Single-cell disulfidptosis regulator patterns guide intercellular communication of tumor microenvironment that contribute to kidney renal clear cell carcinoma progression and immunotherapy

Background Disulfidptosis, an emerging type of programmed cell death, plays a pivotal role in various cancer types, notably impacting the progression of kidney renal clear cell carcinoma (KIRC) through the tumor microenvironment (TME). However, the specific involvement of disulfidptosis within the TME remains elusive. Methods Analyzing 41,784 single cells obtained from seven samples of KIRC through single-cell RNA sequencing (scRNA-seq), this study employed nonnegative matrix factorization (NMF) to assess 24 disulfidptosis regulators. Pseudotime analysis, intercellular communication mapping, determination of transcription factor activities (TFs), and metabolic profiling of the TME subgroup in KIRC were conducted using Monocle, CellChat, SCENIC, and scMetabolism. Additionally, public cohorts were utilized to predict prognosis and immune responses within the TME subgroup of KIRC. Results Through NMF clustering and differential expression marker genes, fibroblasts, macrophages, monocytes, T cells, and B cells were categorized into four to six distinct subgroups. Furthermore, this investigation revealed the correlation between disulfidptosis regulatory factors and the biological traits, as well as the pseudotime trajectories of TME subgroups. Notably, disulfidptosis-mediated TME subgroups (DSTN+CD4T-C1 and FLNA+CD4T-C2) demonstrated significant prognostic value and immune responses in patients with KIRC. Multiple immunohistochemistry (mIHC) assays identified marker expression within both cell clusters. Moreover, CellChat analysis unveiled diverse and extensive interactions between disulfidptosis-mediated TME subgroups and tumor epithelial cells, highlighting the TNFSF12-TNFRSF12A ligand-receptor pair as mediators between DSTN+CD4T-C1, FLNA+CD4T-C2, and epithelial cells. Conclusion Our study sheds light on the role of disulfidptosis-mediated intercellular communication in regulating the biological characteristics of the TME. These findings offer valuable insights for patients with KIRC, potentially guiding personalized immunotherapy approaches.


Introduction
According to Cancer Statistics for, 2023, kidney cancer stands as the second most prevalent tumor within the urinary system, demonstrating an annual growth rate of approximately 1%.Estimates for the year, 2023 project diagnoses of 52,360 cases among males and 29,440 among females in the United States, accounting for nearly half of all urinary system cancers.Additionally, it is anticipated that 14,890 patients will succumb to kidney cancer (1).Within the various histological subtypes of renal cell carcinoma (RCC), kidney renal clear cell carcinoma (KIRC) stands out as the most prevalent, constituting approximately 75%-80% of all RCC cases (2).KIRC tumors exhibit heterogeneity, high metastatic potential, and a poor prognosis (3).Therefore, comprehending the molecular mechanisms underlying KIRC is crucial, holding the potential to significantly enhance survival outcomes for patients.
Programmed cell death (PCD) encompasses metabolic, structural, and functional cellular disorders that result in irreversible damage.It is associated with cellular homeostasis, tissue remodeling, and tumor development (4).Within KIRC, various PCD forms have been identified, including anoikis (5), pyroptosis (6), necroptosis (7), and cuproptosis (8).Disulfidptosis, a recently discovered form of PCD, is characterized by inadequate cellular uptake of cysteine and NADPH supply.Upon NADPH depletion, abnormal disulfide bonds accumulate in the cell cytoskeleton, leading to actin filament disruption and eventual collapse of the actin cytoskeleton, culminating in cell death (9).Actin, a multifunctional cytoskeletal protein involved in cellular morphology, maintenance, differentiation, and intracellular transport, is also linked to disulfide formation (10,11).Alterations in the cytoskeleton of animal and plant cells may positively influence the initiation and regulation of PCD (12,13).Crucially, disulfidptosis-related genes play vital roles in tumor growth, development, invasion, and are closely associated with KIRC's pathogenesis and prognosis (14,15).
The tumor microenvironment (TME) significantly influences KIRC tumor progression and metastasis (16,17).Mass cytometry analysis revealed extensive infiltration of CD8 + PD-1 + T cells within KIRC tumors, indicating heightened lymphocyte presence in the TME (18).Immunotherapy targeting the TME, primarily employing immune checkpoint inhibitors (ICIs) like anti-PD-1/ PD-L1 and CTLA-4 inhibitors (19, 20), has revolutionized cancer treatment.In KIRC, T cell exhaustion is strongly linked to poor prognosis, potentially contributing to the immunosuppressive nature of the TME (21).In addition, the presence of tumorinfiltrating lymphocytes (TILs) is associated with a favorable prognosis in ccRCC.ccRCC also contains a significant number of myeloid-derived suppressor cells (MDSCs), which possess the ability to hinder tumor immune responses (22).Single-cell transcriptomics, unlike traditional bulk analysis, can uncover intercellular communication among different TME cell subtypes, including CAFs, tumor-associated macrophages (TAMs), B cells, and T cells (23,24).Aleksandar et al. identified upregulated tumorspecific macrophage subsets involving TREM2/APOE/C1Q as potential prognostic biomarkers and therapeutic targets for KIRC recurrence (25).Recently, Songyun et al. revealed a strong correlation between disulfidptosis-related genes and tumorinfiltrating immune cells (TIICs), particularly macrophages linked to SLC7A11 and SLC3A2 (26).However, there is limited research on specific disulfidptosis-mediated TME subgroups and their effect on patient prognosis and intercellular communication.
This study investigated the effect of disulfidptosis on major TME components-epithelial cells, fibroblasts, myeloids, B cells, and T cells (27)-utilizing single-cell RNA-seq (scRNA-seq) data from 7 KIRC tumor samples comprising 41,784 cells.Applying nonnegative matrix factorization (NMF) clustering to major KIRC subgroups using disulfidptosis regulatory factors, this study investigated the intricate interplay between these disulfidptosis-mediated TME subgroups and tumor epithelial cells.This study aimed to explore signaling pathways, functional enrichments, transcriptional features, immune characteristics, metabolic pathways, and prognostic implications within these distinct subgroups.Although TME has been shown to play an important role in tumor progression, the signaling molecules involved in intercellular communication in TME are poorly understood.Consequently, this study sheds light on the potential role of disulfidptosis in governing intercellular communication among various TME subgroups and tumor cells, thereby influencing KIRC progression.

