Integrated bioinformatics and machine learning analysis identify CROT as a regulator of immunological features in idiopathic pulmonary fibrosis

Mitochondrial dysfunction and immune dysregulation contribute to the pathogenesis of idiopathic pulmonary fibrosis (IPF). This study systematically identifies mitochondrial-related core regulatory genes and elucidates their potential associations with immunological features in IPF. Two independent IPF cohorts (GSE110147 and GSE32537) were integrated as a training set, followed by a differential expression analysis to identify IPF-specific gene signatures. Weighted gene co-expression network analysis (WGCNA) was used to detect disease-associated modules, which were cross-screened against mitochondrial gene databases. A multi-model framework incorporating Least Absolute Shrinkage and Selection Operator (LASSO) regression, Random Forest (RF), and Support Vector Machine Recursive Feature Elimination (SVM-RFE) identified core candidate genes. An independent dataset (GSE24206) was used for external validation, with diagnostic efficacy evaluated using receiver operating characteristic (ROC) curve analysis. Expression patterns of key genes were validated in bleomycin-induced BEAS-2B cells. Immune cell infiltration was quantified using the CIBERSORT deconvolution algorithm. Our results showed WGCNA, LASSO regression, RF, and SVM-RFE algorithms identified three key genes, ABCD2, CROT, and CYP24A1, which were significantly upregulated in the lung tissues of patients with IPF. Their ROC curves demonstrated excellent diagnostic performance. These findings were confirmed by the results from bleomycin-induced BEAS-2B cells. Functional experiments showed that CROT silencing reduced α-smooth muscle actin expression, increased E-cadherin levels in bleomycin-induced BEAS-2B cells. Our results indicate that CROT is associated with EMT and immune-cell alterations in IPF. Thus, CROT may serve as a potential therapeutic target for regulating the immune response disorder in IPF.

CROT; Idiopathic pulmonary fibrosis; Immune microenvironment; Machine learning; Mitochondria.