Tissue boundaries and cell behavior during neurulation.

We have analyzed the dynamics of the boundaries between the neural plate and the epidermis and between the neural plate and the notoplate. Our experiments confirm that these two boundaries have important roles in neurulation. Measurements of the lengths of neural fold (the boundary between epidermis and neural plate) in embryos of axolotls and newts reveal that neural folds abutting the prospective brain decrease in length while neural folds abutting the prospective spinal cord increase in length during neurulation. We tested the proposition that boundaries of the neural plate with epidermis and with notoplate are essential for proper neurulation. Cuts made along the boundaries with epidermis or with notoplate stop, or greatly diminish, neural plate elongation and tube formation. Explanting the axolotl neural plate without any bordering epidermis stops plate elongation and prevents neural tube closure, but neural plates explanted with a rim of epidermis elongate and close into tubes. Cutting the notoplate boundary stops midline elongation in the newt embryo or diminishes it in the axolotl embryo. We conclude that the notoplate boundary and part of the boundary of the epidermis that abuts the prospective spinal cord organize cell behavior to elongate the neural plate and help close the neural tube. The boundary of the neural plate with the epidermis is essential for tube closure both because it organizes plate elongation in the spinal cord region and because cell behavior becomes organized at the boundary such that neural folds are raised and a rolling moment is produced that helps form the neural tube.

Normal neurulation in amphibians.

How does cell behaviour accomplish neurulation in amphibian embryos? During neurulation, the neural plate (while preserving the same volume) doubles its length, triples its thickness, narrows 10-fold, greatly decreases its surface and rolls into a tube. Cells that compose the neural plate produce these changes in three ways. They change shape, change neighbours and attempt to crawl beneath the contiguous epidermis. Plate width, length and area are decreased and the plate thickens when apical surfaces of plate cells contract radially, but plate length increases and width is further decreased when cells reposition themselves and collect along plate boundaries. Contraction of the apical surfaces of plate cells also helps roll the plate into a tube. Poisson buckling resulting from elongation of plate borders may contribute bending forces that help tube formation. The main folding force in tube formation is a rolling moment toward the midline produced by neural plate cells attempting to crawl beneath the contiguous epidermis. Experiments, observations and computer simulations support these assertions, reveal the organization of cell behaviour and implicate contraction of actin filaments as the main source of the necessary forces.

Evidence that the border of the neural plate may be positioned by the interaction between signals that induce ventral and dorsal mesoderm.

In the early Xenopus embryo, a quadrant of endodermal cells that have descended from the vegetal dorsal localization in the zygote produces signals that pass into the animal hemisphere and induce dorsal mesoderm from the marginal zone. From the remaining three quadrants of the bordering endoderm, signals pass into the animal hemisphere and induce ventral mesoderm in the marginal region. There is evidence that suggests that these same mesoderm-inducing signals continue through the plane of the tissue of the animal hemisphere where they may at least begin the processes of neural and epidermal induction by changing the competence of the prospective ectodermal cells, and possibly influencing the early regional biasing of later expression of at least some gene products, such as Epi-1 whose expression in the future epidermal domain seems specified before gastrulation. We hypothesized that the interaction of the ventral and dorsal signals within the plane of the tissue of the animal hemisphere may position the border of the neural plate. If this is so, then transplantation into the animal pole of cells that signal induction of ventral mesoderm should drive the neural plate boundary back toward the blastopore and shorten the anterior-posterior axis. Removal of cells that induce ventral mesoderm should result in an axis that is longer than normal. Results of our experiments support these predictions. Also, by late pregastrula stage 9, increasing the ventral signals has no effect. Thus the evidence suggests that the position of the anterior neural plate boundary is established before gastrulation begins by the interaction of the signals that induce the ventral and dorsal mesoderm.

Femtosecond pulse generation in a Ti:A1(2)O(3) laser by using second- and third-order intracavity dispersion.