Data source
To explore the intratumor heterogeneity of KIRC, tumor scRNAseq data were gathered from 7 patients with KIRC, aiming to scrutinize the landscape of 24 disulfidptosis regulators.These regulators were pinpointed using a genome-wide CRISPR-Cas9 screen and proteomic analyses conducted by Liu et al. ( 9) on SLC7A11high cells.The complete dataset, comprising a total of 41,784 single-cell scRNA-seq gene expression matrices from tumor samples of 7 patients, was acquired from the Gene Expression Omnibus (GEO) dataset (GSE210038) (www.ncbi.nlm.nih.gov/geo).Additionally, bulk RNA-seq data from 646 patients with KIRC were sourced from The Cancer Genome Atlas (TCGA) database (https:// portal.gdc.cancer.gov/)and the GEO dataset (GSE29609).All data utilized or produced in this study are publicly accessible through previous publications or in the public domain.

Visualization of TME cell types and subtypes in KIRC
Using the "Seurat" package within R software, Seurat objects were individually created based on scRNA-seq data from the 7 patients utilizing the CreateSeuratObject function.Subsequently, these objects were merged using the merge function to construct a comprehensive scRNA-seq gene expression matrix.Cells of subpar quality, including those containing less than 200 genes/cell, more than, 4000 genes/cell, or more than 15% mitochondrial genes, were filtered out.The scRNA-seq data underwent normalization using the NormalizeData function.Moreover, the FindVariableFeatures function was applied to identify the top, 2000 genes for data normalization.Additionally, the principal components (PCs) were computed based on the Seurat objects using the ScaleData and RunPCA functions.For this study, PC=12 and resolution=1.2were chosen.The "t-SNE (t-distributed stochastic neighbor embedding)" algorithm was utilized to condense the topPCs for dimensionality reduction.Finally, the Idents function was employed to label the cells according to the TME cell types or subtypes, and the DimPlot function was used for visualization.

Pseudotime trajectory analysis of disulfidptosis regulators for TME cells
To delve into the relationship between cell pseudotime trajectories and disulfidptosis regulators, the "Monocle" R package was used to analyze single-cell RNA data across all cell types in KIRC (28).Highly variable genes were selected based on the following thresholds: mean_expression ≥ 0.1 and dispersion_empirical ≥ 1 * dispersion_fit.The DDRTree method was employed for dimensionality reduction, with the parameter max_components set to 2. Subsequently, the function "plot_pseudotime_heatmap" was used to illustrate the pseudotime heatmap, showcasing temporal expression patterns of disulfidptosis regulators.Furthermore, the "plot_cell_trajectory" functions were utilized to visually represent pseudotime trajectories of diverse cell types within the KIRC TME.Thus, these analyses effectively demonstrate the dynamic regulation of disulfidptosis and its implications in the complex TME.

NMF of disulfidptosis-mediated disulfidptosis regulators in TME cells
NMF unsupervised clustering represents a widely adopted technique in both data mining and machine learning.It functions by decomposing a data matrix into non-negative basis vectors and coefficient matrices (29).This method proves highly effective in unraveling the intricacies of the TME and identifying heterogeneity among tumor cells (30).To further investigate the influence of disulfidptosis-mediated regulator expression on various TME cell types, the "NMF" algorithm was employed along with the "snmf/r" method.This allowed for a comprehensive dimensionality reduction analysis involving 24 disulfidptosis regulatory factors across all TME cell types, enabling the identification of distinct cell subtypes within these groups.Noteworthy, these steps were performed using approaches that closely align with methodologies used previously (23,31).

