In Vivo and in Vitro Studies on the Hypoblast and Definitive Endoblast of Avian Embryos

/. Embryol. exp. Morph. Vol. 46, pp. 187-205, 1978 Printed in Great Britain © Company of Biologists Limited 1978

In vivo and in vitro studies on the hypoblast and definitive endoblast of avian embryos

ByE. J. SANDERS,1 RUTH BELLAIRS2 AND P. A. PORTCH2 From the Department of Anatomy and Embryology, University College, London

SUMMARY An unusual example of the invasion of one tissue by another occurs during gastrulation in the chick embryo when the definitive endoblast becomes inserted into the hypoblast. The two tissues were examined morphologically by SEM and TEM. They resemble each other in being of an epithelial type, though neither possesses a basal lamina. The definitive endoblast cells are flatter than the hypoblast cells and more closely attached to one another. When they were explanted in hanging drop cultures, the two tissues were found to exhibit differences in their behaviour. In comparison with the definitive endoblast, the hypoblast cells attached more readily to the glass, produced larger ruffle membranes, moved more rapidly, showed poorer contact-inhibition of locomotion and showed a greater tendency to break away from the main explant. When a hypoblast explant was confronted with a definitive endoblast explant, the hypoblast cells became displaced by the definitive endoblast. The hypoblast explant tended to fragment into smaller groups of cells, many of which migrated around the definitive endoblast, thus mimicking the situation in vivo. Control experiments comprised confronting hypoblast with hypoblast, hypoblast with somites, definitive endoblast with definitive endoblast, and definitive endoblast with somites. The hypoblast explants behaved in a consistent manner, always fragmenting when coming into contact with cells from a confronting explant. The definitive endoblast explants showed more contact inhibition of locomotion when confronted with definitive endoblast or with somites than when confronted with hypoblast. It is suggested therefore that the ability of the hypoblast cells to separate from one another may play an important role in the penetration of the hypoblast by the definitive endoblast both in vitro and in vivo.

INTRODUCTION Until recently, it was believed that the endoderm of birds arose in a totally different way from that of amphibians. In amphibians it is invaginated through the blastopore at the same time as the mesoderm. In birds it was considered to arise as a distinct lower layer prior to the imagination of the mesoderm. There was some controversy as to the precise source of this lower layer (see discussion 1 Author's address: Department of Physiology, University of Alberta, Edmonton, Alberta, Canada. 2 Authors' address: Department of Anatomy and Embryology, University College London, Gower Street, London WC1E 6BT, U.K. 188 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH XI XII 2

Hypoblast

Definitive endoblast

Junctional endoblast

Fig. 1. Diagram to illustrate the relations of the hypoblast and the definitive endoblast. Initially, the area pellucida is underlain by hypoblast alone, though subsequently this tissue becomes replaced by definitive endoblast medially and junctional endoblast posteriorly. The definitive endoblast cells invaginate through the primitive streak whilst the junctional endoblast cells migrate from the posterior germ wall. XI, XII = Eyal-Giladi & Kochar stages. 2-5 = Hamburger-Hamilton stages. (Composite diagram, partially after Vakaet, 1970; Rosenquist, 1971; and Fontaine & Le Douarin, 1977.) by Bellairs, 1971) but it was generally accepted that it appeared first of all at the posterior end of the area pellucida, and it was shown by marking experiments that the cells then migrated anteriorly until they formed a continuous sheet, the lower layer of the area pellucida (Spratt & Haas, 1960; Vakaet, 1970). It is now known that this lower layer, the hypoblast (the 'sickle endoblast' of Vakaet, 1970), does not contribute to the embryonic endoderm, but that instead it spreads out to the periphery of the area pellucida and subsequently forms extra-embryonic endoderm (Vakaet, 1962, 1970; Rosenquist, 1971, 1972; Fontaine & Le Douarin, 1977). The hypoblast also contains among its cells the primordial germ cells (the 'endophyll' cells of Vakaet, 1970) which subsequently collect in the anterior germ wall (area opaca) prior to their trans- port to the gonad at a later stage of development (Dubois, 1969; Vakaet & Hertoghs-De Maere, 1973). The embryonic endoderm which we call the definitive endoblast (following Vakaet, 1970) is located around the anterior end of the primitive streak at a later stage (Bellairs, 1953a, b) and like the mesoderm is derived from cells of the epiblast or upper layer, which invaginate through the primitive streak (Modak, 1965, 1966; Nicolet, 1965, 1967, 1970; Hypoblast and definitive endoblast of avian embryos 189 Rosenquist, 1966, 1972; Vakaet, 1970; Fontaine & Le Douarin, 1977). The definitive endoblast inserts itself into the hypoblast and spreads out so that it is like a halo in the central part of the lower layer, whilst the hypoblast moves distally (Fig. 1). This means that the lower layer in the anterior part of the area pellucida consists of two types of cell, and these not only have a different origin but they also have a different fate. This paper is concerned with the relationship between these two types of cell, the hypoblast and the definitive endoblast, which we have examined by transmission and scanning electron microscopy, as well as by time-lapse cinematography. The endoderm at the posterior end of the area pellucida is called the junctional endoblast and is thought to be derived by ingrowth from the germ wall (Vakaet, 1970; Modak, 1966) but we shall not be considering this region in this paper.

