Disulfidptosis-related lncRNA signatures predict prognosis and immune relevance of lung adenocarcinoma

doi:10.21203/rs.3.rs-3083164/v1

Download PDF

Article

Disulfidptosis-related lncRNA signatures predict prognosis and immune relevance of lung adenocarcinoma

https://doi.org/10.21203/rs.3.rs-3083164/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose: Lung adenocarcinoma (LUAD) remains the most common malignancy and has a poor prognosis. Disulfidptosis was identified as a novel type of cell death. Long non-coding RNAs (lncRNAs) play an important role in regulating cell death and LUAD progression. However, the role of disulfidptosis-related lncRNAs (DRLs) in LUAD has not been studied.

Methods: Disulfidptosis-related genes (DRGs) were performed for differential expression, mutation, GO, KEGG, and survival analyses. The univariate Cox regression and least absolute shrinkage and selection operator regression (Lasso) were used to screening the prognostic value of DRLs. Independent prognostic analysis, receiver operating characteristic (ROC) analysis, a nomogram, survival analysis, progression-free survival analysis, Principal Component Analysis (PCA), and tumor mutation difference analysis were used to validate the predictive efficacy of the signature. In addition, the CIBERSORT algorithm was used to assess the differences in immune cells and their immune function.

Results: A total of 21 DRGs differ, and functional enrichment analysis revealed that these 23 DRGs are mainly involved in the structural composition of the cytoskeleton involved in the binding and activation of actin in LUAD. Pearson correlation analysis identified 140 DRLs, LASSO regression analysis obtained 10 DRLs associated with the prognosis of LUAD patients and multifactorial Cox regression analysis was performed to construct a signature consisting of 5 DRLs with independent prognostic significance (LINC01352, AC093673.1, AL606834.1, AL365181.2, MHENCR) to form a prognostic signature. The K-M survival curves showed that the prognosis of patients in the high-risk group was poor, and the ROC curves showed that the area under the curve for the 1, 3, and 5-year survival rates were 0.695, 0.676, and 0.681, respectively, indicating that the signature had the excellent predictive ability. Univariate and multifactorial Cox regression analyses showed that the risk score was an independent prognostic factor. There was a significant correlation between risk score and tumor microenvironment, tumor mutational load, TIDE, and tumor cell stemness. The results of the drug sensitivity analysis showed that patients in the high-risk group had lower semi-inhibitory concentration values for Trametinib, Savolitinib, Ulixertinib, and Crizotinib compared to patients in the low-risk group, while the opposite was true for Tozasertib, Leflunomide, and Ribociclib.

Conclusions: We identified a prognostic signature of disulfidptosis‑related lncRNAs by comprehensive and systematic bioinformatics analysis of LUAD patients. Our developed risk-based signature outperforms standard clinicopathological parameters in predicting survival and prognosis and provided novel insights into the treatment of LUAD.

Biological sciences/Cancer

Biological sciences/Computational biology and bioinformatics

lung adenocarcinoma

disulfidptosis

lncRNA

prognosis

immune microenvironment

LUAD remains the most common malignancy and has a poor prognosis. Disulfidptosis was defined as a new form of death, which ultimately lead to the collapse of the actin network and cell death.
We successfully constructed a disulfidptosis-related lncRNAs model to improve the prognostic evaluation of patients with LUAD
Our developed risk-based signature outperforms standard clinicopathological parameters in predicting survival and prognosis and provided novel insights into the treatment of LUAD

Lung cancer remains one of the most common malignancies in society, with over 1.8 million people affected and an estimated 1.59 million deaths annually(Bade and Dela Cruz, 2020). It is classified into small-cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). Among them, NSCLC is the most predominant variant, mainly including two major histological subtypes: adenocarcinoma (AC) and squamous cell carcinoma (SCC) (de Mello et al., 2020). There is also considerable evidence that lung cancer likely results from a combination of genetic and environmental factors, including smoking, occupational exposures, chronic lung disease, and lung infections(Mattiuzzi and Lippi, 2020). Over the past few decades, significant strides have been made in the study of the pathogenesis, prognostic markers, and therapeutic targets of lung adenocarcinoma (LUAD)(Jia et al., 2019; Wan et al., 2020). However, due to the complexity of tumorigenesis and progression, patients with LUAD have a poor prognosis(Chen et al., 2019). Therefore, it is of great clinical value to do in-depth research into the pathogenesis of lung cancer and the discovery of new biomarkers for early diagnosis, treatment, and prognosis.

In living organisms, most cancer cells take in extracellular cystine mainly through the cystine transporter protein (consisting of the catalytic subunit SLC7A11 and the chaperone subunit SLC3A2), which is reduced intracellularly to cysteine for cellular utilize(Stipanuk, 2004). The SLC7A11 promotes glutathione synthesis and is known to inhibit ferroptosis by taking up cystine(Dixon et al., 2012). However, it has recently been demonstrated that excessive expression of SLC7A11 instead promotes cell death under glucose starvation conditions. The mechanism describes that the enormous accumulation of disulfide molecules in cancer cells leads to abnormal disulfide bonds between actin cytoskeletal proteins, ultimately leading to the collapse of the actin network and cell death. Furthermore, the well-known inhibitors of cell death pathways such as apoptosis, necrosis, oxidative stress, ferroptosis, and cuproptosis did not inhibit cell death, a new form of death defined as disulfidptosis(Liu et al., 2023). At present, there have been only a handful of disulfidptosis-related genes (DRGs) reported to be involved in cancer. It is extremely urgent to find promising disulfidptosis regulators to improve early diagnosis and treatment of cancer.