Identification of the marker genes of disulfidptosis-related cell subtypes in TME
The FindAllMarkers function was utilized to determine the characteristic markers for each NMF cluster and discern different cell subtypes within these types.Genes were filtered based on a logfold change threshold of 0.5 and a p-value below 0.01, with default parameters applied for the remaining settings.Additionally, employing the AddModuleScore function, module scores were computed using differentially expressed genes (DEGs) specific to each NMF cluster.The Dotplot visualization method was then utilized to represent the expression levels of top characteristic genes within each NMF cluster.Furthermore, the FeaturePlot was employed to illustrate the distribution of NMF cluster scores in the TME of KIRC.Supplementary Table S1 provides detailed information on the specific gene sets used for comparing disulfidptosis-mediated TME subgroups.

Functional enrichment analysis for NMF disulfidptosis-related subtypes
The "clusterProfiler" R package was utilized to identify marker genes within NMF clusters based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database.Significance was determined by a corrected p.adjust < 0.05.Additionally, gene set variation analysis (GSVA) using 24 disulfidptosis regulatory factors was conducted to calculate enrichment scores for these NMF clusters.The "scMetabolism" R package, employing VISION methods, enables metabolic activity quantification at a single-cell resolution.This package comprises 85 KEGG pathways and 82 REACTOME pathways, facilitating comprehensive analysis (32).scMetabolism was employed to assess metabolic activity across different NMF clusters, and results were visualized using the DotPlot.metabolismfunction.

SCENIC analysis for NMF disulfidptosisrelated subtypes
SCENIC analysis was employed to investigate the gene regulatory network of transcription factors (TFs) in KIRC (33) using the aertslab database (https://resources.aertslab.org/).The database provided crucial files, namely, "hg19-500bp-upstream-7species.mc9nr.feather"which contained the genomic coordinates of the 500bp upstream regions of all human genes in the human genome based on the hg19 version, and "hg19-tss-centered-10kb-7species.mc9nr.feather"which encompassed the genomic coordinates of the 10 kb regions centered around the transcription start site (TSS) for the same genomic version.These files enabled TSS identification within KIRC scRNA-seq data, allowing recognition of potential TF-target relationships and construction of a co-expression gene network.Further analysis focused on TFs with a Benjamini-Hochberg false discovery rate corrected p-value < 0.05.

Cell-cell communication analysis for NMF disulfidptosis-related subtypes
Utilizing the human and mouse ligand-receptor interaction database, the "CellChat" R package was employed to analyze intercellular communication networks in scRNA-seq data across various cell clusters (34).Initially, overexpressed genes in NMF clusters were identified, focusing on ligands or receptors, and their expression data were projected into a protein-protein interaction network.The collective intercellular communication network was computed using the computeCommunProbPathway and aggregateNet functions.Subsequently, the cell-cell communication network and visualized communication strength were analyzed using the netVisual_circle functions.Finally, leveraging the human ligandreceptor pairs database, the interactions between cell types and visualized ligand-receptor interactions' strength across different cell clusters were examined using the netVisual_bubble function.

Survival analyses with disulfidptosis-related signatures in RNA-seq
Bulk RNA-seq offers substantial clinical information.Thus, bulk RNA-seq data were integrated to explore the effect of disulfidptosis-related clusters on patient prognosis.To pinpoint distinct disulfidptosis-related genes for each NMF cluster, the FindAllmarker function was used, setting a logfc threshold at 0.5.Next, employing the GSVA method, gene scores for these identified genes were calculated from two publicly available KIRC datasets: TCGA-KIRC and GSE29609.Subsequently, this study investigated the association between disulfidptosis-related NMF signatures and the prognosis of patients with KIRC using Kaplan-Meier (K-M) analysis and univariate Cox regression.Cutoff values for different NMF clusters were determined utilizing the "survminer" R package, aiming to elucidate the link between disulfidptosis-related gene signatures and clinical outcomes in patients with KIRC.

Immunotherapeutic analyses for NMF disulfidptosis-related subtypes
The TIDE database offers an integrated analysis of immune dysfunction and exclusion mechanisms in tumor immune evasion, aiding in predicting immunotherapy responses (35).The TIDE database was utilized, and logistic regression analysis was performed to assess immunotherapy response among different disulfidptosis-mediated TME patient subgroups in the TCGA-KIRC and GSE29609 datasets.This study then evaluated the likelihood of implementing immunotherapy responses for specific disulfidptosis-related NMF subtypes.

