Integration of Spatial Transcriptomics and Mendelian Randomization Identifies Candidate Molecular Regulators of...

Endometrial cancer (EC) is a common gynecologic malignancy arising from the epithelial cells of the endometrium. The marked cellular heterogeneity of EC and features of its tumor immune microenvironment (TIME) contribute to disease complexity and have been associated with poor prognosis. This study integrates single-cell RNA sequencing, spatial transcriptomics, and Mendelian randomization (MR) to identify candidate genes associated with EC. Single-cell analysis identified MM0 as a putative stemness-associated transcriptional subpopulation with the highest CytoTRACE-inferred score; irGSEA indicated enrichment of proliferation- and stress-response pathways. Spatial data were analyzed with RCTD, MISTy, and stLearn to estimate spatial associations and pathway activities. MR and colocalization analyses integrating eQTL data and EC GWAS summary statistics prioritized DNAJA4, HSPA6, and LMNA as candidate genes with potential causal associations with EC risk. The expression patterns of these candidate genes were further examined in patient-derived samples. In Ishikawa cells cultured under high-estrogen conditions, siRNA-mediated knockdown of DNAJA4 and HSPA6 significantly suppressed proliferation and migration, accompanied by reduced CDK1 and Cyclin B expression. Collectively, these findings provide insight into EC heterogeneity and support further mechanistic investigation of candidate genes associated with malignant epithelial proliferation.