Single-cell exploration of ovarian aging across vertebrate models

Abstract

Mammalian female reproductive span is thought to be limited by a fixed "ovarian reserve" determined at birth. With age, a dwindling ovarian reserve leads to infertility, culminating in menopause in humans. In addition to infertility, accumulating evidence has shown that age-related ovarian functional decline contributes to multisystem aging and frailty, making post-menopausal women most susceptible to an array of chronic diseases. However, due to limited tissue accessibility and lack of reliable research models, molecular drivers of ovarian aging remain poorly understood. A key barrier in the field has been the limited establishment and benchmarking of preclinical models faithfully recapitulating human ovarian biology. To address this, we curated publicly available single-cell/nucleus ovarian RNA-seq datasets from human, macaque, mouse, and goat, and processed them using a consistent and stringent pipeline. Datasets were then annotated in a harmonized fashion across studies in order to conduct a robust, integrative, cross-species analysis of ovarian aging with single cell resolution. We systematically evaluated cell-type composition, global transcriptional perturbations, gene-level changes, pathway and network features, and drug-response alignments. Across analyses, granulosa and theca cells emerged as the cell-types most affected by aging. We observed limited but promising consistencies across species, including granulosa-specific signature genes (FSHR and OSGIN2) and cell type-linked pathways, with extracellular matrix/adhesion programs in granulosa and ribosomal/mitochondrial programs in theca cells. These convergences suggest that cross-species modeling likely capture core aspects of ovarian aging. Together, our meta-analysis approach may help refine model selection, generate testable hypotheses, and cautiously inform preclinical and translational work in ovarian aging.