Multiplex immunohistochemistry assay
Multiple immunohistochemistry (mIHC) was performed to detect three different antibodies on tissue sections.The primary antibodies used were rabbit monoclonal (1:500 dilution, ab186754, EPR15827(B); Abcam) to DSTN, rabbit monoclonal (1:200 dilution, ab76289, EP2405Y; Abcam) to FLNA, and rabbit monoclonal antibody (1:200 dilution, ab133616, EPR6855; Abcam) to CD4.The cancerous or adjacent normal tissues used for mIHC experiments were obtained from archived paraffin-embedded surgical specimens of patients with KIRC who provided prior informed consent.Employing the TSA method, the primary antibodies were stained.The procedure involved deparaffinization, antigen repair, labeling, inactivation of endogenous peroxidases, and antigen blocking on the first day.The first primary antibody, CD4, was applied, incubating overnight at 4°.On the second day, after treatment with goat antirabbit IgG (H+L; 1:50 dilution) labeled with horseradish peroxidase, Cyanine 5 Tyramide was used for detection.The steps were then repeated with DSTN or FLNA primary antibodies, incubating overnight at 4°.On the third day, following treatment with goat anti-rabbit IgG (H+L; 1:50 dilution) labeled with horseradish peroxidase, the sections were incubated with fluorescein tyramide working solution.Finally, nuclei were stained with DAPI.

Cell culture, transfection, and cell scratch assay
The human KIRC cells (786-O) were cultured in RPMI-1640 (BI, Israel) supplemented with 1% streptomycin and penicillin, along with 10% fetal bovine serum (BI, Israel).siRNA transfection utilized riboFECT ™ CP (RiboBio, China).The target sequences for TNFRSF1A-siRNA were as follows: GAACCTACTTGTAC AATGA.Cultured cells were grown to 95% confluence in 24-well dishes.Subsequently, scratches were vertically created in each well using a 10-mL lance tip.Cells were then washed thrice with phosphate-buffered saline to remove any shed cells.To assess the trauma area at the scratch areas, photographs of the si-TNFRSF1A group and the normal group were captured at 0 and 24 hours, followed by the analysis of the resulting images using ImageJ software.

Statistical analysis
R version 4.2.1 was employed for statistical analysis.To evaluate differences in continuous and categorical variables within cellular subgroups, various tests (Mann-Whitney U test, t-test, Kruskal-Wallis test, and log-rank test) were conducted.Furthermore, to compare distinct characteristics of disulfidptosis-mediated TME subgroups in KIRC, relevant disulfidptosis-related regulators and TME-related genes were obtained from prior literature.The "pheatmap" R package was utilized to visualize the NMF cluster.Statistical significance was determined at a threshold p-value of less than 0.05, indicating significant differences.

Landscape of disulfidptosis regulators in the TME of KIRC
A concise flowchart was presented to illustrate the exploration of disulfidptosis regulators based on scRNA-seq data from patients with KIRC (Figure 1A).The GSE210038 dataset encompassed 41,784 TME cells from 7 patients with KIRC.Following a singlecell analysis workflow, these cells were annotated into major cell types-epithelial cells, stromal cells, myeloids, mast cells, B cells, and T cells (Figure 1B).Additionally, CellChat analysis revealed intercellular interactions among different cell types (Figure 1C).Furthermore, a heatmap was generated to visualize the expression differences of 24 disulfidptosis regulators across diverse cell types (Figure 1D).

Novel disulfidptosis-mediated fibroblasts contributed to the TME of KIRC
The TME is known to be a complex network where stromal cells play a crucial role by releasing various growth factors, cytokines, and signaling molecules.These molecules, such as platelet-derived growth factor, fibroblast growth factor, transforming growth factor-b (TGFb), and interleukin (IL)-6, bind to receptors on tumor cells, promoting tumor growth, differentiation, and metastasis (36).Stromal cells encompass various types, including fibroblasts, endothelial cells, smooth muscle cells, and others (Figure 2A).CAFs in the KIRC TME are often linked with poor prognosis and resistance to immune checkpoint inhibitor therapy (37).To gain insights into the development and dynamics of the TME, pseudotime analysis was employed, which has proven to be a valuable tool in studying cell development, stem cell differentiation, and TME processes (38).By utilizing pseudotime analysis, the essential role of disulfidptosis regulators in the developmental trajectory of TME fibroblasts was unraveled.Specifically, genes related to disulfidptosis -OXSM, GYS1, FLNB, SLC7A11-were expressed at the beginning of the developmental process, while SLC3A2 expression was observed toward the end (Figures 2B, C).These genes may act as crucial mediators in TME subpopulation development and differentiation.Furthermore, intercellular communication and ligand-receptor interactions between subpopulations and tumor cells significantly influence tumor progression (39).CellChat analysis revealed varying ligand-receptor pairs links among clusters-ACTN4+CAF-C1 (n=466), PDLIM1+CAF-C2 (n=389), NDUFA11+CAF-C3 (n=266), Non-Dis-CAF-C4 (n=320), and Unclear-CAF-C5 (n=501)-and epithelial cells.Notably, disulfidptosis-mediated CAFs clusters exhibited stronger communication with tumor cells (Figure 2D).
Additionally, based on DEGs, KEGG enrichment analysis indicated associations of the ACTN4+CAF-C1 cluster with vascular smooth muscle contraction, regulation of actin cytoskeleton, and the PPAR signaling pathway.Meanwhile, the PDLIM1+CAF-C2 cluster showed involvement in the HIF-1 signaling pathway, focal adhesion, and ECM-receptor interaction (Figure 2E, Supplementary Table S2).Previous research underscores the role of iCAF in tumor progression and its influence on immune infiltration in bladder and breast cancers (40).Using Pan-CAF signatures from prior literature (41), a strong correlation existed between PDLIM1+CAF-C2 cell cluster scores and the pan-inflammatory CAF-2 (pan-iCAF-2; Figure 2F).
TFs play a pivotal role in either promoting or repressing downstream genes by recognizing specific DNA sequences.In renal cancer, TFs activation closely relates to somatic gene mutation and tissue-specific cancer risk (42).SCENIC analysis revealed distinct expression patterns of 26 TFs among the five clusters.Notably, TFs such as JUNB, FOS, EGR3, JUND, JUN, and CEBPD were upregulated in the NDUFA11+CAF-C3 cluster (Figure 2G).Given that JunB had a role in promoting cell invasion and angiogenesis in VHL-deficient RCC (43), we hypothesize that a close association between this CAFs cluster and tumor invasion.Finally, pathway heatmaps highlighted significant differences in gene expression levels, particularly in the contractile and RAS pathways, among ACTN4+CAF-C1, NDUFA11+CAF-C3, and unclear-CAF-C5 clusters (Figure 2H).