Non-coding RNAs (ncRNAs) refer to all functional RNAs that cannot be translated into proteins, which can be classified into different subcategories based on their nucleotide size, including small RNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs)(Goodall and Wickramasinghe, 2021). Long non-coding RNAs (lncRNAs) constitute a class of RNAs longer than 200 nucleotides, it exerts regulatory effects on the expression of genes, transcription, and translation by binding DNA, RNA, and proteins(Bridges et al., 2021; Herman et al., 2022). The fundamental role of lncRNAs in regulating numerous physiological and pathological processes, such as cell differentiation, proliferation, apoptosis, and the initiation and progression of cancers, has been increasingly demonstrated in recent years(Liu et al., 2020; Sang et al., 2022). Although it has been extensively studied that lncRNAs are involved in different types of cell death(Jiang et al., 2021), there are no studies on disulfidptosis-related lncRNAs (DRLs) effects. Thus, it is of great significance to explore the role of lncRNAs in the pathogenesis of disulfideosis, particularly in cancer. Interestingly, existing reported that the construction of cuproptosis-related lncRNA signature may contribute to immune status and immunotherapeutic response in lung cancer, thereby establishing a link between cell death-related lncRNAs and immune regulation(Ma et al., 2022). These findings provide a fresh perspective for us to study the relationship between disulfidptosis-related lncRNAs and tumor-associated pathways in the future.

In this study, samples from the TCGA-LUAD data set were randomly divided into a training set and testing set by using the bioinformatics analysis method combined with The Cancer Genome Atlas (TCGA) database data. Statistical analysis of prognostic survival of disulfidptosis-related lncRNAs was conducted by using the training set, to construct a disulfidptosis-related lncRNAs prognostic signature (DLPS). Finally, a DLPS consisting of five DRLs was obtained and the testing set was used to verify the signature. Meanwhile, Independent prognostic analysis, receiver operating characteristic (ROC) analysis, a nomogram, survival analysis, progression-free survival analysis, Principal Component Analysis (PCA), and tumor mutation difference analysis were used to validate the predictive efficacy of the signature and its ability to distinguish between high- and low-risk populations. Then, the immune infiltration, and immune cell function of the prognostic signature were analyzed. In addition, the semi-inhibitory concentrations (IC50) of various drugs were predicted to identify potential therapeutic agents. Based on the systematic findings, this work provides valuable insight into the underlying pathogenesis of LUAD and the basis for the development of novel biomarkers.

Data collection and processing

The transcriptome data for the LUAD cohort and corresponding clinical information in this work were derived from the TCGA database(https://portal.gdc.cancer.gov/) and the design details are shown in Fig. 1. In previous studies, 23 DRGs were collected, including SLC7A11, CAPZB, GYS1, NDUFS1, NDUFA11, NUBPL, NCKAP1, LRPPRC, SLC3A2, RPN1, ACTN4, ACTB, CD2AP, DSTN, FLNA, FLNB, INF2, IQGAP1, MYH10, MYL6, MYH9, PDLIM1, and TLN1(Liu et al., 2023). The somatic mutation and copy number data were available from the TCGA-LUAD cohort database. By using Perl software (v5.30.0.1), the expression matrix and clinical information of 507 cases (cases with complete clinical data and matching expression profiles) were extracted further separating the mRNAs and lncRNAs expression matrices. The ‘caret’ R package was used to randomly assign to a training set (n = 254) and a testing set (n = 253) by the ratio of 1:1.

Differential expression of DRGs

To identify the differential expression of DRGs, by Wilcoxon test to detect the expression levels of DRGs between lung cancer and normal tissues. Error incidence rate (FDR) < 0.01 and | Fold change (FC)|> 2 were used as the screening criterion.

Construction of disulfidptosis-related lncRNAs prognostic signature

To identify disulfidptosis-related lncRNAs, Pearson correlation analysis was performed on 23 DRGs and 16,876 lncRNAs with cutoff values of | Pearson R| > 0.4 and p < 0.001. Univariate Cox regression analysis was used to determine whether these lncRNAs were associated with the outcome of patients in the training set. To avoid overfitting among included lncRNA significantly associated with prognosis, minimum absolute contraction, and selection operator (LASSO) penalty regression analysis was performed(Engebretsen and Bohlin, 2019; Tibshirani, 1997), and five DRLs were identified as independent prognostic risk signature. Then, optimal model parameters risk characteristics, and calculated risk scores. Risk score = lncRNA1*coef-lncRNA1 + lncRNA*coef-lncRNA2 + ... +lncRNAi*coef-lncRNAi. The patients were categorized into high- and low-risk groups based on the median risk score as a threshold. Kaplan-Meier curves and log-rank tests were generated using the ‘survminer’ R package to assess the prognostic significance of five DRLs’ risk signatures. To assess the predictive power of lncRNA-based prognostic risk profiles, the R package ‘timeROC’ was used to measure the ROC 1, 3, and 5-year survival.

Independent prognostic analysis and nomogram construction

We conducted multivariate Cox regression analysis to confirm whether risk score and clinical parameters serve as independent predictors of overall survival (OS). And combined with prognostic and clinical features, a nomogram predicting 1, 3, and 5-year OS of LUAD patients was created using the "RMS" R package(Shinkawa et al., 2021). The prediction accuracy of the nomograms was determined by constructing the calibration curves.

Functional enrichment analysis and principal component analysis

Gene ontology (GO), including biological process (BP), cellular component (CC), and molecular function (MF) categories, was performed using the ‘ggplot2’, ‘colorspace’, ‘stringi’, ‘ggplot2’ R packages(Zhou et al., 2019). Additionally, this package was employed to carry out the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis(Kanehisa et al., 2021). The analysis of gene set enrichment analysis (GSEA) was conducted using several R packages, including ‘org.Hs.eg.db’, ‘limma’, ‘clusterProfiler’, and ‘enrichplot’(Yu et al., 2012). This approach aims to identify enriched pathways in both high- and low-risk groups. The ‘limma’ and ‘scatterplot3d’ R packages were used to perform PCA.

