Multi-omics provide insights into the regulatory network for human germline development

Abstract

Primordial germ cells (PGCs), the precursors of the human germline, are specified during early embryonic development and will ultimately give rise to sperm and eggs. Recent studies have revealed that humans employ different mechanisms for PGC specification compared with model organisms such as mice. For example, humans uniquely require the transcription factor SOX17 for PGC induction, while mice rely on a different transcription factor Sox2. Despite these species-specific differences and the importance of PGCs for fertility, the regulatory machinery of human PGC remains largely unexplored due to the inaccessibility of early human embryos. To address this knowledge gap, our study leverages a comprehensive multi-omics strategy to systematically characterize the regulatory network guiding human PGC development. To capture this dynamic process, we curated and integrated an extensive multi- omics dataset encompassing both in vitro PGC-like cell models and in vivo embryonic PGCs. This dataset includes samples of 581 bulk transcriptomes (RNA-seq), 54 chromatin accessibility profiles (ATAC-seq), 45 histone modification profiles (ChIP-seq), and 69 single-cell transcriptomes (scRNA- seq), covering key developmental stages from pluripotent stem cells to migrating PGCs. All data were uniformly processed and assembled into a specialized database, hPGCdb, which serves as a comprehensive atlas of human PGC development. This resource recapitulates the precisely controlled and stepwise transitions of germ cell specification and enables systematic dissection of both intrinsic gene regulation and extrinsic signaling cues underlying the establishment of the germline.

By analyzing this integrated atlas, we identified the key gene expression programs and transcription factors (TFs) that drive human PGC fate at each stage. Expected key regulators such as PRDM1 (BLIMP1), TFAP2C, and SOX17 were prominent, validating our approach, but importantly we also uncovered novel TF candidates. Many of these novel TFs human PGC-specific upregulation accompanied by increased chromatin accessibility, suggesting roles in activating the germ cell program. Integrating the epigenomic data revealed clear evidence of epigenetic priming: even before germline genes are expressed, their promoters and enhancers in pluripotent stem cells are already in an open chromatin state (often with binding of pioneer factors), preparing these loci for rapid activation upon germ cell fate induction. This coordinated transcriptional activation and chromatin remodeling appears to underlie the initiation of the human germline lineage.

To determine the functional importance of newly identified regulators, we performed loss-of-function experiments using a CRISPR interference (CRISPRi) approach. Among the top novel TF candidates identified from our analysis, SOX4 and TEAD4 were selected for knockdown during the in vitro differentiation of human embryonic stem cells into PGC-like cells. Suppression of either TF led to a significant impairment of PGC formation. These results provide direct evidence that SOX4 and TEAD4 are critical drivers of human germ cell specification, validating the predictive power of our integrative multi- omics strategy in uncovering previously unrecognized germline regulators.

To further elucidate the regulatory network underlying each developmental transition, we leveraged single-cell transcriptomics and advanced gene regulatory network (GRN) modeling. Using the single-cell RNA-seq data, we constructed co-expression networks (via single-cell WGCNA) to identify gene modules associated with germline commitment, and we applied an integrative GRN framework (Dictys) that combines chromatin accessibility with gene expression profiles. These approaches identified not only the known germline regulators but also novel contributors to PGC development, including RNA- binding proteins and signaling pathway components that may modulate post- transcriptional regulation and cell-cell interactions. The resulting networks offer a systems-level view of the gene regulatory hierarchy in hPGC development, revealing how transcription factors, co-factors, and dynamic chromatin states coordinate to control the stepwise differentiation process of the germline lineage.

In addition to intrinsic regulators, we examined the interactions between PGCs and their somatic microenvironment, describing a soma-germline communication network in both in vitro and in vivo contexts. Ligand-receptor analysis discovered that both Syndecan-2 (SDC2) and Laminin A4 (LAMA4) are highly expressed in human PGCs. SDC2 acts as a key receptor, while LAMA4 functions as a critical ligand. Together, they mediate supportive niche signals for PGC specification. Knocking down SDC2 or LAMA4 caused an ~50% reduction in PGC-like cells induction efficiency, confirming their essential roles. Furthermore, we found that NOTCH2 signaling from somatic cells plays a vital role in promoting the later-stage hPGC maturation. Late-hPGCs (which express marker DDX4) uniquely receive NOTCH2 signals from their environment, and experimentally inducing NOTCH2 activity in early PGC-like cells promoted these cells toward a more advanced, DDX4-positive state. These findings demonstrate that extrinsic cues from the cellular niche fundamentally shape PGC development to ensure proper germline progression in both the early and late stage.

Altogether, this work provides an exceptional multi-omics framework for understanding how human germline fate is established, encompassing both the internal gene regulatory programs and the external signaling interactions that drive PGC development. The comprehensive insights gained highlight a distinct divergence between human and mouse PGC regulatory networks, cautioning against direct inference from animal models. By identifying novel germline regulators and intercellular signals, we not only advance the fundamental understanding of human developmental biology but also create a foundation for translational research. The data and resources generated (including hPGCdb) will facilitate future studies in reproductive biology and regenerative medicine.