Further Development 12.2: Organizing the Chick Primitive Streak

Birds and Mammals

The “organizer” of the chick embryo forms from cells initially located just anterior to the posterior marginal zone. The epiblast and middle layer cells in the anterior portion of Koller’s sickle become Hensen’s node, as described earlier. The posterior portions of Koller’s sickle contribute to the posterior portion of the primitive streak. Hensen’s node has long been known to be the avian equivalent of the amphibian dorsal blastopore lip, since (1) it is the region whose cells become the prechordal plate and chordamesoderm, (2) it is the region whose cells can both induce and pattern a second embryonic axis when transplanted into other locations of the gastrula (Figure 1; Waddington 1933,1934; Gallera 1966; Nicolet 1970), and (3) it expresses the same marker genes as Spemann’s organizer in amphibians and the embryonic shield of teleost fishes, such as the transcription factor Goosecoid (Izpisua-Belmonte et al. 1993). Moreover, Hensen’s node can induce neural tissue when grafted into fish, amphibian, or mammalian embryos (Waddington 1936; Kintner and Dodd 1991; Hatta and Takahashi 1996).

**Figure 1** Induction of a new embryo by transplantation of Hensen’s node. (A) A Hensen’s node from a duck embryo is transplanted into the epiblast of a chick embryo. (B) A secondary embryo is induced (as is evident by the neural tube) from host tissues at the graft site. (After Waddington 1933.)

As is the case in all vertebrates, the dorsal mesoderm is able to induce the formation of the central nervous system in the ectoderm overlying it. The cells of Hensen’s node and its derivatives act like the amphibian organizer, and they secrete BMP inhibitors such as Chordin, Noggin, and Nodal. These proteins repress BMP signaling and dorsalize the ectoderm and mesoderm (Figure 2). However, repression of BMP signals by these antagonists does not appear to be sufficient for neural induction (see Stern 2005b). Fibroblast growth factors synthesized in the hypoblast and in Hensen’s node precursor cells just prior to gastrulation appear to be critical for preparing the epiblast to generate neuronal phenotypes. FGFs can block BMP signaling, but this fact alone does not account for the ability of FGFs to induce a transient expression of pre-neural genes in the epiblast (Streit et al. 1998, 2000). These neural genes do not stay active unless they are supported by BMP antagonists (Streit et al. 1998, 2000; Albazerchi and Stern 2006). Thus, FGF signaling inhibits BMPs from inducing the genes that specify ectoderm to become epidermis, and they activate the genes that specify ectoderm to become neural.

**Figure 2** Possible contribution to chick neural induction by the inhibition of BMP signaling. (A) In a neurulating embryo, Noggin protein (purple) is expressed in the notochord and the pharyngeal endoderm. (B) BMP7 expression (dark purple), which had encompassed the entire epiblast, becomes restricted to the non-neural regions of the ectoderm. (C) Similarly, the product of BMP signaling, the phosphorylated form of Smad1 (recognized by antibodies to the phosphorylated form of the protein; dark brown) is not seen in the neural plate. (From Faure et al. 2002. Indigenous patterns of BMP signaling during early chick development. *Dev Biol* 244(1).)

Indeed, fibroblast growth factors play four fundamental roles in cell specification during gastrulation:

As in all vertebrates, FGFs are responsible for specifying the mesoderm. FGFs from the hypoblast (in collaboration with Nodal from the posterior marginal zone) accomplish this specification by activating the Brachyury and Tbx6 genes in the cells passing through the primitive streak (Figure 3; Sheng et al. 2003).
FGFs separate mesoderm formation from neurulation. The mechanism by which FGFs help end mesoderm ingression and stabilize the epiblast appears to be due to a gene called Churchill. While FGFs are rapidly inducing the mesoderm, they are also slowly inducing activation of Churchill in the ectoderm. The Churchill protein (so named because the protein’s two zinc fingers extend like the British prime minister’s famous “V for Victory” symbol) can activate the Smad-interacting protein SIP1. SIP1 controls the genes whose expression is required for ingression of cells through the primitive streak. Thus, once activated, SIP1 helps prospective neural plate cells remain in the epiblast.
FGFs help bring about neurulation in the central ectodermal cells. SIP1, probably through its interaction with Smad1, may make the prospective neural plate cells less sensitive to BMP.
FGFs induce ERNI and Sox3, two pre-neural genes that initiate the signaling cascade leading to the production of neural tissue.

Thus, FGFs appear to be critically important regulators of cell fate in the early chick embryo (Streit et al. 2000; Sheng et al. 2003; Albazerchi and Stern 2006).