Analysis of tumor mutation burden, tumor microenvironment, and immune infiltrate

We generated mutation frequencies and cancer maps in LUAD patients using the ‘maftools’ R package(Mayakonda et al., 2018). To further analyze the data, we divided the patients into two groups based on their tumor mutation burden (TMB) values: low-TMB and high-TMB groups. Subsequently, we conducted a Kaplan-Meier survival curve analysis to evaluate the survival status of the two TMB groups, as well as the combinations of TMB values and risk scores. Using the ‘ESTIMATE’ R package(Yoshihara et al., 2013), tumor microenvironment (TME), including stromal, immune, and ESTIMATE scores, was calculated for each patient based on gene expression. Differences in 22 immunomarkers and immune-related functions were determined for each sample by Single Sample Gene Set Enrichment Analysis (ssGSEA) and Gene Set Variation Analysis (GSVA) using the CIBERSORT(Hanzelmann et al., 2013). We used the Tumor Immune Dysfunction and Exclusion (TIDE) algorithm analysis (http://tide.dfci.harvard.edu/) for the assessment of response to immunotherapy. Examination of differences in immune checkpoint expression between the two groups. Finally, we utilized Spearman correlation analysis to determine the correlation between TMB, TME, TIDE, and risk score. p-values less than 0.05 were considered to be statistically significant.

The role of the risk signature in predicting drug sensitivity

The Wilcoxon rank sum test was used to compare the difference in half-inhibitory concentration (IC50) between the high- and low-risk groups, and the ‘pRRophetic’ R package(Geeleher et al., 2014) was used to analyze the sensitivity of chemotherapeutic drugs, and the results were displayed by the "ggplot2" R package to evaluate the application of prognostic signatures in lung cancer chemotherapy.

Statistical analysis

LASSO, Univariate, and Multivariate regression analysis were used to qualitatively select appropriate molecules and variables. Survival curves were plotted with the Kaplan-Meier method, and the log-rank test was used to test for differences between groups. The Wilcoxon rank-sum test was used to analyze differences between the high- and low-risk groups and differences in clinical phenotypic risk scores. The difference in the distribution of clinical phenotypes between the high- and low-risk groups was analyzed using the chi-squared test, and P < 0.05 was considered statistically significant.

Analysis of DRG expression, gene mutation, and survival in LUAD

We first explored the expression of 23 DRGs in LUAD and paired normal tissues, and the results showed that the mRNA levels of 21 genes were differentially expressed, specifically, the expression of SLC7A11, GYS1, NDUFS1, NDUFA11, NUBPL, NCKAP1, LRPPRC, SLC3A2, RPN1, CD2AP was increased, in contrast, the expression of ACTN4, ACTB, DSTN FLNA, INF2, IQGAP1, MYH10, MYL6, MYH9, PDLIM1, TLN1 was decreased in LUAD (Fig. 2A). We then researched the gene mutation prevalence of DRGs in LUAD. Amongst these, FLNA was found to be the most frequently mutated rate, followed by MYH9 and TLN1 (Fig. 2B). Missense mutations were found to be the most frequently occurring variant classification, while single nucleotide polymorphisms (SNPs) were the most common type of variant. Additionally, the top single nucleotide variant (SNV) was identified as C > A (Fig. 2C). Furthermore, by classifying mutations into transition (Ti) and transversions (Tv), we found that Tv was produced more frequently than Ti across all of the genes (Fig. 2D). Subsequently, we analyzed the copy number variants (CNVs) of DRGs. CNVs were consistently amplified for NUBPL, FLNA, ACTN4, MYL6, TLN1, ACTB, CD2AP and IQGAP1, while largely deleted for INF2, DSTN, SLC7A11, PDLIM1, NDUFA11, GYS1, FLNB and MYH10 (Fig. 2E). Figure 2F shows the location of these 23 DRGs with altered CNVs on the chromosome. Finally, we visualized the complex relationship between DRGs and the prognostic value of LUAD in an integrated landscape using network diagrams (Fig. 2G) and investigated 15 DRGs exhibited a significant association with OS in LUAD. (Fig. 3A-O).

The function enrichment analysis of DRGs

We analyzed these pathways using GO and KEGG databases to clarify the function of DRGs. These DRGs were found to be mainly involved in actin binding and activation, the structural composition of the cytoskeleton, GTPase binding, integrin binding, cytoskeletal movement and composition, and cell adhesion functions (Fig. 4A). Furthermore, the KEGG pathway analysis identified the pathways that were primarily involved with the actin cytoskeletal regulatory pathway, tight junctions, Escherichia coli infection, Salmonella infection, vascular smooth muscle contraction, and cellular iron death (Fig. 4B).

Evaluation and validation of the disulfidptosis-related lncRNA prognostic signature

We first applied co-expression analysis within the TCGA-LUAD cohort to obtain a total of 140 DRLs based on Pearson analysis and visualized the co-expression relationship between DRLs and DRGs (Fig. 5A). The 507 LUAD patients were randomly assigned to the training set (n = 254) and the testing set (n = 253), with no differences in clinical characteristics. Then, lasso regression analysis and univariate Cox regression analysis obtained 10 DRLs with prognostic values in the training set (Fig. 5B-D), of these, 6 were found with a hazard ratio (HR) > 1, indicating poor prognosis, while the remaining 4 were likely protective (Fig. 5B). Finally, a prognosis signature consisting of five DRLs (LINC01352, AC093673.1, AL606834.1, AL365181.2, and MHENCR) was constructed using multivariate Cox regression analysis, and a heatmap shows the correlation between these five lncRNAs and DRGs (Fig. 5E).

Survival analysis of the signature

After constructing a prognostic signature through multivariate Cox regression, we proceeded to examine the prognostic potential of the five DRLs in LUAD. Based on the median risk score = 1, the patients were divided into high- and low-risk groups for survival analysis. We found that the OS and progression-free survival (PFS) of low-risk patients were significantly better than those of high-risk patients. (Fig. 6A-D). Furthermore, mortality increased with increasing risk scores in LUAD patients (Figs. 7A, B). This trend was observed consistently across in the testing set and all sets (Fig. 7D, E, G, H). AC093673.1, AL606834.1, and AL365181.2 were higher in the high-risk training, testing, and overall set (Fig. 7C, F, I), suggesting these DRLs may be poorer prognostic predictors. However, the expression levels of LINC01352 and MHENCR were lower among the high-risk groups, indicating that these DRLs may be protective markers. The survival probability and clinical characteristics of LUAD patients were also compared by age, gender, and grade, and the results indicated that high-risk patients had a shorter OS than low-risk patients, regardless of the clinical variables (Fig. 8A-F). These findings suggest that the signature can be applied to various clinical variables.

