AUTHOR=He Jianbo , Chen Hong , Duan Kaiming , Wushouer Sikandaier , Wang Xiaowei , Li Yaxuan , Qin Xingang 

TITLE=Gene signatures associated with exosomes as diagnostic markers of postpartum depression and their role in immune infiltration

JOURNAL=Frontiers in Endocrinology

VOLUME=Volume 16 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2025.1542327

DOI=10.3389/fendo.2025.1542327

ISSN=1664-2392

ABSTRACT=BackgroundPostpartum depression (PPD) is a significant mental health challenge for new mothers, with diverse and unclear causes. Exosomes significantly contribute to the pathogenesis, identification, treatment outcome determination, and intervention of PPD. However, the functions of exosome-related genes (ERGs) in PPD remain to be fully elucidated. This study examines the potential impact of ERGs on PPD and develops a set of diagnostic tools based on them.MethodsWe acquired and prepared several gene expression datasets from the Gene Expression Omnibus (GEO). Our analysis focused on genes that closely interact with the extracellular matrix. Using advanced techniques, including the limma package, we identified differential expression and conducted enrichment analyses of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). Furthermore, we employed logistic regression, random forest (RF) classifiers, and least absolute shrinkage and selection operator (LASSO) regression to screen critical genes.ResultsWe identified 44 exosome-related differentially expressed genes (ERDEGs) that play key roles in synaptic signal transmission, hormone fluctuations, and inflammatory responses. Ten genes, including TPP2, AKR1B1, CD59, PARK7, PLXNB2, HLA-B, FAH, NDST1, SCARB1, and HNRNPA2B1, were established using logistic regression analysis, RF method, and LASSO regression. In these two sets of data, the manifestations of PARK7 and HNRNPA2B1 differed. The analysis showed that the significant enrichment of gene sets was strongly associated with high-risk scores, particularly in the metabolic (phospholipid metabolism) and neural (mitochondrial translation) pathways. Gene set variation analysis (GSVA) revealed four prominent pathways: MYC targets V2, pancreatic beta cells, unfolded protein response, and oxidative phosphorylation. Single-sample gene set enrichment analysis (GSEA) showed that immune cells demonstrated different degrees of infiltration among at-risk and low-probability risk subsets of immature B cells, regulatory T cells), and T follicular helper cells.ConclusionsERDEGs significantly contribute to PPD occurrence. Our diagnostic model demonstrated high accuracy and potential for use in medical practice. Future research with larger samples is warranted to validate these conclusions and identify effective targets that may affect these pathways during treatment to improve the therapeutic effect.