ABSTRACT
Lithium-induced exogastrulae are abnormal embryos which fail to complete gastrulation and do not form normal neural structures. Scanning electron microscopy has been used to compare the surface structure of the ectoderm cells of exogastrulae with that of the ectoderm cells of normal embryos and has shown that the appearance of ciliated cells is delayed in exogastrulae. In addition, the surface structure of endoderm cells, which remain exposed in these embryos, has been studied.
INTRODUCTION
Recently the scanning electron microscope has been used to great advantage to study the morphological changes which occur during normal early embryogenesis. Tarin (1971) has confirmed that cells at the blastopore groove of amphibian embryos change shape during gastrulation, and has demonstrated the presence of microvilli at the blastopore groove of the gastrula and on the floor of the neural groove of the neurula of Xenopus laevis. More recently, Monroy, Baccetti & Denis-Donini (1976) have used scanning electron microscopy to show that cells lining the blastocoel of Xenopus blastulae form microvilli just before gastrulation, whereas the invaginated chordamesoderm cells form filopodia. Two groups of workers, Billett & Courtney (1973) and Kessel, Beams & Shih (1974), have studied the appearance of ciliated cells in differentiating amphibian ectoderm, using Amblystoma maculatum and Rana pipiens respectively. Their results are comparable – ciliated cells first appear at the neurula stage and increase in number up to the tail-bud stage. Grunz, Multier-Lajous, Herbst & Arkenberg (1976) have used scanning electron microscopy to show differences in the surface structure of ectoderm isolated from Triturus alpestris gastrulae and cultured with and without a vegetalizing inducer.
It has been known for many years (Bäckström, 1954) that treatment of amphibian embryos with lithium chloride produces abnormalities which are to some extent characteristic of the stage at which the embryos are treated. Such embryos are ‘vegetalized’ since lithium treatment causes overproduction of structures derived from ‘vegetal’ endoderm and underproduction of structures derived from ‘animal’ ectoderm. The most extreme example of vegetalization is exogastrulation when embryos fail to gastrulate normally and therefore do not form normal neural structures. Such ‘vegetalized’ embryos may be produced with agents other than lithium chloride such as amines (Stanisstreet, 1974) or theophylline (Pays-de-Schutter, Kram, Hubert & Brachet, 1976) (see Stanisstreet & Osborn, 1976).
In this paper we have studied the exogastrula of Xenopus laevis produced by lithium treatment. As well as the overall morphology of the abnormal embryos, two aspects have been studied in detail: the time of appearance of ciliated cells on exogastrula ectoderm; and the structure of endoderm cells, which are exposed in exogastrulae and can thus be studied without recourse to fracturing the embryos.
MATERIALS AND METHODS
Production of exogastrulae
Embryos were obtained by injecting pairs of adult Xenopus laevis with chorionic gonadotrophin (‘Pregnyl’, Intervet Ltd.) and were staged according to Nieuwkoop & Faber (1956). When the embryos had reached stage 5 (16-cell stage) the jelly coats were removed chemically (Dawid, 1965), and the embryos were washed and subsequently cultured in 10 % Steinberg saline, pH 7 ·3 (Steinberg, 1957). Some of the embryos within one batch were treated with 0-1M lithium chloride in 10 % Steinberg saline for 3 h to produce a high percentage of exogastrulae (Stanisstreet, 1974). After treatment, embryos were washed and kept in 10 % Steinberg saline. Lithium-treated and control embryos were observed and fixed at various times after lithium treatment (Table 1). Throughout the experiment the temperature was controlled at 21 ± °C.
Preparation of embryos for electron microscopy
Immediately prior to fixation, the vitelline membrane was removed with watchmakers’ forceps. The embryos were fixed for 6 h in 2·5 % glutaraldehyde plus 2 % paraformaldehyde with 2·5 mM calcium chloride in 0·1 M cacodylate buffer, pH 7·2 (modified from Karnovsky, 1965) and then washed in changes of cacodylate buffer containing 2·5 mM calcium chloride over 2 h. The embryos were dehydrated in a graded series of alcohols, and the ethanol was then gradually replaced with amyl acetate: occasionally embryos were left at this stage. Embryos were dried using the critical-point method, transferred to the microscope stubs with a fine hair brush as recommended by Tarin (1971), and affixed with adhesive. Finally the embryos were coated with gold-palladium and observed and photographed with a Cambridge ‘Stereoscan’ scanning electron microscope.