Independent analysis of the prognostic factors

To assess whether risk score could serve as an independent prognostic factor, we conducted univariate and multifactorial Cox regression analysis to examine the prognostic significance of DRLs in LUAD. The findings revealed that risk score was undoubtedly an independent prognostic factor for patients with LUAD, as demonstrated by both univariate and multifactorial Cox regression analysis (Fig. 9A, B). Moreover, we tested the accuracy of the signature with DRLs using a time-dependent ROC curve (Fig. 9C), The AUCs of 0.695, 0.676, and 0.681 for the 1, 3, and 5-year survival in the training set, respectively, demonstrated the potential predictive value of the DRLs signature for LUAD. And, the risk score had an AUC of 0.676 (Fig. 9D), indicating superior predictive power over clinical features except staging. To generate an accurate tool to assess the survival of patients with LUAD, we also created a nomogram following the risk score and the above clinical indices (Fig. 9E), the probability of OS at 1, 3, and 5 years for patients was 0.902, 0.703, and 0.548, respectively, while the calibration curves of the prognosis nomogram showed that the predicted and actual survival rates at 1, 3, and 5-years were broadly in agreement (Fig. 9F). Overall, these findings indicate that the risk score may be a strong prognostic factor in the prediction of the clinical outcome of patients with LUAD.

Functional enrichment analysis of differential genes in high- and low-risk groups

We performed PCA analysis and found that the risk lncRNAs were more effective in categorizing LUAD patients into high- and low-risk groups compared to all genes (Fig. 10A-D), suggesting that these lncRNAs could be more reliably used to construct a prognostic signature. We plotted the heat map of the expression levels of the five DRLs in normal and cancerous tissues (Fig. 11A). The heat map of the potential lncRNAs signature profiles showed that AC093673.1, AL606834.1, and AL365181.2 were significantly upregulated in the high-risk group (Fig. 11B). We compared the differential expression of DRGs between the two groups to investigate the relationship between the risk signature and disulfidptosis, most DRGs showed significant differential expression (Fig. 11C). To elucidate the function of the differential genes (DEGs) in the risk groups, we analyzed the relevant pathways using the GO and KEGG database and discovered that these DEGs primarily participate in actin binding and activation, cilia-dependent cell motility, arachidonic acid anabolism, fatty acid synthesis, bile acid synthesis, testosterone reductase activity, endosteroid dehydrogenase activity, immunoglobulin complex, transmembrane receptor protein tyrosine kinase activity, and immunoglobulin receptor binding. These findings suggest that the DEGs are predominantly associated with actin synthesis and function (Fig. 12A). Besides, KEGG pathway analysis revealed that DEGs are mainly involved in cytochrome P450 xenobiotic metabolism, arachidonic acid metabolism, linoleic acid metabolism, retinol metabolism, linoleic acid metabolism, steroid hormone biosynthesis, and bile secretion (Fig. 12B). To better understand the biological processes and pathways associated with the two risk groups identified using the five DRLs signature, we performed a GSEA analysis that showed that pathways such as ascorbate and alginate metabolism, cell cycle, pentose, and glucuronide interconversion, porphyrin and chlorophyll metabolism, and steroid hormone biosynthesis were enriched in the high-risk group, and pathways such as allograft rejection, asthma, autoimmune thyroid disease, and IgA production/vascular smooth muscle contraction of the intestinal immune network were enriched in the low-risk group (Fig. 12C, D). Collectively, the signature of five lncRNAs may have a significant impact on the initiation of disulfidptosis, offering fresh insights to investigate this particular form of cell death in LUAD.

Tumor mutational burden analysis based on disulfidptosis‑related lncRNA signature

Previous studies have shown that TMB can serve as a biomarker for both prognosis and immunotherapy. Therefore, to analyze mutations in the high- and low-risk groups, we downloaded LUAD mutation data from the TCGA database. The top 15 most frequently mutated genes, including TP53, TTN, MUC16, CSMD3, RYR2, LRP1B, ZFHX4, USH2A, KRAS, XIRP2, FLG, SPTA1, NAV3, ZNF536 and COL11A1 (Fig. 13A, B). Additionally, we investigated the differences in TMB between the different risk groups and found significant differences between the two groups (Fig. 13C). Furthermore, we explored the correlation between TMB and risk score, which revealed a positive relationship (Fig. 13D). Subsequently, we conducted a survival analysis of TMB and found that the high TMB group had significantly better OS than the low TMB group (Fig. 13E). Finally, we performed joint TMB risk analysis, dividing the entire patient cohort into four groups ("H-TMB + high risk," "H-TMB + low risk," "L-TMB + high risk," and "L-TMB + low risk") and comparing OS, which showed significant differences in prognosis among the four groups (Fig. 13F).

Analysis of the tumor microenvironment and immune infiltration

The tumor microenvironment consists mainly of cancerous and non-malignant cells, and the dynamic balance between them promotes cancer progression. Therefore, it is imperative to evaluate the tumor microenvironment. Here, we utilized data from the TCGA-LUAD cohort and employed the "ESTIMATE" R package to measure various scores, including immune scores, stromal scores, and estimated scores. Next, through the Wilcoxon rank sum test and Pearson correlation coefficient test, we discovered that the high-risk group exhibited lower scores in all categories and that all scores were negatively associated with risk scores (Fig. 14A-F). Afterward, we assessed the cancer stem cells (CSCs) index, which was notably greater in the high-risk group than in the low-risk group (Fig. 14G). The risk score was positively correlated with the CSCs index (Fig. 14H), indicating that LUAD patients with higher risk scores had a more pronounced stem cell profile with reduced cell differentiation. The TIDE score was higher in the high-risk group than in the low-risk group (Fig. 14I), and the risk score was positively correlated with the TIDE score (Fig. 14J), which suggests that immunotherapy is more effective in low-risk patients. Furthermore, the expression of immune checkpoints was explored, and the results showed that most immune checkpoints were lowly expressed in the high-risk group, suggesting that patients were tolerant to common immune checkpoint therapy in the high-risk group (Fig. 14K). Subsequently, we evaluated the percentage of immune cells according to the CIBERSORT algorithm. Tumor-infiltrating CD4 memory quiescent T cells, regulatory T cells, monocytes, quiescent dendritic cells, quiescent mast cells, and activated dendritic cells were markedly lower in patients from the high-risk cases. In contrast, the percentage of tumor-infiltrating activated mast cells, M0, and M1 macrophages was significantly greater in patients with high risk (Fig. 15A). Finally, analysis of differences in immune-related function by GSVA revealed significantly higher inflammation, parainflammation, and so on, in the low-risk cohort (Fig. 15B).

