Egg and sperm cells (gametes) of the mouse are derived from a founder population of primordial germ cells that are set aside early in embryogenesis. Primordial germ cells arise from the proximal epiblast, a region of the early mouse embryo that also contributes to the first blood lineages of the embryonic yolk sac. Embryonic stem cells differentiate in vitro into cystic structures called embryoid bodies consisting of tissue lineages typical of the early mouse embryo. Because embryoid bodies sustain blood development, we reasoned that they might also support primordial germ cell formation. Here we isolate primordial germ cells from embryoid bodies, and derive continuously growing lines of embryonic germ cells. Embryonic germ cells show erasure of the methylation markers (imprints) of the Igf2r and H19 genes, a property characteristic of the germ lineage. We show that embryoid bodies support maturation of the primordial germ cells into haploid male gametes, which when injected into oocytes restore the somatic diploid chromosome complement and develop into blastocysts. Our ability to derive germ cells from embryonic stem cells provides an accessible in vitro model system for studies of germline epigenetic modification and mammalian gametogenesis.

Epigenetic modifications must underlie lineage-specific differentiation as terminally differentiated cells express tissue-specific genes, but their DNA sequence is unchanged. Hematopoiesis provides a well-defined model to study epigenetic modifications during cell-fate decisions, as multipotent progenitors (MPPs) differentiate into progressively restricted myeloid or lymphoid progenitors. While DNA methylation is critical for myeloid versus lymphoid differentiation, as demonstrated by the myeloerythroid bias in Dnmt1 hypomorphs1, a comprehensive DNA methylation map of hematopoietic progenitors, or of any multipotent/oligopotent lineage, does not exist. Here we examined 4.6 million CpG sites throughout the genome for MPPs, common lymphoid progenitors (CLPs), common myeloid progenitors (CMPs), granulocyte/macrophage progenitors (GMPs), and thymocyte progenitors (DN1, DN2, DN3). Dramatic epigenetic plasticity accompanied both lymphoid and myeloid restriction. Myeloid commitment involved less global DNA methylation than lymphoid commitment, supported functionally by myeloid skewing of progenitors following treatment with a DNA methyltransferase inhibitor. Differential DNA methylation correlated with gene expression more strongly at CpG island shores than CpG islands. Many examples of genes and pathways not previously known to be involved in choice between lymphoid/myeloid differentiation have been identified, such as Arl4c and Jdp2. Several transcription factors, including Meis1, were methylated and silenced during differentiation, suggesting a role in maintaining an undifferentiated state. Additionally, epigenetic modification of modifiers of the epigenome appears to be important in hematopoietic differentiation. Our results directly demonstrate that modulation of DNA methylation occurs during lineage-specific differentiation and defines a comprehensive map of the methylation and transcriptional changes that accompany myeloid versus lymphoid fate decisions.