In Vitro Oogenesis: High-Throughput Epigenetic Screening for Efficient Derivation of Human Germ Cells and Meiotic Induction from induced Pluripotent Stem Cells (iPSCs)

Abstract

In vitro gametogenesis—the development of artificial gametes from pluripotent stem cells—represents a promising approach in reproductive biology with potential implications for understanding infertility, which affects approximately 15% of couples worldwide. While significant progress has been made in mouse models, human in vitro oogenesis has remained challenging due to species-specific differences in germ cell development, complex epigenetic requirements, and technical limitations in recapitulating meiosis outside the body. This thesis explores an integrated approach to address some of these challenges through transcription factor-directed differentiation, epigenetic modulation, and live imaging techniques. We first develop a protocol for inducing human oogonia-like cells (iOLCs) from induced pluripotent stem cells (iPSCs) through systematic screening of germline-associated transcription factors. We identify a combination of five transcription factors—ZNF281, LHX8, SOHLH1, ZGLP1, and ANHX (termed the D5 cocktail)—that enhances DDX4+ iOLC yield when combined with DNA methyltransferase inhibition. These iOLCs express key germline markers, maintain proliferative capacity, and can be maintained feeder-free after isolation, offering an alternative to conventional methods that typically require months of culture in complex environments. Building on this foundation, we investigate the more challenging aspects of meiotic induction. Through a combination of DNA demethylation, retinoid signaling, and overexpression of specific factors (BCL2, HOXB5, and BOLL), we observe expression of early meiotic markers and progression through initial meiotic stages. Single-cell RNA sequencing coupled with computational analyses—including gene regulatory network inference, pseudotime trajectory modeling, and barcode-linked factor enrichment—enables us to map the transcriptional landscapes of differentiating cells and identify key regulators of meiotic commitment. To directly visualize these dynamic cellular processes, we develop CRISPR-mediated knock-in reporter lines for synaptonemal complex proteins using the fluorescent protein mStayGold. Our integrated computational framework leverages deep learning-based cell segmentation, trajectory tracking algorithms, and feature extraction to analyze complex temporal patterns in high-dimensional imaging data. This computational pipeline enables quantitative characterization of protein localization dynamics and cellular transitions during meiotic progression that would be undetectable through traditional analytical approaches. This work contributes to ongoing efforts to model aspects of human oogenesis in vitro and offers potential insights into the regulation of germ cell specification and meiotic entry. While substantial challenges remain in achieving complete meiosis and functional oocyte formation, the approaches described here may provide tools for studying fundamental aspects of human reproduction, exploring factors that influence fertility, and investigating genetic components of reproductive development.