MATERIALS AND METHODS In order to obtain pure hypoblast, the dissection was carried out before the stage at which endoblast inserts. To obtain pure definitive endoblast, it was necessary to dissect the tissue after it had already inserted. The hypoblast was obtained by dissecting the lower layer from chick or quail embryos which had been incubated for about 4-10 h and were therefore at stages X-XIV of Eyal-Giladi & Kochav (1976). The endoblast was obtained by dissecting the lower layer from beneath the region around the anterior part of the primitive streak of embryos which had been incubated for about 24 h and were at stages 3 + to 5 of Hamburger & Hamilton (1951). Roman numerals are used to indicate Eyal-Giladi & Kochav stages, whilst arabic ones are used for Hamburger & Hamilton ones. Somites for use in control experiments were dissected from embryos of stage 12 (Hamburger & Hamilton, 1951) which were first treated with 0-1 % trypsin in Ca2+ and Mg2+-free saline. Twenty-five specimens were prepared for transmission electron microscopy (TEM) and 16 for scanning electron microscopy (SEM). Specimens for TEM were fixed in 2-5% glutaraldehyde in 0-1 M sodium cacodylate for 1-4 h, and then washed three times in 0-1 M sodium cacodylate which contained 0-333 g CaCl2, for a total of \\ h. They were treated with 1 % osmium tetroxide in phosphate buffer for 1 h at 4 °C, then rinsed in phosphate buffer. After dehy- dration in graded ethanols followed by two changes of propylene oxide, they were embedded in araldite. Sections were stained with 2% uranyl acetate at 38 °C for 20 min, then counterstained in lead citrate. Thick sections for light microscopy were stained in toluidine blue. Specimens for SEM were fixed for periods of 4-24 h in 3 % glutaraldehyde, made up in 0-15 M cacodylate buffer, or in half strength Karnovsky's fixative (Karnovsky, 1965). The pH of the fixative was 7-2. After washing in the cacodylate buffer, the specimens were immersed in 1 % osmium tetroxide for 30 min, and then washed again in cacodylate buffer. They were dehydrated in 13 EMB 46 190 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH graded ethanols and dried in a Polaron critical point drying apparatus from liquid CO2. They were mounted on stubs with Uhu glue (Fishmar Ltd., Waterford, Eire) and coated with gold. A total of 278 hanging drop cultures was prepared, and these were of three types. The first consisted of solitary explants of hypoblast or of definitive endoblast or of somites. The second was composed of confronted cultures in which two homologous tissues were grown in close proximity; these were hypoblast with hypoblast, definitive endoblast with definitive endoblast, or somites with somites. The third type consisted of confronted cultures in which two heterologous tissues were grown in close proximity; the combinations were hypoblast with definitive endoblast, hypoblast with somites, or definitive endoblast with somites. In the homologous confronted cultures it was difficult to dis- tinguish between the two explants but the problem was overcome by confronting chick with quail tissues, this technique (Le Douarin, 1969) enabling us to identify the two types of cell correctly on the basis of their nucleolar morphology. In the heterologous confronted cultures the hypoblast, definitive endoblast and somite cells could be distinguished from one another with confidence because of their morphological differences. The culture medium was 9 ml Earle's 199, 1 ml foetal calf serum, 0-5 ml penicillin and streptomycin (5000/*g/ml). Cultures were maintained at 37 °C for periods ranging from 1^ to 3 days. Fixation was generally in formal saline for 24 h. The explants were stained with Harris' haematoxylin, unless the tissues were quail-chick combinations, in which case, Feulgen's stain was used. Time-lapse filming studies were made using a Bolex camera with Wild Variotimer controls. Most of the films were taken at an interval of 12 sec using a x 16 phase contrast objective, giving a film magnification of x 38. Other films were taken at an interval of 2 min, using a x 2-5 objective with dark ground illumination, and a film magnification of x 8. Thirty cultures were grown in sitting drops either in plastic culture dishes, or on coverslips in unsealed Petri dishes. They were maintained in CO2-gassed incubators for \\ days and appeared to develop in much the same way as the tissues grown in hanging drops. The explants grown on plastic were processed for TEM, whereas those on the coverslips were prepared for SEM.

