The neural induction process; its morphogenetic aspects.

This posthumous review of early embryonic inductions concludes: 1) the amphibian egg has only two distinct components, animal and vegetal. Interactions at their mutual boundary forms meso-endoderm. This is "meso-endoderm induction", not just "mesoderm induction". 2) The dorso-ventral polarity of the yolk mass implies a dorsally situated inducing centre. 3) Accumulation of cells into one, two, three or many cell masses [problastopores] along the circumference of the meso-endoderm results in as many axes, implying a self-organizing capacity of meso-endoderm. 4) Induction of the meso-endoderm is slow, spreading cell to cell through the animal moiety from the boundary of the vegetal yolk mass towards the animal pole. 5) Interaction between mesoderm and ectoderm is a separate step leading to cranio-caudal differentiation of the archenteron roof. 6) The initial invaginating endoderm and mesoderm, representing the future pharynx endoderm and prechordal plate mesoderm, first contacts the most posterior presumptive neurectoderm after having passed the still uninvaginated trunk mesoderm. At that moment an antero-posterior level neural induction actually starts. 7) The ectoderm contraction wave coincides spatially and temporally with the induced neural plate. 8) Two successive homoiogenetic waves of inductive activity pass through the presumptive neurectoderm in the anterior direction, the first one, "activation", giving rise to neural differentiation and ultimately forebrain, the second one, "transformation", to more caudal CNS structures. These are separate, successive steps in CNS regional induction. 9) The midbrain represents a secondary formation in the neural plate. 10) The observed changes in morphogenesis may depend upon separate, successive binary decisions via [cell and] nuclear state splitters [involving differentiation waves].

Short historical survey of pattern formation in the endo-mesoderm and the neural anlage in the vertebrates: the role of vertical and planar inductive actions.

After some introductory remarks about vertical versus horizontal inductive interactions and about planar versus homoiogenetic induction, the author discusses: a) the historical development of the more recently studied endo-mesoderm induction in the Urodeles and in the anuran Xenopus laevis, b) the possible causal relationship between endo-mesoderm induction and the initiation of the gastrulation process, and c) the older history of the regional neural induction as initially studied in the Urodeles and only recently analysed in the anuran Xenopus laevis. The essential vertical interaction in the neural induction process both in urodelian and in anuran amphibians is emphasized.

Surface contraction and expansion waves correlated with differentiation in axolotl embryos. II. In contrast to urodeles, the anuran Xenopus laevis does not show furrowing surface contraction waves.

We have observed a number of contraction waves traversing the axolotl (Ambystoma mexicanum) embryo (a urodelan amphibian) from the midblastula transition up to at least neural tube closure, and wished to learn if similar "differentiation waves" appear on the popular laboratory anuran amphibian, the South African clawed toad, Xenopus laevis. Time lapse video microscopy showed that no contraction waves are visible on the surface of Xenopus from gastrulation through neurulation. It is possible that cell intercalations in the double-layered ectoderm of the Xenopus embryo are homologous to the surface waves in the single layered ectoderm of the axolotl embryo. In any case, a simple, universal correspondence between surface waves and induction phenomena and differentiation does not exist.

Vertical versus planar induction in amphibian early development.

In the Urodeles, the archenteron roof invaginates as a single continuous sheet of cells, vertically inducing the neural anlage in the overlying ectoderm during invagination. The induction comprises first the activation process, leading, to forebrain differentiation tendencies, and then the superimposed transformation process, which changes presumptive forebrain development into that of hindbrain and spinal cord acting with a caudally increasing intensity. The activating action, being maximal anteriorly, decreases caudally to nearly zero. In the double-layered Xenopus embryo, the internal mesodermal marginal zone shows much more independent and earlier regional segregation and involution than the external marginal zone in the Urodeles; its prechordal mesoderm already initiating vertical neural induction in overlying ectoderm at stages 10 to 10+ before any visible archenteron invagination. In Xenopus incomplete exogastrulae the prechordal mesoderm involutes normally prior to evagination of the endoderm and mesodem. Artificially produced Xenopus total exogastrulae, made at stage 9 before mesoderm involution, behave just like axolotl total exogastrulae, showing no neural differentiation. The notion of planar neural induction in Xenopus can only be applied in exogastrulae and Keller explants for the transforming action, which is maximal in the caudal archenteron roof. In normal Xenopus development, the formation of the entire nervous system is essentially due to vertical induction by the successively involuting prechordal and notochordal mesoderm. The different behavior of Xenopus embryos in comparison with Urodele embryos can essentially be explained by the double-layered character of the animal moiety of the Xenopus embryo.