Disulfidptosis-mediated macrophages exhibit distinct metabolism features
Within the TME, myeloid cells play a crucial role in both innate and adaptive immunity and demonstrate significant heterogeneity (44).To delve into the functionality of disulfidptosis regulators within myeloid cells, this population was meticulously classified into distinct subgroups-comprising macrophages, monocytes, and other cellular entities (Figure 3A).Notably, our NMF clustering of macrophages uncovered four major clusters: FLNA+Mac-C1 (n=556), DSTN+Mac-C2 (n=326), Non-Dis-Mac-C3 (n=258), and Unclear-Mac-C4 (n=697).CellChat analysis revealed diverse ligand-receptor interactions between these clusters and epithelial cells (Figure 3B).Pseudotime analysis underscored the crucial involvement of disulfidptosis regulators in shaping the trajectory of TME macrophages (Figures 3C, D).
As the TME is closely associated with tumor metabolism and because tumors establish acidic and hypoxic environments by exploiting resources from the surrounding microenvironment, tumor cells undergo metabolic reprogramming to fulfill the specific demands necessary for their growth, proliferation, and survival (45).Lastly, to assess the relationship between the disulfidptosis-mac cluster with metabolic pathways.GSVA and scMetabolism were utilized to gauge single-cell metabolic activity, revealing notable differences in 30 metabolic pathways among the four clusters (Figure 3J).Particularly, the FLNA+Mac-C1 cluster exhibited heightened activation of sulfur metabolism, suggesting its potential significance in regulating disulfidptosis.
Utilizing CellChat, this study unveiled a diverse range of ligandreceptor interactions between disulfidptosis-related T cell clusters and tumor epithelial cells (Figure 5A).SCENIC analysis revealed substantial differences in the expression of TFs among CD4 + T cells, CD8 + T cells, NK cells, and Treg cells within these disulfidptosis clusters (Figure 5B).Additionally, to evaluate the collective effect of disulfidptosis-related T cell subgroups on T cells, notable differences were observed in the average expression of immune genes associated with co-stimulation, co-inhibition, and functional markers.Variations were also noted in the T exhaustion score, T cytotoxic score, T effector score, and T evasion score among these four disulfidptosis-related T cell subgroups (Figure 5C).Furthermore, distinctive ligand-receptor interactions were observed between disulfidptosis-related B cells and T cells (Figure 5D).The heatmap illustrated distinct TFs among the disulfidptosis-related B cell clusters (Figure 5E).These findings strongly indicate robust heterogeneity among disulfidptosis-related T/B cells.Given the pivotal role of CD4 + T cells in tumor immune responses (46), our focus was directed toward investigating the significance of disulfidptosis-mediated CD4 + T cells.

Disulfidptosis-mediated TME patterns contributed to the KIRC prognosis and immunotherapy
For a comprehensive evaluation of disulfidptosis in KIRC at the bulk level, FindAllMarkers was employed to compute all DEGs mediated by disulfidptosis in TME cells (Supplementary Table S3).Subsequently, using GSVA, the scores of different cell subgroups in tumor (n = 531) and normal (n = 76) samples were calculated from the TCGA-KIRC cohort.Consequently, significant score differences between tumor and normal samples were observed within two disulfidptosis-related CD4 + T cell subgroups, namely, DSTN +CD4T-C1 and FLNA+CD4T-C2 (Figure 6A).To validate the differential expression and distribution of these two CD4 + T cell subgroups within the tumor and adjacent nontumor regions, this study conducted multiplexed immunohistochemistry (mIHC) assays.Statistical analysis further indicated a significantly higher proportion of these CD4 + T cell subpopulations in the tumor samples than in the adjacent nontumor samples (p < 0.05; Figure 6B).Subsequent survival analysis, utilizing information from 531 patients with KIRC in the TCGA-KIRC and GSE29609 public cohorts, revealed that patients with low expression of DSTN +CD4T-C1 and FLNA+CD4T-C2 subgroups exhibited better prognoses (Figures 6C, D).Univariate Cox regression analysis consistently showed prognostic significance for patients in both subgroups (Figure 6E).Furthermore, the TIDE database was utilized to investigate the immune response of these two disulfidptosis-related CD4 + T cell subgroups in patients undergoing immunotherapy.TIDE scores indicated that patients with low expression in these subgroups exhibited better responses to immunotherapy (Figures 6F, G).Logistic regression analysis also demonstrated consistent immunotherapy response among patients in both subgroups (Figure 6H).Our ongoing investigation aims to explore the interaction between these two disulfidptosis-related CD4 + T cell subgroups and tumor cells, along with their mechanisms of exerting pro-oncogenic functions.

