Identification of Crucial Genes Associated with Ferroptosis in COPD via Comprehensive Bioinformatics Analysis and the Relevant Experimental Validation

doi:10.21203/rs.3.rs-5315834/v1

Download PDF

Research Article

Identification of Crucial Genes Associated with Ferroptosis in COPD via Comprehensive Bioinformatics Analysis and the Relevant Experimental Validation

https://doi.org/10.21203/rs.3.rs-5315834/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Chronic obstructive pulmonary disease (COPD) is a chronic respiratory disease characterized by partially reversible airway obstruction, with high mortality and disability rates. Smoking is the primary risk factor for COPD. Ferroptosis is a novel form of cell death characterized by iron-mediated lipid peroxidation induced by reactive oxygen species (ROS) generated in the Fenton reaction. Recent studies have shown that ferroptosis in airway epithelial cells may be involved in and mediate the pathogenesis of COPD. This study aimed to identify and validate key genes associated with ferroptosis in COPD via bioinformatics methods.

Methods: Four microarray datasets (GSE10006, GSE20257, GSE11906, and GSE11784) were downloaded from the GEO database. Differential gene expression analysis was conducted separately for each dataset via the limma package in R, resulting in a set of 132 overlapping differentially expressed genes (DEGs). Weighted gene coexpression network analysis (WGCNA) was employed to identify key gene modules associated with COPD. String analysis, Cytoscape, functional enrichment analysis, and construction of protein‒protein interaction (PPI) networkswere utilized to identify hub genes.We subsequently generated a receiver operating characteristic (ROC) curve to predict the risk of COPD occurrence. Concurrently, we conducted differential expression analysis of ferroptosis-related genes across three datasets and identified ferroptosis-related hub genes (FRHGs) that overlapped with pivotal genes related to ferroptosis. These FRHGs were validated via the GSE11784 dataset, followed by validation via in vitro cell experiments (westernblotting, quantitative PCR). Finally, we analyzed immune cell infiltration and performed consistent clustering analysis on the basis of gene set enrichment analysis (GSEA) scores.

Results: We identified four potential hub genes associated with ferroptosis in COPD (NQO1, AKR1C3, GPX2, and CBR1), identifying new therapeutic targets for clinical treatment and diagnosis. Additionally, on the basis of these four FRHGs, we found that acetaminophen and glycidamide were highly relevant drug targets.

Conclusion: This study identified 4 FRHGs as potential biomarkers for COPD diagnosis and treatment. We predict COPD occurrence through bioinformatics analysis and various machine learning algorithms. Moreover, cell experiments revealed significant upregulation trends of the FRHGs identified in this study in COPD disease models, suggesting new avenues for clinical diagnosis and treatment strategies.

COPD

ferroptosis

bioinformatics

machine learning

immune infiltration

Chronic obstructive pulmonary disease (COPD) is a chronic respiratory disease characterized by incomplete and reversible airflow obstruction, with high mortality and disability rates [1]. Compared with healthy individuals, patients with COPD have a greater risk of developing other diseases, such as lung cancer [2]. Risk factors for COPD include genetic factors, smoking, and airway inflammation [3–5], with smoking being the primary risk factor. Increasing evidence suggests that the pathogenesis of COPD involves various biological functions, including cell proliferation, apoptosis, autophagy, and ferroptosis [6–9]. Although COPD is a common chronic disease, it can progress to pulmonary heart disease and respiratory failure. According to the World Health Organization (WHO), COPD is associated with a significant incidence and mortality rate and ranks as the third leading cause of death globally [10–13]. It is estimated that approximately 384 million people worldwide suffer from chronic obstructive pulmonary disease [14]. The onset of COPD is related to environmental factors and family history, with recent research focusing on inflammatory damage, oxidative stress, protease/antiprotease imbalance, and autoimmune responses. Among these factors, oxidative stress is a major contributor to COPD and involves an imbalance between oxidation and antioxidant processes, leading to lung tissue damage induced directly or indirectly by reactive oxygen species (ROS).

Ferroptosis is an iron-dependent form of regulated cell death first proposed by Dixon et al. in 2012 [34]. Unlike other forms of cell death, such as apoptosis, necrosis, and autophagy, the morphological features of ferroptosis include high mitochondrial membrane density, reduced or absent mitochondrial cristae, and mitochondrial membrane folding or wrinkling. Its biochemical processes involve abnormal iron metabolism, oxidative‒reductive imbalance, and lipid peroxidation [15], with excessive iron-dependent lipid oxidation accumulation as its primary characteristic. This is manifested by abnormal metabolism of intracellular lipid peroxides catalyzed by excess iron ions, generation of reactive oxygen species (ROS), and regulated cell death mediated by excessive oxidation of polyunsaturated fatty acids (PUFAs). Excess iron promotes ROS production through the Fenton reaction and iron-binding proteins, thereby promoting ferroptosis [16, 17]. The lipid repair enzyme glutathione peroxidase 4 (GPx4), a member of the selenium-dependent protein family, has been shown to negatively regulate ferroptosis by directly reducing lipid hydroperoxides [18, 19].

