Over the last century, mammalian embryos have been used extensively as a common animal model to investigate fundamental questions in the field of developmental biology. More recently, the establishment of transgenic and gene-targeting systems in laboratory mice has enabled researchers to unveil the genetic mechanisms underlying complex developmental processes (Mak, 2007). However, our understanding of cell–cell interactions and their molecular basis in the early stages of mammalian embryogenesis is still very fragmentary. One of the major problems is the difficulty of precise manipulation and limited accessibility to mammalian embryos via uterus wall. Unfortunately, existing tissue and organotypic culture systems per se do not fully recapitulate three-dimensional, dynamic processes of organogenesis observed in vivo. Although transgenic animal technology and virus-mediated gene delivery are useful to manipulate gene expression, these techniques take much time and financial costs, which limit their use. Whole-embryo mammalian culture system was established by New and colleague in the 1970s, and was modified thereafter by several researchers (reviewed by New, 1971, 1978; Sturm and Tam, 1993; Hogan et al., 1994; Eto and Osumi-Yamashita, 1995; Tam, 1998). At first, the whole-embryo culture system was used in the field of teratology, and then applied in a variety of developmental biology fields, because this