Machine learning identified the application of Disulfidoptosis-Related Ferroptosis Genes Guinness in Lung Adenocarcinoma

doi:10.21203/rs.3.rs-5044174/v1

Download PDF

Research Article

Machine learning identified the application of Disulfidoptosis-Related Ferroptosis Genes Guinness in Lung Adenocarcinoma

https://doi.org/10.21203/rs.3.rs-5044174/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Disulfidoptosis, a recently recognized form of Regulated Cell Death (RCD), is known to play a crucial role in the development and progression of lung adenocarcinoma (LUAD). This study employed correlation analysis to identify 21 ferroptosis-related genes associated with disulfidoptosis, subsequently utilizing these genes to distinguish between two distinct groups characterized by contrasting prognostic and immune cell infiltration features. A risk model consisting of 17 DFG-related genes (SOD1, AKT1S1, TMPRSS11E, EPHX3, CPS1, PAK1, PSMB1, CDH17, NLRP2, HSD17B14, MRPL16, LACTB2, DEDD2, MCEMP1, LCAL1, VSTM2L, FLNC) was constructed through univariate Cox analysis and Lasso regression analysis. Within the TCGA cohort, the prognostic significance of these risk features was established as independent factors. Furthermore, a nomogram was developed to predict individual overall survival rates at 1, 3, and 5 years; with an AUC value of 0.833 for the first year, indicating its marked clinical utility. Conversely, high-scoring attributed encompass tumors in advanced stages and lower levels of cell differentiation. Furthermore, our observations suggested that patients in the high-risk group may potentially derive benefits from the administration of drugs such as 5-Fluorouracil.In conclusion, DFG holds potential for assessing patient prognosis, immune characteristics, and treatment outcomes. The findings of this study offer novel insights for clinical application and immune-based therapies.

disulfidoptosis

ferroptosis

lung adenocarcinoma

immune therapy

prognosis

drug sensitivity

Lung cancer stands as the second most prevalent malignancy worldwide and also serves as the foremost cause of cancer-related mortality. Within this category, lung adenocarcinoma emerges as the most frequently encountered histological subtype, comprising over 40% of all lung cancer cases^1,2. The TNM staging system has traditionally served as a crucial prognostic factor for lung cancer³. However, recent studies indicate that due to tumor heterogeneity, patients with the same stage of lung cancer may experience varying prognoses⁴. Furthermore, with the growing utilization of immunotherapy in lung cancer treatment, the identification of patients who are likely to respond favorably to immune-based therapies has become a pressing clinical concern^5,6. Unfortunately, the anatomical TNM staging system has limitations in accurately selecting patients who will benefit from immune treatments. Therefore, the development of a biomarker capable of accurately predicting long-term prognosis and immune therapy response in lung cancer patients holds significant importance for the advancement of precision medicine^7,8.

SLC7A11, a member of the cystine/glutamate antiporter family, plays a crucial role in mediating amino acid transport across the plasma membrane. It is involved in maintaining cellular redox homeostasis, regulating ferroptosis, and influencing cancer drug resistance^9–11. Recent studies have revealed that cells expressing high levels of SLC7A11 can undergo rapid cell death mediated by disulfide stress, resulting from the accumulation of excessive cystine facilitated by SLC7A11. This phenomenon, typically occurring under glucose deprivation, represents a novel form of regulated cell death known as disulfidoptosis¹². Notably, previous research has shown that downregulation of SLC7A11 can induce ferroptosis by inhibiting cysteine metabolism pathways^13,14. Ferroptosis, an iron-dependent form of regulated cell death, is triggered by excessive lipid peroxidation on the cell membrane, hindering antioxidant reactions. Key features of ferroptosis include phospholipid oxidation, the presence of iron for redox reactions, and impaired lipid peroxide repair mechanisms. Ferroptosis assumes a pivotal role in the pathogenesis and advancement of diverse malignancies, exerting its influence on the formulation of tumor treatment approaches^15–17 .

In this study, we utilized sophisticated bioinformatics analysis techniques to discern DFG-related genes, subsequently constructing and validating a prognostic DFG score based on the identification of distinctive clustering features. Moreover, we investigated the impact of the DFG score on somatic mutations, potential pathways, immune characteristics, cancer stem cell (CSC) indices, and drug sensitivity. The findings strongly suggest that the DFG score demonstrates exceptional prognostic predictive capabilities and holds significant clinical relevance in assessing immune efficacy and drug sensitivity across diverse LUAD subtypes. Ultimately, our investigation unveiled a compelling association between the low-risk group and favorable prognostic outcomes, accompanied by enhanced immune infiltration.

2.1. Processing and identification of data related to disulfidptosis

The disulfidoptosis gene, SLC7A11, was obtained from relevant studies, specifically from references^11,18,19. Additionally, we accessed the FerrDb database (http://www.zhounan.org/) to retrieve genes associated with ferroptosis. Through correlation analysis, we initially identified 21 ferroptosis-related genes that exhibited a high correlation with disulfidoptosis(|r|0.3, p < 0.05), as previously described in studies²⁰. RNA-Seq data for gene expression, as well as corresponding clinical and mutation data for LUAD, were downloaded from TCGA database (https://portal.gdc.cancer.gov/). The downloaded expression data from the TCGA dataset were represented as FPKM values.

2.2. Detection of Somatic Copy Number Variations (CNV) and Mutations

The somatic mutation data for TCGA-LUAD were obtained from the Genomic Data Commons (GDC) portal (https://portal.gdc.cancer.gov/), while copy number variations (CNVs) were acquired from UCSC Xena (https://xenabrowser.net). Analysis of CNV frequency for the DFGs was performed using the "maftools" package in R software, and the results were visualized in a bar plot. The positions of the 21 DFGs on chromosomes were displayed using the "RCircos" package in R.