Numerous studies indicate that ferroptosis plays a crucial role in the development of diseases such as acute lung injury, asthma, chronic obstructive pulmonary disease, Alzheimer's disease, and renal failure. Prolonged exposure to particulate matter in smoke, including iron particles, leads to the accumulation of reactive oxygen species in the respiratory epithelium. This accumulation causes membrane damage, resulting in the release of damage-associated molecular patterns (DAMPs). Consequently, the balance of iron within the body is altered, leading to oxidative stress, inflammation, and ultimately ferroptosis. These processes can exacerbate the progression of emphysema and airway narrowing in COPD patients [20, 21].

2.1 Microarray Data Collection and Processing

Microarray data were retrieved from the NCBI Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) database via the R package GEOquery. Four datasets (GSE10006, GSE20257, GSE11906, and GSE11784) were obtained. GSE10006, GSE20257, and GSE11906 included gene expression data from patients with chronic obstructive pulmonary disease (COPD) and normal controls. GSE11784 served as an external validation dataset for verifying hub genes related to ferroptosis (FRHGs) and exploring the impact of smoking on the differential expression of FRHGs. A total of 564 ferroptosis-related genes were obtained from the FerDB database (http://www.zhounan.org/ferrdb/).

Data preprocessing is crucial, as it significantly affects all subsequent analyses. Following data download, arrays were standardized and quantile-normalized via the R package affy of the BioConductor project. For genes measured by multiple probes, the median value was taken as the gene expression value. Probes were converted to gene symbols, and gene expression matrix files were prepared. The data were log2 transformed.

Principal component analysis (PCA) was used to validate data reproducibility, and PCA plots were generated via R packages. The limma package in R was used to define differentially expressed genes (DEGs) with adjusted p values less than 0.05 and absolute fold changes (FCs) greater than 1 [22]. Heatmaps and volcano plots were created via the R packages heatmap and ggplot2, respectively [23]. Detailed information about the datasets used in the study can be found in Table 1, and the research design flowchart is depicted in Fig. 1.

2.2 Construction of protein‒protein interaction (PPI) networks and analysis of DEGs

To explore the potential interactions of the DEGs, we first mapped them onto a protein‒protein interaction (PPI) network. The STRING database (https://string-db.org/) was utilized to analyze the PPI network of the DEGs. The analyzed STRING data were subsequently imported into Cytoscape software (version 3.9.1). The Cytohubba plugin in Cytoscape, which employs the maximum clique centrality (MCC) algorithm, was used to identify the top 10 hub genes on the basis of their scores and visualize them.

Furthermore, we matched these identified hub genes with those in the human ferroptosis-related gene database to identify genes overlapping with both ferroptosis-related genes and hub genes. This integrated approach allowed us to explore and visualize the interactions among DEGs, identify hub genes within the PPI network, and identify genes associated with both ferroptosis and hub networks. This section outlines the methodology for constructing and analyzing PPI networks of DEGs, leveraging STRING database analysis and visualization via Cytoscape software.

2.3. Identification of significant module genes via weighted gene coexpression network analysis (WGCNA)

In this study, we employed WGCNA to identify significantly correlated module genes associated with COPD. Initially, we calculated the median absolute deviation (MAD) for each gene and excluded the bottom 50% with the smallest MAD. Using the goodSamplesGenes function in WGCNA, we subsequently filtered the expression matrix of differentially expressed genes (DEGs) to construct a scale-free coexpression network. Next, employing the pickSoftThreshold function, we computed adjacency relationships on the basis of coexpression similarity to determine the soft-thresholding power β. The adjacency relationships were then transformed into a topological overlap matrix (TOM), and gene dissimilarity (1-TOM) was calculated accordingly. Modules were identified via hierarchical clustering and dynamic tree cutting methods. For the classification of genes with similar characteristics, we utilized a minimum module size (gene count) of 50 on the basis of TOM dissimilarity measures in the gene dendrogram and initiated average linkage hierarchical clustering. Finally, we selected a cut height in the module dendrogram and merged modules on the basis of the estimated dissimilarity of feature genes. Concurrently, visualization techniques were applied to the network of feature genes.

2.4. Functional enrichment analysis

KEGG analysis provides a knowledge base for systematic gene function analysis. Using the clusterProfiler package in R, we performed GO and KEGG pathway analyses of FRHGs and visualized the results [24, 25]. The GO analysis included the identification of biological process (BP), cellular component (CC), and molecular function (MF) terms. The adjusted significance criteria were set at a p value < 0.05 and a q value < 0.2. Enrichment analysis was conducted via the GOplot package in R, which integrates LogFC values obtained from previous differential gene expression analyses. The GSVA package was subsequently used for gene set enrichment analysis (GSEA), which focuses on relevant pathways and validates enrichment results on the basis of the normalized enrichment score (NES), gene ratio, and P value, with |NES| > 1 and FDR q < 0.25 considered significant enrichment.