We investigate the effects of third-order dispersion on pulse width by using a femtosecond Ti:A1(2)O(3) laser with independently adjustable second- and third-order intracavity dispersion compensation. A novel technique for compensating third-order dispersion is demonstrated using Gires-Tournois interferometers that are fabricated monolithically by using multilayer dielectric films. With the addition of intracavity third-order compensation, pulse-width reduction from 45 to 28 fs is achieved. The dispersion compensation effect produced by the Gires-Tournois interferometers is measured in situ using frequency-domain dispersion measurement techniques.

The origins of neural crest cells in the axolotl.

We address the question of whether neural crest cells originate from the neural plate, from the epidermis, or from both of these tissues. Our past studies revealed that a neural fold and neural crest cells could arise at any boundary created between epidermis and neural plate. To examine further the formation of neural crest cells at newly created boundaries in embryos of a urodele (Ambystoma mexicanum), we replace a portion of the neural folds of an albino host with either epidermis or neural plate from a normally pigmented donor. We then look for cells that contain pigment granules in the neural crest and its derivatives in intact and sectioned host embryos. By tracing cells in this manner, we find that cells from neural plate transplants give rise to melanocytes and (in one case) become part of a spinal ganglion, and we find that epidermal transplants contribute cells to the spinal and cranial ganglia. Thus neural crest cells arise from both the neural plate and the epidermis. These results also indicate that neural crest induction is (at least partially) governed by local reciprocal interactions between epidermis and neural plate at their common boundary.

The restriction of the heart morphogenetic field in Xenopus laevis.

We have examined the spatial restriction of heart-forming potency in Xenopus laevis embryos, using an assay system in which explants or explant recombinates are cultured in hanging drops and scored for the formation of a beating heart. At the end of neurulation at stage 20, the heart morphogenetic field, i.e., the area that is capable of heart formation when cultured in isolation, includes anterior ventral and ventrolateral mesoderm. This area of developmental potency does not extend into more posterior regions. Between postneurula stage 23 and the onset of heart morphogenesis at stage 28, the heart morphogenetic field becomes spatially restricted to the anterior ventral region. The restriction of the heart morphogenetic field during postneurula stages results from a loss of developmental potency in the lateral mesoderm, rather than from ventrally directed morphogenetic movements of the lateral mesoderm. This loss of potency is not due to the inhibition of heart formation by migrating neural crest cells. During postneurula stages, tissue interactions between the lateral mesoderm and the underlying anterior endoderm support the heart-forming potency in the lateral mesoderm. The lateral mesoderm loses the ability to respond to this tissue interaction by stages 27-28. We speculate that either formation of the third pharyngeal pouch during stages 23-27 or lateral inhibition by ventral mesoderm may contribute to the spatial restriction of the heart morphogenetic field.

The role of the dorsal lip in the induction of heart mesoderm in Xenopus laevis.

We have examined the tissue interactions responsible for the expression of heart-forming potency during gastrulation. By comparing the specification of different regions of the marginal zone, we show that heart-forming potency is expressed only in explants containing both the dorsal lip of the blastopore and deep mesoderm between 30 degrees and 45 degrees lateral to the dorsal midline. Embryos from which both of these 30 degrees-45 degrees dorsolateral regions have been removed undergo heart formation in two thirds of cases, as long as the dorsal lip is left intact. If the dorsal lip is removed along with the 30 degrees-45 degrees regions, heart formation does not occur. These results indicate that the dorsolateral deep mesoderm must interact with the dorsal lip in order to express heart-forming potency. Transplantation of the dorsal lip into the ventral marginal zone of host embryos results in the formation of a secondary axis; in over half of cases, this secondary axis includes a heart derived from the host mesoderm. These findings suggest that the establishment of heart mesoderm is initiated by a dorsalizing signal from the dorsal lip of the blastopore.

Neural fold formation at newly created boundaries between neural plate and epidermis in the axolotl.