2.3. Subgroup Classification and Clinical Feature Comparison of DFGs-Related Molecular Clusters:

Utilizing the "ConsensusClusterPlus" ²¹package, a consensus clustering analysis was performed on LUAD samples from the TCGA dataset, employing the 21 DFG-related genes as the basis for analysis. The number of clusters was determined between 2 and 10, based on cumulative distribution function (CDF) and area under the curve (AUC) analysis. The "pheatmap" package was then used to visualize the differences in RNA levels of DFG genes among the classified clusters, as well as the clinical parameters such as age, gender, TNM staging, and grading. Subsequently, the "clusterSurv" package was employed to compare the overall survival (OS) time between the divided clusters, allowing for the evaluation of the subgroups' prognosis.

Exploration of Biological Activity and Pathway Enrichment Analysis of DFGs To explore the biological activity of DFGs, The "LIMMA" package was employed to discern the differentially expressed genes (DEGs) linked to the DFG subtypes. Subsequently, the single-sample Gene Set Enrichment Analysis (ssGSEA) software package and Gene Set Variation Analysis (GSVA) method were employed to calculate cluster values for immune cells and immune-related pathways, followed by pathway enrichment analysis for both groups. The "estimate" package was utilized to calculate immune-related functional scores for 23 immune-related functions. Lastly, the "ggplot2" package was utilized to generate the principal component analysis (PCA) graph, illustrating the clustering of genes.

2.4. Construction and Evaluation of Prognostic Models Related to Disulfidptosis

Using the "caret" package, samples from the TCGA dataset were randomly allocated into training and testing cohorts at an equitable 1:1 ratio. Cox regression models and the "clusterSur" package were employed to identify the optimal prognostic factors and construct a signature. The risk score was calculated utilizing the following formula: Risk_score = S(Expi * Coefi), wherein Coefi denotes the risk factor associated with each gene, and Expi represents the corresponding expression level.

The patients were categorized into the high/low-risk groups .The "clusterSUR" package is utilized for survival analysis, encompassing OS and PFS. Additionally, univariate independent prognostic analysis is utilized to assess the autonomous prognostic significance of risk prediction markers. Moreover, the pheatmap package is employed to generate a patient survival status plot and a gene expression heatmap based on risk scores. The time-dependent ROC curves, using 1-year, 3-year, and 5-year survival rates, are employed to evaluate the predictive ability of features in identifying high/low risk The SISTERPLOT3D package is applied for principal component analysis to explore the distribution of patients with different risk scores. The rms package is used to construct a nomogram incorporating risk score, gender, age, and T/N/M stage. Calibration curves are graphed to illustrate the disparities between the projected and observed outcomes of the nomogram.. The C-index curve is employed to validate the accuracy of the signature in predicting the survival of LUAD patients. Lastly, patients are stratified into stages I-II and III-IV to assess the applicability of the signature in different stages.

2.5. Enrichment analysis

Using |log2FoldChange(FC)| > 1 as the criterion and adjusting for p < 0.05, DEGs that distinguish high/low-risk patients are determined. By integrating the ClusterProfiler²² package, we perform Gene Ontology (GO) enrichment analysis based on the TREZID of DEGs obtained using the org.Hs.eg.db program. Subsequently, gene set enrichment analysis (GSEA) is conducted to analyze pathway activities in the high and low-risk groups, using the clusterProfiler package.

2.6. Immunological function analysis and tumor mutational burden (TMB) analysis

The GSEABase, GSVA, and other R packages are employed to analyze differences in immune-related functions among patients. In the study of TMB, the maftools package is employed to investigate the association between the risk score and tumor mutational burden (TMB). Notably, the survminer package is employed to investigate the impact of TMB on the prognosis of high/low risk

2.7. Immunotherapy analysis and analysis of tumor microenvironment differences

We obtained tumor immune dysfunction and exclusion (TIDE) data for NSCLC from the TIDE website (http://tide.DFGci.harvard.edu/). The TIDE algorithm accurately predicts the efficacy of immunotherapy drugs received by patients using TIDE scores, which have been demonstrated to be positively correlated with the response to anti-tumor immune drugs (anti-PD1, anti-CTLA4)²³. We also explore the presence of differences in immune cells and immune functions between the low-risk and high-risk groups, focusing on 22 infiltrating immune cell types. Additionally, our study investigates the correlation between the stemness index and risk scores.

Drug screening: The oncoPredict²⁴ package is used to screen potential effective drugs for patients in the high/low risk groups from a pool of over 300 drugs, with a pFilter threshold of 0.001.

3.1. Study on DFG gene variations in lung adenocarcinoma

In this study, the occurrence rate of copy number variations (CNVs) and somatic mutations in 21 DFGs in lung adenocarcinoma was summarized. Among the 616 samples, 77 (12.5%) showed genetic variations, with a frequency of only 2% in the major mutated genes ABCC1 and GCLC (Fig. 1a). Figure 1b illustrates the CNV alterations of DFGs on chromosomes. Further investigation of the frequency of CNV mutations has unveiled that CNV deletions exhibit a higher prevalence in the genes SLC7A5 and SRXN1,while G6PD and MAFG genes were more frequently expressed in the form of CNV amplification. The CNV amplification and deletion frequencies of the PGD gene were roughly equivalent (Fig. 1c). The aforementioned analysis indicates that genetic and expression changes in DFGs in lung adenocarcinoma are highly heterogeneous, suggesting the significant role of DFG expression imbalance in the development of LUAD²⁵.

3.2. Tumor classification and related research based on DFGs

To further investigate the clinical value and functional biological patterns of these DFGs, we conducted cluster analysis and grouped all tumor samples in the TCGA LUAD cohort based on the expression levels of 21 DFGs. The expression levels of DFG-related genes were compared with the expression of LUAD subtypes to explore the relationship between them. After setting the clustering value (K) from 2 to 10, the best clustering stability was observed when K = 2 (Fig. 2a-c). Therefore, the TCGA LUAD cohort was divided into Cluster A (n = 446) and Cluster B (n = 71) based on DFGs. A heatmap was used to display the gene expression of 517 LUAD patients along with patient information such as TNM staging, age (≤ 65 years or > 65 years), and gender. As shown in the figure, Cluster B exhibited higher overall DFG expression levels than Cluster A and was more closely associated with unfavorable prognostic indicators (such as higher TNM staging, grading, and older age) (Fig. 2d). PCA analysis, conducted using DFG expression levels, yielded noteworthy disparities in the transcriptional profiles observed within the two DFG clusters (Fig. 2e). Kaplan-Meier analysis of Cluster A and B revealed that Cluster A had a more favorable overall survival (OS), while Cluster B showed a survival disadvantage consistent with clinical parameters, with a statistically significant difference (P = 0.015, Fig. 2f).