2.5. Hub Gene ROC Curve Construction

To assess the diagnostic value of each hub gene in distinguishing COPD patients from normal controls, ROC curves for each hub gene were plotted via R software. These curves were used to determine the sensitivity and specificity of the target genes, and the results were quantified via the area under the ROC curve (AUC). Genes with an AUC > 0.6 were considered diagnostic genes.

2.6. Immune cell infiltration

Immune cell infiltration analysis is commonly used to investigate disease pathogenesis. The CIBERSORT algorithm calculates the proportions of various immune cells in samples [26]. We employed this algorithm to analyze immune cells from healthy subjects and COPD patients across three datasets (GSE10006, GSE11906, and GSE20257) and visualized the immune cell compositions of both groups via box plots. Differences in immune cell proportions were assessed via the Wilcoxon rank-sum test, with results considered significant at P < 0.05.

2.7. Targeted drug prediction

Drugs that interact with robust DEGs were identified via the DSigDB through Enrichr (https://maayanlab.cloud/Enrichr/). We used the DSigDB database to predict potential targeted drugs related to key genes that may contribute to treating COPD by modulating ferroptosis.

2.8. Validation of dataset GSE11784

The expression of the screened target genes was tested via the GSE11784 dataset to further validate the authenticity and feasibility of the identification of the hub genes via the above bioinformatics methods. Subsequently, GSEA was performed to identify functional enrichment in COPD-affected smokers and nonsmoking healthy individuals.

2.9. Cell model construction and experimental validation

The human bronchial epithelial cell line (BEAS-2B) was sourced from the Department of Pathogenic Biology, School of Basic Medical Sciences, Jilin University, and maintained in culture in DMEM (Cytiva). Approximately 120,000 cells were seeded per well in a six-well plate and cultured in a cell culture incubator until they reached a cell growth density of 80%-90%, followed by CS stimulation. Cigarette smoke was introduced into DMEM using a negative pressure device, and the absorbance (OD) was measured at 320 nm. An OD of 0.38 ± 0.02 was defined as the 100% cigarette smoke (CS) concentration [27]. One group of BEAS-2B cells was treated with culture medium containing 8% cigarette smoke (CS) and designated the COPD group. Another group of BEAS-2B cells was cultured in maintenance medium for 24 hours, after which protein extraction was performed for Western blot (WB) experiments. To replicate the above experimental method, a cell model was constructed. The RNA extracted from this model will then be subjected to quantitative polymerase chain reaction (qPCR) to validate gene expression levels, presumably focusing on the identified hub genes.

3.1 Identification of Ferroptosis Key Genes in COPD

We downloaded the GSE10006, GSE20257, and GSE11906 datasets from the GEO database. The expression matrices of the three datasets were normalized via R software. Differential expression analysis was performed via the limma package, with criteria set at |log2 (FC)| > 1 and P value < 0.05 for selection. From the GSE10006 dataset, 780 DEGs were identified (366 upregulated, 414 downregulated); from GSE20257, 286 DEGs (179 upregulated, 107 downregulated); and from GSE11906, 343 DEGs (178 upregulated, 165 downregulated). Volcano plots of the DEGs in the above 3 datasets are shown in Fig. 2A-2C. Venn diagrams were subsequently generated (Fig. 2D), revealing a total of 132 genes with consistent differential expression across all three datasets used in this study. http://www.zhounan.org/ferrdb/A total of 564 ferroptosis-related genes were obtained from the FerDB database ().

3.2. Results of protein‒protein interaction network construction

A Venn diagram analysis identified GPX2, IL1B, PIR, NQO1, SRXN1, AKR1C3, AKR1C2, CBR1, SLC7A11 and DUOX2 as key genes associated with ferroptosis (Fig. 3A). Using STRING, we established interactions among the identified DEGs to construct a protein–protein interaction (PPI) network (Fig. 3B). Cytoscape software (version 3.9.1) was used to import and analyze the STRING data. The top 10 genes on the basis of their scores were designated hub genes (Fig. 3C). Ultimately, we identified 10 hub genes: NQO1, AKR1C3, AKR1B10, ALDH3A1, AKR1B1, IL1B, CBR1, CCL2, GPX2, and MRC1 (Fig. 3D). By plotting a Venn diagram to intersect the PPI hub genes with the previously identified ferroptosis-related hub genes, the following hub genes were obtained: GPX2, NQO1, AKR1C3, CBR1, and IL1B (Fig. 3E).