According to a recent model, the cortical tractor model, neural fold and neural crest formation occurs at the boundary between neural plate and epidermis because random cell movements become organized at this site. If this is correct, then a fold should form at any boundary between epidermis and neural plate. To test that proposition, we created new boundaries in axolotl embryos by juxtaposing pieces of neural plate and epidermis that would not normally participate in fold formation. These boundaries were examined superficially and histologically for the presence of folds, permitting the following observations. Folds form at each newly created boundary, and as many folds form as there are boundaries. When two folds meet they fuse into a hollow "tube" of neural tissue covered by epidermis. Sections reveal that these ectopic folds and "tubes" are morphologically similar to their natural counterparts. Transplanting neural plate into epidermis produces nodules of neural tissue with central lumens and peripheral nerve fibers, and transplanting epidermis into neural plate causes the neural tube and the dorsal fin to bifurcate in the region of the graft. Tissue transplanted homotypically as a control integrates into the host tissue without forming folds. When tissue from a pigmented embryo is transplanted into an albino host, the presence of pigment allows the donor cells to be distinguished from those of the host. Mesenchymal cells and melanocytes originating from neural plate transplants indicate that neural crest cells form at these new boundaries. Thus, any boundary between neural plate and epidermis denotes the site of a neural fold, and the behavior of cells at this boundary appears to help fold the epithelium. Since folds can form in ectopic locations on an embryo, local interactions rather than classical neural induction appear to be responsible for the formation of neural folds and neural crest.

The specification of heart mesoderm occurs during gastrulation in Xenopus laevis.

The establishment of heart mesoderm during Xenopus development has been examined using an assay for heart differentiation in explants and explant combinations in culture. Previous studies using urodele embryos have shown that the heart mesoderm is induced by the prospective pharyngeal endoderm during neurula and postneurula stages. In this study, we find that the specification of heart mesoderm must begin well before the end of gastrulation in Xenopus embryos. Explants of prospective heart mesoderm isolated from mid- or late neurula stages were capable of heart formation in nearly 100% of cases, indicating that the specification of heart mesoderm is complete by midneurula stages. Moreover, inclusion of pharyngeal endoderm had no statistically significant effect upon either the frequency of heart formation or the timing of the initiation of heartbeat in explants of prospective heart mesoderm isolated after the end of gastrulation. When the superficial pharyngeal endoderm was removed at the beginning of gastrulation, experimental embryos formed hearts, as did explants of prospective heart mesoderm from such embryos. These results indicate that the inductive interactions responsible for the establishment of heart mesoderm occur prior to the end of gastrulation and do not require the participation of the superficial pharyngeal endoderm.

A survey by scanning electron microscopy of the extracellular matrix and endothelial components of the primordial chick heart.

Scanning electron microscopy was used to study the morphogenesis of the primitive embryonic chick heart (stage 5 late primitive streak through stage 9+). Components of the developing heart (myocardium, endocardial endothelium, and extracellular matrix) were viewed from the ventral surface after removal of the endoderm. The myocardial component of the heart can first be seen by light microscopy at stage 5 as two darker oval-shaped areas located on either side of the embryonic axis in the cranial region of the embryo. Scanning electron microscopy demonstrates that as early as stage 6 an area of extracellular matrix, enriched in comparison to more lateral and medial splanchnic mesoderm, can be identified ventral to the myocardial primordium. As heart formation progressed we observed primordial endothelial elements in the splanchnic mesoderm lateral to the myocardial primordia. By late stage 7 these lateral primordial elements had anastomosed into small, loose plexuses. This process of anastomosis progressed rapidly, and by stage 8 the entire cranial surface of the myocardial primordium was covered with vascular plexuses. By late stage 8 the progressive fusion of these plexuses resulted in the formation of large multiple tubular elements near the midline. More medially the fusion of tubular elements resulted in a continuous endothelial sheet at the midline.

Features of embryonic induction.