The ssGSEA score indicated that Group A exhibits abundant infiltration of innate immune cells, such as activated B cells, activated dendritic cells, natural killer cells, and mast cells (Fig. 2g). To explore the pathways associated with these DFGs, we subsequently performed GSVA analysis. As shown in Fig. 2h, asthma was more significantly enriched in Cluster A, while pathways such as pentose phosphate pathway and glucose metabolism were significantly enriched in Cluster B.

The tumor microenvironment (TME) refers to the localized milieu composed of tumor cells, immune cells, cytokines, stromal cells, and chemokines. As shown in Fig. 2I, Cluster A had higher StromalScore, immune score, and ESTIMATEScore, with statistically significant differences (p < 0.001), and was associated with lower tumor purity. Cluster B showed poorer immune infiltration and higher tumor purity, indicating a worse prognosis.

3.3. Prediction and Validation of Risk Models

Based on 282 DEGs, we constructed a DFG score and used the LASSO method for Cox regression analysis to identify 17 gene features corresponding to the optimal λ value (Fig.3A, B). The risk score was computed using the following formula: Risk score = （0.330615644515484* SOD1）+（0.737805752695113* AKT1S1）+（0.110975880002471* TMPRSS11E）+（-0.290569613562898* EPHX3）+（0.0981752522575458* CPS1）+（0.261430872004936*PAK1）+（0.477994179749537*PSMB1）+（0.121881061745916* CDH17）+（-0.225076862993535* NLRP2）+（-0.201271409455139* HSD17B14）+（0.423290355450028* MRPL16）+（0.30479628765554* LACTB2）+（0.368597261418397* DEDD2）+（-0.148283206012694* MCEMP1）+（-0.180119867428623* LCAL1）+（0.123696864043202* VSTM2L）+（0.183396161574755* FLNC）.

We compared the differences in PFS between the high- and low-risk subgroups in both the training set (n = 249) and validation set (n = 249). It was observed that, the high-risk group exhibited poorer prognosis (Fig. 3c).

The results of univariate Cox regression analysis indicated that the risk score could independently predict adverse survival outcomes in the TCGA cohort (Fig. 3d, e). The predictive accuracy of the risk score was evaluated using the ROC curve. Furthermore, the C-index of the risk score demonstrated a significantly higher value compared to that of other clinical features, such as age, gender, and TNM stage in predicting survival outcomes (Fig. 3f). Time-dependent ROC curve analysis and calculation of the AUC at different time points were performed to assess the accuracy of the prediction model. The AUC values of the model were 0.822, 0.760 and 0.691 at 1, 3, and 5 years, respectively (Fig. 3g). These findings suggest that the prediction model exhibits promising prognostic value for both short-term and long-term follow-up periods.

By incorporating factors such as age, gender, and TNM stage into a nomogram, we developed a reliable tool for predicting 1-, 3-, and 5-year overall survival (OS) in LUAD patients (Fig. 4a, b). Further analysis indicated significant differences in OS between high- and low-risk patients in different stages (I-II vs. III-IV) (P < 0.05) (Fig. 4c, d), demonstrating the ability of the model to compare survival outcomes among patients with different stages. Finally, principal component analysis was performed to observe the distribution of all genes, DFG genes, DFG-related genes, and risk genes in patients, revealing that DFG-associated risk genes can be confidently utilized to construct a robust signature(Fig. 4e-h).

3.4. Functional Enrichment Analysis

The results of the GO analysis demonstrated that DFG-related genes were predominantly enriched in molecular functions, including serine endopeptidase activity and sulfur compound binding. In terms of cellular components, they showed a primary association with immune globulin complexes and lumen structures. Regarding biological processes, their main connections lay in antibacterial responses, immune reactions, and lysosomal activities (Fig. 5a, b). Subsequently, in our subsequent investigation, we conducted Gene Set Enrichment Analysis (GSEA) on the high and low-risk groups. In the pathway analysis, pathways such as the cell cycle, Huntington's disease, ribosome, spliceosome, and steroid hormone biosynthesis were prominently enriched in the high-risk group (Fig. 5c). In contrast, the gene sets of the low-risk group exhibited strong associations with immune responses, being predominantly enriched in pathways like allograft rejection, asthma, graft-versus-host disease, intestinal immune network for IgA production, and systemic lupus erythematosus (Fig. 5d).

3.5. Immune Correlation Comparison

To assess immune correlation, we initially scrutinized the distribution of somatic cell mutations within the TCGA-LUAD cohort, differentiating between the high/low score group. Figure 6a demonstrates that the higher risk group had a more widespread distribution of tumor mutational burden (TMB) compared to the low-risk group (93.28% vs. 88.41%). Among the high/low-risk groups, the most prominent somatic cell mutations were Tp53 and TTN²⁶, suggesting that patients with high TMB may benefit from immunotherapy²⁷. Survival analysis further demonstrated a positive association between a low DFG score and better survival in the high TMB (H-TMB) subgroup, while a high DFG score combined with the low TMB (L-TMB) subgroup showed lower survival probability (Fig. 6b, c).

In our current research, we observed a notable disparity in TIDE scores between the low-risk group and the high-risk group, with the former exhibiting higher scores (Fig. 6d). However, further investigation is needed to determine if patients with low TIDE scores receive more benefits from immunotherapy compared to those with high scores. Analysis of immune-related functional differences indicated that the low-risk group exhibited a higher infiltration of immune cells (Fig. 6e). Additionally, the analysis of stem cell relevance revealed a positive correlation between patient risk scores and stem cell indices (p < 0.001) (Fig. 6f), suggesting that higher risk scores correspond to lower cell differentiation levels²⁸.