Retinoic acid causes an anteroposterior transformation in the developing central nervous system.

All-trans retinoic acid (RA) is well known as a biologically active form of vitamin A and a teratogen. The identification of nuclear receptors for this ligand suggests strongly that it is an endogenous signal molecule, and measurements of RA and teratogenic manipulations suggest further that RA is a morphogen specifying the anteroposterior axis during limb development. Besides the limb, RA and other retinoids affect development of other organs, including the central nervous system (CNS). None of these other effects has been investigated in detail. Our purpose here was to begin analysing the effects of RA on CNS development in Xenopus laevis. We find that RA acts on the developing CNS, transforming anterior neural tissue to a posterior neural specification. These and other findings raise the possibility that RA mediates an inductive interaction regulating anteroposterior differentiation within the CNS. Following recent reports implicating transforming growth factor-beta 2-like and fibroblast growth factor-like factors in mesoderm induction, this indicates that a different type of signal molecule (working through a nuclear receptor, not a plasma membrane receptor) might mediate inductive cell interactions during early embryonic development.

Inductive interactions in early amphibian development and their general nature.

After a short discussion on cell interactions in general and inductive interactions in particular, the almost completely epigenetic nature of amphibian development is emphasized. In the symmetrized egg undergoing cleavage a large-scale inductive interaction occurs which leads to the formation of the meso-endoderm. Meso-endoderm formation gives rise to the morphogenic process of gastrulation. In the ensuing triple-layered embryo inductive interactions are strongly enhanced. The following large-scale inductive interaction leads to the formation of the neural anlage. This is again followed by the morphogenetic process of neurulation or neural tube formation. Subsequent interactions between the germ layers of the triple-layered embryo give rise to the formation of the regional pattern of organ anlagen. Finally, the most promising approaches to the nature of inductive interactions for mesoderm and endoderm formation are discussed.

Neural induction, a two-way process.

A quantitative analysis of Triturus alpestris recombinates of newly invaginated "caudal" archenteron roof with either competent gastrula ectoderm or non-competent neurula ectoderm has demonstrated that neuralisation evoked by the underlying archenteron roof affects, in its turn, the regional differentiation of the latter. It does so by enhancing its notochordal differentiation; so, neural induction is clearly a two-way process. Somite differentiation is not significantly different in the two series. This is especially remarkable since both series show positive correlations between the amounts of notochordal and somite structures. The explantation is that the correlation coefficients differ considerably in the two series. The possible reason for this is discussed. Moreover the recombinates with competent gastrula ectoderm demonstrate that the size of the neural structures is determined by the total mass of the axial mesoderm and not by that of the notochord or that of the somites separately.

A diffusion model for mesoderm induction in amphibian embryos.

In this paper we try to answer the question whether diffusion is a possible mechanism to explain mesoderm induction in Amphibians. First the embryological data are discussed and a hypothesis for mesoderm formation is set forth. The blastula being essentially a hollow sphere, we assume that the induction mechanism in an embryo at the blastula stage can be simulated by diffusion-reaction processes on spherical surfaces. A model is constructed for the simple case when the source is held constant with respect to time, the decay proportional to the concentration and the diffusion coefficient a constant. From simulation we find a (best) value for the decay constant to be 6 x 10(-5)/sec and for the diffusion constant to be 0.24 x 10(-6) cm2/sec. The relation between the parameters is derived from an analytic solution for the diffusion process on a spherical surface with a continuously producing point source and the concentration proportional to the decay. The form and regulative properties of the steady concentration gradient are discussed.

Embryological evidence for a possible polyphyletic origin of the recent amphibians.

