Machine Learning-Based Screening and Experimental Validation of PIK3C2A as a Candidate Angiogenesis-Related Molecule in Pulmonary Hypertension

Pulmonary hypertension (PH) is a progressive cardiopulmonary disease characterized by pulmonary vascular remodeling and dysregulated angiogenesis. However, the angiogenesis-related molecular mechanisms underlying PH remain incompletely understood, and reliable diagnostic biomarkers are still lacking. Therefore, To identify angiogenesis-related key molecules in PH through integrated transcriptomic analyses and experimental validation, we aimed to establish a candidate molecular signature and evaluate the discriminative potential of PIK3C2A. Specifically, GSE15197 was used as the primary bulk-transcriptomic discovery dataset, whereas GSE1909 for cross-species validation, and human pulmonary artery single-cell RNA sequencing data (GSE210248) for cellular characterization. Moreover, Functional enrichment, Weighted Gene Co-expression Network Analysis (WGCNA), immune infiltration analysis, and three machine learning algorithms—logistic regression, random forest (RF), and Least Absolute Shrinkage and Selection Operator (LASSO) were performed to identify model genes. In addition, Single-cell analyses were performed using GSE210248, followed by quantitative real-time polymerase chain reaction (qRT-PCR) validation of candidate-gene expression in monocrotaline (MCT) and Sugen5416 plus hypoxia (SuHx) rat models. As a result, Forty-three angiogenesis-related differentially expressed genes (ARDEGs) were identified and were enriched in terms and pathways related to angiogenesis, vascular development, extracellular matrix organization, and HIF-1 signaling. Four model genes (DCN, DKK1, PIK3C2A, and UNC5B) were identified and showed moderate-to-high discrimination between PH and control samples in GSE15197. Among these genes, PIK3C2A showed its highest expression in endothelial cells, while complementary immune-infiltration and CellChat (version 2.1.1) analyses characterized the broader immune–stromal context of the angiogenesis-related signature. qRT-PCR further confirmed significant downregulation of PIK3C2A in the lungs and pulmonary arteries of MCTand SuHxinduced PH models (p < 0.05). In conclusion, An angiogenesis-related candidate molecular signature comprising DCN, DKK1, PIK3C2A and UNC5B was identified in PH. Among these genes, PIK3C2A showed preferential endothelial expression and was downregulated in both the MCT and SuHx rat PH models, supporting its further evaluation as a candidate molecule associated with PH.