The patterned distribution of different organs in the amphibian embryo begins with the establishment of two domains, the animal and vegetal regions, that differ in developmental potency. Differences amplify as inductive interactions occur across boundaries between areas of different potency. Embryonic induction establishes a temporally and spatially dynamic area of developmental potency - a morphogenetic field. The final arrangement and differentiation of cell types within the field emerge from subsequent interactions occurring primarily within the field. These principles are illustrated in a review of the induction of the lens and the heart. Recent studies show that the induction of the lens of the eye and the induction of the heart begin early in development. Most of lens inductions occurs before the formation of the optic vesicle, and the heart appears to be part of a complex of dorsal structures whose formation is dependent upon the establishment of the dorsoventral axis. Suppressive as well as inductive tissue interactions occur during the determination of both of these organs, affecting their position and time of appearance. The complex processes of induction defined by the past nine decades of experimental work present many challenging questions that can now be addressed, especially in terms of the molecular events, cellular behaviour and regulatory physiology of the responding tissue.

Somitomeres: mesodermal segments of vertebrate embryos.

Well before the somites form, the paraxial mesoderm of vertebrate embryos is segmented into somitomeres. When newly formed, somitomeres are patterned arrays of mesenchymal cells, arranged into squat, bilaminar discs. The dorsal and ventral faces of these discs are composed of concentric rings of cells. Somitomeres are formed along the length of the embryo during gastrulation, and in the segmental plate and tail bud at later stages. They form in strict cranial to caudal order. They appear in bilateral pairs, just lateral to Hensen's node in the chick embryo. When the nervous system begins to form, the brain parts and neuromeres are in a consistent relationship to the somitomeres. Somitomeres first appear in the head, and the cranial somitomeres do not become somites, but disperse to contribute to the head the same cell types contributed by somites in the trunk region. In the trunk and tail, somitomeres gradually condense and epithelialize to become somites. Models of vertebrate segmentation must now take into account the early presence of these new morphological units, the somitomeres. Somitomeres were discovered in the head of the chick embryo (Meier, 1979), with the use of stereo scanning electron microscopy. The old question of whether the heads of the craniates are segmented is now settled, at least for the paraxial mesoderm. Somitomeres have now been identified in the embryos of a chick, quail, mouse, snapping turtle, newt, anuran (Xenopus) and a teleost (the medaka). In all forms studied, the first pair of somitomeres abut the prosencephalon, but caudal to that, for each tandem pair of somitomeres in the amniote and teleost, there is but one somitomere in the amphibia. The mesodermal segments of the shark embryo are arranged like those of the amphibia.

Mesodermal metamerism in the teleost, Oryzias latipes (the medaka).

Previous studies of the metameric pattern in mesodermal tissues of chick, mouse, turtle, and amphibian embryos have indicated that segmental characteristics exist along the entire length of the embryo. This paper describes this phenomenon in a fish embryo, for some differences in the cranial segmental plan exist between the anamniote and the amniote embryos hitherto studied. Embryos of the cyprinodont, Oryzias latipes, were fixed at various times, the examined by means of stereo scanning electron microscopy. As in other vertebrate embryos, the first indication of mesodermal metamerism in this fish embryo is the occurrence of somitomeres, which are orderly, tandemly arranged units of uncondensed mesenchymal cells in the paraxial mesoderm. As many as ten somitomeres can be observed caudal to the last formed somite to the elongating tail region. In addition, 7 somitomeres are present rostral to the first definitive somite, which is segment number eight. As in other vertebrate embryos examined, somitomeres in Oryzias embryos are circular, bilaminar arrays of paraxial mesoderm that form before any indications of segmentation can be seen with the light microscope. In the trunk region these mesodermal units condense to give rise to definitive somites, but in the head they eventually disperse. Despite a fundamentally different mode of gastrulation and a relatively small number of cells in the newly formed somitomeres, cranial segmentation in Oryzias embryos was found to be more similar in number to the metameric pattern of the embryos of the bird, reptile, and mammal than to the situation found in the two amphibians studied thus far.

Neurulation and the cortical tractor model for epithelial folding.