RESULTS (A) Morphology of the lower layers in vivo Figure 2 shows the hypoblast of a chick blastoderm at stage X. The cells are rounded, and are in close contact with one another only in restricted regions, though filopodial-like connexions extend across the intercellular spaces. In these very young embryos large intercellular spaces are sometimes present through which the ventral surface of the epiblast cells can be seen. Cells abutting Fig. 2. SEM to show part of the ventral surface of the hypoblast of an embryo of stage X. The cells are in close contact only in restricted regions, x 1200. Fig. 3. SEM to show hypoblast cells from the same embryo as in Fig. 2. Long filopodial extensions (/.) from the hypoblast cells (hyp.) are in contact with the ventral surface of the epiblast (ep.). x 1200. Fig. 4. TEM through the hypoblast of an embryo of stage XII. Note the filopodia (/.), the intracellular yolk (y.), and the large spaces between the cells, x 4000. Fig. 5. TEM to show the contacts between two hypoblast cells in an embryo of stage XII. Note the filopodia (/.) and specialized contacts (s.c). x 50000. 13-2 192 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH this space usually possess lamellipodial extensions (Fig. 3). By stage XII these cells are more closely packed but there is little if any alteration in their general morphology. The surface of each cell shows many bulges, caused by the intracellular yolk spheres. Figure 4 is a TEM section through the hypoblast at stage XII. The same features are visible, e.g. the loose contacts between the cells, the long filopodial connexions and the large intracellular yolk spheres. There is no basal lamina. The contacts between adjacent cells seem to be of three types: by filopodia (Fig. 5), by short regions of closely apposed smooth cell membrane (not illustrated), or by occasional specialized contacts of the type illustrated in Fig. 5. The latter have been described by others as an early stage in the formation of desmosomes (e.g. Trinkaus & Lentz, 1967, who investigated Fundulus). Micro- filaments are present within the cells but they appear to be randomly orientated and have not been seen running as thick bands. Microtubules were not common and, as already reported by Sanders & Zalik (1970), tend to be randomly orientated. Figure 6 is an SEM micrograph of the definitive endoblast of a stage-5 embryo. The cells are in close contact with one another, their borders being marked by ridges. Some of the cells are rich in globular projections (Fig. 7). The definitive endoblast cells are flatter and thinner than those of the hypoblast, and TEM micrographs show that they are poor in intracellular yolk spheres in comparison with the hypoblast cells (Fig. 8). The specialized contacts between the cells are similar to those between hypoblast cells.

(B) Morphology of the explants After the hypoblast and definitive endoblast have been explanted in hanging drop cultures, they each spread out as a sheet of cells (Fig. 9). The two cultures