The markedly different mode of mesoderm formation in anuran and urodelan amphibians (which is related to the early double-layered nature of the anuran blastula wall in contrast to its single-layered nature in the urodeles), but particularly the fundamentally different place and mode of origin of the primordial germ cells in the two groups of amphibians, strongly pleads in favour of a very ancient bifurcation in the phylogenetic history of the two groups, even suggesting a polyphyletic origin from different ancestral fishes.

The induction of the primordial germ cells in the urodeles.

Xenoplastic recombinates of animal ectodermal caps with the ventral vegetative yolk mass of blastulae of stage 81/2-83/4 ofA. mexicanum, T. alpestris, T. cristatus carnifex andP. waltlii have demonstrated unambiguously that in the urodeles the primordial germ cells-together with other ventro-caudal mesodermal structures-develop from the animal ectodermal moiety of the blastula under an inductive influence emanating from the ventral vegetative yolk mass. Similar recombinates of3H-labeled and unlabeled ectodermal and endodermal components fully support this conclusion.Recombinates of the ventral vegetative yolk mass with different regions of the animal ectodermal hemisphere show that primordial germ cells can be formed by any region of the animal ectodermal hemisphere, including those regions which in normal development will never form them. The number of primordial germ cells formed differs significantly among the various regions, that of the ventral peripheral region being the highest and that of the central, animal region the lowest. The capacity for primordial germ cell formation shows two increasing gradients, one animal-vegetative and the other dorse-ventral (in the peripheral zone). Although accurate measurements could not be made, there seems to be a relation between the number of primordial germ cells formed and the amount of ventro-caudal mesoderm induced.The experiments, moreover, show that notochord differentiation largely or entirely suppresses primordial germ cell formation. Notochord differentiation shows a similar animalvegetative, but an opposite ventro-dorsal increase in frequency (in the peripheral zone) as compared with the capacity for primordial germ cell formation. The notochord-forming gradient in the peripheral regions is mainly due to the inductive action already exerted by the dorsal vegetative yolk mass in the intact blastula prior to isolation and recombination (see control explants). The ventro-dorsal decline in primordial germ cell formation in the peripheral regions is very probably due only to the inhibition of primordial germ cell formation by notochord differentiation (as an expression of dorsal mesoderm induction). Therefore, in the animal ectodermal moiety of the blastula there exists only an animal-vegetative gradient in mesodermal competence.These results make it very likely that in urodeles the primordial germ cells do not arise from predetermined elements such as those demonstrated in anurans, but develop from common, totipotent animal ectodermal cells. The discrepancy in the mode of origin of the primordial germ cells between anurans and urodeles could be due only to pronounced differences in the time of appearance of the germinal cytoplasm (in anurans during oogenesis, in urodeles possibly during determination of the primordial germ cells within the ventro-caudal mesoderm).The differences in site and mode of origin of the primordial germ cells between urodeles and anurans favor a dual phylogenetic origin of the two groups.

The formation of the mesoderm in urodelean amphibians : V. Its regional induction by the endoderm.

Dorsal (D), lateral (L and R), and ventral (V) portions of the endoderm of blastulae ofAmbystoma mexicanum of different age (stages 8+ to 10-) were combined with ectodermal caps of stage 8+ blastulae. All V and most L and R portions induced only ventrocaudal mesodermal structures - "ventral" type of mesoderm induction. Almost all D portions induced much more voluminous structures of predominantly axial character - "dorsal" type of mesoderm induction. The difference in mesoderm-inducing capacity of the dorsal as against the lateral and ventral endoderm is probably purely quantitative in character. The dorsal endoderm exhibits a pronounced dominance in mesoderm-inducing capacity. During the early symmetrization of the amphibian egg it is apparently especially the presumptive dorsal endoderm that becomes endowed with strong mesoderm-inducing properties.A comparison of the results obtained with endodermal portions of blastulae of different age showed that the mesoderm-inducing capacity first begins to decline in the dorsal endoderm (around stage 9), subsequently in the lateral, and finally in the ventral endoderm (at stage 10-). At stage 10- the dorsal endoderm no longer has mesoderm-inducing capacities.In the recombinates there is a striking correspondence between the regional differentiation of the mesoderm and that of the endoderm. The latter differs markedly from the presumptive significance of the various endodermal regions in the normal embryo.Primordial germ cells, which constitute a characteristic component of the "ventral" type of mesoderm induction, can be induced not only by ventral, but also by lateral and to some extent even by dorsal endoderm. The development of primordial germ cells from the ectodermal component of the various recombinates indicates that in the urodeles the origin of the primordial germ cells differs markedly from that in the anurans.

