In the early mammalian embryo, X chromosome inactivation (XCI) achieves dosage parity between males and females for X-linked genes. During mouse development, imprinted paternal XCI is observed first and switches to random XCI in the epiblast but not placental lineages. The mechanism by which this epigenetic switch occurs is currently unknown. Here, we establish an ex vivo model for imprinting and identify a novel trans-acting regulatory factor for imprinted XCI. Using an induced trophoblast stem cell (iTS) model, we show that embryonic stem (ES) cells transdifferentiated into trophoblasts retain partial memory of the XCI imprint. Cdx2, a stem cell factor that determines commitment to the extraembryonic lineage, directly binds Xist and activates expression of Xist RNA in extrembryonic cells. Cdx2 competes with Oct4, a stem cell factor that determines commitment to the embryonic lineage, for overlapping binding sites within Xist. We propose that mutually exclusive binding between Cdx2 and Oct4 in Xist underlies the switch between imprinted and random XCI in the early mouse embryo. XCI is epigenetically regulated and is tightly linked to changes in pluripotency and cell fate decisions in the early mouse embryo (Monk and Harper 1979). Between embryonic days 0.5 and 5.5 (E0.5-E5.5), two forms of XCI occur sequentially in the mouse. "Imprinted XCI" is a germline-determined process during which silencing occurs exclusively on the paternal X chromosome (X P ) (Takagi and Sasaki 1975;Takagi 2003). Evidence of imprinted XCI is observed by the two-cell stage, where repetitive elements on X P are transcriptionally suppressed relative to those on the maternal X (X M ) (Huynh and Lee 2003; Namekawa et al. 2010). Silencing progressively encompasses genic elements on X P during the next several divisions until the early mouse blastocyst stage (Okamoto et al. 2005;Kalantry et al. 2009;Namekawa et al. 2010).In the later mouse blastocyst, embryonic (epiblast) and extraembryonic lineages (trophectoderm, primitive endoderm) become evident for the first time and it is during this time that evidence of "random XCI" is first detected. Whereas the extraembryonic cells retain imprinted XCI, the embryonic lineage reactivates X P at E4.5 (Mak et al. 2004;Okamoto et al. 2004) and undergoes a second round of XCI (Harper et al. 1982;Tan et al. 1993), this time in a "random" way such that X P and X M have an equal chance of becoming the inactive X (Xi). Random XCI is essential for in vivo differentiation of the epiblast to the three germ lineages (ectoderm, mesoderm, and endoderm) and for the ex vivo differentiation of epiblast-derived embryonic stem (ES) cells. Recent work in stem cell engineering shows that mouse XCI is also intimately linked to the reprogramming process ex vivo in the derivation of mouse induced pluripotent stem (iPS) cells (Maherali et al. 2007).How and why the early mouse embryo switches from imprinted to random XCI present two intriguing questions. Still unknown are specific factors that dictate the decision to ...