FIGURES 6-10 Fig. 6. SEM to show the ventral surface of the invaginated definitive endoblast located just ventral to the node. The cells are in close apposition to one another, the borders between them frequently being marked by microvilli. Some of the cells are rich in globular protrusions, x 3200. Fig. 7. SEM of invaginated definitive endoblast from an embryo of stage 5 to show cells rich in globular projections, x 3000. Fig. 8. TEM through the definitive endoblast of an embryo of stage 6. x 5000. Figs. 9-15 are light micrographs of fixed and stained explants Fig. 9. Confronted cultures of hypoblast (hyp.) from stage XII embryo and definitive endoblast (e) from a stage-5 embryo which have not yet made contact with one another. Fixed after 18 h in culture. The hypoblast cells are larger than those of the definitive endoblast. x 50. Fig. 10. Hypoblast cells from a stage-XII embryo and definitive endoblast cells from a stage-5 embryo lying against one another. The intracellular yolk spheres are larger in the hypoblast cells than in the definitive endoblast. The hypoblast cells also possess larger lamellae, x 150. Hypoblast and definitive endoblast ofavian embryos 193 194 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH are distinguishable when examined by phase contrast or in fixed and stained preparations by the fact that the hypoblast cells are larger than the definitive endoblast cells, contain more intracellular yolk spheres and have larger lamellae (Fig. 10). They are also morphologically different from the cells which grow out from somite explants (Fig. 15), which tend to be fibroblast-like. By about 24 h after explantation, vacuoles have begun to appear in the cytoplasm of hypoblast and endoblast cells. These are not signs of ill-health or debility in the cells and they are not banished or avoided by frequently replacing the medium. Rather they are evidence that the intracellular yolk spheres, which stain heavily at the time of explantation, are undergoing digestion (Figs. 11, 12, 13). When cultures of hypoblast and definitive endoblast were maintained for 3 or 4 days they remained healthy, and tracts of microfilaments became visible in their lamellae (Fig. 12). When cultures of hypoblast cells were examined by SEM the intracellular yolk spheres could be seen bulging beneath the cell surface (Fig. 16). Many microvilli were present both on the upper surface of the cells and along the boundaries between the cells. Each cell appeared to be in close contact with its neighbour and not loosely apposed as in vivo. When explants of the hypoblast and definitive endoblast were examined by TEM (Figs. 17 and 18) it was found that each possessed dense bands of microfilaments and the same type of specialized cell contacts between the cells. In many ways these contacts resembled those seen in the hypoblast cells in vivo (see Fig. 5). The two types of explant differed from one another, however, in the thickness of their lamellae, the hypoblast lamellae being only about 0-3 jam whereas the definitive endoblast lamellae were about 0-6 /on. Micro- tubules were infrequent in both types of lamellae.

(C) Behaviour of the explants In addition to the morphological differences, the two types of explant showed some behavioural differences. First, the hypoblast explants settled on the glass and produced good outgrowths much more readily than the definitive endoblast explants. Second, when cinefilms of the two types of explant were com-

FlGURES 11-14 Fig. 11. Hypoblast cells from a stage-XII embryo fixed after 24 h in culture. Note the intracellular yolk spheres and the vacuoles. x 150. Fig. 12. Hypoblast cells from a stage-XIV chick embryo, fixed after 3 days in culture, during which the medium was changed three times. The vacuoles around the nuclei mark the site of digested yolk spheres. Note that the lamellae are bigger than those illustrated in Fig. 11, and bundles of microfilaments are visible, x 200. Fig. 13. Definitive endoblast cells from a stage-5 embryo, fixedafte r 24 h in culture. The vacuoles mark the site of digested yolk spheres, x 150. Fig. 14. Confronted culture of hypoblast from a stage-XII embryo and definitive endoblast from a stage-5 embryo. Definitive endoblast cells (arrowed) have penetrated among the hypoblast cells, x 100. Hypoblast and definitive endoblast of avian embryos 195 196 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH

!l5

Fig. 15. Somite cells from a stage-12 embryo fixed after 18 h in culture, x 120. Fig. 16. SEM micrographof hypoblast cells from a stage-XIII embryo explanted in a sitting drop, and fixed after 24 h in culture. Note that the cells are in much closer contact with one another than the hypoblast cells shown in Figs. 2 and 3. The borders between the cells are marked by microvilli, and the intracellular yolk spheres are visible as bulges beneath the cell surface, x 1200. Fig. 17. TEM micrograph of a section through an explant of hypoblast cells taken from a stage-XIT embryo and grown in culture for 18 h. x 37000. Fig. 18. TEM micrographof a section through an explant of definitive endoblast taken from a stage-5 embryo and grown in culture for 18 h. x 27800. Hypoblast and definitive endoblast of avian embryos 197 pared, the hypoblast was found to produce bigger regions of ruffling than the definitive endoblast cells whose ruffles were small in amplitude and more restricted in length. Third, the hypoblast cells moved more rapidly than the definitive endoblast cells. Random measurements taken on three separate explants of each type were 1-8 //in per min for the hypoblast cells, and 0-24 /

Table 1. Summary of the behaviour of explants in confronted culture

Confrontation Behaviour Hypoblast-hypoblast Interpenetration of cells Endoblast-endoblast Barrier formation, no mixing Somite-somite Barrier formation, no mixing Hypoblast-endoblast Interpenetration of cells Hypoblast-somite Interpenetration of cells Endoblast-somite No mixing

F/^5. 19-24 are stills from time-lapse cine films taken of explanted tissues Fig. 19. Region of confrontation of two explants of hypoblast cells. The cells from the two explants are intermingled, no clear border being visible between them. x340. Fig. 20. Region of confrontation between two explants of definitive endoblast (a) and (b). Note the 'barrier' region in which the cells have become spindle shaped and closely aligned in parallel rows at right angles to their original path of direction. x340.