To assess the reproducibility of the intragroup data, principal component analysis (PCA) was conducted on three datasets (Fig. 4A-C), which revealed excellent data consistency. The volcano plots of differentially expressed genes (DEGs) from these three datasets (highlighting selected key genes) are shown in Figs. 4D-F. The heatmaps for GSE10006, GSE11906, and GSE20257 revealed significant upregulation of NQO1, CBR1, GPX2, and AKR1C3 in the COPD group.

3.3. Significant module gene identification in COPD via WGCNA

We first calculated the coexpression correlation coefficients among genes on the basis of their measured expression levels. According to the calculation results in Fig. 5A, genes were clustered to generate a gene dendrogram. Different branches of the clustering tree represent different gene modules, distinguished by different colors. Genes were categorized into modules on the basis of their expression patterns via weighted correlation coefficients. WGCNA was employed to identify important module genes associated with COPD. Supplementary Fig. 5B shows the selection of soft thresholding and gene clustering trees.

The red and yellow modules presented the most significant positive correlations with COPD. The correlation between module membership in the red/yellow modules and gene significance in COPD patients is illustrated in Supplementary Fig. 5C. For further analysis, 1950 genes were selected from these two modules. Intersection of genes associated with ferroptosis (n = 564) and COPD-related module genes (n = 1950) yielded 75 overlapping hub genes, as shown in Fig. 2C. The intersection with the top ten hub genes obtained from constructing the PPI network revealed seven hub genes: GPX2, AKR1B10, ALDH3A1, NQO1, AKR1C3, AKR1B1, and CBR1. Further intersection with the five genes obtained from differential expression analysis revealed the four final hub genes associated with ferroptosis in COPD: AKR1C3, NQO1, GPX2, and CBR1 (Fig. 5D).

3.4. GO and KEGG enrichment analyses of the differentially expressed FRHGs

GO and KEGG enrichment analyses were conducted for four key genes associated with ferroptosis. The results of the KEGG enrichment analysis revealed that these DEGs are notably involved in metabolic processes such as folate biosynthesis, chemical carcinogenesis (reactive oxygen species), arachidonic acid metabolism, fluid shear stress, and atherosclerosis (Fig. 6A-B). The GO enrichment analysis highlights categories including prostaglandin biosynthetic process, prostaglandin metabolic process, unsaturated fatty acid biosynthetic process, eicosanoic acid biosynthetic process, quinone or similar compounds that act on NAD(P)H oxidoreductase activity, antioxidant activity, CH-OH donor oxidoreductase activity that acts on NAD or NADP as acceptors, and other oxidoreductase activities (Fig. 6C-D).

3.5. ROC curve of the hub genes

We generated ROC curves from data from COPD patients and healthy subjects. These results indicate that these four genes have significant diagnostic value for smoking-related COPD. Compared with those in the NC group, the expression of four genes was upregulated in the COPD group (Fig. 7A). ROC curve analysis revealed that each gene has good predictive performance, as follows (Fig. 7B-D): in the GSE10006 dataset, NQO1 (AUC: 0.99), AKR1C3 (AUC: 1.00), CBR1 (AUC: 0.93), and GPX2 (AUC: 1.00). The GSE11906 dataset included NQO1 (AUC: 1.00), AKR1C3 (AUC: 0.93), CBR1 (AUC: 0.91), and GPX2 (AUC: 1.00). In the GSE20257 dataset, NQO1 (AUC: 0.99), AKR1C3 (AUC: 1.00), CBR1 (AUC: 0.93), and GPX2 (AUC: 1.00) were identified. The 20257 dataset included NQO1 (AUC: 0.99), AKR1C3 (AUC: 1.00), CBR1 (AUC: 0.92), and GPX2 (AUC: 0.99).

3.6. Immune infiltration analysis

Bar charts were generated to visualize the proportions of immune cells in each sample. The bar chart (Fig. 8A) shows that, compared with those in the control group, the proportions of resting mast cells, resting dendritic cells, eosinophils, M0 macrophages, activated memory CD4 + T cells, and M2 macrophages were greater in the COPD group, whereas the proportions of neutrophils, resting NK cells, CD8 + T cells, activated dendritic cells, resting memory CD4 + memory T cells, follicular helper T cells, activated mast cells, and activated gamma delta T cells were lower in the COPD group than in the control group. Correlation analysis revealed that resting dendritic cells had the highest positive correlation with M0 macrophages, whereas resting memory CD4 T cells had the lowest correlation with naive B cells (Fig. 8B). Additionally, we found that four central DEGs were positively correlated with immune cell infiltration (Fig. 8C).

3.7. Potential therapeutic drug prediction

Drugs that interact with robust DEGs were identified via the DSigDB through Enrichr. We used the DSigDB database to predict potential targeted drugs related to key genes that may contribute to treating COPD by modulating ferroptosis. We ultimately predicted 352 targeted drugs, with the top 10 predicted drugs on the basis of the comprehensive scoring listed in Table 2. Importantly, increasing the dosage of acetaminophen may increase the risk of developing COPD. [28]. This research provides further evidence linking acetaminophen use to increased risks of asthma and chronic obstructive pulmonary disease (COPD), as well as decreased lung function (Fig. 9A-B). Currently, there are no studies linking glycidamide with COPD. However, its comprehensive score suggests potential new avenues and targets for COPD treatment research.