Disulfidptosis-mediated TME of KIRC enhances intercellular communication
The interaction between tumor cells and cells in the TME significantly contributes to tumor progression (48).Through cellchat analysis, we aimed to elucidate the comprehensive ligandreceptor interactions between disulfidptosis-mediated TME subgroups and tumor epithelial cells, involving various signaling pairs such as TNFSF12-TNFRSF12A, TNF-TNFRSF1A, OSM-(OSMR+IL6ST), OSM-(LIFR+IL6ST), HBEGF-EGFR, and EREG-EGFR.Of particular interest, the interactions of signaling pairs TNFSF12-TNFRSF12A mediated the interaction of both DSTN +CD4T-C1 and FLNA+CD4T-C2 subgroups with tumor epithelial cells (Figure 7A).To further investigate this interaction's functional significance, knockdown experiments targeting TNFRSF12A were conducted using siRNA in the KIRC cell line 786-O.Remarkably, after TNFRSF12A knockdown, significant inhibition of tumor cell migration was observed (p < 0.05; Figure 7B).These findings suggest that interactions between the DSTN+CD4T-C1 and FLNA+CD4T-C2 subgroups with tumor cells through the TNFSF12-TNFRSF12A signaling pairs play a crucial role in promoting tumor migration.This suggests that the progression of KIRC could be influenced by the interplay between disulfidptosis-mediated TME and tumor cells.