We present here a new model for epithelial morphogenesis, which we call the 'cortical tractor model'. This model assumes that the motile activities of epithelial cells are similar to those of mesenchymal cells, with the added constraint that the cells in an epithelial sheet remain attached at their apical circumference. In particular, we assert that there is a time-averaged motion of cortical cytoplasm which flows from the basal and lateral surfaces to the apical region. This cortical flow carries with it membrane and adhesive structures that are inserted basally and resorbed apically. Thus the apical seal that characterizes epithelial sheets is a dynamic structure: it is continuously created by the cortical flow which piles up components near where they are recycled in the apical region. By use of mechanical analyses and computer simulations we demonstrate that the cortical tractor motion can reproduce a variety of epithelial motions, including columnarization (placode formation), invagination and rolling. It also provides a mechanism for driving active cell rearrangements within an epithelial sheet, while maintaining the integrity of the apical seal. Active repacking of epithelial cells appears to drive a number of morphogenetic processes. Neurulation in amphibians provides an example of a process in which all four of the above morphogenetic movements appear to play a role. Here we reexamine the process of neurulation in amphibians in light of the cortical tractor model, and find that it provides an integrated view of this important morphogenetic process.

Morphogenesis of the head of a newt: mesodermal segments, neuromeres, and distribution of neural crest.

Segmentation of the mesoderm in the head of a newt embryo is revealed by scanning electron microscopy. By the end of gastrulation, the newt embryo is already segmented from one end to the other, with additional segments added later by the tail bud. This metameric segmentation appears long before the first "somite" can be seen in the late neurula by light microscopy. The six segments found in the newt head look much like the six most-cranial segments described decades ago in shark embryos. Mesodermal segments in the newt head are similar to somitomeres in amniote embryos, but in amniote embryos, the numbers and relationships of head segments are quite different from those of the newt. In both amniote and newt, the first segment abuts the prosencephalon, but for each more caudal head segment, where the newt embryo has one segment, the amniote has two. Although the pattern and distribution of cranial neural crest is quite similar in newt and amniote embryos, there are different relationships between migrating crest masses and mesodermal segments due to the doubling of most of the cranial segments in amniotes. It now appears that all vertebrate embryos, regardless of their mode of gastrulation, form similar mesodermal segments from one end of the embryo to the other, and this metameric pattern is established during gastrulation.

Further evidence that formation of the neural tube requires elongation of the nervous system.

In previous papers, we have correlated rapid elongation of the midline of the neural plate with the time of closure of the plate into a tube in the newt embryo and at one stage of the chick embryo. We proposed a model in which stretching of the midline of the plate causes the plate to buckle out of the plane and roll into a tube. In this paper, I show for another stage of development in the chick embryo, the period of closure of the brain tube, that rapid elongation of the nervous system accompanies closure of the tube. If elongation of the brain plate causes formation of the tube, then treatments that stop tube formation should also stop brain elongation. I tested this hypothesis by using low fluences of UV irradiation, known to stop tube formation (Davis, '44), and measuring the effects on elongation of the brain plate. The open plates of UV-irradiated embryos failed to elongate normally. Furthermore, photoreactivation with longer wavelengths of light reversed the UV effects and allowed closure of the tube in UV-irradiated embryos. These embryos elongated their brains.

Experimental studies of the origin and expression of metameric pattern in the chick embryo.

On either side of Hensen's node of the fully extended primitive streak of the chick embryo (stage 4) the mesoderm is already organized into circular domains called somitomeres. As Hensen's node regresses, paraxial somitomeres are added in tandem and are early morphological representatives of metameric pattern in the mesoderm. These organized circular domains of mesenchyme cells are best visualized with stereo pair scanning electron microscopy. Experiments suggested that a prepattern of segmentation exists in and around the fully extended primitive streak. Streaks divested of Hensen's node can generate some paraxial somites, but only if surgically split down the midline. We assessed metameric pattern formation in nodeless streaks, both severed and unsevered down the midline. Operated blastoderms were cultured 15 hours, fixed, dissected, processed for scanning electron microscopy, and photographed in stereo. Split nodeless streaks produced a cranial to caudal sequence of somitomeric development. This sequence is similar to the sequential maturational events seen in the segmental plate of older embryos. The least mature somitomeres, toward the posterior end of the severed edge, appear as circular domains of radially oriented cells, looking much like the first somitomeres to emerge near Hensen's node of the stage 4 streak. More cranially along the severed edge, somitomeres are morphologically more mature, being more condensed, with cells oriented about a central myocoele. At the most cranial end of the severed piece, somitomeres are the most mature, having contracted about their centers to create intersomitomeric gaps that permit their identification with light microscopy as individual "somites." Embryos from which the node was removed, but the streak left intact, generated only the most primitive somitomeric pattern repetitively along either side of the primitive groove. We conclude that regression of Hensen's node provides for the timely initiation of morphogenesis of somitomeres from a prepattern of segmentation that already exists.

