Brazil Nut Effect Drives Pattern Formation in Early Mammalian Embryos

2021 
The formation of three-dimensional ordered spatial patterns, which is essential for embryonic development, tissue regeneration, and cancer metastasis, is mainly guided by the chemical concentration gradient of morphogens. However, since no chemical concentration gradient has been observed in the early embryonic development (pre-implantation) of mammals, the pattern formation mechanism has been unsolved for a long time. During the second cell fate decision of mouse embryos, the inner cell mass (ICM) segregates into topographically regionalized epiblast (EPI) and primitive endoderm (PrE) layers. Here, we report that the segregation process of PrE/EPI precursors coincides with an emerged periodic expansion-contraction vibration of the blastocyst cavity, which induces phase transition in the ICM compartment to a higher fluidity state and generates directional tissue flows. By experiments and modeling, we demonstrate that the spatial segregation of PrE and EPI precursors is mediated by a "Brazil nut effect"-like viscous segregation mechanism in which PrE precursors with low affinity gradually migrate to the surface of ICM along with the tissue flow, while EPI precursors with high affinity remains inside ICM under cavity vibration. Artificially manipulation of the frequency and amplitude of cavity vibration could control the process of spatial separation as well as lineage specification of PrE/EPI. Furthermore, disruption of the cavity vibration in the initial stage after segregation could reverse the ICM cells back to a mixed state. Therefore, this study reveals a fundamental mechanism that guarantees the robustness of cell segregation and pattern formation without specific morphogens in early mammalian embryos. Our model also emphasizes a conserved function of cavity structure that widely exists in organisms as an energy reservoir and converter between different forms, such as chemical and mechanical energy.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    50
    References
    1
    Citations
    NaN
    KQI
    []