Molecular Signatures of Preeclampsia Subtypes Determined Through Integrated Weighted Gene Co-expression Network Analysis and Differential Gene Expression Analysis of Placental Transcriptomics

Luhao Han¹Olivia Holland^1,2*Fabricio Da Silva Costa^3,4Anthony Perkins^1,5

• ¹School of Pharmacy and Medical Sciences, Griffith University, Gold Coast, Queensland, Australia
• ²Women-Newborn-Children-Services, Gold Coast University Hospital, Gold Coast, Queensland, Australia
• ³School of Medicine and Dentistry, Griffith University, Gold Coast, Queensland, Australia
• ⁴Maternal Fetal Medicine Unit, Women-Newborn-Children Services, Gold Coast University Hospital, Gold Coast, Queensland, Australia
• ⁵School of Health, University of the Sunshine Coast, Sunshine Coast, Queensland, Australia

Background: Preeclampsia (PE) is a multisystemic pregnancy syndrome that presents in different clinical subtypes. While placental dysfunction is a critical feature of PE, its contribution to different PE subtypes remains unclear. This study aims to use integrated bioinformatics analysis of placental transcriptomics to investigate subtype-specific molecular mechanisms associated with PE.Methods: A systematic search of the Gene Expression Omnibus (GEO) repository identified two datasets (GSE234729, n=123; GSE75010, n=157) for integrated Weighted Gene Co-expression Network Analysis (WGCNA) and differential gene expression analysis. We constructed co-expression networks and identified gene modules correlated with three PE subtypes (severe, early-onset and late-onset). Differential gene expression analysis was conducted using the "limma" R package. Differentially expressed genes (DEGs) overlapping with PE subtype-correlated WGCNA modules underwent Gene Ontology (GO) enrichment analysis. Consistently dysregulated genes were validated in an additional external dataset (GSE25906) and RT-PCR analysis of placental samples from 21 PE cases and 21 uncomplicated controls.Results: We identified distinct molecular signatures associated with each PE subtype. The green gene module was positively correlated with severe PE (r = 0.63, p = 4e-15), containing 179 DEGs primarily involved in lipid metabolism and hypoxia response processes. Early-onset PE had two highly significant gene modules: the yellow module (r = 0.73, p = 4e-15) with 112 DEGs enriched in biological processes related to gonadotrophin secretion and lipid storage, and the black module (r = -0.55, p = 5e-08) with 47 DEGs significantly enriched in chronic inflammation responses. Late-onset PE showed moderate correlation with the ivory module (r = 0.46, p = 5e-05), containing 23 DEGs enriched in p38MAPK stress-response signalling. Cross-subtype analysis identified 20 consistently dysregulated genes across three PE subtypes, with four upregulated genes (LEP, FSTL3, HTRA4, and HK2) confirmed in the external dataset GSE25906. However, RT-PCR validation showed only moderate upregulation without statistical significance.Conclusion: Though placental dysfunction occurs across all subtypes with a core set of upregulated genes, variation exits in placental gene expression patterns among PE subtypes. Severe and early-onset PE exhibit large molecular perturbations, while late-onset PE presents more subtle alterations. Aberrant placental lipid storage may contribute to disease severity and early manifestation.