Model by which FGFs regulate mesoderm formation and neurulation. (A) Stage XI, where the hypoblast (green) secretes Fgf8, which induces pre-neural genes ERNI and Sox3 (blue) in the epiblast. The cells in this domain, however, remain uncommitted. Nodal, expressed in the posterior epiblast, cannot function; it is inhibited by the Cerberus protein secreted by the hypoblast. (B) At around stage 1, the hypoblast is displaced from the posterior edge by the endoblast (secondary hypoblast; gold), allowing Nodal to function. Nodal plus Fgf8 induces Brachyury and Tbx6expression to specify the mesoderm and initiate the ingression of mesoderm cells through the primitive streak (red). (C) At stage 4, continued Fgf8 expression activates Churchill in the epiblast. (D) By end of stage 4, Churchill protein induces Sip1, which blocks Brachyury and Tbx6, preventing further ingression of epiblast cells through the streak. The remaining epiblast cells can now become sensitized to neural inducers from Hensen’s node (purple).

**Figure 3** Model by which FGFs regulate mesoderm formation and neurulation. (A) Stage XI, where the hypoblast (green) secretes Fgf8, which induces pre-neural genes *ERNI* and *Sox3* (blue) in the epiblast. The cells in this domain, however, remain uncommitted. Nodal, expressed in the posterior epiblast, cannot function; it is inhibited by the Cerberus protein secreted by the hypoblast. (B) At around stage 1, the hypoblast is displaced from the posterior edge by the endoblast (secondary hypoblast; gold), allowing Nodal to function. Nodal plus Fgf8 induces *Brachyury* and *Tbx6*expression to specify the mesoderm and initiate the ingression of mesoderm cells through the primitive streak (red). (C) At stage 4, continued Fgf8 expression activates *Churchill* in the epiblast. (D) By end of stage 4, Churchill protein induces Sip1, which blocks *Brachyury* and *Tbx6*, preventing further ingression of epiblast cells through the streak. The remaining epiblast cells can now become sensitized to neural inducers from Hensen’s node (purple). (After Sheng et al. 2003. Churchill, a zinc finger transcriptional activator, regulates the transition between gastrulation and neurulation. *Cell* 11(5).)

Anterior-posterior polarity

While they are still in the epiblast, but close to the primitive streak, the mesoderm cells appear to receive instructions that tell them exactly where they are along the anterior-posterior axis. The entire length of the notochord at the midline is derived from cells that are present in Hensen’s node by the full primitive streak stage. A population of progenitor cells remains in the node; their descendants gradually leave as the node regresses, laying down the chordamesoderm and the ventral midline of the neural tube (the future floor plate of the spinal cord) (Selleck and Stern 1991; Psychoyos and Stern 1996; Tzouanacou et al. 2009). Therefore, anterior-posterior identities along the axis from the hindbrain to the tail are specified as a function of the time of emergence from the primitive streak and Hensen’s node.

It has been proposed that the length of time cells are resident in the primitive streak region determines which Hox genes are expressed by the cells. This pattern of Hox gene expression can also be under the influence of the FGF and retinoic acid gradients (Gaunt 1991; Wilson et al. 2009). As we will detail later in the chapter, Hox genes are the vertebrate homologues of the homeotic (Hom-C) genes of Drosophila. And just as in Drosophila embryos, vertebrate Hox genes specify the identity of cells along the anterior-posterior axis. In the case of vertebrates, however, there are four gene clusters (HoxA, HoxB, HoxC, and HoxD) instead of just one, and rather than individual Hox genes appearing at particular segmental levels, there is a nested set of Hox gene expression. For example, the mesodermal precursor cells are patterned along the anterior-posterior axis by the HoxB genes, which appear to inform the cells when to leave the epiblast and ingress into the primitive streak. “Anterior” Hox genes (which are identified with lower numbers, e.g., Hoxb4) are expressed early and extend farther anterior than genes such as Hoxb9, which is expressed later and does not extend as far into the embryo’s anterior (Iimura and Pourquié 2006; Figure 4). Thus, the more posterior cells express more Hox genes than the more anterior cells do.

**Figure 4** Hox gene activation begins when the mesodermal precursor cells are still in the epiblast. These genes are activated in an anterior-to-posterior fashion. Migration into the primitive streak is regulated by the Hox gene expression pattern (the most posterior Hox gene having preference). (A) *Hoxb4* expression in stage-4, -5, and -6 chick embryos. (B) *Hoxb9* expression in stage-7, -8, and -8+ chick embryos. Note that the anterior border of expression is more posterior than the anterior border of *Hoxb4* expression. (From Iimura,T and O. Pourquié. 2006. Collinear activation of Hoxb genes during gastrulation is linked to mesoderm cell ingression. *Nature 442*(7102).)

Literature Cited

Albazerchi, A. and C. D. Stern. 2007. A role for the hypoblast (AVE) in the initiation of neural induction, independent of its ability to position the primitive streak. Dev. Biol. 301: 489–503.