Differentiation of the metameric pattern in the embryonic axis of the mouse. II. Somitomeric organization of the presomitic mesoderm.

The formation of the embryonic axis is brought about by the continuous recruitment of cells from the primitive streak, and at later stages from the tail bud. Presumptive somitic cells are first incorporated into presomitic mesoderm before they emerge as metamerically arranged somites. When the presomitic mesoderm was examined in stereo with the scanning electron microscope (SEM), mesenchymal cells were found to be already organized into segmental units. These segmental units are called somitomeres because of their striking similarity to structures in the embryonic axis of the chick embryo described by Meier [16]. Cells within the somitomere are arranged in concentric whorls about a core center, bisected by a medio-lateral seam which subdivides the cell population into anterior and posterior halves. The concentric configuration of the cells is most easily observed along the medial face of the presomitic mesoderm when it is generally wedge-shaped. Even tough the units are tandemly contiguous, somitomeric interfaces are distinguished by abrupt change in cellular orientation. Despite a nearly two-fold fluctuation in the overall size of the presomitic mesoderm during embryonic development, a relatively constant number of somitomeres (six) is found in tandem sequence. Somitomeric maturation culminating in somite formation involves compaction of the cell population, more orderly alignment of cells, reduction in extracellular space, and changes in the shape of the somitomere concomitant with neurulation. Though the more mature somitomere is about 70% the size of the most recently formed somitomere at the caudal end of the presomitic mesoderm, the average size of each somitomere is adjusted proportionally to the overall length of the presomitic mesoderm. In vitro culture of the presomitic mesoderm shows a direct developmental lineage between the somitomere and the somite, suggesting that somite formation is a morphologic manifestation of a somitomeric pattern laid down at an earlier stage in development. The somitomeric pattern in the paraxial mesoderm is the earliest recognizable morphologic evidence of metamerism in the embryonic axis. This pattern is later emulated by other tissues that are topographically associated with the paraxial mesoderm.

Cephalic flexure formation in the chick embryo.

The cephalic flexure, found in all vertebrate brains, is a ventrally directed bend through the mesencephalon, and a ventral bulging and elongation of the prosencephalon. Most sources say the cephalic flexure is caused by differential growth. We have measured the changing angle of flexure through time and find that flexure occurs between chick embryo stages 10 to 15. We measured, during these stages, the lengths, thicknesses, and volumes of the floor and roof of the mesencephalon and of the prosencephalon. As expected, during flexure the mesencephalic roof elongates much more than the floor. Both roof and floor increase in thickness, and mesencephalic roof volume increases twice as much as floor volume. However, prosencephalon, which does not bend, also has differential growth between roof and floor, but the growth is taken up in complex changes of shape other than flexure. There are sufficient numbers of mitoses in the brain to account for the observed tissue growth, assuming accompanying cell enlargement. We deleted brain parts adjacent to the mesencephalon before flexure and the mesencephalon bent, so migration of cells from or to these adjacent parts to contribute to the differential growth of the mesencephalon is unlikely. We reduced cerebrospinal fluid pressure during flexure by explanting heads to the chorioallantoic membrane, or into New cultures. The mesencephalon of explanted heads bends, but the prosencephalon fails to elongate. We conclude that differential growth may be necessary for mesencephalic flexure in the chick embryo, but other factors that decide the disposition of the products of growth in space must determine the shape.