan by restraining the expression of related genes and activating the patient’s own immune system.^[11] However, not all patients can benefit from immunotherapy due to the complex tumor immune microenvironment (TIME).^[12,13] Immune tolerance, escape and silencing are the most common problems leading to immunotherapy failure.^[14] Therefore, a detailed assessment of the TIME is important and crucial for guiding clinical treatment.

Programmed cell death plays important roles in biological processes, influencing cell growth, proliferation, and migration.^[15,16] Multiple previous studies have reported the effects of therapeutic agents targeting regulators of cell death.^[17,18] For instance, while strengthening the activity of ferroptosis in engineered mice, the progression of pancreatic cancer can be significantly inhibited.^[19] In addition, in both head and neck squamous cell carcinoma and glioblastoma, enhanced sensitivity to necroptosis is able to promote tumor metastasis.^[20,21] In contrast, the same kind of cell death model may also be thought to be a protective mechanism in other tumors.^[22,23] Due to differential response in different tumors, the relationship between cell death and tumors requires further exploration. Similarly, cell death participates in the translocation of immune cells and regulation of the TIME.^[24,25] As indicated by Seifert et al^[26], pancreatic oncogenesis can be promoted by necrosomes through Mincle-induced immune suppression. In addition, the occurrence of liver cancer is associated with the necroptosis microenvironment.^[27] Therefore, treatment focused on regulating cell death shows great potential in tumor therapy, especially immunotherapy.

Recently, Liu et al^[28] discovered and determined a new cell death pattern in high-SLC7A11 cells. Under glucose starvation, overabundant intracellular disulfides accumulate in SLC7A11^high cells, leading to an uncharacterized form distinct from apoptosis and ferroptosis. After being named disulfdptosis, the authors found that F-actin collapsed in glucose starvation and aberrant disulfide bonds were induced in an SLC7A11-dependent manner. Additionally, glucose transporter inhibitors can suppress SLC7A11^high tumor growth by downregulating the activity of disulfdptosis. Considering these promising results, disulfdptosis seems to display great prospects in the treatment of tumors; nevertheless, due to limited studies, the underlying mechanism of disulfdptosis in cancers and their correlation remain unknown. Hence, we conducted this pancancer analysis for primary exploration based on multiomics data from various databases.

2. Methods and materials

2.1. Obtaining disulfidptosis-related genes (DRGs) and the pancancer dataset

As determined by Liu et al^[28], 10 DRGs were selected for this analysis, including GYS1, NDUFS1, NDUFA11, NUBPL, LRPPRC, SLC7A11, SLC3A2, RPN1, NCKAP1, and OXSM, which are considered regulators of the Rac1-WRC-Arp2/3 signaling axis. Pancancer-related data, including the RNA-seq matrix, clinicopathological features, somatic mutation information, and copy number variations (CNVs), were downloaded from the TCGA database through the UCSC Xene online database. After annotating the gene symbols through the Ensembl website, the DRG expression matrix of the merged pancancer cohort was established using the “limma” R package.

2.2. Expression levels of DRGs and activity of disulfidptosis between tumor and normal samples

The total expression of DRGs in each TCGA tumor was calculated, and the differential expression between tumor and normal samples was compared by the Wilcox test according to the log2-fold change value. In addition, the coexpression of 10 DRGs was analyzed with the “Corrplot” R package. A protein–protein interaction network of DRGs was constructed via the STRING database (http://string-db.org/). In addition, the activity of disulfidptosis in each tumor sample was calculated by single-sample gene set enrichment analysis (ssGSEA) referring to the DRG set and was compared in TCGA tumors.

2.3. Correlation analysis between DRGs and clinical features

After calculating the expression of 10 DRGs of patients, patients were divided into low- and high-expression groups based on the median expression value. K–M survival analysis was conducted to compare the prognosis mainly focused on overall survival (OS) in each low-DRG and high-DRG group. Moreover, univariate Cox hazard regression analysis was also performed to evaluate the relationship between DRGs and OS. Subsequently, the correlation between DRGs and clinicopathological features was assessed in some tumors, including colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), rectum adenocarcinoma (READ) and stomach adenocarcinoma (STAD).

