Placental PRDX2-VEGFA/ENG network rewiring links oxidative stress to angiogenic failure in preeclampsia: A multi-cohort transcriptomic study.

BACKGROUND: Preeclampsia (PE) affects 2-8% of pregnancies and causes > 76,000 maternal deaths annually. The mechanistic link between placental oxidative stress and angiogenic failure remains poorly defined, and no validated multi-gene biomarker panel exists for early detection. In particular, how antioxidant enzymes such as PRDX2 interact with angiogenic regulators (VEGFA, ENG) to drive network rewiring in the PE placenta has not been characterized across independent cohorts. This study aimed to characterize the transcriptomic interdependencies between oxidative stress defense, cytokine signaling, and angiogenic regulation across five independent PE placental GEO cohorts. METHODS: Five GEO datasets (GSE44711, GSE10588, GSE25906, GSE74341, GSE75010) comprising 143 PE cases and 153 controls were analyzed. A 17-gene panel (oxidative stress: CAT, NFE2L2, PRDX1, PRDX2, GSR, GCLM, NOX4, EPAS1; cytokine: IFNG, CCL5, STAT1, NFKB1, IL4, IL13; angiogenic: VEGFA, ENG, CDX2) underwent differential expression, pathway enrichment, Spearman correlation, ROC with 5-fold cross-validation, and co-expression network analysis. RESULTS: We identified 801 significant differentially expressed probes (|log2FC=>1, padj<0.05). Angiogenesis pathway was strongly downregulated (mean enrichment score -55.99, p = 0.022) while JAK-STAT signaling was activated (+179.3). Of 64 oxidative stress-cytokine gene pairs, 28 were significantly correlated (p < 0.05). The strongest correlations were PRDX2-ENG (r = -0.777), PRDX2-VEGFA (r = -0.697), GSR-STAT1 (r = -0.656), and CAT-STAT1 (r = -0.598). PE-specific network rewiring was evident. Single-gene ROC identified IFNG as the best individual biomarker (AUC=0.677). A 17-gene logistic regression panel achieved AUC= 0.781 (CV-AUC=0.729 ± 0.145) with specificity 82.2%. Co-expression analysis revealed CAT and IL13 as top hub genes (degree 11 each) and a four-gene module (PRDX2-IL13-VEGFA-ENG) as the most PE-relevant. CONCLUSIONS: This multi-cohort transcriptomic study uncovers a novel PRDX2-driven oxidative stress-angiogenic axis (PRDX2-ENG, r = -0.777; PRDX2-VEGFA, r = -0.697). The 17-gene panel (AUC=0.781, 82.2% specificity) provides a promising foundation for biomarker development, warranting prospective validation in high-risk pregnancies.