DEVELOPMENTAL BIOLOGY 3230
Amphibian Axis Formation
Historically, animal development has been thought to be either mainly mosaic and lineage dependent or regulative and dependent on position. Animals like Drosophila and C. elegans fit the mainly mosaic model. If you ablate a C.elegans founder cell then all cells that normally arise from that lineage are missing from the embryo. In contrast, if you ablate a cell in the early embryo of an animal like the frog you often see no effect on development. The remaining cells can alter their fate based on their new neighbor relationships (position within the embryo) and compensate for the missing cell's progeny. However, as we begin to understand the molecular mechanisms underlying a cells potency and progressive cell fate restrictions we see that there is no fundamental difference between mosaic and regulative development. All animals use a spectrum of molecular mechanisms to regulate different stages of cell specification and differentiation.
In the figure at right showing the first cleavage of a frog embryo you see the results of isolating the blastomeres. Normally, the first cleavage bisects the "gray crescent". Isolated blastomeres give rise to perfectly normal duplicated embryos. However, if you experimentally alter the first cleavage so that one blastomere gets all the gray crescent cytoplasm and the other gets none, then the isolated blastomeres behave very differently. The one receiving the gray crescent material develops into a normal embryo, but the other forms a disorganized mass of tissue called a "belly piece" because it contains mainly ventral tissues. This suggests that some important cytoplasmic determinants are localized in the early vertebrate embryo. We are interested in the same questions that we addressed in Fly development. Where do the first asymmetries arise in vertebrate development? How are the anterior-posterior and dorsal-ventral axis established? How does embryonic patterning arise? When you look at a frog egg you can clearly see one maternally derived asymmetry. The animal-vegetal axis is obvious due to the pigmentation and yolk differences that are determined during oogenesis.
Except for the isolation experiment described above, most experiments on frog blastula and early gastrula stage embryos suggested that cell fate was not determined until the mid gastrula stage. The transplantation experiments shown at right illustrate this. Heterotopic transplantations of early gastrula tissue resulted in a normal embryo. The transplanted tissue "depended" on its new host neighbors to tell it how to develop (Dependent "conditional" development). However, the same heterotopic transplants performed at the late gastrula stage gave very different results. Now the transplanted tissue exhibited "independent development" and developed "autonomously" based on its identity in the donor. You should not be surprised by these results as we have discussed the progressive cell determination that occurs during development. The question is exactly when and how does this determination occur? What is the source of the patterning and cell determination signals. Is it extrinsic or intrinsic or both?
The key experiments were done by Hans Spemann and his graduate student Hilde Mangold in 1924. They found that there was one region of the late blastula and early gastrula that behaved in an extraordinary way when transplanted to a ectopic host position. The dorsal lip of the blastopore (the region fated to give rise to dorsal mesoderm) not only did not change fate when transplanted, but induced surrounding host cells to change their fate. The fate changes induced by the dorsal lip transplant were dramatic; an entirely new embryonic axis was induced. The dorsal lip of the blastopore was named the "organizer" because of its unique inductive abilities.
Proof that the blastomeres of the dorsal lip were inducing host cells to take on new fates came from transplants using "marked" cells. Spemann and Mangold actually performed their transplantation studies using two different species of newt embryos that were differently pigmented so the cells of each species could be unambigously identified. In the figures at right the transplanted dorsal lip cells are labeled in red. Notice that the secondary invagination includes many unlabeled host cells (B) and later the induced secondary axis includes many structures formed by unlabeled host cells (somites, neural tube, etc.). In (C) you can see why the dorsal lip tissue was called the organizer. In rare cases a complete twin embryo resulted from the transplant of the dorsal lip.

These classic experiments, for which Spemann received the Noble prize in 1935, focused the attention of a generation of developmental biologists on the "organizer". Mangold died in a freak accident in 1924, as a graduate student, before her experimental results were published.

What cells have "organizer" properties? How do these cells acquire these amazing "organizing" abilities? What are the molecular signals from the organizer cells that induce the axial patterns of cell fate restrictions? The search was on for the molecules responsible for organizer identity and function.
The dorsal lip of the blastopore is a constantly changing population of cells. The first cells of the involuting marginal zone (IMZ) to migrate through the dorsal lip migrate the farthest anteriorly to give rise to foregut endoderm and head mesoderm. As gastrulation procedes the mesodermal and endodermal precursors migrate less and less to give rise to progressively more posterior structures. Thus, do not think of the dorsal lip as fixed population of cells, rather think of the dorsal lip as a "cellular wave" made up of contiually changing population of cells that pass through during gastrulation.
In the 60s Nieuwkoop and Nakamura performed isolation and re-combination experiments just as we described earlier in the course for sea urchins. When animal cap cells are isolated in culture they give rise to ectoderm; marginal zone cells give rise to mesoderm, and vegetal cells give rise to endoderm. However, when animal cap cells are combined with vegetal cells the animal cap is converted to mesoderm! These and other experiments suggested that "mesoderm fate" does not exist in the early blastula, but is induced at the equatorial region by an inductive interaction between the vegetal cells and animal cells.