2.4. Immunohistochemistry

To preliminarily corroborate the association of survival-related DRGs and the above 5 kinds of tumors, immunohistochemistry (IHC) images were obtained via the human protein atlas (HPA) database (https://www.proteinatlas.org), including tumor cells and the corresponding normal cells.

2.5. Assessment of the tumor microenvironment

To further assess the TIME, the “estimate” R package was applied to calculate the immune score, stromal score, ESTIMATE score, and tumor purity of each patient. Spearman correlation analysis was then performed to explore the relationship between DRGs and TIME scores. Similarly, the immune cell infiltration state was assessed by various platforms via the TIMER database based on the Spearman correlation test. Furthermore, the expression levels of DRGs in different immune subtypes were compared in COAD, ESCA, HNSC, READ, and STAD. In addition, the expression levels of approximately 10 DRGs in cells around malignant cells were analyzed based on single-cell RNA-seq datasets from the TISCH database.

2.6. Tumor mutation burden (TMB), stemness, and microsatellite instability (MSI)

The correlation between DRG expression and TMB was calculated with the Spearman correlation test, as shown in the radar plot. In addition, the stem cell features of the DNA stem score (DNAss) and RNA stem score (RNAss) for cancers were extracted from the UCSC Xene database. Moreover, the association with MSI was also assessed by correlation analysis.

2.7. Prediction of clinical treatment

To explore the roles of DRGs in the immunotherapeutic response, the expression relationship between DRGs and immune inhibitor-related genes was examined. Additionally, immunotherapy-related data were obtained from the TCIA database, and the effects were evaluated. Similarly, to predict and select potential drugs for tumors, the CellMiner and Enrichr databases were applied to identify potential agents.

3. Results

3.1. Multiomics landscape of DRG expression levels in TCGA cohorts

The expression differences in DRGs among 33 kinds of tumors are shown in Figure 1A, which indicates that the RPN1 gene had the highest average expression and that SLC7A11 had the lowest average expression. Based on the Wilcoxon test, the differences in log2-fold change values between tumor and normal samples were visualized by the heatmap shown in Figure 1B. Referring to these results, NDUFS1 was downregulated in tumor samples, whereas SLC3A2 was higher in most cancers. The protein–protein interaction network diagram revealed the interactive relationship of these DRGs. In addition, referring to the results regarding disulfidptosis activity based on ssGSEA, patients in the TCGA-DLBC cohort had the highest scores; nevertheless, LUAD patients exhibited the lowest activation levels. (Fig. 1C) The correlation plot reflects the internal relationship of these 10 DRGs, and the associated ratios are also displayed in the plot. (Fig. 1D) Similarly, as indicated by the heatmap of CNVs in Figure 1E, the CNVs of NDUFA11 and GYS1 exhibited many copy number deletions in most cancers; however, SLC3A2 showed a greater increase. (Fig. 1F)

Figure 1.

Multiomics landscape of DRG expression levels in TCGA cohorts. (A) Expression of DRGs in the TCGA cohort, (B) comparative expression of DRGs between tumor and normal samples, (C) protein–protein interaction network, (D) coexpression relationship of DRGs, (E) activity score of disulfidptosis in tumors based on ssGSEA, (F) preferred copy number variations in tumors. DRGs = disulfidptosis-related genes, ssGSEA = single-sample gene set enrichment analysis.