3.8. GSE11784 confirms the expression and diagnostic value of FRHGs

The expression of the screened target genes was detected via the GSE11784 dataset. The results indicate that the differential expression of four iron-related hub genes (NQO1, AKR1C3, GPX2, and CBR1) between COPD patients and nonsmoking healthy individuals is consistent with our predictions (Fig. 10A-D).

GSEA revealed significant positive correlations between the four hub genes and the HALLMARK gene sets "XENOBIOTIC_METABOLISM", "FATTY_ACID_METABOLISM", "P53_PATHWAY", "REACTIVE_OXYGEN_SPECIES_PATHWAY", and "MYOGENESIS" in relation to COPD (Fig. 10E). These five gene sets are implicated in the occurrence and progression of ferroptosis. These findings further substantiate the associations of these five hub genes with ferroptosis in COPD patients.

3.9. Validation of NQO1, AKR1C3, CBR1 and GPX2 in a cellular model

To validate the authenticity and feasibility of the identified hub genes, we established an epithelial cell model and performed Western blot and qPCR analyses. Western blotting was used to assess the expression levels of CBR1, NQO1, AKR1C3, and GPX2 in the cells. Compared with those in the NC group, the protein expression levels of CBR1, NQO1, AKR1C3, and GPX2 in the COPD group were significantly greater (Fig. 11A-B). qPCR was used to measure the expression levels of CBR1, NQO1, AKR1C3, and GPX2 mRNAs in BEAS-2B cells from both groups. Compared with those in the NC group, the mRNA levels of these hub genes in the COPD group were significantly greater (Fig. 11C).

In this study, we employed a series of bioinformatics methods, combining Limma differential analysis and WGCNA, to analyze three airway epithelial cell microarray datasets from the GEO database (GSE10006, GSE11906, and GSE20257). We identified four ferroptosis-related hub genes (NQO1, AKR1C3, CBR1, and GPX2) that were consistently upregulated in the COPD group. Among these four ferroptosis-related genes, AKR1C3 and NQO1 are ferroptosis suppressor genes, while GPX2 is classified as a ferroptosis marker gene. CBR1 is categorized as an unclassified ferroptosis-related gene [29]. Importantly, AKR1C3 and NQO1 are both ROS detoxification enzymes. This finding was further confirmed through validation with the GSE11784 dataset and in vitro experiments (western blot and qPCR), which validated the high expression trend of these ferroptosis-related hub genes in COPD patients. To elucidate the role of these four ferroptosis-related DEGs in COPD, KEGG pathway and GO analyses were conducted.

Chronic obstructive pulmonary disease (COPD) is characterized by airway inflammation, irreversible airflow obstruction, and destruction of the lung parenchyma or emphysema [30]. COPD ranks as the third leading cause of morbidity and mortality worldwide [31]. The pathogenesis of COPD is incompletely understood but involves oxidative stress, protease/antiprotease imbalance, and an abnormal inflammatory response to harmful gases such as cigarette smoke (CS), which are major risk factors for this disease [32]. A study analyzing bronchoalveolar lavage fluid from rats exposed to cigarette smoke and COPD patients revealed that cigarette smoke particles may alter iron homeostasis, trigger metal accumulation, affect oxidative stress and inflammation, and contribute to lung damage in smokers with COPD [33]. Since ferroptosis was discovered in 2012, it has received widespread attention[34]. As ferroptosis progresses, intracellular iron stores decrease, leading to an increase in Fe²⁺ levels. Excess Fe²⁺ catalyzes the Fenton reaction, producing large amounts of ROS, hydroxyl radicals, and other oxidative species. These oxidative species may bind to polyunsaturated fatty acids, resulting in substantial lipid peroxidation. Iron chelators, such as the ferroptosis inhibitor Fer1, can reduce iron levels and lipid peroxidation, thereby inhibiting ferroptosis[35, 36].

