Abstract
Preeclampsia (PE), a devastating pregnancy complication affecting 5% of gravidas worldwide, exhibits poorly characterized connections between mitochondrial dysfunction and immune dysregulation. This study aims to identify integrated mitochondrial-immune biomarkers for preeclampsia by multi-omics analysis of severe PE cohorts, enabling mechanistic insights and diagnostic potential. We developed a novel computational framework integrating multi-omics analysis (GSE10588 transcriptomics), machine learning (LASSO-SVM algorithm), and molecular dynamics simulation. This pipeline systematically identified mitochondrial-immune hubs from 1,589 PE-related DEGs and 188 mitochondrial regulators, followed by in silico validation of therapeutic targets. Fifteen candidate genes were identified by intersecting 1,589 DEGs and 188 MRGs. PPIs revealed associations with the citrate cycle (TCA cycle). Machine learning prioritized ACO1 and OGDH as biomarkers, showing opposite expression trends (ACO1: higher in controls; OGDH: elevated in PE). Both localized to chromosomes 9 and 7, respectively, with nuclear predominance. Enrichment linked them to cytokine-cytokine receptor and neuroactive ligand-receptor pathways. Immune infiltration highlighted correlations with activated NK cells and CD8+ T cells. Regulatory networks implicated lncRNAs (e.g., KCNQ1OT1), miRNAs (e.g., hsa-miR-214-3p), and TFs (e.g., TFAP2A). Drug predictions and docking identified devimistat and acetylcysteine as potential binders. This study establishes ACO1 and OGDH as context-specific mitochondrial-immune coordinators in preeclampsia. We propose a computationally derived dual-target therapeutic strategy, wherein ACO1 agonism aims to restore metabolic homeostasis while OGDH inhibition targets pathological overactivation. These findings, originating from in silico analyses, require preclinical validation through experimental models. Additionally, the mitochondrial-immune scoring system serves as a candidate tool for PE subtyping.