Discussion
Disulfidptosis, a novel manifestation of disulfide stressmediated PCD, has garnered attention due to its correlation with cancer pathogenesis (26, 49).Notably, studies using disulfidptosis scores have identified the CD96 gene as an independent prognostic marker for esophageal squamous cell carcinoma (ESCC).Knocking down CD96 not only significantly suppresses ESCC cell proliferation but also promotes apoptosis (50).Prognostic markers consisting of disulfidptosis-related lncRNAs can predict the survival of patients with different colorectal cancers and the use of targeted therapies and immunotherapies accordingly to the patient's condition (51).Additionally, a prognostic model (Enet), based on disulfidptosis-related genes, aids in predicting the prognosis for patients with thyroid carcinoma ( 52).Yet, limited research exists on the role of disulfidptosis in cancer at the singlecell level.This study performed a comprehensive investigation into the regulatory factors for disulfidptosis in various cell types within the KIRC TME, exploring different disulfidptosis-mediated TME subgroups' functions and determining cell-cell interactions between these subtypes.This study aimed to provide detailed insights into how these distinct TME subgroups influenced by disulfidptosis influence the prognosis of patients with KIRC.
Tumor tissue consists of tumor cells and various stromal cells, together forming a complex TME that plays a crucial role in tumor growth, invasion, and immune evasion (53).Components of tumor stroma induce inflammatory responses and angiogenesis, promoting tumorigenesis (54).Although immune cells, including therapy efficacy by remodeling the extracellular matrix, establishing a barrier hindering drug or immune cell infiltration into tumor tissues (60).CAFs activation during cancer progression involves pathways including recruitment and activation of resident fibroblasts, epithelial-mesenchymal transition (EMT), endothelial-mesenchymal transformation (EndMT), and differentiation of bone marrowderived mesenchymal cells (61)(62)(63)(64)(65). Through secretion of chemokines and effector molecules such as c-x-c chemokine ligand 5 (CXCL5), IL-1b, matrix metalloproteinases (MMPs), and collagen, CAFs contribute to immunosuppression and tumor angiogenesis (66).Moreover, different CAFs subpopulations modulate gene expression, regulate regulatory T-cell accumulation, and influence immune and cancer-related pathways by interacting with tumor cells (67).Despite five identified CAFs subtypes, pan-myCAF, pan-dCAF, pan-iCAF, pan-iCAF-2, and pan-pCAF (41).Limited research exists on the potential regulatory role of disulfidptosis mechanisms in CAFs.Our study revealed that disulfidptosis-related fibroblasts exhibit extensive communication with tumor epithelial cells compared to non-disulfidptosis-mediated fibroblasts.
This study revealed a strong correlation between the expression of ACTN4+CAF-C1 and increased levels of MMPs, particularly MMP9 and MMP14.This suggests that CAFs potentially contribute to the formation of the TME conducive to tumor metastasis through the secretion of MMP9 and MMP14 (68,69).Furthermore, KEGG pathway analysis highlighted the involvement of CAFs in various pathways such as vascular smooth muscle contraction, actin cytoskeleton regulation, and the PPAR signaling pathway.Additionally, PDLIM1+CAF-C2 displayed heightened expression levels of SERPINE1, CTHRC1, THBS2, SULF1, TGFB1, FN1, and BGN.Pathway analysis also indicated the participation of CAFs in the HIF-1 signaling pathway, focal adhesion, and ECMreceptor interaction.CAFs may participate in the HIF-1 signaling pathway and ECM pathways through the secretion of TGFB1, FN1, and BGN (70)(71)(72).Tumors are known to exploit ECM remodeling to create microenvironments conducive to tumorigenesis and metastasis (73).Consequently, we suggests that disulfidptosis-mediated CAFs may potentially enhance tumor invasion and metastasis by influencing tumor cell motility and ligand-receptor interactions.
CAFs release pro-inflammatory cytokine IL-1 and the chemokine CXCL1 to recruit macrophages, driving their differentiation into protumorigenic macrophages (M2-like TAMs) (74).Macrophages have a crucial role in immune regulation and controlling inflammation.However, in the context of cancer, their functionality becomes altered.Upon interaction with the TME, macrophages induce immunosuppression, which hampers the effector T cell response (75).Metabolic processes profoundly influence TAMs, regulating cancer development and immune responses involving glucose, glutamine, and fatty acid metabolism (76).Under glucose-deficient conditions, elevated expression of SLC7A11 leads to rapid depletion of intracellular NADPH, causing abnormal accumulation of disulfides such as cysteine, which triggers disulfide stress and rapid cell death.Through NMF clustering, extensive crosstalk occurred between disulfidoptosis-mediated macrophages and tumor cells.Notably, disulfidoptosis-mediated macrophages, particularly the FLNA+Mac-C1 subtype, exhibit significant activation of metabolic pathways including sulfur metabolism, the pentose phosphate pathway, oxidative phosphorylation, glycolysis, and gluconeogenesis.Furthermore, the study identified interactions between disulfidoptosis-mediated CD4 + T cells, CD8 + T cells, Treg cells, B cells, and tumor cells were identified, each demonstrating distinct functional characteristics.These findings collectively underscore the significant roles of disulfidoptosis-related macrophages and T cells in tumor biology.
In macrophages, higher activity of TFs including TGIF1, JUND, FOSB, CEBPB, FOSL2, XBP1, ZBTB1, CREM, BCL3, NFE2L2, NFIL3, EP300, RUNX1, and HIF-1A was observed in the FLNAmac-C1 subtype.BCL3 and EP300 as prognostic factors for KIRC (82, 83), and upregulated RUNX1 is closely associated with renal cancer progression (84).Moreover, macrophage HIF-1a has been identified as an independent prognostic indicator for renal cancer, being associated with highly invasive or deteriorating renal tumors (85).This led to the hypothesis that a close association between the FLNA-mac-C1 subtype and cancer progression.Furthermore, diverse TF characteristics were noted for disulfidptosis-mediated B and T cell subgroups.In summary, disulfidptosis-mediated TME subgroups likely regulate distinct TF networks, reshaping the TME.Lastly, cellular network analysis revealed significant connectivity and communication between these disulfidptosis-mediated TME subgroups and tumor cells.
The effectiveness of tumor therapy heavily relies on the microenvironment, particularly the tumor immune microenvironment, intricately linked to treatment prognosis (86).Immunotherapy has recently emerged as a promising strategy for cancer treatment, becoming a pivotal component in many cancer treatment regimens (87).Its therapeutic mechanism involves reactivating the immunosuppressive environment caused by cancer cells and boosting the immune cell-mediated anti-tumor response (88).Chen et al. studied the tumor immune microenvironment infiltration characteristics of disulfidptosis-related genes in breast cancer and reported that TNFRSF14, among disulfidptosis-related genes, serves as a key regulatory gene.Targeting TNFRSF14 alongside immune checkpoint inhibition was observed to inhibit tumor proliferation and induce disulfidptosis in tumor cells (89).Alterations in characteristic tumor biomarkers are closely related to tumor occurrence and development (90).By detecting the expression levels of key biomarkers and taking into account the clinical characteristics of patients, it is important to predict the prognosis of patients (91).Furthermore, emerging nanoformulations and tumor antigen vaccines have demonstrated high specificity and efficiency (92, 93).Guan L et al. synthesized mesoporous organosilicon nanoparticles exhibiting robust antitumor abilities and biosafety (94).Nevertheless, because of the complexity of the TME, the development of more targeted drugs tailored to different patients remains necessary.
Understanding the complex intrinsic regulatory patterns of disulfidptosis in the TME of KIRC, a comprehensive analysis using bulk RNA-seq and scRNA-seq data from TCGA and GEO was conducted.This analysis aimed to investigate the relationship between different disulfidptosis-mediated TME subgroups, prognosis, and immunotherapy.Consequently, significant variations in patient prognosis were observed based on different disulfidptosis factors.Specifically, the integrated scores of DSTN+CD4T-C1 and FLNA +CD4T-C2 subgroups were upregulated in tumor tissues and associated with unfavorable prognosis.These findings were corroborated by employing mIHC on tissue sections from patients with KIRC to confirm the expression and location of these two cell subgroups.Numerous studies indicate that CD4 + T cells, a major T lymphocyte subpopulation, play a crucial role in tumor immunity and exhibit diverse functions (95).Notably, CD4 + T cells effectively impede tumor cell division by arresting their cell cycle at the G1/S phase, inhibiting tumor growth (96).Furthermore, an independent study noted that CD4 + T cells isolated from bladder tumors exhibited substantial cytotoxicity when cultured in vitro, inducing apoptosis in tumor cells (97).Our investigation revealed that that DSTN+CD4T and FLNA+CD4T cells interact with tumor cells through the TNF-TNFRSF1A signaling pair, thereby facilitating tumor migration.The TME might influence CD4 + T cells to produce cytokines with protumorigenic functions, promoting tumor survival (98).Additionally, because of their plasticity, CD4 + T cells can convert into Treg cells secreting IL-10 and TGF-b, suppressing immune responses and aiding in tumor immune escape (99).These findings align with our conclusions and support the development of therapeutic strategies targeting these CD4 + T cell subsets associated with tumors.
However, it is important to note that this study was analyzed based on scRNA-seq data from public databases, and a self-test cohort of clinical samples is still needed to provide more realistic and accurate results.Besides, this study's scope is confined to in vitro cellular experiments, susceptible to various confounding factors and biases.Further studies using flow cytometric analysis and sorting experiments and in vivo animal models are required to comprehensively investigate the interaction between these two cellular subpopulations and tumor cells through the TNF-TNFRSF1A signaling pair, as well as the potential influence of other cell-cell communications on tumor cells.The TNF-TNFRSF1A signaling pair strongly correlates with responses to and resistance against anti-PD-1 treatment (100).The results of the present study indicate significant differences in immune responses to immune checkpoint blockade therapy among patients from different subgroups, underscoring the crucial role of disulfidptosis in patients with KIRC and warranting further investigation.Therefore, this study emphasizes the importance of exploring disulfidptosis concerning prognosis and immunotherapy in patients with KIRC.

