Loss of Oct4 expression during the development of murine embryoid bodies

Abstract

We describe the internal organization of murine embryoid bodies (EBs) in terms of the structures and cell types formed as Oct4 expression becomes progressively lost. This is done by making the EBs from iPS cells carrying an inducible and permanent Oct4 reporter (Oct4-MerCreMer;mTmG). When these EBs are treated with tamoxifen, the Oct4 expressing cells switch from a red to a green fluorescence color, and this is maintained thereafter by their progeny. We show that there is no specific pattern in which Oct4 is downregulated, rather it appears to be spatially random. The earliest cells to lose Oct4 expression are internal and stain positive for -fetoprotein (AFP) indicating that they are visceral endoderm. However, GATA4, characteristic of primitive endoderm, was found in Oct4-expressing cells at this stage. This indicates that the first formed visceral endoderm does not arise from primitive endoderm, a difference from normal embryonic development. Contrary to previous reports, our EBs did not form a layer of primitive endoderm, or visceral endoderm, around the outside. Markers of the early body axis, BRACHYURY (T) and FOXA2, behaved somewhat differently from each other. BRA, which marks the early mesoderm, node and notochord, arises in Oct4 expressing cells on days 3-4. FOXA2, which marks the floor plate of the neural tube and definitive endoderm, as well as the node and notochord, arises at the same time but mostly in cells that have already lost Oct4 expression. Although there is usually a concentration of T or FOXA2 cells in one region of the EB, the morphology is not predictable and there are also scattered cells expressing these markers. Several clumps of cardiomyocytes are visible by day 7 of EB development, and we show that the cells forming these clumps lose Oct4 expression between days 3 and 5. Overall, our results indicate that EBs recapitulate normal development quite well in terms of the tempo of events and the appearance of specific markers, but they do not resemble embryos in terms of their morphology and structure, which is contrary to the previous reports.