The Limma R package offers a comprehensive solution for analyzing gene expression data[37]. WGCNA, a systems biology technique for analyzing gene association patterns between samples, was used to construct gene coexpression networks and identify related gene clusters, calculating the correlations between gene modules and phenotypes. We identified 1950 significant module genes and used Limma analysis and WGCNA to screen for connecting genes, shared genes, and related gene clusters associated with COPD and ferroptosis. The intersection of ferroptosis-related hub genes (n = 5) and COPD-associated module genes (n = 1950) revealed four common risk genes related to both COPD and ferroptosis. We subsequently performed GSEA to identify functional enrichments between COPD smokers and nonsmoking healthy individuals. Next, we generated ROC curves using data from COPD patients and healthy controls. The results indicate that the AUC values of these 4 hub genes (NQO1, AKR1C3, CBR1, and GPX2) in three gene sets (GSE10006, GSE11906, and GSE20257) are greater than 0.9 (AUC > 0.6 indicates diagnostic significance). These findings demonstrate the significant diagnostic value of NQO1, AKR1C3, CBR1, and GPX2 for COPD. We subsequently conducted immune infiltration analysis, and the results revealed that the expression of the 4 hub genes was positively correlated with immune cell infiltration. A type 2 microenvironment enriched with eosinophils in spatially restricted areas has been identified in COPD [38]. The proportion of T cells in the lungs of current smokers and COPD patients decreases, whereas the proportion of macrophages increases [39]. The responses of macrophages to various tissue microenvironments and external stimuli are plastic [40]. We found a significant increase in the proportion of M2 macrophages, which primarily participate in anti-inflammatory responses. In the COPD group, there was a notable increase in the number of M2 macrophages, suggesting their potential role in the pathogenesis of COPD. Finally, we utilized the Enrichr platform with the DsigDB database to predict target drugs for the selected hub genes (NQO1, AKR1C3, CBR1, GPX2). The top 10 drugs were identified on the basis of comprehensive scoring. Among them, acetaminophen and glycidamide are noted for their high research value and clinical significance in the treatment of COPD.

GPX2 is a major isoform of cigarette smoke-induced pulmonary glutathione peroxidase (GPX) and is regulated by nuclear factor erythroid 2-related factor 2 (Nrf2). Its primary function is to detoxify hydrogen peroxide or organic hydroperoxides, thereby protecting biological membranes and cellular components from oxidative stress[41]. Tian et al. demonstrated that knockdown of GPX2 partially reversed the increase in lipid ROS and iron levels induced by erastin, supporting the potential role of GPX2 as a driver of ferroptosis [42]. Moreover, NQO1 is a multifunctional antioxidant enzyme that plays a crucial role in protecting cells from oxidative damage through proteasomal degradation, exogenous detoxification, p53 regulation, superoxide scavenging, and maintenance of endogenous antioxidants [43]. CBR1 is a widely expressed NADPH-dependent enzyme belonging to the short-chain reductase/dehydrogenase family, which catalyzes the reduction of a broad range of exogenous and endogenous carbonyl compounds, and existing studies suggest that chrysin, a CBR1 inhibitor, can induce ferroptosis by targeting and inhibiting CBR1[44]. Additionally, AKR1C3 has been shown to participate in the detoxification of toxic lipid metabolites produced from the oxidation of various polyunsaturated fatty acids, such as 4-hydroxynonenal [45].

This study ultimately identified 4 hub genes associated with ferroptosis (NQO1, AKR1C3, GPX2, and CBR1), providing important insights for the diagnosis, treatment, and targeted drug screening of COPD. However, the study relied primarily on bioinformatics analysis, with validation through in vitro experiments. Data were sourced from the DEO database, which lacks sufficient COPD-related information, thereby presenting certain limitations. Future research could involve the construction of animal models for in vivo experimental validation and further exploration of the mechanisms of ferroptosis in COPD.

1. Ethics approval and consent to participate:

Not applicable.

2. Consent for publication:

Not applicable.

3. Availability of data and materials

The datasets generated and/or analysed during the current study are available in the [GEO] repository [(https://www.ncbi.nlm.nih.gov/geo/)] and [FerrDb] database [ (http://www.zhounan.org/ferrdb/)].

4. Competing interests

The authors declare that they have no competing interests.

5. Funding

This research was supported by Science and Technology Development Plan Project of Jilin Province(Grant No. 20200901007SF).

6. Authors' contributions

YQ and GCC did all the experimental parts. WF and WGQ did the plotting of Figures 1-5 and the tabulation, SWH and LHQ did the plotting of Figures 6-11, and YQ & QMR did the data analysis for this paper. YQ was the main contributor to writing this manuscript. All authors read and approved the final manuscript.

7. Ackowledgements

We acknowledge GEO database for providing their platforms and contributors for uploading their meaningful datasets.