Dale and Slack extended the results from Nieuwkoop by combining cells from a 32 cell stage frog blastula. The systematically combined vegetal pole blastomeres from different dorso-ventral positions with animal cap blastomeres. Remember that the DV axis is known from the site of sperm entry and the appearance of the gray crescent. They found that the dorsal most vegetal cell had the greatest ability to induce dorsal mesoderm, while the ventral most vegetal cell had the greatest ability to induce ventral mesoderm. They concluded that even at this very early stage the vegetal cells must be secreting a "mesoderm inducing" signal. They suggested a dorsal to ventral gradient in dorsal mesoderm inductive ability. The region of the dorsal most vegetal cells that induced dorsal mesoderm (the organizer) was named the Nieuwkoop Center in honor of the developmental biologist Nieuwkoop.
The work of Nieuwkoop, and Dale and Slack, and others is summarized in the model at right. They hypothesized that only ectodermal and endoderm fate existed in the earliest embryo. Mesodermal fate is induced by signals from the vegetal cells to the marginal zone cells. Nieuwkoop's Center represented a special zone represented by the dorsal most vegetal cells that had the ability to induce dorsal mesoderm or organizer function. The next obvious question was to find and characterized these postulated mesoderm inducing signals.
Frog eggs (eg. Xenopus) can be very large and can be generated in large numbers in the lab. Protein and mRNA that is regionally localized or enriched has been identified and characterized as candidate molecules involved in organizer specification and function. Another important technique to identify molecules involved in organizer specification and function was to make frog cDNA libraries. These libraries could then be sub divided into pools that were then injected into frog blastomeres. In this way single cDNAs could be identified that affected organizer specification and function. In addition, information about conserved molecular signaling pathways was being gained from genetic experiments in yeast, worm, and fly model systems.

Slowly but surely a molecular model of organizer specification and function has emerged. Maternal mRNA have been found localized to the vegetal pole. Nodal related, VegT and Veg1 mRNA is differentially localized and /or translated in the vegetal region to give a graded TGF-beta signal. Further, beta-catenin was found to be enriched in the dorsal blastomeres. Beta-catenin was identified as a component of the Wg/Wnt signaling pathway (see conserved pathways) and can act as a transcriptional regulator.
The molecular model for organizer specification is shown at right. The animal vegetal axis is determined during oogenesis. Maternal mRNAs and proteins are localized to the vegetal pole of the egg. There is no pre-existing dorsal ventral or anterior posterior axis. The sperm can fertilized the egg anywhere on in the animal hemisphere. Sperm entry initiates the developmental program, including a cortical rotation towards the site of sperm entry. This cortical rotation is thought to bring Disheveled protein (Dsh) to the dorsal side of the embryo. (The cortical rotation causes the appearance of the gray crescent as the darkly pigment animal cortex rotates away from the equator and exposes the underlying "grey" cytoplasm.) Initially, beta catenin is uniformly distributed in the egg. Glycogen synthase kinase-3 is also uniformly distributed (probably in an inactive form). On activation of the developmental program, GSK-3 is activated and begins phosphorylating beta-catenin. This phosphorylation targets beta-catenin for degradation. Dsh inhibits GSK-3 and prevents the degradation of beta catenin. Since Dsh is enriched in the dorsal region of the embyro, it is the dorsal region of the embryo that becomes enriched for stable beta catenin.
As cleavage progresses, it will be the dorsal most blastomeres that get the highest concentration of beta catenin.
Beta-catenin acts together with a widely expressed transcriptional repressor, Tcf-3, to activate transcription of the siamois gene. The siamois protein acts together with a transcription factor induced by the TGF-beta signaling pathway (Veg1, VegT and/or Nodal) to activate the transcription of the goosecoid gene. The goosecoid protein is thought to be the determinant of organizer function. There may be some homology between fly and vertebrate goosecoid function.
VegT and Vg1 act synergistically with beta catenin to activate Xenopus nodal related proteins. This leads to to a gradient of Xnr proteins across the endoderm, with highest concentrations dorsally. Ventral mesoderm expresses BMP4 and Xwnt-8, but high concentrations of Xnr repress BMP4 and Xwnt-8 in the overlying mesodermal cells so that BMP4 and Xwnt-8 are expressed in a ventral to dorsal gradient.
What are the molecules secreted by the organizer that pattern the embryo? Scientists were surprised to find that the molecules secreted by the organizer acted to inhibit the functions of BMP4 and Xwnt8. The organizer secrets Noggin, Chordin, Follistatin that bind to BMP4 and prevent BMP4 from binding to its receptor. This leads to an effective ventral to dorsal gradient of BMP4. The same is true for Xwnt-8. Frzb secreted by the organizer competitively binds to Xwnt-8 and prevents Xwnt-8 signanling. Thus both BMP-4 and Xwnt-8 are functional expressed in a ventral to dorsal gradient. It turns out that these two signaling molecules specify ventral mesoderm and endoderm at high concentrations and intermediate mesoderm and endoderm at low concentrations. Surprisingly, the default fate of ectoderm is neural! BMP-4 and Xwnt-8 act as anti-neuralizing agents and specify ventral mesodermal and endodermal fate and epidermal ectodermal fate. Dorsal endoderm, dorsal mesoderm and neural ectoderm is the fate of cells NOT exposed to threshold values of BMP-4 and Xwnt-8.
