Reconciling different models of forebrain induction and patterning: a dual role for the hypoblast

Several models have been proposed for the generation of the rostral nervous system. Among them, Nieuwkoop's activation/transformation hypothesis and Spemann's idea of separate head and trunk/tail organizers have been particularly favoured recently. In the mouse, the finding that the visceral endoderm (VE) is required for forebrain development has been interpreted as support for the latter model. Here we argue that the chick hypoblast is equivalent to the mouse VE, based on fate, expression of molecular markers and characteristic anterior movements around the time of gastrulation. We show that the hypoblast does not fit the criteria for a head organizer because it does not induce neural tissue from na?ve epiblast, nor can it change the regional identity of neural tissue. However, the hypoblast does induce transient expression of the early markers Sox3 and Otx2. The spreading of the hypoblast also directs cell movements in the adjacent epiblast, such that the prospective forebrain is kept at a distance from the organizer at the tip of the primitive streak. We propose that this movement is important to protect the forebrain from the caudalizing influence of the organizer. This dual role of the hypoblast is more consistent with the Nieuwkoop model than with the notion of separate organizers, and accommodates the available data from mouse and other vertebrates.     

Activation of epiblast gene expression by the hypoblast layer in the prestreak chick embryo

Axis formation is a highly regulated process in vertebrate embryos. In mammals, inductive interactions between an extra-embryonic layer, the visceral endoderm, and the embryonic layer before gastrulation are critical both for anterior neural patterning and normal primitive streak formation. The role(s) of the equivalent extra-embryonic endodermal layer in the chick, the hypoblast, is still less clear, and dramatic effects of hypoblast on embryonic gene expression have yet to be demonstrated. We present evidence that two genes later associated with the gastrula organizer (Gnot-1 and Gnot-2) are induced by hypoblast signals in prestreak embryos. The significance of this induction by hypoblast is discussed in terms of possible hypoblast functions and the regulation of axis formation in the early embryo. Several factors known to be expressed in hypoblast, and retinoic acid, synergistically induce Gnot-1 and Gnot-2 expression in blastoderm cell culture. The presence of retinoic acid in prestreak embryos has not yet been directly demonstrated, but exogenous retinoic acid appears to mimic the effects of hypoblast rotation on primitive streak extension, raising the possibility that retinoid signaling plays some role in the pregastrula embryo.     

Distinct steps of neural induction revealed by Asterix, Obelix and TrkC, genes induced by different signals from the organizer

The amniote organizer (Hensen's node) can induce a complete nervous system when grafted into a peripheral region of a host embryo. Although BMP inhibition has been implicated in neural induction, non-neural cells cannot respond to BMP antagonists unless previously exposed to a node graft for at least 5 hours before BMP inhibitors. To define signals and responses during the first 5 hours of node signals, a differential screen was conducted. Here we describe three early response genes: two of them, Asterix and Obelix, encode previously undescribed proteins of unknown function but Obelix appears to be a nuclear RNA-binding protein. The third is TrkC, a neurotrophin receptor. All three genes are induced by a node graft within 4-5 hours but they differ in the extent to which they are inducible by FGF: FGF is both necessary and sufficient to induce Asterix, sufficient but not necessary to induce Obelix and neither sufficient nor necessary for induction of TrkC. These genes are also not induced by retinoic acid, Noggin, Chordin, Dkk1, Cerberus, HGF/SF, Somatostatin or ionomycin-mediated Calcium entry. Comparison of the expression and regulation of these genes with other early neural markers reveals three distinct "epochs", or temporal waves, of gene expression accompanying neural induction by a grafted organizer, which are mirrored by specific stages of normal neural plate development. The results are consistent with neural induction being a cascade of responses elicited by different signals, culminating in the formation of a patterned nervous system.     

