Placental transcriptomics identifies a candidate HLA-DQA2-FGL2 immune signature in gestational diabetes mellitus

Gestational diabetes mellitus (GDM) is a major public health challenge characterized by placental immunometabolic dysregulation. This study aimed to identify molecular signatures associated with GDM-related placental pathology and evaluate their diagnostic and therapeutic implications.

Two bulk placental transcriptomic datasets (GSE70493 and GSE263483) and one single-cell dataset (GSE173193) were analyzed. Differential expression analysis, WGCNA, and machine-learning feature selection (LASSO and SVM-RFE) were used to identify candidate gene signatures. A two-gene logistic regression model was constructed and externally validated. Immune-cell correlation rewiring and single-cell in silicoperturbation analysis explored immune-network alterations and regulatory roles of identified genes. A drug-prioritization framework was used for therapeutic screening.

Intersecting WGCNA modules with differentially expressed genes identified 29 candidates enriched in antigen presentation and mononuclear phagocyte functions. Dual-algorithm selection converged on HLA-DQA2 and FGL2. The two-gene model achieved an AUC of 0.786 in the discovery cohort and 0.813 in the external validation cohort. Immune rewiring analysis indicated a shift toward M1 macrophage polarization. Single-cell perturbation analysis supported regulatory roles of HLA-DQA2 and FGL2 within the macrophage lineage. Drug analysis reaffirmed insulin, metformin, and glyburide as stable therapeutic anchors.

The HLA-DQA2/FGL2 signature captures key aspects of placental immunometabolic dysregulation in GDM, including altered antigen presentation and macrophage-associated immune rewiring. This integrated framework provides candidate biomarkers for GDM risk stratification and generates computational hypotheses for future mechanistic and therapeutic studies.