The formation of the mesoderm in urodelean amphibians : IV. Qualitative evidence for the purely "ectodermal" origin of the entire mesoderm and of the pharyngeal endoderm.

Xenoplastic recombinations of animal and vegetative parts ofAmbystoma mexicanum and Triturus alpestris blastulae, and similar recombinations of parts of3H-thymidinelabelled and unlabelledAmbystoma mexicanum blastulae demonstrate convincingly that the vegetative part (zone IV, see Nieuwkoop, 1969a) of such a recombinate does not contribute to mesoderm formation, but exclusively forms endodermal derivatives. In contrast, the animal cap of the blastula (zones I.II)-which only gives rise to atypical ectoderm if isolated-not only furnishesall the ecto-neurodermal derivatives, butall the mesodermal structures of the developing recombinate as well, and finally to a varying extent forms additional endodermal structures in the recombinate.In the recombinates endodermization of the ectodermal cap occurred at the anterior end of the invaginated archenteron-corresponding to the presumptive pharyngeal endoderm -, and along the dorsal side of the endodermal tube, while an endoderm-like epithelium is formed at the boundary between the caudal endoderm and the ectoderm (proctodaeum formation). These results suggest that in normal development also endodermization occurs in the "ectodermal half" of the egg. This occurs particularly on the dorsal side, leading to the formation of the presumptive pharyngeal endoderm situated above the dorsal blastoporal groove.These experiments show that the vegetative "half" of the amphibian blastula is firmly determined as the future endoderm, whereas the animal "half" is still virtually undetermined and pluripotent.

Mesoderm formation in the anuranXenopus laevis (Daudin).

UsingXenopus blastulae of stage 9-, recombinates were made of the animal, ectodermal cap (zones I.II) and the vegetative, endodermal yolk mass (zone IV) (see Fig. 1). For the experiments either the entire ectodermal cap (A.B), the single outer layer (A) or the stratified inner layer (B) were used.A comparison of the quantitative composition of the recombinates and the corresponding isolates-on the basis of absolute values expressed in units of section surface area-demonstrates unequivocally that the entire mesoderm originates from the ectodermal "half" of the anuran egg under an inductive influence emanating from the endodermal "half". This holds for recombinates of the vegetative yolk mass with the entire ectodermal cap as well as with its outer or inner layer alone.A comparison of mesoderm formation in recombinates of the entire ectodermal cap or with its outer or inner layer with the vegetative yolk mass shows that in all cases mesoderm formation is proportional to the amount of ectoderm available. In addition, the outer layer of the ectoderm is partially endodermized which may be brought in relation with the fact that in normal development an endodermal lining extends upwards from the endodermal mass, which, among other things, covers the prechordal mesoderm on the outside.The outer layer of the ectoderm has markedly lower neural competence than the inner layer, from which in normal development the bulk of the neural material arises.

The formation of the mesoderm in urodelean amphibians : III. The vegetalizing action of the Li ion.

Strong Li treatment leads to auto-activation of blastula and gastrula ectoderm due to lethal and sublethal cytolysis, resulting in adirect meso- and endodermization of the ectoderm. In contrast, a mild Li treatment-0.04 M LiCl in Holtfreter solution for only two hours-which does not show cytolizing effects, has no or only a weak mesodermizing effect upon blastula ectoderm. However, when the explant contains endoderm as well the same treatment leads to a marked transformation of ectoderm into mesoderm and subsequently into endoderm (vegetalization). Thisindirect action of the Li ion under "physiological" conditions apparently represents an enhancement of the normally occurring mesoderm formation from the ectodermal component of the blastula under the inducing action of the endodermal component and its subsequent transformation into endodermal structures, probably due to a rise in competence. The question is raised in how far the direct meso-and endodermization of gastrula ectoderm by heterogeneous inductors is open to the same criticsm as in the case of auto-activation of the ectoderm by a strong Li treatment. Finally, the experiments of Ave, Kawakami and Sameshima (1968) are briefly discussed.