When confronted cultures are of hypoblast and definitive endoblast the typical pattern of behaviour is as follows: some of the hypoblast cells migrate around the periphery of the endoblast, either singly or in groups, and they may completely envelop it (Fig. 23). Meanwhile at the junctional region an exchange of cells and some interpenetration may take place (Figs. 14, 21 and 25G-I). Parallel rows of aligned, spindle-shaped cells are seldom formed. Low power Hypoblast and definitive endoblast of avion embryos 199

Fig. 21. Region of confrontation of explant of hypoblast {h) and an explant of definitive endoblast (e). The definitive endoblast cells are passing between the hypoblast (arrowed), x 340. Fig. 22. Region of confrontation between explants of hypoblast and somites. The somite cells (arrowed) are passing between the hypoblast cells, x 340. Fig. 23. Dark ground view of confronted explants of hypoblast (h) and definitive endoblast (e). The hypoblast is becoming displaced around the periphery of the definitive endoblast. x 70. Fig. 24. Dark ground view of confronted hypoblast (h) and somites (s). Some of the somite explants have become surrounded by hypoblast cells, x 70. 200 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH

L Fig. 25. Tracings of the outlines of cells in four cine films. Intervals shown in hours in bottom right hand corner. (A-C) Three sequences from confronted cultures of hypoblast and hypoblast. Note the intermingling of the cells. (D-F) The three sequences from confronted cultures of definitive endoblast and definitive endoblast. Note a 'barrier' region of aligned cells is formed and there is no intermingling of cells. (G-l) Three sequences from confronted cultures of definitive endoblast (left) and hypoblast (right). Note the intermingling of cells. (J-L) Three sequences from confronted cultures of hypoblast (left) and somites (right). Note the intermingling of cells. cinematography shows that the sheet of definitive endoblast penetrates into the hypoblast apparently by displacing the hypoblast cells. The definitive endoblast displaces the hypoblast either as a sheet or, if the hypoblast fragments, by penetration between the hypoblast cells. Having penetrated, the definitive endoblast cells move freely within the spaces between the hypoblast cells with considerable ruffling. Contact inhibition is shown by definitive endo- Hypoblast and definitive endoblast of avian embryos 201 blast cells unable to penetrate the hypoblast; these cells stopped moving and retracted upon touching the hypoblast sheet. In order to decide whether the behaviour at the border between hypoblast and definitive endoblast was due to specific properties of the hypoblast or of the definitive endoblast, further experiments were carried out in which each was confronted with somites. This series of experiments was controlled by confronting somite explants with somites (for a full description, see Bellairs, Sanders & Portch, in preparation). It was found that when somites were confronted with somites a barrier region formed though not such a sharply denned one as when definitive endoblast and definitive endoblast were confronted. Similarly, when definitive endoblast and somites were confronted, there was no intermingling of cells, though multilayering sometimes took place. But when hypoblast and somites were confronted the two types of cell were found to interpenetrate (Figs. 22 and 25 J-L). The hypoblast explant tended to break up into clumps, and individual somite cells could be seen squeezing among them and even over them. Hypoblast cells tended to migrate around the explanted somites, surrounding them in much the same way as they surrounded the definitive endoblast explants (Fig. 24). Thus it appears that the hypoblast explants tend to break up into individual cells and to be displaced by other cells with which they come into contact.

