Integrating single-cell analysis and machine learning to identify COPD hub genes and explore the intervention mechanism of effective component compatibility of Bufei Yishen formula III

Abstract

Ethnopharmacological relevance

Chronic obstructive pulmonary disease (COPD) is a prevalent condition that poses a major threat to public health. Endothelial cells play a critical role in COPD pathogenesis. Recent evidence suggests that the effective-component combination of Bufei Yishen formula III (ECC-BYF III) significantly improves clinical symptoms and quality of life in COPD patients.

Aim of the study

To identify endothelial cell-associated hub genes in COPD using single-cell analysis and machine learning, and to elucidate the intervention mechanism underlying ECC-BYF III.

Materials and methods

Single-cell analysis was used to identify altered cellular subtypes in COPD samples. High-dimensional weighted gene co-expression network analysis (hdWGCNA) and multiple machine learning algorithms were applied to identify COPD-related hub genes. These genes were validated using receiver operating characteristic (ROC) curves, independent datasets, qRT-PCR in human umbilical vein endothelial cells (HUVECs) and a rat model of COPD.

Results

Single-cell analysis revealed nine distinct cell subtypes, with endothelial cells markedly reduced in COPD samples compared to controls. Cell communication and pseudotime trajectory analysis highlighted the role and developmental trajectory of endothelial cells in COPD. Differential expression analysis and hdWGCNA identified 269 endothelial cell–associated genes, from which six hub genes were selected via machine learning. qRT-PCR confirmed that CD74, AQP1, SOX4, and ANXA1 were significantly dysregulated in both HUVECs and COPD rat models, consistent with the data analysis results. Notably, ECC-BYF III intervention reversed these gene expression abnormalities. Molecular docking demonstrated that the components of ECC-BYF III exhibited strong binding affinities for the hub genes.

Conclusions

Four hub genes (CD74, SOX4, AQP1, and ANXA1) involved in the pathogenesis of endothelial cells in COPD were identified. ECC-BYF III was shown to modulate their expression, supporting its potential as a therapeutic agent in traditional Chinese medicine (TCM) for COPD. These findings offer novel insights into the mechanisms of COPD and open avenues for TCM treatment.