KMT2D loss rewires estrogen-PI3K crosstalk in endometrioid adenocarcinoma

BACKGROUND: Endometrioid adenocarcinoma is defined by pervasive PI3K-pathway dysregulation and estrogen-driven growth. More than 80% of tumors harbor PI3K-pathway alterations, and approximately 20% exhibit loss-of-function mutations in the histone methyltransferase KMT2D, often at high variant allele frequencies, suggesting biological relevance rather than mutational noise. While estrogen enhances PI3K signaling through genomic and non-genomic mechanisms, the epigenetic factors governing this interaction remain poorly understood in endometrial cancer. In ER+ breast cancer, KMT2D governs regulatory crosstalk between these two pathways, raising the possibility that loss-of-function mutations in endometrial cancer alter estrogen-responsive transcription and shift oncogenic dependencies. METHODS: KMT2D was depleted via siRNA in 12Z-ESR1 endometrial epithelial cells, followed by estradiol or vehicle treatment (n=3/group). Bulk RNA-sequencing assessed transcriptional changes. To evaluate in vivo function, a conditional Kmt2d-SET-fl/fl mouse model, yielding a catalytically impaired protein, was analyzed by uterine RNA-seq. Clinical relevance was assessed using an integrated dataset of whole-exome sequencing and transcriptomics from endometrioid adenocarcinomas. Tumors were stratified by KMT2D loss-of-function status for differential expression and pathway enrichment analyses. RESULTS: KMT2D knockdown markedly altered estradiol-driven transcription, shifting ER-related signaling pathways that remained intact in control cells. In Kmt2d-SET-deficient mouse uteri, PI3K-Akt pathway genes were significantly upregulated even without estrogen stimulation or neoplastic change, demonstrating that KMT2D enzymatic activity constrains baseline PI3K signaling in vivo. In patient tumors, KMT2D-mutant cancers showed enrichment of oncogenic pathways consistent with convergence of estrogen and PI3K programs. Across models, KMT2D loss amplified PI3K pathway activity and altered estrogen receptor-directed transcription, suggesting increased reliance on PI3K signaling. CONCLUSION: This supports a model in which KMT2D loss disrupts ER-regulated transcriptional programs, enhances PI3K pathway activation, and creates a potentially targetable dependency in endometrioid adenocarcinoma. Given the mixed clinical efficacy and toxicity of PI3K inhibitors, identifying molecular contexts that heighten PI3K dependence is critical. This work highlights KMT2D loss as a potential biomarker for PI3K-targeted therapy and provides new insight into epigenetic regulation of hormone-oncogenic crosstalk.