DISCUSSION The main point to be discussed is that definitive endoblast cells were found to penetrate among hypoblast cells when the two tissues were confronted in culture and that at the same time, the hypoblast cells migrated around the periphery of the definitive endoblast until they surrounded it. The latter phenomenon is similar to the envelopment of one tissue by another which may occur when fragments are explanted in contact with one another and which has been attributed to differences in the adhesiveness of the two tissues (Wiseman, Steinberg & Phillips, 1972). The relationship seen in our cultures is similar to that in the living embryo in which the definitive endoblast cells insert them- selves into the hypoblast so that ultimately the definitive endoblast lies cen- trally in the area pellucida with the hypoblast surrounding it. We will consider the significance of these findings in relation to the morphology of the two tissues and in terms of the concepts of contact inhibition of locomotion and of differential adhesion. (a) Morphological aspects. Both the hypoblast and the definitive endoblast are epithelial-type tissues, but unlike true epithelia each lacks a clearly defined basal lamina. This may be an important factor in the ability of the definitive endoblast cells to leave the primitive streak and to penetrate into the hypoblast. Another factor is probably the nature of the intercellular contacts. We have seen both by SEM and TEM that the hypoblast cells at Stage XII 202 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH are in contact with one another only along limited regions of their surfaces and that otherwise large gaps separate them. Similar findings were reported by Sanders (1973) who concluded that neither desmosomes nor extensive tight junctions were present between these cells. By comparison, the definitive endoblast cells in situ at stage 4 are more closely attached to one another. Large intercellular gaps are infrequent and tight junctions are present between the cells. These differences between the hypoblast and definitive endoblast cells have also been noted by Vakaet & Hertoghs-De Maere (1973). It seems likely that these differences in the nature of the cell contacts might partially account for the fact that in our cultures the hypoblast cells separated easily from one another whilst the definitive endoblast cells did not. A similar conclusion was reached by Mareel, Yakaet & De Ridder (1968) who found that grafts of sarcoma cells would readily adhere to the hypoblast but not to the definitive endoblast. (b) Contact inhibition. Extensive invasion of one type of tissue by another is unusual in confronted cultures as is well known from the classical work of Abercrombie & Heaysman (1954) on normal fibroblasts migrating in vitro. These authors noted that when two migrating cells came into contact, movement of each was inhibited at least temporarily. They called this process contact inhibition and attributed to it the failure of confronted cultures of fibroblast cells to penetrate among one another and over one another. By contrast to these results however, they also discovered that when a fibroblast colony was confronted by a sarcoma explant, cells from the two populations invaded, and mingled with each other. They concluded that deficiencies in contact inhibition were present in the sarcoma cells (Abercrombie, Heaysman & Karthauser, 1957; Abercrombie & Heaysman, 1976). Reduction in contact inhibition may also occur in other situations. Armstrong & Armstrong (1974), using labelled cells, have shown that when fragments of mesonephric mesenchyme are fused together in vitro, cells move across the region of contact and mingle with the cells in the other fragment. These authors have suggested that contact inhibition operates with reduced efficiency in solid tissues. Similarly there is evidence that contact inhibition is poor in some embryonic cells (Trinkaus, 1973). Our results give some support to this view and to the idea that contact inhibition is not an 'all or none' phenomenon (Vaughan & Trinkaus, 1966). In the present experiments both hypoblast and definitive endoblast exhibited some of the features usually associated with contact inhibition. Each possessed ruffled membranes, and each showed some cessation of movement when they first made contact with other cells. In addition, when they pulled away from the other cells after contact, long retraction fibres were sometimes formed. The definitive endoblast cells appear to be more contact inhibited than the hypoblast cells. When definitive endoblast cells were confronted with other definitive endoblast cells they formed a barrier region which was similar to the Hypoblast and definitive endoblast of avion embryos 203 barrier region between contact inhibited fibroblasts (Elsdale, 1968). We suggest, therefore, that there is a high level of contact inhibition between definitive endoblast cells. By contrast, the confronted cultures of hypoblast to hypoblast did not normally form such a barrier, and cells continued to move relative to one another; this may imply that contact inhibition is lower between hypoblast cells than between definitive endoblast cells. In the mixed cultures of hypoblast and definitive endoblast, the invasion of hypoblast by the definitive endoblast suggests a low level of contact inhibition between the two types of cell. Wiseman, Gorbsky & Melester (1976) speculated that cellular movements within tissue masses is due to a reduction of contact inhibition. Thus the definitive endoblast cells behave differently depending on whether they are confronted with other definitive endoblast cells or with hypoblast cells. These findings are similar to those of Abercrombie & Heaysman (1976) who reported differences in contact inhibition of certain cells when they were in homologous, from when they were in heterologous, populations, and of Heaysman (1970) who showed that contact inhibition need not be reciprocal. By contrast, the hypoblast tended to behave in the same way in confronted cultures, irrespective of whether it was confronted by other hypoblast cells, by definitive endoblast or by somites. In each situation it tended to fragment, permitting the penetration of the sheet of cells by the other tissue. We suggest therefore that fragmentation of the hypoblast sheet is a feature of the tissue rather than of the properties of the confronting cells. The precise stimulus which results in this fragmentation is not known. (c) Cell spreading and adhesion. A further factor which may be important in the invasion of the hypoblast by the definitive endoblast is the differences in spreading and/or adhesiveness of the two types of cell, both in relation to the glass substrate, and to the other cells. We have not carried out any direct measurements of the relative adhesiveness but there is indirect evidence from the present investigation that the hypoblast cells may be more adhesive to the glass substrate than are the definitive endoblast cells. Thus the hypoblast explants attach more readily to the glass and their lamellae cover a larger area than do the definitive endoblast cells. This is reflected in the fact that the hypoblast cells become thinner dorso-ventrally than the definitive endoblast cells, although at the time of explantation the hypoblast cells are the thicker. It also seems possible that the hypoblast cells may be more adhesive to the glass than to one another and that this, combined with the poor cell attachments may be the main reason why clumps of hypoblast cells pull away from one another in vitro. The resulting fragmentation of the sheet facilitates the penetration of the definitive endoblast. We have no evidence as to whether the definitive endoblast cells are more adhesive to the glass or to each other, but some indications may be obtained by comparing them with the hypoblast cells. We suggest that since the explants 204 E. J. SANDERS, R. BELLAIRS AND P. A. PORTCH of definitive endoblast seldom fragment the cells are more tightly adherent than those of the hypoblast explants. We have already seen that the definitive endoblast cells are probably less adherent to the glass substrate than are the hypoblast cells. (d) In conclusion, it appears that the following may play a role in the invasion of the hypoblast by the endoblast: the poor adhesion of the hypoblast cells to one another together with their good adhesion to the glass may result in gaps appearing between them. The definitive endoblast cells migrating forward as a sheet are not greatly contact inhibited by the hypoblast and extend forward into the gaps. In the hypoblast, meanwhile, the relatively high mobility of the hypoblast cells leads to the migration of some of them around the periphery of the definitive endoblast. It seems likely that some of these features may play a role in the penetration of the hypoblast by the definitive endoblast which takes place during morphogenesis in the embryo.