3.6. Drug Sensitivity Analysis

Regarding drug sensitivity analysis, the high risk (HR) group demonstrated greater sensitivity to drugs such as Cytarabine, 5-Fluorouracil, PD0325901, BI-2536, and Dasatinib. Conversely, drugs such as SB216763, Doramapimod, PF-4708671, GSK269962A, and SB505124 exhibited higher sensitivity in the low risk (LR) group. These results indicate a relationship between DFGs and drug sensitivity (Fig. 7a-m).

As the most prevalent form of lung cancer, the diagnosis of early-stage non-small cell lung cancer (NSCLC) primarily relies on pathological examinations, such as bronchoscopic lung biopsy, percutaneous puncture biopsy, and cytology of exfoliated cells²⁹. Unfortunately, most patients are diagnosed at advanced stages with highly invasive tumors. Despite significant advancements in lung cancer treatment, the 5-year survival rate for LUAD remains low at 15% due to factors like the development of resistance to conventional chemotherapy and radiotherapy^30,31. In recent years, immune checkpoint inhibitors (ICIs) have demonstrated effectiveness as an immunotherapeutic approach for lung cancer^32–34. High expression of PD-L1 (tumor cell expression ≥ 50%) has been associated with improved response to ICIs. However, in clinical practice, most patients with high PD-L1 expression show limited sensitivity to ICIs, while only a small subset of patients with low or negative PD-L1 expression may derive therapeutic benefits from immunotherapy^35,36. Therefore, there is a need to explore novel and more accurate biomarkers and treatment strategies to guide precision medicine in clinical settings.

In the latest research, it has been discovered that SLC7A11 can uptake cystine under glucose starvation conditions, inducing disulfide stress and resulting in a distinct form of regulated cell death (RCD) called disulfidptosis, which is different from ferroptosis, copper-induced cell death, and apoptosis^12,37. This process has been demonstrated to play a role in combating intracellular infections by flexibly applying and interconnecting different cell death pathways³⁸. Similar to ferroptosis, the occurrence of disulfidptosis is also associated with changes in cellular redox status and can promote tumor cell death by altering the conformation of cellular cytoskeletal proteins^39,40. Our study has also found a strong correlation between disulfidptosis and ferroptosis. Numerous studies have shown that SLC7A11 plays a regulatory role in lung cancer by modulating ferroptosis^14,41–43. The significant enrichment of DFG-related genes in oxidative-reduction and energy metabolism pathways, along with the strong expression correlation between disulfidptosis and ferroptosis, provides support for this viewpoint. Interestingly, it has been reported that metabolic therapy using glucose transporter protein (GLUT) inhibitors can induce disulfidptosis and inhibit tumor growth⁴⁴. Thus, achieving a balance between ferroptosis and disulfidptosis may serve as a novel strategy to improve treatment efficacy and survival rates in LUAD patients.

To comprehensively explore the intricate relationship between disulfidptosis and ferroptosis, as well as their regulatory implications in tumor initiation and progression, our study delved into the distinctions among diverse DFG scoring groups at the genetic, transcriptional, and immunological levels. Furthermore, we investigated their interconnections specifically within LUAD. This endeavor bears paramount significance as it unravels the potential interplay between disulfidptosis and ferroptosis, thus providing deeper insights into antitumor immune activity within the tumor microenvironment (TME) and mechanisms that impede tumor immune evasion. Moreover, it offers valuable guidance for the selection of appropriate and effective immunotherapeutic strategies.

In the subsequent stages of our research, we employed unsupervised clustering techniques to categorize patients into two distinct subtypes, aiming to unveil novel LUAD characteristic patterns and lay a foundation for guiding tumor progression and treatment decisions. The two identified disulfide-ferroptosis clusters were named DFGcluster A and DFGcluster B.

Notably, these two subgroups exhibit significant disparities in OS, TME and immune infiltration levels. DFGcluster A is positively associated with immune cell activation and a favorable prognosis, while DFGcluster B, characterized by elevated levels of disulfide-driven ferroptosis gene expression, demonstrates higher tumor purity and relatively poorer immune infiltration. The variations in immune properties likely play a pivotal role in the observed major discrepancies in overall survival between the two groups. Our ssGSEA analysis reveals that DFGcluster A displays more abundant characteristics of natural killer (NK) cell, B cell, and dendritic cell (DC) infiltrations. NK cells have been substantiated to play a crucial role in augmenting antibody and T cell responses in the context of antitumor immunity⁴⁵, while T cells act as vital complements to T cell-mediated antitumor immunity and serve as essential indicators for evaluating the prognosis of NSCLC^46,47. Meanwhile, DCs play a pivotal role in initiating and regulating immune responses, forming the cornerstone of successful generation of antitumor immune responses⁴⁸. The GSVA analysis results may shed light on the inferior prognosis of DFGcluster B from the perspective of energy metabolism^49–51.

Considering the heterogeneity and complexity of individuals, a model comprising 17 genes (SOD1, AKT1S1, TMPRSS11E, EPHX3, CPS1, PAK1, PSMB1, CDH17, NLRP2, HSD17B14, MRPL16, LACTB2, DEDD2, MCEMP1, LCAL1, VSTM2L, FLNC) was employed to quantitate the disulfidptosis/ferroptosis patterns in each patient based on the expression of DFGs. While most research on SOD1 has focused on amyotrophic lateral sclerosis (ALS), experimental and bioinformatics studies also reveal elevated SOD1 levels in NSCLC cell lines and tissues, significantly correlated with overall survival (OS) in LUAD patients. Deletion of SOD1 remarkably reduces tumor burden in patients and inhibits the growth of KRAS-mutant NSCLC cells in vitro, suggesting SOD1 as a potentially effective biological biomarker for diagnosing LUAD patients^52–54. AKT1S1, acting as an intermediary in the AKT and mTORC1 signaling pathways, exhibits elevated phosphorylation of PRAS40 encoded by AKT1S1 in specimens from prostate cancer and lung cancer⁵⁵. Elevated expression of TMPRSS11E has been demonstrated to be associated with a worse prognosis in patients with NSCLC⁵⁶. CPS1 expression is inversely correlated with the prognosis of LUAD patients⁵⁷.PAK1 is closely associated with chemoresistance and an unfavorable prognosis in patients with NSCLC⁵⁸.In gene set enrichment analysis, PSMB1 participates in the degradation of AXIN protein, which could potentially promote tumor progression through activation of the WNT signaling pathway⁵⁹.CDH17 is upregulated in LUAD samples and functions as an independent prognostic biomarker for LUAD outcomes⁶⁰.NLRP2 inhibits cell proliferation and migration in LUAD cells through the regulation of epithelial-mesenchymal transition (EMT)⁶¹.LCAL1 is likely to promote lung cancer survival by suppressing AMP-activated protein kinase (AMPK) activity⁶². On the other hand, FING, functioning as a regulator of NSCLC cell migration, could potentially influence patient survival via autophagy^63,64. However, there is currently no research on the mechanistic roles of EPHX3, HSD17B14, MRPL16, LACTB2, DEDD2, MCEMP1, and VSTM2L in the occurrence and development of LUAD. This study may offer new insights for future investigations into the roles of these genes in LUAD.