Rabe KF, Watz H. Chronic obstructive pulmonary disease. Lancet 2017;389:1931-40.
Silvestri GA, Young RP. Strange Bedfellows: The Interaction between COPD and Lung Cancer in the Context of Lung Cancer Screening. Ann Am Thorac Soc2020;17:810-2.
Yuan C, Chang, Lu G, et al. Genetic polymorphismand chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 2017;12:1385-93.
Hikichi M, Mizumura K, Maruoka S, et al. Pathogenesisof chronic obstructive pulmonary disease (COPD) induced by cigarette smoke. J Thorac Dis 2019;11:S2129-40.
Csoma B, Bikov A, Nagy L, et al. Dysregulation of theendothelial nitric oxide pathway is associated with airway inflammation in COPD. Respir Res 2019;20:156.
Zheng M, Hong W, Gao M, et al. Long Noncoding RNA COPDA1 Promotes Airway Smooth Muscle Cell Proliferation in Chronic Obstructive Pulmonary Disease. Am J Respir Cell Mol Biol 2019;61:584-96.
Zong D, Li J, Cai S, et al. Notch1 regulates endothelial apoptosis via the ERK pathway in chronic obstructive pulmonary disease. Am J Physiol Cell Physiol 2018;315:C330-40.
Gouzi F, Blaquière M, Catteau M, et al. Oxidative stress regulates autophagy in cultured muscle cells of patients with chronic obstructive pulmonary disease. J Cell Physiol 2018;233:9629-39.
Yoshida M, Minagawa S, Araya J, et al. Involvement of cigarette smoke-induced epithelial cell ferroptosis in COPD pathogenesis. Nat Commun 2019;10:3145.
André S, Conde B, Fragoso E, et al. COPD and Cardiovascular Disease. Pulmonology. 2019;25(3):168–76. https://doi.org/10.1016/j.pulmoe.2018.09.006.
Boehm A, Pizzini A, Sonnweber T, et al. Assessing global COPD awareness with Google Trends. Eur Respir J. 2019;53(6):1900351. https://doi. org/10.1183/13993003.00351-2019. Published 2019 Jun 27.
Henrot P, Prevel R, Berger P, Dupin I. Chemokines in COPD: from implication to therapeutic use. Int J Mol Sci. 2019;20(11):2785. https://doi.org/10.3390/ijms20112785.
Halpin DMG, Celli BR, Criner GJ, et al. It is time for the world to take COPD seriously: a statement from the GOLD board of directors.Eur Respir J. 2019;54(1):1900914. https://doi.org/10.1183/13993003.00914-2019.
Sandelowsky H, Weinreich UM, Aarli BB, et al. COPD - do the right thing. BMC Fam Pract. 2021;22(1):244. https://doi.org/10.1186/s12875-021-01583-w.
X. Liu, Y. Ma, L. Luo, et al., Dihydroquercetin suppresses cigarette smoke induced ferroptosis in the pathogenesis of chronic obstructive pulmonary disease by activating Nrf2-mediated pathway, Feb, Phytomedicine 96 (2022), 153894,https://doi.org/10.1016/j.phymed.2021.153894.
Dixon SJ, Stockwell BR. The role of iron and reactive oxygen species in cell death. Nat Chem Biol. (2014) 10:9–17. doi: 10.1038/nchembio.1416
Kehrer JP. The Haber-Weiss reaction and mechanisms of toxicity. Toxicology.(2000) 149:43–50. doi: 10.1016/S0300-483X(00)00231-6
Friedmann Angeli, J. P. et al. Inactivation of the ferroptosis regulator Gpx4 triggers acute renal failure in mice. Nat. Cell Biol. 16, 1180–1191 (2014).
Imai, H. et al. Early embryonic lethality caused by targeted disruption of the mouse PHGPx gene. Biochem. Biophys. Res. Commun. 305, 278–286 (2003).
Ghio AJ, Hilborn ED, Stonehuerner JG, et al. Particulate matter in cigarette smoke alters iron homeostasis to produce a biological effect. Am
Berg S, Luneberg E, Frosch M. Development of an amplification and hybridization assay for the specific and sensitive detection of Mycoplasma
Davis S, Meltzer PS. GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor. Bioinformatics 2007;23:1846-7.
Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 2016;32:2847-9.
Guangchuang Yu, Li-Gen Wang, Yanyan Han, Qing-Yu He. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS: A Journal of Integrative Biology. 2012, 16(5):284-287.
Luo W, Brouwer C. Pathview: an R/Biocondutor package for pathway-based data integration and visualization. Bioinformatics, 2013, 29(14):1830-1831, doi: 10.1093/bioinformatics/btt285
Newman AM, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015 May;12(5):453-7
Z. Wang, K. Fang, G. Wang, X. Guan, Z. Pang, Y. Guo, Y. Yuan, N. Ran, Y. Liu, F. Wang, Protective effect of amygdalin on epithelial–mesenchymal transformation in experimental chronic obstructive pulmonary disease mice, Phytother. Res. 33 (3) (2019) 808–817
Tricia M. McKeever,Sarah Lewis,Henriëtte A. Smit, et al. The Association of Acetaminophen, Aspirin, and Ibuprofen with Respiratory Disease and Lung Function. American Journal of Respiratory and Critical Care Medicine. 2005;171 (9):966-971. doi:10.1164/rccm.200409-1269oc
Zhou N, Bao J. FerrDb: a manually curated resource for regulators and markers of ferroptosis and ferroptosis-disease associations. Database (Oxford) 2020;2020:baaa021.
Vestbo J, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med. 2013;187(4):347–365.
GBD Chronic Respiratory Disease Collaborators. Prevalence and attributable health burden of chronic respiratory diseases, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Respir Med. 2020;8(6):585–596.
Barnes PJ, et al. Cellular and molecular mechanisms of chronic obstructive pulmonary disease. Clin Chest Med. 2003;22(4):672–686.
Ghio AJ, Hilborn ED, Stonehuerner JG, et al. Particulate matter in cigarette smoke alters iron homeostasis to produce a biological effect. Am J Respir Crit Care Med. 2008;178(11):1130–1138. doi:10.1164/rccm.200802-334°C
Dixon, S. J. et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell 149, 1060–1072 (2012).
Zhou H, Li F, Niu JY, et al. Ferroptosis was involved in the oleic acid-induced acute lung injury in mice. Sheng Li Xue Bao. 2019;71(5):689–697. doi:10.13294/j.aps.2019.0070
Wen X, Wu J, Wang F, Liu B, Huang C, Wei Y. Deconvoluting the role of reactive oxygen species and autophagy in human diseases. Free Radic Biol Med. 2013;65:402–410. doi:10.1016/j.freeradbiomed.2013.07.013
M.E. Ritchie, B. Phipson, D. Wu, Y. Hu, C.W. Law, W. Shi, G.K. Smyth, Limma powers differential expression analyses for RNA-sequencing and microarray studies, Nucleic Acids Res. 43 (2015) e47.
Kim SH, Hong JH, Yang WK, et al. Herbal combinational medication of Glycyrrhiza glabra, Agastache rugosa containing glycyrrhizic acid, tilianin inhibits neutrophilic lung inflammation by affecting CXCL2, interleukin-17/STAT3 signaling pathways in a murine model of COPD. Nutrients. 2020;12(4):926. doi:10.3390/nu12040926
Ren Z, Hu M, Wang Z, et al. Ferroptosis-related genes in lung adenocarcinoma: prognostic signature and immune, drug resistance, mutation analysis. Front Genet. 2021;12:672904. doi:10.3389/fgene.2021.672904
Cruz T, López-Giraldo A, Noell G, et al. Multilevel immune response network in mild-moderate chronic obstructive pulmonary disease (COPD). Respir Res. 2019;20(1):152. doi:10.1186/s12931-019-1105-z
Singh A, Rangasamy T, Thimmulappa RK, et al. Glutathione peroxidase 2, the major cigarette smoke-inducible isoform of GPX in lungs, is regulated by Nrf2. Am J Respir Cell Mol Biol 2006;35:639-50.
Tian Q, Zhou Y, Zhu L, et al. Development and Validation of a Ferroptosis-Related Gene Signature for Overall Survival Prediction in Lung Adenocarcinoma. Front Cell Dev Biol 2021;9:684259.
Ling Zhou,Chen Yang,Weilan Zhong, et al. Chrysin induces autophagy-dependent ferroptosis to increase chemosensitivity to gemcitabine by targeting CBR1 in pancreatic cancer cells. Biochemical Pharmacology. 2021;193 (0):114813-114813. doi:10.1016/j.bcp.2021.114813
Zhu K, Li Y, Deng C, et al. Significant association of PKM2 and NQO1 proteins with poor prognosis in breast cancer. Pathol Res Pract 2020;216:153173.
Burczynski ME, Sridhar GR, Palackal NT, et al. The reactive oxygen species--and Michael acceptor-inducible human aldo-keto reductase AKR1C1 reduces thealpha, beta-unsaturated aldehyde 4-hydroxy-2-nonenal to 1,4-dihydroxy-2-nonene. J Biol Chem 2001;276:2890-7.