3.2. Correlation between DRGs and clinical features

Based on the uni-Cox survival regression analysis, the hazard ratio for 10 DRGs in 33 different tumors regarding OS was evaluated and summarized in a forest plot. (Fig. 2A) In addition, 6 K–M survival curves for COAD, HNSC and READ were selected, and the P values were < .05. Among them, HNSC patients with low LRPPRC or SLC3A2 expression had a better prognosis than those with high-expression. For patients with COAD, the upregulated expression of OXSM enhanced OS, but the higher expression of PRN1 increased the risk. Similar results were also obtained in READ patients, in which the high-LRPPRC or high-NUBPL groups exhibited longer survival times. (Fig. 2B) Moreover, the remaining results of the K–M survival comparison are also summarized and attached (Figure S1, Supplemental Digital Content, http://links.lww.com/MD/L259, which demonstrates the results of K–M survival analysis).

Figure 2.

Association between DRGs and clinicopathological features. (A) Forest plot of univariate Cox regression analysis of DRGs in tumors, (B) K–M survival comparison analysis, (C) correlation between DRGs and pathological lymph node metastasis in COAD, ESCA, HNSC, READ, and STAD. COAD = colon adenocarcinoma, DRGs = disulfidptosis-related genes, ESCA = esophageal carcinoma, HNSC = head and neck squamous cell carcinoma, READ = rectum adenocarcinoma, STAD = stomach adenocarcinoma.

Subsequently, we also compared the differences in pathologic lymph node metastasis in 5 tumors of COAD, ESCA, HNSC, READ, and STAD. As indicated by Figure 2C, the expression of LRPRC and SLC3A2 exhibited significant differences in pathologic lymph node metastasis in COAD, and SLC7A11 appeared to have different expression levels in HNSC. Moreover, in different N stages of COAD and ESCA, the expression of OXSM also exhibited significant differences. (P < .0001 and P < .01, respectively).

In addition, IHC images of COAD, HNSC, READ and their related normal samples were downloaded from the HPA database for external validation. (Figure S2, Supplemental Digital Content, http://links.lww.com/MD/L261, which displays the IHC results for prognostic DRGs via the HPA database) As indicated by these IHC results, LRPPRC and SLC3A2 were also considered useful diagnostic biomarkers and prognostic predictors.

3.3. Assessment of immune infiltration

According to the “estimate” R package, the immune score, stromal score, ESTIMATE score and tumor purity of each patient were assessed, and the correlation analysis was constructed and compared based on Spearman correlation analysis. (Fig. 3A) In addition, the immune cell infiltration status was explored with multiple algorithms via the TIMER database. Specifically, we focused on the relationships among B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages and dendritic cells in different tumors (Figure S3, Supplemental Digital Content, http://links.lww.com/MD/L262, which reveals immune cell infiltration via the TIMER database). As indicated by the correlation analysis, the CD8 + T cells exhibited similar infiltration tendencies in COAD and READ while considering most DRGs; however, HNSCC seemed to exhibit the opposite association results. In addition, we also compared the differences in DRG expression among different immune subtypes. Referring to the results mentioned in Figure 3B, among the above 5 tumors, LRPPRC exhibited different expression levels among different immune subtypes, including immune subtype C1 (wound healing), C2 (IFN-gamma dominance), C3 (inflammation), C4 (lymphocyte depletion) and C6 (TGF-beta dominance).

Figure 3.

Assessment of tumor immune microenvironment. (A) Stromal scores, immune scores, estimate scores and tumor purity in tumors, (B) correlation between DRGs and immune subtypes in COAD, ESCA, HNSC, READ, and STAD. COAD = colon adenocarcinoma, DRGs = disulfidptosis-related genes, ESCA = esophageal carcinoma, HNSC = head and neck squamous cell carcinoma, READ = rectum adenocarcinoma, STAD = stomach adenocarcinoma.

Similarly, to investigate the expression level of DRGs in cells among tumor cells, we used the TISCH database to visualize UMAP and violin plots. (Fig. 4) Referring to the results of 4 scRNA-seq cohorts (GSE146771, GSE139324, GSE173950, and GSE167297), the NDUFA11 gene was commonly expressed in most kinds of immune and stromal cells in the TIME of the above tumors. In addition, RPN1 and SLC3A2 were also found to display higher expression in plasma cells.

Figure 4.

Expression of DRGs in cells based on single-cell RNA sequencing via the TISCH database. DRGs = disulfidptosis-related genes.

