Endoplasmic reticulum stress and unfolded protein response play roles in recurrent pregnancy loss: A bioinformatics study

Abstract

This study aims to explore whether endoplasmic reticulum stress (ERS) and unfolded protein response (UPR) processes could be potential targets for preventive, diagnostic, and therapeutic for recurrent pregnancy loss (RPL). RPL datasets GSE165004 and GSE26787 were sourced from the GEO database, and ERS- and UPR-related gene sets were obtained from the MsigDB database. After differentially expressed genes (DEGs) identification, key genes were screened from intersecting DEGs in RPL-ERS and RPL-UPR datasets. The z-score algorithm was conducted to obtain phenotype scores. Functional enrichment and machine learning analyses were performed to assess gene function and diagnostic value evaluation. Interaction networks were conducted to investigate upstream regulated relationships of the key genes. Immune infiltration and single-cell RNA sequencing (scRNA-seq) were assessed to explore ERS and UPR functions at the cellular level. Totally 25 key genes RPL-ERS DEGs and 16 key genes RPL-UPR DEGs were identified. Among them, six key genes (NFYB, EXOSC2, UBQLN2, RNF139, DERL1, and FBXO27) were validated to show consistent expression trends in both RPL datasets. Functional enrichment highlighted their involvement in the immunity of RPL. Machine learning indicated the significant diagnostic value of these validated genes for RPL, with an accuracy rate of > 80 %. scRNA-seq analysis revealed elevated ERS and UPR expressions in monocytes/macrophages in RPL samples. In conclusion, ERS and UPR processes are associated with RPL occurrences, and were mainly upregulated in monocytes/macrophages within RPL samples. ERS and UPR processes may serve as potential targets for the prevention, diagnosis, and treatment of RPL.

Keywords: Bioinformatics; Differentially expressed genes; Endoplasmic reticulum stress; Recurrent pregnancy loss; Unfolded protein response.