Neural induction requires BMP inhibition only as a late step, and involves signals other than FGF and Wnt antagonists

A dominant molecular explanation for neural induction is the 'default model', which proposes that the ectoderm is pre-programmed towards a neural fate, but is normally inhibited by endogenous BMPs. Although there is strong evidence favouring this in Xenopus, data from other organisms suggest more complexity, including an involvement of FGF and modulation of Wnt. However, it is generally believed that these additional signals also act by inhibiting BMPs. We have investigated whether BMP inhibition is necessary and/or sufficient for neural induction. In the chick, misexpression of BMP4 in the prospective neural plate inhibits the expression of definitive neural markers (Sox2 and late Sox3), but does not affect the early expression of Sox3, suggesting that BMP inhibition is required only as a late step during neural induction. Inhibition of BMP signalling by the potent antagonist Smad6, either alone or together with a dominant-negative BMP receptor, Chordin and/or Noggin in competent epiblast is not sufficient to induce expression of Sox2 directly, even in combination with FGF2, FGF3, FGF4 or FGF8 and/or antagonists of Wnt signalling. These results strongly suggest that BMP inhibition is not sufficient for neural induction in the chick embryo. To test this in Xenopus, Smad6 mRNA was injected into the A4 blastomere (which reliably contributes to epidermis but not to neural plate or its border) at the 32-cell stage: expression of neural markers (Sox3 and NCAM) is not induced. We propose that neural induction involves additional signalling events that remain to be identified.     

Anterior neural induction by nodes from rabbits and mice

The organizer of vertebrate embryos represents the major regulatory center for the formation of the embryonic axis during gastrulation. The early blastopore lip of amphibia and Hensen's node of the chick at the full-length primitive streak stage possess both a head- and a trunk-inducing potential. In mice, a head-inducing activity was identified in the extraembryonic, anterior visceral endoderm (AVE) by tissue ablation and genetic experiments. Evidence for a similar activity in the AVE from the rabbit was obtained by transplanting below the avian epiblast. However, it was still unclear whether the AVE is the exclusive origin of anterior neural induction or if this activity is recapitulated by the node and/or its derivatives. We report here that nodes from both rabbit and mouse embryos can induce a complete neural axis including forebrain structures upon grafting to chick hosts. Thus, in rabbits and mice not only the AVE, but also the node, possesses a potential for the induction of anterior neural tissue.     

Roles of organizer factors and BMP antagonism in mammalian forebrain establishment

A critical question in mammalian development is how the forebrain is established. In amphibians, bone morphogenetic protein (BMP) antagonism emanating from the gastrula organizer is key. Roles of BMP antagonism and the organizer in mammals remain unclear. Anterior visceral endoderm (AVE) promotes early mouse head development, but its function is controversial. Here, we explore the timing and regulation of forebrain establishment in the mouse. Forebrain specification requires tissue interaction through the late streak stage of gastrulation. Foxa2(-/-) embryos lack both the organizer and its BMP antagonists, yet about 25% show weak forebrain gene expression. A similar percentage shows ectopic AVE gene expression distally. The distal VE may thus be a source of forebrain promoting signals in these embryos. In wild-type ectoderm explants, AVE promoted forebrain specification, while anterior mesendoderm provided maintenance signals. Embryological and molecular data suggest that the AVE is a source of active BMP antagonism in vivo. In prespecification ectoderm explants, exogenous BMP antagonists triggered forebrain gene expression and inhibited posterior gene expression. Conversely, BMP inhibited forebrain gene expression, an effect that could be antagonized by anterior mesendoderm, and promoted expression of some posterior genes. These results lead to a model in which BMP antagonism supplied by exogenous tissues promotes forebrain establishment and maintenance in the murine ectoderm.     