3.4. Association of DRGs with TMB, stemness and MSI

The radar plot of DRG expression and TMB is shown in Figure 5A, suggesting a potential correlation between DRGs and TMB. Similarly, the DNAss and RNAss of each sample were determined according to the dataset from the UCSC Xene database. As shown in Figure 5B, these 10 DRGs were positively associated with DNAss in ovarian cancer (OV), indicating that the increased expression of DEGs may promote the proliferation of tumor stem cells. Similar results were also obtained for RNAss, in which DRGs also exhibited a positive association with the RNAss of OV. In addition, as shown in these plots, the LRPPRC gene was positively correlated with 32 kinds of tumors except thymoma. Furthermore, the radar plots of MSI revealed that 9 DRGs were positively associated with MSI in STAD, and 8 of them showed significant differences; nevertheless, NUBPL exhibited a negative association with a P value < .05. (Fig. 5C)

Figure 5.

Correlation analysis of DRGs with tumor mutation burden (A), stemness (B) and microsatellite instability (C). DRGs = disulfidptosis-related genes.

3.5. Immunotherapy and molecular agent therapy

After pooling these 10 DRGs in the TISDB database, the correlation between DRGs and ICI-related genes was analyzed and displayed in heatmaps. (Figure S4, Supplemental Digital Content, http://links.lww.com/MD/L260, which summarizes the correlation with checkpoint-related genes via the TISDB database) According to the above results, patients with lower DRG expression may exhibit a better response to immunoinhibitors. Subsequently, the correlation between DRG expression and the IPS of PD-1 and CTLA4 was also reflected in Figure 6A. According to these IPS results, the increasing expression of DRGs decreases the average IPS in most types of cancers, indicating a weaker immunotherapeutic response in tumors. These associations of ICI inhibitors support the TISDB results that the expression of DRGs may be negatively correlated with the effects of immunotherapy.

Figure 6.

Prediction of clinical treatment. (A) Correlation between DRGs and IPS for immune checkpoint inhibitors in tumors, (B) drug sensitivity for agents predicted by the CellMiner database. DRGs = disulfidptosis-related genes.

Subsequently, we also explored and predicted potential molecular agents for tumor therapy. As shown in Figure 6B, the top 16 potential drugs were identified based on the CellMiner database. Among them, most correlation tests reflect a negative association with DRGs, indicating that the increased expression of DRGs may decrease drug sensitivity. However, there was a contrary result between NDUFA11 and hydroxyurea, with a median strength correlation and a significant difference. In addition, as indicated by the DsigDB database, several related molecular drugs were predicted for potential treatment and are summarized in Table 1.

Table 1.

Results about potential agents predicted by DsigDB via enrichr database.

Term	P value	Odds ratio	Combined score	Genes
5109870 MCF7 UP	.000071	50.56632653	483.3988941	GYS1;OXSM;SLC7A11
F0447-0125 PC3 UP	.000183	127.8910256	1100.917922	SLC3A2;SLC7A11
Semustine PC3 UP	.000220	115.9709302	976.5813448	SLC3A2;SLC7A11
5707885 PC3 UP	.000220	115.9709302	976.5813448	SLC3A2;SLC7A11
Metformin hydrochloride	.000272	103.8645833	852.6509586	NDUFA11;NDUFS1
Securinine PC3 UP	.000294	99.7	810.6171267	SLC3A2;SLC7A11
Pyrvinium PC3 UP	.000306	97.74019608	790.9491627	SLC3A2;SLC7A11
Trimipramine PC3 UP	.000318	95.85576923	772.1076137	SLC3A2;SLC7A11
Zidovudine PC3 UP	.000329	94.04245283	754.0431617	NCKAP1;SLC7A11
Alexidine PC3 UP	.000446	80.35483871	619.922544	SLC3A2;SLC7A11

