Single-Nucleus Transcriptomics Reveals Granulosa Cell Heterogeneity and Microenvironmental Remodeling Across Bovine Ovarian States

Ovarian function is essential for fertility in dairy cattle, yet the cellular and molecular features associated with physiological ovarian states and ovarian dysfunction remain incompletely characterized. In this study, serum and follicular-fluid hormone measurements showed distinct endocrine profiles among ovarian states, and the follicular-fluid estrogen to progesterone was highest during the follicular phase and lowest in luteal and luteal cystic ovaries. Then, single-nucleus RNA sequencing was performed on ovarian tissues from 18 Holstein cows representing follicular, luteal, mid-gestation pregnancy, inactive, and luteal cystic states. After quality control, 154,054 nuclei were retained for cell-type annotation, granulosa cell (GC) subclustering, trajectory inference, co-expression analysis, ligand-receptor and ligand-target prediction, and transcriptome-based metabolic flux estimation. Twenty-seven ovarian cell clusters and five GC subtypes were identified, with state-associated variation in relative nuclear composition and transcriptional profiles. Luteal cystic ovaries showed a higher relative representation of immune cells and enrichment of inflammation-related transcriptional signatures, whereas inactive ovaries exhibited lower levels of predicted intercellular communication. GC analyses indicated differences among ovarian states in transcriptional programs related to proliferation, steroidogenesis, extracellular-matrix organization, inflammation, and metabolism. Transcriptome-based metabolic inference further suggested subtype-associated variation in tricarboxylic acid cycle, lipid, polyamine, phosphoinositide, and gamma-aminobutyric acid-related pathways. This study provides a multi-state single-nucleus transcriptomic resource for bovine ovarian research and identifies candidate cell populations and molecular features for future experimental validation.