TMB represents the mutation count per megabase of sequenced DNA in specific cancers. It is widely believed that an increase in TMB expression raises the probability of triggering T-cell responses. Research suggests that TMB may possess clinical diagnostic importance in lung cancer and has the potential to serve as a biological biomarker for immunotherapy^65,66.In our study, a combination of low DFG scores and high TMB is indicative of a significantly extended OS.Moreover, mutations are another crucial factor that substantially affects the effectiveness of immunotherapy. Studies have highlighted that TP53 mutations, among the most frequent changes in human tumors, are frequently linked to increased invasiveness and unfavorable prognosis, and have been validated as effective biomarkers for predicting PD-1 immunotherapy response^67–69.Furthermore, TTN mutations are notably linked to prolonged OS in LUAD patients^70,71, consistent with our research results. In the investigation of immune infiltration, differences were observed in monocytes and dendritic cells between LR and HR groups, showing an inverse correlation with the risk score^48,72. On the contrary, Macrophage M1 positively correlates with the risk score, possibly due to its inferior phagocytic and antigen-presenting abilities, ascribed to higher PD1/PD-L1 expression⁷³. Ii is noteworthy that typically, highly infiltrated CD8 + T cells are commonly linked to a favorable prognosis. However, in our study, CD8 + T cells were found to be positively correlated with the risk score. This could be attributed to the high infiltration of CD8 + T cells, promoting tumor cell progression, mutation, and evolution, resulting in tumor immune escape⁷⁴. Similar research on melanoma and colorectal cancer corroborates this perspective^75,76. Moreover, it is crucial to consider the potential mechanisms of HSD17B14, DEDD2, and MCEMP1 in their association with Tregs cells^77,78. Additionally, we discovered that MCEMP1 exerts a negative regulation on various T cells but a positive regulation on macrophages, which could offer an immunological rationale for the gene's function in LUAD^79,80. Overall, our research results indicate that DFG scores can serve as a reliable scoring system for evaluating immunotherapy in LUAD patients.

Our research reveals a positive correlation between the cancer stem cell index and the risk score. Higher CSC indices often indicate lower tumor differentiation, pointing to a poorer prognosis and chemotherapy response⁸¹. Further drug sensitivity studies confirm that HR patients exhibit higher sensitivity to chemotherapy and targeted therapy, with significantly lower sensitivity scores for chemotherapeutic drugs like Cisplatin and 5-Fluorouracil compared to LR patients. Therefore, the DFG score can be considered a predictive factor for assessing drug sensitivity in LUAD patients. Considering the favorable performance of DFGs in predicting the prognosis of LUAD patients and their potential benefits in immunotherapy and drug sensitivity, DFGs can serve as valuable references for guiding individualized treatment plans for LUAD patients.

Nevertheless, our study should recognize certain limitations. Firstly, despite some findings being experimentally validated, exploring the underlying mechanisms between disulfidptosis and ferroptosis necessitates further validation with more experimental data. Secondly, similar to many publications^20,82,83, the data in our study were derived from public databases, lacking validation from animal experiments and clinical specimens. Therefore, more prospective and fundamental research is required to enhance the relevance and detail of this study.

Disclosure

TCGA belongs to public databases. The patients involved in the database have obtained ethical approval. Users can download relevant data for free for research and publish relevant articles. Our study is based on open-source data, so there are no ethical issues and other conflicts of interest.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

Author contributions

LZC: Data analysis, Writing- Original draft preparation. YRD: Writing- Original draft preparation. PXM and LYX collected, processed and analyzed the data. ZX: Conception and design, Final approval of the version to be published.