Screening of potential drugs for LUAD

Pharmacotherapy plays a crucial role in treating LUAD, and drug sensitivity analysis provides a tailored approach to identifying effective drug treatment mechanisms. We screened 12 potential drugs that differed from high-risk and low-risk patients (Fig. 16A-L). In high-risk patients, Trametinib, Savolitinib, Ulixertinib, and Crizotinib were found to be effective. In turn, in low-risk patients, Tozasertib, leflunomide, and ribociclib were found to be more responsive to treatment.

LUAD poses a serious global threat to human health and increasingly affects younger individuals(Lamort et al., 2022). Despite current advances in treatment options and outcomes, the clinical prognosis for LUAD remains bleak. Due to the limited diagnostic value of traditional lung tumor markers, including CEA, CA125, and CYFRA21-1, LUAD is usually diagnosed at an advanced stage when it is detected(Feng et al., 2016; Song et al., 2022a; Song et al., 2022b). With the rapid development and widespread application of high-throughput sequencing technology, the role of lncRNAs has been gradual to participate in tumor diagnosis, treatment, and prognosis(Jin et al., 2020). However, so far, only a few DRLs have been identified to play a role in cancer. Therefore, we construct a disulfidptosis-related lncRNAs prognostic signature based on the TCGA database, which was not only beneficial for improving the early detection of LUND patients but also helpful for exploring potential therapeutic targets.

In this study, we initially evaluated the mRNA levels of 23 DRGs in LUAD and normal tissues, and 21 DRGs showed significant differences. Among them, the expression of SLC7A11, GYS1, NDUFS1, NDUFA11, NUBPL, NCKAP1, LRPPRC, SLC3A2, RPN1, and CD2AP were upregulated, and the expression of ACTN4, ACTB, DSTN, FLNA, INF2, IQGAP1, MYH10, MYL6, MYH9, PDLIM1, and TLN1 were downregulated. Cancer is a disease driven by multiple mutations that contribute to uncontrolled cell proliferation and abnormal cell behavior. These mutation genes are referred to as "drivers", genes that promote tumorigenesis when their mutated forms affect the homeostasis of a range of critical cellular functions(Martinez-Jimenez et al., 2020). We then performed mutation analysis and found that FLNA was the most mutated gene, followed by MYH9 and TLN1. Functional enrichment analysis revealed that these 23 DRGs are involved in actin binding and activation, the structural composition of the cytoskeleton, GTPase binding, integrin binding, cytoskeletal movement and composition, and cell adhesion. Meanwhile, the expression of 15 genes was correlated with the OS of patients with LUAD.

Next, we conducted Lasso regression and Cox regression analysis on LUAD data in TCGA to obtain five DRLs (LINC01352, AC093673.1, AL606834.1, AL365181.2, and MHENCR) with independent prognostic values, which were used to construct prognosis signature. Notably, there is limited research available on the impact of these five DRLs on LUAD, with only a few studies reporting their effects on other types of cancer. For example, downregulation of LINC01352 upregulates miR-135b, resulting in decreased production of adenomatous polyposis coli (APC), activating the Wnt/beta protein signaling pathway and promoting cancer progression(Huang et al., 2020). MHENCR plays the role of an oncogene in melanoma by activating the miR-425/489 mediated PI3K-Akt pathway and may become a therapeutic target for melanoma(Chen et al., 2017). MHENCR can also inhibit the proliferation, migration, and invasion of colorectal cancer cells by regulating miR-532-3p, making it a novel prognostic marker for colorectal cancer(Zhou et al., 2022). In addition, high expression of AL606834.1 was associated with poor prognosis and unfavorable immune response in LUAD patients(Ma and Zheng, 2021). However, there has been no research reported on the clinical significance of AC093673.1 and AL365181.2, indicating the need for further investigation. In this study, we found that the OS of the high-risk group was significantly shorter than those of the low-risk group, and the patient mortality rate increased with increasing risk scores. The ROC curve indicated that the risk assessment signature had high accuracy in predicting prognosis. The PCA analysis demonstrates the scientific validity of the signature group. Furthermore, the independent prognostic analysis revealed that both clinical staging and risk score were independent prognostic factors. The calibration curves exhibited remarkable accuracy in predicting the survival of LUAD patients using clinical staging and risk scores. The above results indicate that the risk signature constructed from the five DRLs outperformed other clinical risk factors, while an increasing risk score was significantly associated with the progression of LUAD. Although conclusive evidence linking disulfidptosis to LUAD development is still lacking, the present study helps us to further explore the role of DRLs in LUAD. Meanwhile, the DLPS furthers our understanding of LUAD and provides clues to prognosis and novel therapeutic targets.