A novel role for retinoids in patterning the avian forebrain during presomite stages

Retinoids, and in particular retinoic acid (RA), are known to induce posterior fates in neural tissue. However, alterations in retinoid signalling dramatically affect anterior development. Previous reports have demonstrated a late role for retinoids in patterning craniofacial and forebrain structures, but an earlier role in anterior patterning is not well understood. We show that enzymes involved in synthesizing retinoids are expressed in the avian hypoblast and in tissues directly involved in head patterning, such as anterior definitive endoderm and prechordal mesendoderm. We found that in the vitamin A-deficient (VAD) quail model, which lacks biologically active RA from the first stages of development, anterior endodermal markers such as Bmp2, Bmp7, Hex and the Wnt antagonist crescent are affected during early gastrulation. Furthermore, prechordal mesendodermal and prospective ventral telencephalic markers are expanded posteriorly, Shh expression in the axial mesoderm is reduced, and Bmp2 and Bmp7 are abnormally expressed in the ventral midline of the neural tube. At early somite stages, VAD embryos have increased cell death in ventral neuroectoderm and foregut endoderm, but normal cranial neural crest production, whereas at later stages extensive apoptosis occurs in head mesenchyme and ventral neuroectoderm. As a result, VAD embryos end up with a single and reduced telencephalic vesicle and an abnormally patterned diencephalon. Therefore, we propose that retinoids have a dual role in patterning the anterior forebrain during development. During early gastrulation, RA acts in anterior endodermal cells to modulate the anteroposterior (AP) positional identity of prechordal mesendodermal inductive signals to the overlying neuroectoderm. Later on, at neural pore closure, RA is required for patterning of the mesenchyme of the frontonasal process and the forebrain by modulating signalling molecules involved in craniofacial morphogenesis.     

A new method to transfect the hypoblast of the chick embryo reveals conservation of the regulation of an Otx2 enhancer between mouse and chick extraembryonic endoderm

Background

The mouse anterior visceral endoderm (AVE) and the chick hypoblast are thought to have homologous roles in the early stages of neural induction and primitive streak formation. In mouse, many regulatory elements directing gene expression to the AVE have been identified. However, there is no technique to introduce DNA into the chick hypoblast that would enable a comparison of their activity and this has hampered a direct comparison of the regulation of gene expression in the mouse and chick extraembryonic endoderm.

Results

Here we describe a new method to introduce DNA into the chick hypoblast, using lipofectamine-mediated transfection. We show that the hypoblast can be easily transfected and that it starts to express a luciferase reporter within 2 hours of transfection. The validity of technique is tested by following the movement and fate of hypoblast cells, which reveals their translocation to the anterior germinal crescent. We then introduce a vector containing GFP driven by the mouse VEcis-Otx2 enhancer (which directs gene expression to the mouse AVE) and we detect activity in the hypoblast.

Conclusion

The new technique for delivering expression constructs to the chick hypoblast will enable studies on gene activity and regulation to be performed in this tissue, which has proved difficult to transfect by electroporation. Our findings also reveal that regulatory elements that direct gene expression to the mouse AVE are active in chick hypoblast, supporting the idea that these two tissues have homologous functions.     

Anterior identity is established in chick epiblast by hypoblast and anterior definitive endoderm

Previous studies of head induction in the chick have failed to demonstrate a clear role for the hypoblast and anterior definitive endoderm (ADE) in patterning the overlying ectoderm, whereas data from both mouse and rabbit suggest patterning roles for anterior visceral endoderm (AVE) and ADE. Based on similarity of gene expression patterns, fate and a dual role in 'protecting' the prospective forebrain from caudalising influences of the organiser, the chick hypoblast has been suggested to be the homologue of the mouse anterior visceral endoderm. In support of this, when transplanted to chick embryos, the rabbit AVE induces anterior markers in the chick epiblast. To reevaluate the role of the hypoblast/ADE (lower layer) in patterning the chick ectoderm, we used rostral blastoderm isolates (RBIs) as an assay, that is, rostral regions of blastoderms transected at levels rostral to the node. RBIs are, therefore, free from the influences of Hensen's node and ingressing axial mesoderm - tissues that are able to induce Ganf, the earliest specific marker of anterior neural plate. We demonstrate, using such RBIs (or RBIs dissected to remove the lower layer with or without tissue replacement), that the hypoblast/ADE (lower layer) is required and sufficient for patterning anterior positional identity in the overlying ectoderm, leading to expression of Ganf in neuroectoderm. Our results suggest that patterning of anterior positional identity and specification of neural identity are separable events operating to pattern the rostral end of the early chick embryo. Based on this new evidence we propose a revised model for establishing anteroposterior polarity, neural specification and head patterning in the early chick that is consonant with that occurring in other vertebrates.     

