Plasticizer-related toxicogenomic signatures in gestational diabetes: Mendelian randomization, placental single-cell transcriptomics and trophoblast validation.

Plasticizer exposure has been associated with gestational diabetes mellitus (GDM), but the placental toxicogenomic mechanisms that may connect environmental chemical signals to maternal dysglycaemia remain unresolved. We integrated plastic-associated compound target annotation, placental bulk transcriptomics, cis-eQTL-based Mendelian randomization (MR), Bayesian colocalization, machine-learning prioritization, pathway analysis, placental single-cell transcriptomics and trophoblast validation. Candidate targets were intersected with GDM-related placental expression signals and tested against FinnGen GDM summary statistics using blood-derived cis-eQTL instruments. Among expression-supported targets, 15 genes showed nominal MR associations with GDM and 11 loci showed colocalization support. LASSO regression and random forest prioritization converged on AKT1, CLEC7A and CTSS. Expression-direction checks showed concordance between bulk placental expression and MR direction for AKT1 and CTSS, whereas CLEC7A showed a discordant genetic direction and was therefore interpreted cautiously. Functional analyses placed these genes in immune, redox and metabolic pathways, while single-cell data localized them to trophoblast and immune-related placental compartments. In HTR-8/SVneo trophoblast cells, di-(2-ethylhexyl) phthalate (DEHP) exposure altered CLEC7A expression, providing targeted evidence that CLEC7A is phthalate-responsive at the mRNA level under the tested conditions. The two-dose design did not support a formal concentration-response conclusion. The findings define a toxicogenomic hypothesis in which plasticizer-related stress intersects with placental immunometabolic regulation in GDM. The study does not establish direct exposure causality, but it prioritizes testable placental targets for exposure-resolved toxicological and epidemiological validation.