Recent studies have demonstrated that TMB, TIDE, and TME can serve as reliable indicators for monitoring the effectiveness of immunotherapy in LUAD(Bagaev et al., 2021; Li et al., 2021), Moreover, these factors also play a crucial role in assessing the patient's prognosis(Altorki et al., 2019; Remark et al., 2015). Accordingly, we explored whether the risk signature could be used to predict the efficacy of immunotherapy by evaluating the difference in TMB and TIDE scores between the high- and low-risk groups. The results showed that the risk signature was able to predict the response to immunotherapy. Accumulated evidence showed that patients with higher TMB and lower TIDE status were prominently correlated with greater clinical response to immunotherapy(Jiang et al., 2018; Yarchoan et al., 2017). Subsequently, we explored the relationship between DLPS and TME heterogeneity in LUAD and found that low numbers of stromal cells and immune cell infiltration, as well as high expression of tumor cell, and stem cell-related genes in the high-risk group, were strongly correlated with both risk factors. The TME has been recognized as a significant promoter of tumor growth, progression, and resistance to therapy(Hinshaw and Shevde, 2019). Although immune checkpoint inhibitors have rapidly become a mainstay treatment option for lung cancer in the past decade, their objective benefit rate in lung cancer patients is less than 20%(Sasikumar and Ramachandra, 2022). Therefore, we investigated the association between high- and low-risk groups and different immune checkpoints. Surprisingly, most of the immune checkpoint genes were under-expressed in the high-risk group, indicating that some immune checkpoint blockade may be beneficial for individual high-risk patients. The analysis of immune infiltration showed that the high-risk group had lower levels of immune cell infiltration, including CD4 memory resting T cells, regulatory T cells, and dendritic cells. However, there was a significant increase in the percentage of activated mast cells, M0 macrophages, and M1 macrophages, indicating the potential contribution of these immune cells to the development and progression of LUAD. It was shown that combined treatment with interferon-γ and β-glucan repolarized M2 macrophages into M1 macrophages and enhanced antitumor activity, and the respective antitumor and pro-tumor effects of tumor-associated macrophages reflect the complexity of macrophage function in malignancy(Wu et al., 2021). For instance, in renal cell carcinoma, high levels of resting dendritic cells were positively associated with tumor prognosis(Pan et al., 2020). Dendritic cells and T follicular helper cells play an essential role in antitumor immunity(Cui et al., 2021). CD8⁺ T cells and γ δ T cells activated by CD4⁺ Th1 cells are usually associated with a type I immune response and are associated with a positive prognosis in patients with lung cancer(Bremnes et al., 2016). Finally, in our findings, the high-risk group was associated with reduced semi-inhibitory concentrations of commonly used chemotherapeutic agents (Trametinib, Savolitinib, etc.) against LUAD, suggesting the additional potential of DLPS in predicting chemotherapeutic efficacy.

Disulfidptosis has shown great therapeutic promise for anti-tumor response and, importantly, has opened new avenues for studying the clinical utility of cancer drugs. However, numerous outstanding questions in suspense still need to be further studied. What specific substances are oxidized in the disulfide process, and by what means does it interact with other forms of cell death at the molecular level? Therefore, further research on all of the issues is vital and warranted to clarify. At the same time, additional studies on disulfidptosis can help us to more adequately unravel the underlying pathogenesis and progression of related diseases. Disulfidptosis inhibitors may alleviate the progression of the disease and serve as a starting point for future basic and clinical research in the treatment and prevention of tumors. As the rate of early detection of lung cancer increases, further novel therapies will be explored in addition to existing treatments such as surgery, chemotherapy, radiotherapy, molecular targeting, and immunotherapy.

Our study, despite some notable findings, has several limitations. Firstly, we did not evaluate the predictive accuracy of the DLPS on external clinical data. Moreover, we only analyze TCGA public data with a single sample size, which is prone to errors. In addition, there is a lack of experimental verification of the results of disulfidptosis-related lncRNAs, which requires further research and verification.

We identified a prognostic signature of disulfidptosis‑related lncRNAs by comprehensive and systematic bioinformatics analysis of LUAD patients. Our developed risk-based signature outperforms standard clinicopathological parameters in predicting survival and prognosis and provided novel insights into the treatment of LUAD.

Author contributions WQL wrote the manuscript. YW and YL analyzed the data and created the figures. XYL and YYW reviewed and edited the manuscript. DSL and YYS designed the experiments and revised the manuscript. All the authors reviewed and approved the final version of the manuscript.

Funding This work was supported by Shandong Provincial Natural Science Foundation (Grant No. ZR2021MH082 and ZR2020MH011). This work was also supported by the Key Research Project of Yantai City (Grant No. 2020XDRH111).

Data availability The original contributions presented in the study are included in the article material. And further inquiries can be directed to the corresponding author (Dr. Yeying Sun).

Conflict of interest The authors declare no competing financial interests.