Tables 1 to 2 are available in the Supplementary Files section

No competing interests reported.

Tables.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification of Crucial Genes Associated with Ferroptosis in COPD via Comprehensive Bioinformatics Analysis and the Relevant Experimental Validation

Status:

Version 1

Abstract

Figures

1. Introduction

2. Data and methods

2.1 Microarray Data Collection and Processing

2.2 Construction of protein‒protein interaction (PPI) networks and analysis of DEGs

2.3. Identification of significant module genes via weighted gene coexpression network analysis (WGCNA)

2.4. Functional enrichment analysis

2.5. Hub Gene ROC Curve Construction

2.6. Immune cell infiltration

2.7. Targeted drug prediction

2.8. Validation of dataset GSE11784

2.9. Cell model construction and experimental validation

3. Results

3.1 Identification of Ferroptosis Key Genes in COPD

3.2. Results of protein‒protein interaction network construction

3.3. Significant module gene identification in COPD via WGCNA

3.4. GO and KEGG enrichment analyses of the differentially expressed FRHGs

3.5. ROC curve of the hub genes

3.6. Immune infiltration analysis

3.7. Potential therapeutic drug prediction

3.8. GSE11784 confirms the expression and diagnostic value of FRHGs

3.9. Validation of NQO1, AKR1C3, CBR1 and GPX2 in a cellular model

4. Discussion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1