The figure at right shows the gradient of Xwnt8 from the marginal zone mesoderm opposed by a gradient of Frzb expressed by the organizer. Frzb mimics the binding site of the Frizzled protein.
This is a summary of dorsal ventral and anterior posterior patterning by the organizer. BMP and Wnt gradients are opposed by proteins secreted by the organizer. Noggin, Chordin, and Follistatin bind and inactivate BMP-4, while Frzb binds and inactivates Xwnt8. There is another molecule, secreted by the foregut endoderm, cerberus, that inhibits Wnt function in the most anterior region and specifies the anterior brain. FGF and retinoic acid also seem to be secreted by the organizer and establish important morphogen gradients. FGF may be important for maintaining proliferation. The RA gradient is an important regulator of the Hox genes. Note that as cells migrate through the dorsal lip they are "exposed" to the morphogen gradients and acquire an "inductive" capacity (memory of morphogen exposure at the dorsal lip). The arrows illustrate that the inductive signaling from the dorsal mesoderm to the overlying ectoderm occurs over a short distance.
When the molecules specifying the dorso-ventral axis in the frog were identified there was a surprising degree of similarity to dorso-ventral patterning in the fly--except that the pattern was inverted! The fly used dpp to specify Dorsal, while in the frog BMP specified Ventral. In 1822 an anatomist name Geoffroy had suggested that the nerve cord of arthropods (located on the ventral side of the animal) was homologous to the vertebrate spinal cord. In the illustration at right you can see that in many regards the arthropod body plan looks inverted compared to the vertebrate. The recent molecular homologies in DV patterning suggest that Geoffroy was right. All bilarteral phyla shared a common ancester that used the BMP/dpp molecular pathway to regulate dorso-ventral patterning.
The illustration at right shows how anterior-posterior patterning occurs in the ectoderm. BMP induces ectoderm to take on an epidermal fate. In ventral regions where the BMP signal is high the ectoderm becomes epidermal. In regions where BMP is blocked by chordin, noggin and follistatin secreted from the organizer the ectoderm takes on a neural fate. Where BMP is blocked and levels of retinoci acid and FGF are high the ectoderm takes on a more posterior neural fate. Remember that the Wnt gradient is also important. The the foregut endoderm secretes Cerberus that binds Wnt and induces the most anterior neural fate. Remember that foregut endoderm was part of the very first cells to form the organizer and migrated the farthest anterior.
The experiments at right show that the inductive abilities of the dorsal lip changes with time. When young dorsal lip cells are transplanted into the blastocoel cavity they induce anterior structures, but when the dorsal lip from a late gastrula is transplanted into a young gastrula blatocoel it induces posterior structures. This again illustrates the point I tried to make earlier. The population of cells that make up the dorsal lip (organizer) is continually changing through gastrulation. Cells are exposed to a morphogen gradient and specified as "anterior" or "posterior" and have inductive capabilities appropriate for their AP position.

The experiments below by Spemann and Mangold further illustrate this by testing the inductive abilities of archenteron roof tissure (endoderm and mesoderm) taken from various anterior to posterior positions in an embryo just after the end of gastrulation. The archenteron tissue is again transplanted into the blastocoel of an early gastrula. Note that the "extra" structures that are induced correlate with the anterior posterior position of the donor archenteron tissue. Thus, even AFTER cells are no longer part of the ORGANIZER they still retain a memory of their exposure to morphogen gradients and have specific inductive abilities based on when they were part of the organizer.
What specifies the anterior-posterior regional identity of cells? Just as we learned in Drosophila AP patterning, the vertebrate Hox genes are differentially expressed along the anterior-posterior axis of the developing animal. Again, the Hox genes are roughly expressed in a 3'-5' anterior-posterior order along the body. The conservation between fly and mouse patterning via Hom/Hox genes is amazing. The morphogen gradients established by the organizer must directly or indirectly regulate Hox gene expression. The rentinoic acid gradient is a directly regulator of Hox gene expression. Many of the Hox genes have regulatory domains that contain retinoic acid response elements (RERs). The organizer is thought to also be the source of the retinoic acid gradient.
In the vertebrate lineage there has been a two-fold genome wide duplication, leading to a 4 fold expansion of the Hox genes when compared to the fly. In addition there has been some expansion of in the number of Hox genes in each cluster so that instead of 8 genes in the Hom cluster there are up to 13 in each of the four Hox clusters.
When we discuss limb patterning and anterior-posterior pattering in the nervous system I will describe experiments that directly demonstrate the effects of retinoic acid on Hox expression.