Conflict of interest

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

Cappa C, Gregson K, Wardlaw T, et al. Birth registration: a child′s passport to protection[J]. Lancet Glob Health, 2014, 2(2): e67⁃68. DOI: 10.1016/S2214⁃109X(13)70180⁃3.
Sung, H. et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71, 209-249, doi:10.3322/caac.21660 (2021).
Barta, J. A., Powell, C. A. & Wisnivesky, J. P. Global Epidemiology of Lung Cancer. Annals of global health 85, doi:10.5334/aogh.2419 (2019).
Woodard, G. A., Jones, K. D. & Jablons, D. M. Lung Cancer Staging and Prognosis. Cancer treatment and research 170, 47-75, doi:10.1007/978-3-319-40389-2_3 (2016).
Hensing, T., Chawla, A., Batra, R. & Salgia, R. A personalized treatment for lung cancer: molecular pathways, targeted therapies, and genomic characterization. Advances in experimental medicine and biology 799, 85-117, doi:10.1007/978-1-4614-8778-4_5 (2014).
Madden, K. & Kasler, M. K. Immune Checkpoint Inhibitors in Lung Cancer and Melanoma. Seminars in oncology nursing 35, 150932, doi:10.1016/j.soncn.2019.08.011 (2019).
Wu, M. et al. Systemic Immune Activation and Responses of Irradiation to Different Metastatic Sites Combined With Immunotherapy in Advanced Non-Small Cell Lung Cancer. Frontiers in immunology 12, 803247, doi:10.3389/fimmu.2021.803247 (2021).
Pan, S. et al. Identification of cuproptosis-related subtypes in lung adenocarcinoma and its potential significance. Frontiers in pharmacology 13, 934722, doi:10.3389/fphar.2022.934722 (2022).
Wang, Y. et al. Molecular subtypes based on cuproptosis-related genes and immune profiles in lung adenocarcinoma. Frontiers in genetics 13, 1006938, doi:10.3389/fgene.2022.1006938 (2022).
Lin, W. et al. SLC7A11/xCT in cancer: biological functions and therapeutic implications. American journal of cancer research 10, 3106-3126 (2020).
Liu, J., Xia, X. & Huang, P. xCT: A Critical Molecule That Links Cancer Metabolism to Redox Signaling. Molecular therapy : the journal of the American Society of Gene Therapy 28, 2358-2366, doi:10.1016/j.ymthe.2020.08.021 (2020).
Reck, M., Remon, J. & Hellmann, M. D. First-Line Immunotherapy for Non-Small-Cell Lung Cancer. J Clin Oncol 40, 586-597, doi:10.1200/jco.21.01497 (2022).
Liu, X. et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nature cell biology 25, 404-414, doi:10.1038/s41556-023-01091-2 (2023).
Yang, J. et al. Metformin induces Ferroptosis by inhibiting UFMylation of SLC7A11 in breast cancer. Journal of experimental & clinical cancer research : CR 40, 206, doi:10.1186/s13046-021-02012-7 (2021).
Wang, X. et al. Stem Cell Factor SOX2 Confers Ferroptosis Resistance in Lung Cancer via Upregulation of SLC7A11. Cancer research 81, 5217-5229, doi:10.1158/0008-5472.Can-21-0567 (2021).
Jiang, X., Stockwell, B. R. & Conrad, M. Ferroptosis: mechanisms, biology and role in disease. Nature reviews. Molecular cell biology 22, 266-282, doi:10.1038/s41580-020-00324-8 (2021).
Chen, X., Kang, R., Kroemer, G. & Tang, D. Broadening horizons: the role of ferroptosis in cancer. Nature reviews. Clinical oncology 18, 280-296, doi:10.1038/s41571-020-00462-0 (2021).
Tang, D., Kang, R., Berghe, T. V., Vandenabeele, P. & Kroemer, G. The molecular machinery of regulated cell death. Cell research 29, 347-364, doi:10.1038/s41422-019-0164-5 (2019).
Zhao, S. et al. Crosstalk of disulfidptosis-related subtypes, establishment of a prognostic signature and immune infiltration characteristics in bladder cancer based on a machine learning survival framework. Frontiers in endocrinology 14, 1180404, doi:10.3389/fendo.2023.1180404 (2023).
Wang, T. et al. Disulfidptosis classification of hepatocellular carcinoma reveals correlation with clinical prognosis and immune profile. International immunopharmacology 120, 110368, doi:10.1016/j.intimp.2023.110368 (2023).
Zhao, C., Zhang, Z. & Jing, T. A novel signature of combing cuproptosis- with ferroptosis-related genes for prediction of prognosis, immunologic therapy responses and drug sensitivity in hepatocellular carcinoma. Frontiers in oncology 12, 1000993, doi:10.3389/fonc.2022.1000993 (2022).
Wilkerson, M. D. & Hayes, D. N. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics (Oxford, England) 26, 1572-1573, doi:10.1093/bioinformatics/btq170 (2010).
Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics : a journal of integrative biology 16, 284-287, doi:10.1089/omi.2011.0118 (2012).
Jiang, P. et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 24, 1550-1558, doi:10.1038/s41591-018-0136-1 (2018).
Maeser, D., Gruener, R. F. & Huang, R. S. oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Briefings in bioinformatics 22, doi:10.1093/bib/bbab260 (2021).
Xie, C. & Tammi, M. T. CNV-seq, a new method to detect copy number variation using high-throughput sequencing. BMC bioinformatics 10, 80, doi:10.1186/1471-2105-10-80 (2009).
Mogi, A. & Kuwano, H. TP53 mutations in nonsmall cell lung cancer. Journal of biomedicine & biotechnology 2011, 583929, doi:10.1155/2011/583929 (2011).
Ma, W., Pham, B. & Li, T. Cancer neoantigens as potential targets for immunotherapy. Clinical & experimental metastasis 39, 51-60, doi:10.1007/s10585-021-10091-1 (2022).
Wang, G. et al. Cuproptosis regulator-mediated patterns associated with immune inﬁltration features and construction of cuproptosis-related signatures to guide immunotherapy. Frontiers in immunology 13, 945516, doi:10.3389/fimmu.2022.945516 (2022).
Gridelli, C. et al. Non-small-cell lung cancer. Nature reviews. Disease primers 1, 15009, doi:10.1038/nrdp.2015.9 (2015).