4. Discussion

Cell death plays important roles in tumor-related biological processes, including proliferation, invasion and metastasis.^[29] Referring to previous studies, different types of cell death models have been investigated, and their influence on multiple tumors has been determined.^[30–33] As a novel breakthrough point for the treatment of tumors, cell death-related genes can be considered potential therapeutic targets, which have been utilized and applied for individual treatment.^[34] Among them, immunotherapy has great potential effects on tumors, and specifically, ICI therapy is recognized as a mature technology that prolongs survival time, reduces metastasis and improves prognosis.^[9–11] However, not all patients can benefit from immunotherapy because of the different and complex states of the TIME, which may lead to immune escape and treatment failure. Importantly, consisting of immune cells and stromal cells, the TIME can also be regulated and influenced, as indicated by previous evidence-based studies.^[12–14] Disulfidptosis, as reported by Liu et al^[28], has been determined to be a novel form of cell death that can possibly be applied for tumor therapy. Herein, given the therapeutic potential, we conducted this analysis to explore the effectiveness of disulfiptosis in 33 kinds of tumors based on pancancer analysis.

According to our analysis, these 10 DRGs exhibited different average expression levels in tumors, which also exhibited significant differences compared to normal samples. The results of ssGSEA revealed the potential activity of disulfiptosis in different tumors, and the heatmap reflected the superiority of CNV for cancers. Given this, the expression landscape of DRGs was primarily displayed in our research. Similarly, survival analysis was also conducted to assess the effects on prognosis (especially in OS) for these DRGs. Based on the forest plot of the hazard ratio, the SLC7A11 gene was considered the most relevant survival biomarker among 33 tumors, which enhanced risk in 9 tumors but reduced risk in OV. Referring to a previous study, 3 prognostic genes (including SLC7A11) were applied to establish the signature for ovarian cancer, which exhibited satisfactory predictive effects.^[35] As a mediator of cystine-glutamate exchange (especially cellular uptake of L-alanosine), SLC7A11 can regulate intracellular glutathione levels and serve as a predictor of glutathione-mediated resistance to geldanamycin.^[36] K–M analysis comparing the low-DRG and high-DRG groups also supported the results that the low-SLC7A11 and high-SCL7A11 groups exhibited significant differences in survival status in 8 types of tumors, which was more than other DRGs. In summary, SLC7A11 may be regarded as the most prognosis-related DRG to predict the survival status in pancancers.

Subsequently, we also evaluated the relationship between TMB and DRGs. Previous studies have indicated that somatic mutations are strongly associated with prognosis, TIME and clinical therapeutic response.^[37–39] According to our analysis, we investigated the correlation between DRGs and TMB based on pancancer analysis. Specifically, LRPPRC showed a positive association with TMB in gastrointestinal cancers, including COAD, STAD and READ. Similarly, as demonstrated by Yang et al^[40], LRPPRC was considered the key functional and therapeutic target for P53 mutation-induced chemoresistance in colorectal cancer by specifically binding to MDR1 through the miR-34a/LRPPRC/MDR1 axis. In addition, we also found that SLC7A11 was positively correlated with the TMB of COAD, ESCA, HNSC and STAD, coinciding with a previous study showing that SLC7A11 was associated with KRAS mutation.^[41] Similarly, referring to findings about tumor stemness, most DRGs were positively associated with RNAss, especially the LRPPRC gene, indicating that the increased expression of DRGs may contribute to more stem-like cells in tumors.^[42] In addition, in our analysis, the negative association with MSI in colorectal tumors revealed a higher sensitivity for immunotherapeutic response when these DRGs were downregulated.^[43,44] Considering these correlations with TMB, stemness and MSI, we were able to predict the potential effectiveness of chemotherapy and immunotherapy.

