Exploration of mitochondrial-related biomarkers and regulatory mechanisms in pulmonary hypertension

Multiple mechanisms of mitochondrial dysfunction have been implicated in the pathogenesis of pulmonary hypertension (PH). Thus, this study aims to identify biomarkers associated with mitochondrial-related genes (MRGs) in patients with PH. The GSE33463 (blood training set), GSE15197 (tissue training set), GSE113439 (external validation set) and MRGs were all acquired from public databases. Differential expression analysis was undertaken to selected differentially expressed genes (DEGs) from blood and tissue training set, respectively. Then, the DEGs were intersected with MRGs to identify potential candidate genes, followed by detecting their functions via enrichment analysis. Moreover, crucial genes were obtained by overlapping the signature genes from least absolute shrinkage and selection operator (LASSO) regression and support vector machine recursive feature elimination (SVM-RFE). Based on the receiver operating characteristic (ROC) curves of crucial genes, the genes with area under curves (AUC) value ≥ 0.7 were deemed as biomarkers. Gene Set Enrichment Analysis (GSEA), multiple regulatory networks and drug prediction were applied to explore biomarkers’ molecular mechanism. A total of 144 DEGs (78 up-DEGs and 66 down-DEGs) were obtained by intersecting 751 DEGs1 in blood training set and 6,155 DEGs2 in tissue training set. Next, taking the intersection of 144 DEGs and 1,136 MRGs, 7 candidate genes (ALAS2, CPT2, TST, SLC25A39, COQ10B, MRPS30 and MCL1) were acquired and enriched in multiple mitochondria related pathways, like mitochondrial inner membrane. MRPS30, ALAS2 and TST were recognized as biomarkers depending on the 4 crucial genes (CPT2, MRPS30, ALAS2 and TST) through overlapping signature genes by two machine learning algorithms. The immune analysis showed that the percentage of 7 kinds of immune cells (e.g. B cells naive, monocytes, neutrophils) were differed markedly at control and PH groups, and monocytes were related to all three biomarkers. In 3 datasets, the expression of ALAS2 was up-regulated in PH, while MRPS30 was down-regulated in PH. And bisphenol A was predicted by ALAS2, MRPS30 and TST. Three biomarkers (ALAS2, MRPS30 and TST) associated with MRGs was identified and verified, which provided a new perspective to probe the mechanism of MRGs in PH.