Chen, Z., Fillmore, C. M., Hammerman, P. S., Kim, C. F. & Wong, K. K. Non-small-cell lung cancers: a heterogeneous set of diseases. Nature reviews. Cancer 14, 535-546, doi:10.1038/nrc3775 (2014).
Seguin, L., Durandy, M. & Feral, C. C. Lung Adenocarcinoma Tumor Origin: A Guide for Personalized Medicine. Cancers 14, doi:10.3390/cancers14071759 (2022).
Duma, N., Santana-Davila, R. & Molina, J. R. Non-Small Cell Lung Cancer: Epidemiology, Screening, Diagnosis, and Treatment. Mayo Clinic proceedings 94, 1623-1640, doi:10.1016/j.mayocp.2019.01.013 (2019).
Zhou, Y. et al. Nivolumab plus ipilimumab versus pembrolizumab as chemotherapy-free, first-line treatment for PD-L1-positive non-small cell lung cancer. Clinical and translational medicine 10, 107-115, doi:10.1002/ctm2.14 (2020).
Huang, M. et al. Cost-effectiveness of pembrolizumab versus chemotherapy as first-line treatment in PD-L1-positive advanced non-small-cell lung cancer in the USA. Immunotherapy 11, 1463-1478, doi:10.2217/imt-2019-0178 (2019).
Reck, M. et al. Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer. N Engl J Med 375, 1823-1833, doi:10.1056/NEJMoa1606774 (2016).
Mok, T. S. K. et al. Pembrolizumab versus chemotherapy for previously untreated, PD-L1-expressing, locally advanced or metastatic non-small-cell lung cancer (KEYNOTE-042): a randomised, open-label, controlled, phase 3 trial. Lancet (London, England) 393, 1819-1830, doi:10.1016/s0140-6736(18)32409-7 (2019).
Zhong, Z. et al. NFATc1-mediated expression of SLC7A11 drives sensitivity to TXNRD1 inhibitors in osteoclast precursors. Redox biology 63, 102711, doi:10.1016/j.redox.2023.102711 (2023).
Doerflinger, M. et al. Flexible Usage and Interconnectivity of Diverse Cell Death Pathways Protect against Intracellular Infection. Immunity 53, 533-547.e537, doi:10.1016/j.immuni.2020.07.004 (2020).
Tang, D., Chen, X., Kang, R. & Kroemer, G. Ferroptosis: molecular mechanisms and health implications. Cell research 31, 107-125, doi:10.1038/s41422-020-00441-1 (2021).
Zhang, R., Kang, R. & Tang, D. Reductive cell death: the other side of the coin. Cancer gene therapy, doi:10.1038/s41417-023-00612-3 (2023).
Zhang, W. et al. RBMS1 regulates lung cancer ferroptosis through translational control of SLC7A11. The Journal of clinical investigation 131, doi:10.1172/jci152067 (2021).
Lu, X. et al. MiR-27a-3p Promotes Non-Small Cell Lung Cancer Through SLC7A11-Mediated-Ferroptosis. Frontiers in oncology 11, 759346, doi:10.3389/fonc.2021.759346 (2021).
Hu, K. et al. Suppression of the SLC7A11/glutathione axis causes synthetic lethality in KRAS-mutant lung adenocarcinoma. The Journal of clinical investigation 130, 1752-1766, doi:10.1172/jci124049 (2020).
Zheng, P., Zhou, C., Ding, Y. & Duan, S. Disulfidptosis: a new target for metabolic cancer therapy. Journal of experimental & clinical cancer research : CR 42, 103, doi:10.1186/s13046-023-02675-4 (2023).
Shimasaki, N., Jain, A. & Campana, D. NK cells for cancer immunotherapy. Nature reviews. Drug discovery 19, 200-218, doi:10.1038/s41573-019-0052-1 (2020).
Salcher, S. et al. High-resolution single-cell atlas reveals diversity and plasticity of tissue-resident neutrophils in non-small cell lung cancer. Cancer cell 40, 1503-1520.e1508, doi:10.1016/j.ccell.2022.10.008 (2022).
Engelhard, V. et al. B cells and cancer. Cancer cell 39, 1293-1296, doi:10.1016/j.ccell.2021.09.007 (2021).
Wculek, S. K. et al. Dendritic cells in cancer immunology and immunotherapy. Nature reviews. Immunology 20, 7-24, doi:10.1038/s41577-019-0210-z (2020).
Koppula, P., Zhuang, L. & Gan, B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein & cell 12, 599-620, doi:10.1007/s13238-020-00789-5 (2021).
Pavlova, N. N., Zhu, J. & Thompson, C. B. The hallmarks of cancer metabolism: Still emerging. Cell metabolism 34, 355-377, doi:10.1016/j.cmet.2022.01.007 (2022).
Kroemer, G. & Pouyssegur, J. Tumor cell metabolism: cancer's Achilles' heel. Cancer cell 13, 472-482, doi:10.1016/j.ccr.2008.05.005 (2008).
Wang, X. et al. SOD1 regulates ribosome biogenesis in KRAS mutant non-small cell lung cancer. Nature communications 12, 2259, doi:10.1038/s41467-021-22480-x (2021).
Zhang, L., Zhang, Z. & Yu, Z. Identification of a novel glycolysis-related gene signature for predicting metastasis and survival in patients with lung adenocarcinoma. Journal of translational medicine 17, 423, doi:10.1186/s12967-019-02173-2 (2019).
Liu, S. et al. SOD1 Promotes Cell Proliferation and Metastasis in Non-small Cell Lung Cancer via an miR-409-3p/SOD1/SETDB1 Epigenetic Regulatory Feedforward Loop. Frontiers in cell and developmental biology 8, 213, doi:10.3389/fcell.2020.00213 (2020).
Qi, Z. et al. PRAS40 hyperexpression promotes hepatocarcinogenesis. EBioMedicine 51, 102604, doi:10.1016/j.ebiom.2019.102604 (2020).
Li, S. et al. Elevated expression of the membrane-anchored serine protease TMPRSS11E in NSCLC progression. Carcinogenesis 43, 1092-1102, doi:10.1093/carcin/bgac069 (2022).
Pham-Danis, C. et al. Urea Cycle Sustains Cellular Energetics upon EGFR Inhibition in EGFR-Mutant NSCLC. Molecular cancer research : MCR 17, 1351-1364, doi:10.1158/1541-7786.Mcr-18-1068 (2019).
Chen, M. J., Wu, D. W., Wang, Y. C., Chen, C. Y. & Lee, H. PAK1 confers chemoresistance and poor outcome in non-small cell lung cancer via β-catenin-mediated stemness. Scientific reports 6, 34933, doi:10.1038/srep34933 (2016).
Zhang, J. Y. et al. The Silence of PSMC6 Inhibits Cell Growth and Metastasis in Lung Adenocarcinoma. BioMed research international 2021, 9922185, doi:10.1155/2021/9922185 (2021).
Li, F., Wan, B. & Li, X. Q. Expression Profile and Prognostic Values of CDH Family Members in Lung Adenocarcinoma. Disease markers 2022, 9644466, doi:10.1155/2022/9644466 (2022).