The filming apparatus was purchased with the aid of a grant to one of us (R. B.) from the Medical Research Council, and the cost of the investigation was partly borne by grants from the Medical Research Council of Canada (to E.J.S.) and from the University of Kuwait (to R.B.). We also wish to thank Dr Alan Boyde for the use of his scanning electron microscope facilities which were partly provided by the Medical Research Council. We are most grateful to Miss Doreen Bailey and to Miss E. Maconochie for their skilled technical assis- tance and to Mrs J. Astafiev for drawing Figs. 1 and 25. We would also like to acknowledge the helpful discussions we have had with Dr J.E.M. Heaysman and Mr M. Abercrombie, F.R.S.

REFERENCES ABERCROMBIE, M. & HEAYSMAN, J. E. M. (1954). Observations on the social behaviour of cells in tissue culture. II. 'Monolayering' of fibroblasts. Expl Cell. Res. 6, 293. ABERCROMBIE, M. & HEAYSMAN, J. E. M. (1976). Invasive behaviour between sarcoma and fibroblast populations in cell culture. /. natn. Cancer Inst. 56, 561-569. ABERCROMBIE, M., HEAYSMAN, J. E. M. & KARTHAUSER, H. (1957). Social behaviour of cells in tissue culture. III. Mutual influence of sarcoma cells and fibroblasts. Expl Cell Res. 13, 276. ARMSTRONG, P. B. & ARMSTRONG, M. T. (1974). Are cells in solid tissue immobile? Meso- nephric mesenchyme studied in vitro. Devi Biol. 35, 197-209. BELLAIRS, R. (1953 a). Studies on the development of the foregut in the chick blastoderm. I. The presumptive foregut area. /. Embryol. exp. Morph. 1, 115-124. BELLAIRS, R. (19536). Studies on the development of the foregut in the chick blastoderm. II. The morphogenetic movements. /. Embryol. exp. Morph. 3, 242-250. BELLAIRS, R. (1971). Developmental Processes in Higher Vertebrates. London: Logos Press. DIPASQUALE, A. (1975). Locomotory activity of epithelial cells in culture. Expl Cell Res. 94, 191-215. DUBOIS, R. (1969). Donnees nouvelles sur la localisation des cellules germinales primordiales dans le germe non incube de poulet. C.r. hebd. Seanc. Acad. Sci., Paris 269, 205-208. ELSDALE, T. R. (1968). Parallel orientation of fibroblasts in vitro. Expl Cell Res. 51, 439-450. EYAL-GILADI, H. & KOCHAV, S. (1976). From cleavage to primitive streak formation: A complementary normal table and a new look at the first stages of the development of the chick. I. General morphology. Devi Biol. 49, 321-337. FONTAINE, J. & LE DOUARIN, N. M. (1977). Analysis of endoderm formation in the avian blastoderm by the use of quail-chick chimaeras. /. Embryol. exp. Morph. 41, 209-222. Hypoblast and definitive endoblast of avion embryos 205 HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. /. Morph. 88, 49-92. HEAYSMAN, J. E. M. (1970). Non-reciprocal contact inhibition. Experientia 26, 1344. KARNOVSKY, M. J. (1965). A formaldehyde-glutaraldehyde fixative of high osmolarity for use in electron microscopy. /. Cell Biol. 27, 137A. LE DOUARIN, N. (1969). Particularity du noyau interphasique chez la caille japonaise {Coturnix coturnix japonica). Utilisation de ces particularity's comme 'marguage biologique' dans les recherches sur les interactions tissulaires et les migrations cellulaires au cours de 