Conclusion
For the first time, this study utilized the scRNA-seq analysis method to identify disulfidptosis-mediated TME subgroups.Additionally, combining bulk RNA-seq enabled us to elucidate the role of disulfidptosis-mediated cell-cell communication in regulating tumor growth and anti-tumor immune modulation.

1
FIGURE 1 Overview of disulfidptosis regulators in the single-cell data for kidney renal clear cell carcinoma (KIRC).(A) Flowchart of the workflow used in this study.(B) Main cell type annotations using the Seurat uniform manifold approximation and projection (UMAP) plot of 41,784 cells.(C) Cell-cell communications between the main cell types by CellChat analysis.(D) Heatmap distribution of disulfidptosis regulators in T cells, epithelial cells, stromal cells, myeloids, mast cells, and B cells.

2 3
FIGURE 2 Disulfidptosis regulators modifying the features of fibroblast cells.(A) Presence of fibroblasts in stromal cells.(B) Heatmap of pseudotime trajectory analysis revealing the role of disulfidptosis-related genes in fibroblast cells (2,398 cells).(C) Trajectory analysis of fibroblast cells.(D) Cell-cell communications from disulfidptosis-related fibroblast subgroups to epithelial cells.(E) Activation of KEGG signaling pathways by the main disulfidptosis fibroblast subgroups, as depicted in the heatmap based on differentially expressed genes (DEGs; p < 0.05).(F) Correlation between different disulfidptosis-related fibroblast subgroups and different cancer-associated fibroblasts (CAFs) cluster characteristics (p < 0.05).(G) Comparison of activities of transcription factors (TFs) among the five disulfidptosis-related fibroblast subgroups, illustrated in the heatmap based on the average area under the curve (AUC) values (Kruskal-Wallis test, p < 0.001).The activity of TFs was evaluated using AUCell.(H) Heatmap showing the different average expression levels of common signaling pathway genes in the 5 disulfidptosis-related fibroblast subgroups, including Proinflammatory, RAS, Contractile, Neo-Angio, TGFb, MMPs, and ECM.

7
FIGURE 7 Cell-cell communications from disulfidptosis-related tumor microenvironment (TME) subgroups to epithelial cells.(A) Receptor-ligand pairs from DSTN/FLNA-related TME subgroups to epithelial cells.(B) Scratch healing assay suggesting a significant reduction in 796-O cell migration in silencing TNFRSF1A (scale bar, 400 mm; p = 0.027).All data are expressed as mean ± standard deviation of three independent experiments.