Li, T. et al. NLRP2 inhibits cell proliferation and migration by regulating EMT in lung adenocarcinoma cells. Cell biology international 46, 588-598, doi:10.1002/cbin.11755 (2022).
Li, J. Y. & Luo, Z. Q. LCAL1 enhances lung cancer survival via inhibiting AMPK-related antitumor functions. Molecular and cellular biochemistry 457, 11-20, doi:10.1007/s11010-019-03507-w (2019).
Pal, J. et al. Systematic analysis of migration factors by MigExpress identifies essential cell migration control genes in non-small cell lung cancer. Molecular oncology 15, 1797-1817, doi:10.1002/1878-0261.12973 (2021).
Wu, Z. X., Huang, X., Cai, M. J., Huang, P. D. & Guan, Z. Development and Validation of a Prognostic Index Based on Genes Participating in Autophagy in Patients With Lung Adenocarcinoma. Frontiers in oncology 11, 799759, doi:10.3389/fonc.2021.799759 (2021).
Chan, T. A. et al. Development of tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic. Ann Oncol 30, 44-56, doi:10.1093/annonc/mdy495 (2019).
Hu, C. et al. Genomic profiles and their associations with TMB, PD-L1 expression, and immune cell infiltration landscapes in synchronous multiple primary lung cancers. Journal for immunotherapy of cancer 9, doi:10.1136/jitc-2021-003773 (2021).
Olivier, M., Hollstein, M. & Hainaut, P. TP53 mutations in human cancers: origins, consequences, and clinical use. Cold Spring Harbor perspectives in biology 2, a001008, doi:10.1101/cshperspect.a001008 (2010).
Aubrey, B. J., Strasser, A. & Kelly, G. L. Tumor-Suppressor Functions of the TP53 Pathway. Cold Spring Harbor perspectives in medicine 6, doi:10.1101/cshperspect.a026062 (2016).
Dong, Z. Y. et al. Potential Predictive Value of TP53 and KRAS Mutation Status for Response to PD-1 Blockade Immunotherapy in Lung Adenocarcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research 23, 3012-3024, doi:10.1158/1078-0432.Ccr-16-2554 (2017).
Chen, J. et al. Deciphering Prognostic Value of TTN and Its Correlation With Immune Infiltration in Lung Adenocarcinoma. Frontiers in oncology 12, 877878, doi:10.3389/fonc.2022.877878 (2022).
Wang, Z., Wang, C., Lin, S. & Yu, X. Effect of TTN Mutations on Immune Microenvironment and Efficacy of Immunotherapy in Lung Adenocarcinoma Patients. Frontiers in oncology 11, 725292, doi:10.3389/fonc.2021.725292 (2021).
Ugel, S., Canè, S., De Sanctis, F. & Bronte, V. Monocytes in the Tumor Microenvironment. Annual review of pathology 16, 93-122, doi:10.1146/annurev-pathmechdis-012418-013058 (2021).
Cai, H., Zhang, Y., Wang, J. & Gu, J. Defects in Macrophage Reprogramming in Cancer Therapy: The Negative Impact of PD-L1/PD-1. Frontiers in immunology 12, 690869, doi:10.3389/fimmu.2021.690869 (2021).
Kadara, H. et al. Whole-exome sequencing and immune profiling of early-stage lung adenocarcinoma with fully annotated clinical follow-up. Ann Oncol 28, 75-82, doi:10.1093/annonc/mdw436 (2017).
Tumeh, P. C. et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature 515, 568-571, doi:10.1038/nature13954 (2014).
Galon, J. et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313, 1960-1964, doi:10.1126/science.1129139 (2006).
Tanaka, A. & Sakaguchi, S. Targeting Treg cells in cancer immunotherapy. European journal of immunology 49, 1140-1146, doi:10.1002/eji.201847659 (2019).
Tanaka, A. & Sakaguchi, S. Regulatory T cells in cancer immunotherapy. Cell research 27, 109-118, doi:10.1038/cr.2016.151 (2017).
Hu, G., Sun, N., Jiang, J. & Chen, X. Establishment of a 5-gene risk model related to regulatory T cells for predicting gastric cancer prognosis. Cancer cell international 20, 433, doi:10.1186/s12935-020-01502-6 (2020).
Choi, Y. J. et al. Lung-specific MCEMP1 functions as an adaptor for KIT to promote SCF-mediated mast cell proliferation. Nature communications 14, 2045, doi:10.1038/s41467-023-37873-3 (2023).
Malta, T. M. et al. Machine Learning Identifies Stemness Features Associated with Oncogenic Dedifferentiation. Cell 173, 338-354.e315, doi:10.1016/j.cell.2018.03.034 (2018).
Zhang, H. et al. A novel defined cuproptosis-related gene signature for predicting the prognosis of lung adenocarcinoma. Frontiers in genetics 13, 975185, doi:10.3389/fgene.2022.975185 (2022).
Ma, S. et al. A cuproptosis-related long non-coding RNA signature to predict the prognosis and immune microenvironment characterization for lung adenocarcinoma. Translational lung cancer research 11, 2079-2093, doi:10.21037/tlcr-22-660 (2022).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Machine learning identified the application of Disulfidoptosis-Related Ferroptosis Genes Guinness in Lung Adenocarcinoma

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials and Methods

2.1. Processing and identification of data related to disulfidptosis

2.2. Detection of Somatic Copy Number Variations (CNV) and Mutations

2.3. Subgroup Classification and Clinical Feature Comparison of DFGs-Related Molecular Clusters:

2.4. Construction and Evaluation of Prognostic Models Related to Disulfidptosis

2.5. Enrichment analysis

2.6. Immunological function analysis and tumor mutational burden (TMB) analysis

2.7. Immunotherapy analysis and analysis of tumor microenvironment differences

3. Results

3.1. Study on DFG gene variations in lung adenocarcinoma

3.2. Tumor classification and related research based on DFGs

3.3. Prediction and Validation of Risk Models

3.4. Functional Enrichment Analysis

3.5. Immune Correlation Comparison

3.6. Drug Sensitivity Analysis

4 Discussion

Declarations

References

Additional Declarations

Status:

Version 1