1'ontogenese. Bull. Biol. Fr. Belg. 103, 435-452. MAREEL, M., VAKAET, L. & DE RIDDER, L. (1968). Grafting of HeLa cells in young chick blastoderm. Europ. J. Cancer 4, 249-252. MODAK, S. P. (1965). Sur l'origine de l'hypoblaste chez les oiseaux. Experientia 21, 273. MODAK, S. P. (1966). Analyse experimental de l'origine de l'endoblaste embryonnaire chez les oiseaux. Rev. Suisse de Zool. 73, 877-908. NICOLET, G. (1965). Etude autoradiographique de la destination des cellules invaginees au niveau du noeud de Hensen de la ligne primitive achevee de l'embryon de Poulet. Ada Embryol. Morph. exp. 8, 213-220. NICOLET, G. (1967). La chronologie d'invagination chez le Poulet: etude a l'aide de la Thymidine tritiee. Experientia 23, 576-577. NrcoLET, G. (1970). Analyse autoradiographique de la localization des differentes ebauches presomptives dans la ligne primitive de l'embryon de Poulet. /. Embryol. exp. Morph. 23, 79-108. ROSENQUIST, G. C. (1966). A radioautographic study of labeled grafts in the chick blastoderm. Carnegie Inst. Contrib. Embryol. 38, 111-121. ROSENQUIST, G. C. (1971). The location of pregut endoderm in the chick embryo at the primitive streak stage as determined by radioautographic mapping. Devi Biol. 26, 323-335. ROSENQUIST, G. C. (1972). Endoderm movements in the chick embryo between the early short streak and head process stages. /. exp. Zool. 180, 95-104. SANDERS, E. J. (1973). Intercellular contact in the unincubated chick embryo. Z. Zellforsch. mikrosk. Anat. 141, 459-468. SANDERS, E. J. & ZALIK, S. E. (1970). The occurrence of microtubules in the pre-streak chick embryo. Protoplasma 71, 203-208. SPRATT, N. T. & HAAS, H. (1960). Importance of morphogenetic movements in the lower surface of the young chick blastoderm. /. exp. Zool. 144, 257-275. TRINKAUS, J. P. (1973). Surface activity and locomotion of Fundulus deep cells during blastula and gastrula stages. Devi Biol. 30, 68-103. TRINKAUS, J. P. & LENTZ, T. L. (1967). Surface specializations of Fundulus cells and their relation to cell movements during gastrulation. /. Cell Biol. 32, 139-153. VAKAET, L. (1962). Some new data concerning the formation of the definitive endoblast in the chick embryo. /. Embryol. exp. Morph. 10, 38-57. VAKAET, L. (1970). Cinephotomicrographic investigations of gastrulation in the chick blastoderm. Archs Biol. (Liege). 81, 387-426. VAKAET, L. & HERTOGHS-DE MAERE, E. (1973). Jonctions intercellulairs dans le feuillet inferieur du jeune blastoderme de Poulet, etudees au microscope electronique a baiayage. C. r. Seanc. Soc. Biol. 167, 1300-1301. VAUGHAN, R. & TRINKAUS, J. P. (1966). Movements of epithelial sheets in vitro. J. Cell Sci. 1, 407-413. WESTON, J. A. & ABERCROMBIE, M. (1967). Cell mobility in fused homo- and heteronomic tissue fragments. /. exp. Zool. 164, 317-324. WISEMAN, L. L., GORBSKY, G. J. & MELESTER, T. S. (1976). Is the movement of single cells within solid tissue masses induced by trypsinization? Expl Cell Res. 103, 426-431. WISEMAN, L. L., STEINBERG, M. S. & PHILLIPS, H. M. (1972). Experimental modulation of intercellular cohesiveness: reversal of tissue assembly patterns. Devi Biol. 28, 498-517.

(Received 25 January 1978, revised 3 April 1978) !4 EMB 46