Anterior endoderm and head induction in early vertebrate embryos

Early work on the formation of the vertebrate body axis indicated the existence of separate head- and trunk-inducing regions in Spemann's organizer of the amphibian gastrula. In mammals some head-organizing activity may be located in anterior visceral (extraembryonic) endoderm (AVE). By analogy, the equivalent structure in the Xenopus laevis gastrula, the anterior endoderm, has been proposed to be the amphibian head organizer. Here we review recent data that challenge this notion and indicate that the involvement of AVE in head induction seems to be an exclusively mammalian characteristic. In X. laevis and chick, it is the prechordal endomesoderm that is the dominant source of head-inducing signals during early gastrulation. Furthermore, head induction in mammals needs a combination of signals from anterior primitive endoderm, prechordal plate, and anterior ectoderm. Thus, despite the homology of vertebrate anterior primitive endoderm, a role in head induction seems not to be conserved.     

Posteriorization by FGF, Wnt, and retinoic acid is required for neural crest induction.

The neural crest is a unique cell population induced at the lateral border of the neural plate. Neural crest is not produced at the anterior border of the neural plate, which is fated to become forebrain. Here, the roles of BMPs, FGFs, Wnts, and retinoic acid signaling in neural crest induction were analyzed by using an assay developed for investigating the posteriorization of the neural plate. Using specific markers for the anterior neural plate border and the neural crest, the posterior end of early neurula embryos was shown to be able to transform the anterior neural plate border into neural crest cells. In addition, tissue expressing anterior neural plate markers, induced by an intermediate level of BMP activity, was transformed into neural crest by posteriorizing signals. This transformation was mimicked by bFGF, Wnt-8, or retinoic acid treatment and was also inhibited by expression of the dominant negative forms of the FGF receptor, the retinoic acid receptor, and Wnt signaling molecules. The transformation of the anterior neural plate border into neural crest cells was also achieved in whole embryos, by retinoic acid treatment or by use of a constitutively active form of the retinoic acid receptor. By analyzing the expression of mesodermal markers and various graft experiments, the expression of the mutant retinoic acid receptor was shown to directly affect the ectoderm. We thereby propose a two-step model for neural crest induction. Initially, BMP levels intermediate to those required for neural plate and epidermal specification induce neural folds with an anterior character along the entire neural plate border. Subsequently, the most posterior region of this anterior neural plate border is transformed into the neural crest by the posteriorizing activity of FGFs, Wnts, and retinoic acid signals. We discuss a unifying model where lateralizing and posteriorizing signals are presented as two stages of the same inductive process required for neural crest induction.     

Visceral endoderm mediates forebrain development by suppressing posteriorizing signals

The anterior visceral endoderm (AVE) has attracted recent attention as a critical player in mouse forebrain development and has been proposed to act as "head organizer" in mammals. However, the precise role of the AVE in induction and patterning of the anterior neuroectoderm is not yet known. Here we identified a 5'-flanking region of the mouse Otx2 gene (VEcis) that governs the transgene expression in the visceral endoderm. In transgenic embryos, VEcis-active cells were found in the distal visceral endoderm at 5.5 days postcoitus (dpc), had begun to move anteriorly at 5.75 dpc, and then became restricted to the AVE prior to gastrulation. The VEcis-active visceral endoderm cells exhibited ectodermal morphology distinct from that of the other endoderm cells and consisted of two cell layers at 5.75 dpc. In the Otx2(-/-) background, the VEcis-active endoderm cells remained distal even at 6.5 dpc when a primitive streak was formed; anterior definitive endoderm was not formed nor were any markers of anterior neuroectoderm ever induced. The Otx2 cDNA transgene under the control of the VEcis restored these Otx2(-/-) defects, demonstrating that Otx2 is essential to the anterior movement of distal visceral endoderm cells. In germ-layer explant assays between ectoderm and visceral endoderm, the AVE did not induce anterior neuroectoderm markers, but instead suppressed posterior markers in the ectoderm; Otx2(-/-) visceral endoderm lacked this activity. Thus Otx2 is also essential for the AVE to repress the posterior character. These results suggest that distal visceral endoderm cells move to the future anterior side to generate a prospective forebrain territory indirectly, by preventing posteriorizing signals.     