Altorki, N.K., Markowitz, G.J., Gao, D., Port, J.L., Saxena, A., Stiles, B., McGraw, T., Mittal, V., 2019. The lung microenvironment: an important regulator of tumour growth and metastasis. Nat Rev Cancer 19, 9-31.
Bade, B.C., Dela Cruz, C.S., 2020. Lung Cancer 2020: Epidemiology, Etiology, and Prevention. Clin Chest Med 41, 1-24.
Bagaev, A., Kotlov, N., Nomie, K., Svekolkin, V., Gafurov, A., Isaeva, O., Osokin, N., Kozlov, I., Frenkel, F., Gancharova, O., Almog, N., Tsiper, M., Ataullakhanov, R., Fowler, N., 2021. Conserved pan-cancer microenvironment subtypes predict response to immunotherapy. Cancer Cell 39, 845-865 e847.
Bremnes, R.M., Busund, L.T., Kilvaer, T.L., Andersen, S., Richardsen, E., Paulsen, E.E., Hald, S., Khanehkenari, M.R., Cooper, W.A., Kao, S.C., Donnem, T., 2016. The Role of Tumor-Infiltrating Lymphocytes in Development, Progression, and Prognosis of Non-Small Cell Lung Cancer. J Thorac Oncol 11, 789-800.
Bridges, M.C., Daulagala, A.C., Kourtidis, A., 2021. LNCcation: lncRNA localization and function. J Cell Biol 220.
Chen, H., Carrot-Zhang, J., Zhao, Y., Hu, H., Freeman, S.S., Yu, S., Ha, G., Taylor, A.M., Berger, A.C., Westlake, L., Zheng, Y., Zhang, J., Ramachandran, A., Zheng, Q., Pan, Y., Zheng, D., Zheng, S., Cheng, C., Kuang, M., Zhou, X., Zhang, Y., Li, H., Ye, T., Ma, Y., Gao, Z., Tao, X., Han, H., Shang, J., Yu, Y., Bao, D., Huang, Y., Li, X., Zhang, Y., Xiang, J., Sun, Y., Li, Y., Cherniack, A.D., Campbell, J.D., Shi, L., Meyerson, M., 2019. Genomic and immune profiling of pre-invasive lung adenocarcinoma. Nat Commun 10, 5472.
Chen, X., Dong, H., Liu, S., Yu, L., Yan, D., Yao, X., Sun, W., Han, D., Gao, G., 2017. Long noncoding RNA MHENCR promotes melanoma progression via regulating miR-425/489-mediated PI3K-Akt pathway. Am J Transl Res 9, 90-102.
Cui, C., Wang, J., Fagerberg, E., Chen, P.M., Connolly, K.A., Damo, M., Cheung, J.F., Mao, T., Askari, A.S., Chen, S., Fitzgerald, B., Foster, G.G., Eisenbarth, S.C., Zhao, H., Craft, J., Joshi, N.S., 2021. Neoantigen-driven B cell and CD4 T follicular helper cell collaboration promotes anti-tumor CD8 T cell responses. Cell 184, 6101-6118 e6113.
de Mello, R.A., Neves, N.M., Tadokoro, H., Amaral, G.A., Castelo-Branco, P., Zia, V.A.A., 2020. New Target Therapies in Advanced Non-Small Cell Lung Cancer: A Review of the Literature and Future Perspectives. J Clin Med 9.
Dixon, S.J., Lemberg, K.M., Lamprecht, M.R., Skouta, R., Zaitsev, E.M., Gleason, C.E., Patel, D.N., Bauer, A.J., Cantley, A.M., Yang, W.S., Morrison, B., Stockwell, B.R., 2012. Ferroptosis: An Iron-Dependent Form of Nonapoptotic Cell Death. Cell 149, 1060-1072.
Engebretsen, S., Bohlin, J., 2019. Statistical predictions with glmnet. Clin Epigenetics 11, 123.
Feng, Y.C., Feng, M., Zhang, D., Huang, Y.C., 2016. Retrospective Study to Determine Diagnostic Utility of 6 Commonly Used Lung Cancer Biomarkers Among Han and Uygur Population in Xinjiang Uygur Autonomous Region of People's Republic of China. Medicine 95.
Geeleher, P., Cox, N., Huang, R.S., 2014. pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PLoS One 9, e107468.
Goodall, G.J., Wickramasinghe, V.O., 2021. RNA in cancer. Nat Rev Cancer 21, 22-36.
Hanzelmann, S., Castelo, R., Guinney, J., 2013. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 14, 7.
Herman, A.B., Tsitsipatis, D., Gorospe, M., 2022. Integrated lncRNA function upon genomic and epigenomic regulation. Mol Cell 82, 2252-2266.
Hinshaw, D.C., Shevde, L.A., 2019. The Tumor Microenvironment Innately Modulates Cancer Progression. Cancer Res 79, 4557-4566.
Huang, P., Xu, Q., Yan, Y., Lu, Y., Hu, Z., Ou, B., Zhang, H., Mao, K., Zhang, J., Wang, J., Xiao, Z., 2020. HBx/ERalpha complex-mediated LINC01352 downregulation promotes HBV-related hepatocellular carcinoma via the miR-135b-APC axis. Oncogene 39, 3774-3789.
Jia, K., Wu, Y., Huang, J., Wu, H., 2019. Survival-Associated Alternative Splicing Events in Pan-Renal Cell Carcinoma. Front Oncol 9, 1317.
Jiang, N., Zhang, X.Y., Gu, X.J., Li, X.Z., Shang, L., 2021. Progress in understanding the role of lncRNA in programmed cell death. Cell Death Discov 7.
Jiang, P., Gu, S., Pan, D., Fu, J., Sahu, A., Hu, X., Li, Z., Traugh, N., Bu, X., Li, B., Liu, J., Freeman, G.J., Brown, M.A., Wucherpfennig, K.W., Liu, X.S., 2018. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 24, 1550-1558.
Jin, K.T., Lu, Z.B., Lv, J.Q., Zhang, J.G., 2020. The role of long non-coding RNAs in mediating chemoresistance by modulating autophagy in cancer. RNA Biol 17, 1727-1740.
Kanehisa, M., Furumichi, M., Sato, Y., Ishiguro-Watanabe, M., Tanabe, M., 2021. KEGG: integrating viruses and cellular organisms. Nucleic Acids Res 49, D545-D551.
Lamort, A.S., Kaiser, J.C., Pepe, M.A.A., Lilis, I., Ntaliarda, G., Somogyi, K., Spella, M., Behrend, S.J., Giotopoulou, G.A., Kujawa, W., Lindner, M., Koch, I., Hatz, R.A., Behr, J., Sotillo, R., Schamberger, A.C., Stathopoulos, G.T., 2022. Prognostic phenotypes of early-stage lung adenocarcinoma. Eur Respir J 60.
Li, Z., Pan, C., Wang, Z., Deng, X., Zhu, Q., Wu, W., Chen, L., 2021. LncRNA PCBP1-AS1 correlated with the functional states of cancer cells and inhibited lung adenocarcinoma metastasis by suppressing the EMT progression. Carcinogenesis.
Liu, X., Nie, L., Zhang, Y., Yan, Y., Wang, C., Colic, M., Olszewski, K., Horbath, A., Chen, X., Lei, G., Mao, C., Wu, S., Zhuang, L., Poyurovsky, M.V., James You, M., Hart, T., Billadeau, D.D., Chen, J., Gan, B., 2023. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol 25, 404-414.
Liu, Y.Q., Chen, X.J., Chen, X.L., Liu, J.Q., Gu, H., Fan, R.T., Ge, H., 2020. Long non-coding RNA HOTAIR knockdown enhances radiosensitivity through regulating microRNA-93/ATG12 axis in colorectal cancer. Cell Death Dis 11.