Subsequently, we also constructed an assessment of the TIME for tumors. Referring to the estimate algorithm, which provides the quantitative immune landscape in cancers, tumor purity increased with the upregulation of DRGs, whereas the stromal, immune and ESTIMATE scores decreased when the gene expression increased. Additionally, the results of immune cell infiltration analysis suggested that COAD and READ had similar relationships with the infiltration of CD8 + T cells in most DRGs, which may provide a potential view and guideline for immunotherapy.^[45] Moreover, LRPPRC was always expressed differently among different immune subtypes, which suggests that LRPPRC is strongly associated with wound healing, IFN-gamma dominance, inflammation, lymphocyte depletion and TGF-beta dominance in these 5 kinds of tumors. The results of scRNA-seq also support the conclusion that several DRGs display similar expression levels in immune or stromal cells around malignant cells. Given this, the potential commonality of biological processes in COAD, ESCA, HNSC, READ, and STAD can possibly be revealed.

For clinical treatment, we first assessed the correlation between DRGs and ICI genes. In general, the negative correlation with ICIs for most DRGs among tumors indicated that upregulated DRGs may decrease the effectiveness of ICI therapy.^[46] However, for NCKAP1, NDUFS1, and SLC7A11, the positive association with CD274 suggested that higher expression of these 3 DRGs was able to increase the sensitivity of the PD-1 immunotherapeutic response.^[47] Similarly, the relationship between DRG expression and IPS also supported the above results that patients with lower DRG expression exhibited better immunotherapeutic responses to ICI therapy in most situations, except for NDUFS1. Referring to these results, individual immunotherapy for tumor patients can be precisely guided and advised to gain more benefits. In contrast, as indicated by the CellMiner database, the sensitivity increased with the gradually reduced expression of DRGs, which suggested a better drug sensitivity for chemotherapeutic or small molecular agents. According to the analysis of clinical treatment, we found that NCKAP1 was most associated with immunotherapy and chemotherapy. As reported by a previous study, this gene regulates the dynamics of the actin cytoskeleton to influence cancer metastasis and invasion by forming a regulatory complex.^[48] Similarly, as determined by Sertel et al^[49], NCKAP1 was identified to play crucial roles in artesunate resistance toward tumor cells. In addition, RAC1-targeted NCKAP1 reportedly promoted the progression of Braf; Pten-driven melanoma in mouse models.^[50] Considering the essential roles of NCKAP1 in both chemotherapy and immunotherapy, this DRG can be regarded as a potential therapeutic target in tumors. Hence, based on the above results, a synthetic treatment for patients can be further formulated even though immunotherapy promotes few benefits.

However, several limitations likewise appeared in our research due to the lack of external validation with sequencing data and clinicopathological information. Tumor cohorts with large samples are needed in the future for further assessment and corroboration. In addition, although we performed a pancancer analysis to primarily assess the potential effects of disulfidptosis in tumors based on multiple datasets, as a novel kind of cell death, the underlying mechanism of disulfidptosis in each cancer (e.g., related signaling pathways) remains unknown and requires further detailed exploration. Despite these limitations, this multiomics landscape analysis to evaluate and explore the role of disulfidptosis in pancancer can be considered advisable and meaningful as a primary investigation of this new biological process.

5. Conclusion

In this study, we conducted a big data analysis to evaluate the effectiveness of disulfidptosis in tumors, including expression, prognosis, TIME, TMB, stemness, MSI and clinical treatment, based on multiple datasets obtained from various databases. Given these findings, a primarily synthetic landscape of disulfidptosis in tumors was established and provided guidance for further exploration and investigation.

Author contributions

Conceptualization: Juntao Huang.

Data curation: Juntao Huang.

Formal analysis: Juntao Huang.

Funding acquisition: Yi Shen.

Investigation: Juntao Huang, Ziqian Xu, Yi Shen.

Methodology: Juntao Huang, Ziqian Xu, Yi Shen.

Project administration: Ziqian Xu, Dahua Chen.

Resources: Dahua Chen.

Software: Dahua Chen.

Supervision: Chongchang Zhou.

Validation: Chongchang Zhou.

Visualization: Chongchang Zhou.

Writing – original draft: Juntao Huang, Ziqian Xu, Yi Shen.

Writing – review & editing: Juntao Huang, Yi Shen.