Endoderm induction by the organizer-secreted factors chordin and noggin in Xenopus animal caps.

Spemann's organizer has potent neural inducing and mesoderm dorsalizing activities in the Xenopus gastrula. A third activity, the organizer's ability to induce a secondary gut, has been difficult to analyze experimentally due to the lack of early gene markers. Here we introduce endodermin, a pan-endodermal gene marker, and use it to demonstrate that chordin (Chd), a protein secreted by the organizer region, is able to induce endodermal differentiation in Xenopus. The ability of chd, as well as that of noggin, to induce endoderm in animal cap explants is repressed by the ventralizing factor BMP-4. When FGF signaling is blocked by a dominant-negative FGF receptor in chd-injected animal caps, neural induction is inhibited and most of the explant is induced to become endoderm. The results suggest that proteins secreted by the organizer, acting together with known peptide growth factors, regulate differentiation of the endodermal germ layer.     

Dickkopf1 and the Spemann-Mangold head organizer

Work in amphibians indicates that inhibition of Wnt and BMP signals is essential for head development and that head induction by the Spemann-Mangold organizer may be mediated by secreted Wnt antagonists. Wnts are potent posteriorizing factors and antagonize the Spemann-Mangold organizer. Dickkopf1 (dkk1) encodes a secreted effector expressed in head organizing centers of Xenopus, mouse and zebrafish. It acts as a Wnt inhibitor and is able together with BMP inhibitors to induce the formation of ectopic embryonic heads in Xenopus. It anteriorizes both mesendoderm and neuroectoderm, promoting prechordal plate and forebrain fates. Injection of inhibitory antibodies leads to microcephaly and cyclopia. Dkk1 thus is an essential mediator of the vertebrate head organizer.     

Neural crest induction by paraxial mesoderm in Xenopus embryos requires FGF signals

At the border of the neural plate, the induction of the neural crest can be achieved by interactions with the epidermis, or with the underlying mesoderm. Wnt signals are required for the inducing activity of the epidermis in chick and amphibian embryos. Here, we analyze the molecular mechanisms of neural crest induction by the mesoderm in Xenopus embryos. Using a recombination assay, we show that prospective paraxial mesoderm induces a panel of neural crest markers (Slug, FoxD3, Zic5 and Sox9), whereas the future axial mesoderm only induces a subset of these genes. This induction is blocked by a dominant negative (dn) form of FGFR1. However, neither dnFGFR4a nor inhibition of Wnt signaling prevents neural crest induction in this system. Among the FGFs, FGF8 is strongly expressed by the paraxial mesoderm. FGF8 is sufficient to induce the neural crest markers FoxD3, Sox9 and Zic5 transiently in the animal cap assay. In vivo, FGF8 injections also expand the Slug expression domain. This suggests that FGF8 can initiate neural crest formation and cooperates with other DLMZ-derived factors to maintain and complete neural crest induction. In contrast to Wnts, eFGF or bFGF, FGF8 elicits neural crest induction in the absence of mesoderm induction and without a requirement for BMP antagonists. In vivo, it is difficult to dissociate the roles of FGF and WNT factors in mesoderm induction and neural patterning. We show that, in most cases, effects on neural crest formation were parallel to altered mesoderm or neural development. However, neural and neural crest patterning can be dissociated experimentally using different dominant-negative manipulations: while Nfz8 blocks both posterior neural plate formation and neural crest formation, dnFGFR4a blocks neural patterning without blocking neural crest formation. These results suggest that different signal transduction mechanisms may be used in neural crest induction, and anteroposterior neural patterning.     

