Abstract
OBJECTIVE: Recent studies suggest a correlation between endoplasmic reticulum (ER) stress and gestational diabetes mellitus (GDM). Nevertheless, the potential causal role of ER stress-related genes remains largely unexplored. This study aims to generate hypotheses regarding these connections through an integrative analysis of multiomics data. METHODS: We utilized an exploratory list of genes related to ER stress and integrated quantitative trait loci (QTL) for gene expression (eQTLs), DNA methylation (mQTLs), and protein levels (pQTLs). Genome-wide association studies (GWAS) summary statistics for GDM were obtained from the publicly accessible FinnGen database, with replication attempted using data from the GWAS Catalog. The summary-data-based Mendelian Randomization (SMR) method was applied to explore the genetic ties between these genes and GDM, followed by colocalization analysis to pinpoint overlapping causal genetic variants. Placental endothelial transcriptome data (GSE103552) were used for validation. RESULTS: The SMR and colocalization identified potential causal links for 27 mQTLs, 8 eQTLs, and 4 pQTLs with GDM risk. Integration of evidence across mQTL and eQTL levels suggested potential causal roles for the NUP133, VHL, TAPBP, and GPX1 genes in GDM. Notably, NUP133 shows suggestive colocalization evidence at the eQTL level. Analysis relating methylation to expression suggested hypermethylation at the CpG site cg17439967 may upregulate NUP133, potentially associating with reduced GDM risk. Transcriptomic validation in placental endothelial cells further showed differential expression of these four genes between GDM and controls. CONCLUSION: Our findings provide suggestive genetic evidence linking specific ER stress-related genes, particularly NUP133, with GDM risk, highlighting potential pathways that warrant further investigation.