The Formation of the Mesoderm in Urodelean Amphibians : II. The origin of the dorso-ventral polarity of the mesoderm.

Experiments are described in which in early to late blastulae ofAmbystoma mexicanum (stages 7-8/9 Harrison) the animal, ectodermal "half" (zones I.II) was combined with the vegetative, endodermal yolk mass (zone IV) in various orientations, viz. in random orientation or with the dorso-ventral axes of the two components in identical, opposite or perpendicular orientation (0°, 180°, or 90° translocation respectively). The results demonstrate unequivocally that the dorso-ventral polarity of the induced mesoderm, and thus of the embryo, depends exclusively upon the inherent dorso-ventral polarity of the endoderm, whereas the grey crescent, a considerable part of which is located in the animal, ectodermal "half', plays no causal role whatsoever.The results also show that the dorso-ventral polarity is inherent in the entire endodermal mass, but that the subsequent regional differentiation of the endoderm depends upon stimulating influences emanating from the surrounding mesoderm, the later nutritive yolk representing that part of the endoderm which normally does not come under the influence of the mesoderm, and therefore fails to receive the necessary stimulus for further differentiation.On the basis of these findings "Schultze's Umkehrexperiment" as studied byPENNERS andSCHLEIP, PENNERS, andPASTEELS are reinterpreted, whileDALCQ andPASTEELS' general developmental theory as well asCURTIS' cortical grafting experiments are critically discussed.

The formation of the mesoderm in urodelean amphibians : I. Induction by the endoderm.

The blastula [stage 8+ to 8/9 (Harrison)] ofAmbystoma mexicanum was subdivided into four successive animal-vegetative zones and the relative amounts of cellular material present in the successive zones were determined. The developmental capacities of the isolates I, II, III, IV, and I.II and III.IV as well as of the various recombinates of three of the four and of all four zones were studied, and their quantitative composition at the end of the culture period was determined. To this end the embryos were allowed to develop for only 5 to 6 days, during which period the primary organization and initial differentiation was accomplished, but without the appearance of marked changes in the volumes of the different components, which would have occurred upon extensive decomposition of intracellular yolk and subsequent cytoplasmic growth during a longer period of development.Comparing the differentiation of the recombinates with that of the corresponding isolates - in particular the recombinate I.II.IV with the isolates I, II and IV - it was concluded that the mesoderm arises as a result of an interaction between the pigmented, ectodermal and the unpigmented, endodermal "halves" of the egg, which initially [before stage 7 (Harrison)] constitute the only two components of the egg. A comparison of the quantitative composition of the recombinates with that of the corresponding isolates yielded strong arguments in favour of the statement thatthe mesoderm develops exclusively from the ectodermal "half" of the egg under the influence of an inductive action from the part of the endodermal "half". This statement was further corroborated by arguments collected from the literature.Whereas neither the endoderm nor the ectoderm alone are initially able to differentiate beyond a certain point - so-called atypical ectodermal and endodermal differentiation respectively - their interaction product, the mesoderm, apparently contains the information needed for differentiation into the characteristic mesodermal structures. Influences emanating from the differentiating mesoderm then enable both the ectoderm and the endoderm to proceed further on their path of differentiation.The role of the blastocoelic cavity - a cavity with a negative morphogenetic function - in thespatial interaction between the two primary components of the egg was elucidated. In the light of the conclusions mentioned above the centrifugation experiments ofPASTEELS (1953, 1954) were reinterpreted, whileSCHULTZE'S "Umkehrexperiment" byPENNERS andSCHLEIP (1928),PENNERS (1929) andPASTEELS (1938, 1939) andCURTIS' cortical grafting experiments (1960, 1962) were briefly discussed. The hypothesis was then advanced that the inductive interactions taking place in the early embryo preferentially spread through the most superficial layer of the egg, where the cells are tightly connected with each other. Finally, thetemporal aspects of mesoderm induction were discussed in relation to observations collected from the literature.Some parallels were indicated between the morphogenetic events taking place in early amphibian development, and recent biochemical observations on RNA and protein synthesis before the onset of gastrulation.Finally a general picture was drawn of the development of the amphibian egg on the basis of the principle of a stepwise increase in multiplicity by means of inductive interactions.