FGF signaling and the anterior neural induction in Xenopus

We previously showed that FGF was capable of inducing Xenopus gastrula ectoderm cells in culture to express position-specific neural markers along the anteroposterior axis in a dose-dependent manner. However, conflicting results have been obtained concerning involvement of FGF signaling in the anterior neural induction in vivo using the same dominant-negative construct of Xenopus FGF receptor type-1 (delta XFGFR-1 or XFD). We explored this issue by employing a similar construct of receptor type-4a (XFGFR-4a) in addition, since expression of XFGFR-4a was seen to peak between gastrula and neurula stages, when the neural induction and patterning take place, whereas expression of XFGFR-1 had not a distinct peak during that period. Further, these two FGFRs are most distantly related in amino acid sequence in the Xenopus FGFR family. When we injected mRNA of a dominant-negative version of XFGFR-4a (delta XFGFR-4a) into eight animal pole blastomeres at 32-cell stage, anterior defects including loss of normal structure in telencephalon and eye regions became prominent as examined morphologically or by in situ hybridization. Overexpression of delta XFGFR-1 appeared far less effective than that of delta XFGFR-4a. Requirement of FGF signaling in ectoderm for anterior neural development was further confirmed in culture: when ectoderm cells that were overexpressing delta XFGFR-4a were cocultured with intact organizer cells from either early or late gastrula embryos, expression of anterior and posterior neural markers was inhibited, respectively. We also showed that autonomous neuralization of the anterior-type observed in ectoderm cells that were subjected to prolonged dissociation was strongly suppressed by delta XFGFR-4a, but not as much by delta XFGFR-1. It is thus indicated that FGF signaling in ectoderm, mainly through XFGFR-4, is required for the anterior neural induction by organizer. We may reconcile our data to the current "neural default model," which features the central roles of BMP4 signaling in ectoderm and BMP4 antagonists from organizer, simply postulating that the neural default pathway in ectoderm includes constitutive FGF signaling step.     

Getting your head around Hex and Hesx1: forebrain formation in mouse

An increasing amount of evidence suggests that in mouse there are two signalling centres required for the formation of a complete neural axis: the anterior visceral endoderm (AVE), and the node and its derivatives. Embryological and genetic studies suggest that the AVE has a head-inducing activity. In contrast, the node appears to act first as a head inducer in synergy with the AVE initiating anterior neural patterning at early stages of mouse development, and later, node derivatives are necessary for maintenance and embellishment of anterior neural character. Hex and Hesx1 are homeobox genes that are expressed in relevant tissues involved in anterior patterning. The analysis of the Hex and Hesx1 mutant mice has revealed that the lack of these genes has little or no effect on the early steps of anterior neural induction. However, both genes are required subsequently for the proper expansion of the forebrain region. We suggest that disturbance in the specification of an Fgf8 signalling centre in the anterior neural ridge may account for the anterior defects observed in these mutants.     

Hensen's node induces neural tissue in Xenopus ectoderm. Implications for the action of the organizer in neural induction.

The development of the vertebrate nervous system is initiated in amphibia by inductive interactions between ectoderm and a region of the embryo called the organizer. The organizer tissue in the dorsal lip of the blastopore of Xenopus and Hensen's node in chick embryos have similar neural inducing properties when transplanted into ectopic sites in their respective embryos. To begin to determine the nature of the inducing signals of the organizer and whether they are conserved across species we have examined the ability of Hensen's node to induce neural tissue in Xenopus ectoderm. We show that Hensen's node induces large amounts of neural tissue in Xenopus ectoderm. Neural induction proceeds in the absence of mesodermal differentiation and is accompanied by tissue movements which may reflect notoplate induction. The competence of the ectoderm to respond to Hensen's node extends much later in development than that to activin-A or to induction by vegetal cells, and parallels the extended competence to neural induction by axial mesoderm. The actions of activin-A and Hensen's node are further distinguished by their effects on lithium-treated ectoderm. These results suggest that neural induction can occur efficiently in response to inducing signals from organizer tissue arrested at a stage prior to gastrulation, and that such early interactions in the blastula may be an important component of neural induction in vertebrate embryos.