Ma, S.Z., Zhu, J., Wang, M.M., Zhu, J.F., Wang, W.C., Xiong, Y.L., Jiang, R.M., Seetharamu, N., Abrao, F.C., Puthamohan, V.M., Liu, L., Jiang, T., 2022. A cuproptosis-related long non-coding RNA signature to predict the prognosis and immune microenvironment characterization for lung adenocarcinoma. Transl Lung Cancer R 11, 2079-+.
Ma, Y., Zheng, W., 2021. H3K27ac-induced lncRNA PAXIP1-AS1 promotes cell proliferation, migration, EMT and apoptosis in ovarian cancer by targeting miR-6744-5p/PCBP2 axis. J Ovarian Res 14, 76.
Martinez-Jimenez, F., Muinos, F., Sentis, I., Deu-Pons, J., Reyes-Salazar, I., Arnedo-Pac, C., Mularoni, L., Pich, O., Bonet, J., Kranas, H., Gonzalez-Perez, A., Lopez-Bigas, N., 2020. A compendium of mutational cancer driver genes. Nat Rev Cancer 20, 555-572.
Mattiuzzi, C., Lippi, G., 2020. Cancer statistics: a comparison between World Health Organization (WHO) and Global Burden of Disease (GBD). Eur J Public Health 30, 1026-1027.
Mayakonda, A., Lin, D.C., Assenov, Y., Plass, C., Koeffler, H.P., 2018. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res 28, 1747-1756.
Pan, Q.F., Wang, L.W., Chai, S.S., Zhang, H., Li, B., 2020. The immune infiltration in clear cell Renal Cell Carcinoma and their clinical implications: A study based on TCGA and GEO databases. J Cancer 11, 3207-3215.
Remark, R., Becker, C., Gomez, J.E., Damotte, D., Dieu-Nosjean, M.C., Sautes-Fridman, C., Fridman, W.H., Powell, C.A., Altorki, N.K., Merad, M., Gnjatic, S., 2015. The non-small cell lung cancer immune contexture. A major determinant of tumor characteristics and patient outcome. Am J Respir Crit Care Med 191, 377-390.
Sang, L.J., Yang, L.J., Ge, Q.W., Xie, S.S., Zhou, T.H., Lin, A.F., 2022. Subcellular distribution, localization, and function of noncoding RNAs. Wires Rna 13.
Sasikumar, P.G., Ramachandra, M., 2022. Small Molecule Agents Targeting PD-1 Checkpoint Pathway for Cancer Immunotherapy: Mechanisms of Action and Other Considerations for Their Advanced Development. Front Immunol 13, 752065.
Shinkawa, H., Tanaka, S., Takemura, S., Amano, R., Kimura, K., Kinoshita, M., Takahashi, K., Matsuzaki, S., Kubo, S., 2021. Nomograms predicting extra- and early intrahepatic recurrence after hepatic resection of hepatocellular carcinoma. Surgery 169, 922-928.
Song, H.B., Liu, D.C., Wang, L.W., Liu, K.S., Chen, C., Wang, L., Ren, Y., Ju, B., Zhong, F.H., Jiang, X.Y., Wang, G.S., Chen, Z.S., Zou, C., 2022a. Methyltransferase like 7B is a potential therapeutic target for reversing EGFR-TKIs resistance in lung adenocarcinoma. Mol Cancer 21.
Song, P., Li, W., Wu, X., Qian, Z., Ying, J., Gao, S., He, J., 2022b. Integrated analysis of single-cell and bulk RNA-sequencing identifies a signature based on B cell marker genes to predict prognosis and immunotherapy response in lung adenocarcinoma. Cancer Immunol Immunother 71, 2341-2354.
Stipanuk, M.H., 2004. Sulfur amino acid metabolism: pathways for production and removal of homocysteine and cysteine. Annu Rev Nutr 24, 539-577.
Tibshirani, R., 1997. The lasso method for variable selection in the Cox model. Stat Med 16, 385-395.
Wan, Q., Sang, X., Jin, L., Wang, Z.C., 2020. Alternative Splicing Events as Indicators for the Prognosis of Uveal Melanoma. Genes-Basel 11.
Wu, M.F., Lin, C.A., Yuan, T.H., Yeh, H.Y., Su, S.F., Guo, C.L., Chang, G.C., Li, K.C., Ho, C.C., Chen, H.W., 2021. The M1/M2 spectrum and plasticity of malignant pleural effusion-macrophage in advanced lung cancer. Cancer Immunol Immunother 70, 1435-1450.
Yarchoan, M., Hopkins, A., Jaffee, E.M., 2017. Tumor Mutational Burden and Response Rate to PD-1 Inhibition. N Engl J Med 377, 2500-2501.
Yoshihara, K., Shahmoradgoli, M., Martinez, E., Vegesna, R., Kim, H., Torres-Garcia, W., Trevino, V., Shen, H., Laird, P.W., Levine, D.A., Carter, S.L., Getz, G., Stemke-Hale, K., Mills, G.B., Verhaak, R.G., 2013. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 4, 2612.
Yu, G.C., Wang, L.G., Han, Y.Y., He, Q.Y., 2012. clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. Omics 16, 284-287.
Zhou, D., Liao, Z., Chen, X., Fan, Y., Zuo, H., 2022. LncRNA MHENCR Predicts Poor Outcomes in Patients with Colorectal Carcinoma and Modulates Tumorigenesis by Impairing MiR-532-3p. Tohoku J Exp Med 259, 77-84.
Zhou, Y., Zhou, B., Pache, L., Chang, M., Khodabakhshi, A.H., Tanaseichuk, O., Benner, C., Chanda, S.K., 2019. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun 10, 1523.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Disulfidptosis-related lncRNA signatures predict prognosis and immune relevance of lung adenocarcinoma

Status:

Version 1

Abstract

Figures

Headings

Introduction

Materials and methods

Data collection and processing

Differential expression of DRGs

Construction of disulfidptosis-related lncRNAs prognostic signature

Independent prognostic analysis and nomogram construction

Functional enrichment analysis and principal component analysis

Analysis of tumor mutation burden, tumor microenvironment, and immune infiltrate

The role of the risk signature in predicting drug sensitivity

Statistical analysis

Results

Analysis of DRG expression, gene mutation, and survival in LUAD

The function enrichment analysis of DRGs

Evaluation and validation of the disulfidptosis-related lncRNA prognostic signature

Survival analysis of the signature

Independent analysis of the prognostic factors

Functional enrichment analysis of differential genes in high- and low-risk groups

Tumor mutational burden analysis based on disulfidptosis‑related lncRNA signature

Analysis of the tumor microenvironment and immune infiltration

Screening of potential drugs for LUAD

Discussion

Conclusions

Declarations

References

Additional Declarations

Status:

Version 1