Bioinformatics-based identification of ferroptosis biomarkers and a diagnostic model with in vivo validation in chronic obstructive pulmonary disease

Chronic obstructive pulmonary disease (COPD) is difficult to recognize early, and its oxidative stress, iron metabolism disorder, and lipid peroxidation are closely associated with ferroptosis. Screening stable ferroptosis-related biomarkers may help elucidate the molecular mechanisms of COPD and improve auxiliary diagnosis. This study aimed to screen key ferroptosis biomarkers in COPD based on bioinformatics, machine learning, and single-cell transcriptomic analyses, construct a combined diagnostic model, and validate their expression in cigarette smoke-exposed mice.

Two lung tissue bulk transcriptomic datasets, GSE178513 (3 normal, 6 COPD) and GSE106986 (5 normal, 14 COPD), and a single-cell transcriptomic dataset GSE279570 (7 normal, 11 COPD) were obtained from the Gene Expression Omnibus (GEO). Combined with 484 ferroptosis-related genes from the FerrDb database, differential analysis, functional enrichment analysis, and weighted gene co-expression network analysis were performed to screen candidate genes. Seurat, Monocle2, and CellChat were used to analyze cell clustering, pseudotime trajectories, and cell communication among 64620 high-quality cells. The least absolute shrinkage and selection operator (LASSO) regression was used to select feature genes, and the combined model was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis. Male C57BL/6 mice aged 6 to 8 weeks were randomly divided into control and COPD groups (n=8 per group) using a random number table. The COPD group was exposed to cigarette smoke for 12 consecutive weeks. Lung function testing, hematoxylin-eosin staining, RT-qPCR, Western blotting, and immunohistochemistry were used to assess lung function, lung histopathology, and key gene expression.

Five ferroptosis-related key genes were identified: SPATA2, CEMIP, FABP4, DPP4, and IL6. Their expression in COPD lung tissues was significantly higher than in normal controls (P all<0.05). Single-cell analysis revealed that IL6 and FABP4 increased along the pseudotime trajectory from monocytes to type 2 conventional dendritic cells, and enhanced communication between macrophages and Club cells was observed in the COPD group (P<0.01). The area under the curve of the 5-gene combined model was 0.921. Compared with the control group, mice in the COPD group showed decreased forced expiratory volume in 0.3 s (FEV0.3) (P=0.0314), increased forced vital capacity (FVC) (P=0.0287), a reduced FEV0.3/FVC ratio (P=0.0048), and emphysema-like pathological changes. The mRNA and protein expression of all five genes was increased (P<0.05).

SPATA2, CEMIP, FABP4, DPP4, and IL6 are candidate biomarkers associated with ferroptosis, inflammation, and abnormal lipid metabolism in COPD. The 5-gene combined model showed good discriminatory performance in the training data, and the abnormally high expression was validated in vivo in mice, providing clues for mechanistic studies and the screening of auxiliary diagnostic biomarkers for COPD.