RESULTS
The surface structure of normal embryos of Xenopus laevis has been comprehensively described by Tarin (1971) and since our observations confirmed his, the control embryos will be described only for comparison with the abnormal embryos. For convenience, abnormal embryos were classified as early exogastrulae (EO), exogastrulae (El) and late exogastrulae (E2, E3, E4). As can be seen from Table 1, these stages correspond to certain normal stages of Nieuw-koop & Faber (1956). At each stage little variability in the presence or absence of cilia was noted, but it is possible that some variability in the time at which cilia appeared was masked by the time intervals used.
Twelve hours
At 12 h the control embryos had reached stage (late gastrula). The lip of the blastopore formed a complete circle with a diameter equal to approximately half that of the whole embryo, within which yolky cells were visible (Fig. 1). As reported by Tarin (1971) the cells of the blastopore groove were covered with many microvilli (Fig. 2).
In the lithium-treated embryos (E0), the blastopore formed a complete but much larger circle, so that more yolky cells were visible (Fig. 3). Such embryos were more susceptible to fracturing at the yolk plug during critical-point drying, possibly indicating a change in the degree of adhesiveness of the cells in this region. Like the normal gastrulae, the cells of the ‘blastopore lip’ of these embryos showed microvilli (Fig. 4).
Twenty-five hours
By 25 h the control embryos had reached stage 19–20 (neurulae). The neural folds were closing and the embryo was covered with ectoderm (Fig. 5). Higher power examination showed that the ectoderm cells had discrete borders and small marginal pits (Fig. 6). Some differentiation of the ectoderm was apparent – as reported by Tarin (1971) the floor of the neural groove showed microvilli (Fig. 7) and in the lateral ectoderm occasional sunken cells with a few short cilia were apparent (Fig. 8).
Floor of neural groove of Xenopus neurula showing microvilli. (× 1700)
The lithium-treated embryos had reached the exogastrula stage (El). Exogastrulae were dumbell-shaped and consisted of a sphere covered with pigmented ectoderm from which protruded a sphere of yolky endoderm cells (Fig. 9). Examination of the ectoderm of exogastrulae showed a similar general picture to neurula ectoderm (Fig. 10), but ciliated cells were not present and there appeared to be more marginal pits. Examination of endoderm cells showed a different cell surface structure: endoderm cells were large and ‘raspberry-like’ and often showed a central depression (Fig. 11). It is possible that this ‘raspberry’ appearance was due to the collapse of the cell membrane onto the underlying large yolk platelets. Under higher power examination some cells showed microvilli (Fig. 12).
Thirty-one hours
At 31 h the control embryos had reached the late neurula stage (stage 22–23). The most noticeable change in the structure of the ectoderm was the increase in the occurrence of cilia. Ciliated cells were more frequent, forming an approximately regular pattern (Fig. 13) and each ciliated cell had more and longer cilia. The appearance of the endoderm cells of late exogastrulae (E2) was similar to their appearance in the previous stage (El), although in a few areas tube-like structures were visible (Fig. 14). In the ectoderm, marginal pits were still frequent, and cilia were not apparent.
Thirty-five hours
By this time the control embryos were at stage 24–25. Ciliated cells formed a more regular array, and cilia were longer (15). The ectoderm of exogastrulae (E3) now showed the presence of some ciliated cells (Fig. 16), and thus appeared to differentiate some 10 h later than normal ectoderm.
Fifty-one hours
At 51 h the control embryos had reached stage 32–33, and were highly ciliated. The cell borders were raised and discrete, and very few marginal pits were present (Fig. 17). The ectoderm of later exogastrulae (E4) presented a picture similar to that of stage 32–33 embryos. More ciliated cells were present than in the previous stage, and the cilia were longer (Fig. 18). Thus it appears that the differentiation of exogastrula ectoderm is delayed rather than inhibited completely.
DISCUSSION
In the preparation of delicate material for scanning electron microscopy the possibilities of the obliteration of surface detail and generation of artifactual structures must be considered. Whilst the fixation of material appears to present few problems, the method of drying the specimens can alter the final appearance (Billett & Courtney, 1973). Tarin (1971), who first used the scanning electron microscope to observe amphibian embryos, used air drying from fluoracil. Billett & Courtney (1973) compared two methods of drying specimens, air drying from acetone and freeze substitution, and found the latter to be a better method. We have used the more recent method of critical-point drying, which appears to give good preservation of surface structure.
Our observations on normal embryos of Xenopus laevis support those of previous workers: both in the appearance of microvilli at gastrulation and neurulation, and in the progressive appearance of ciliated cells in the ectoderm from neurulation onwards. Light microscopical and stereoscan observations of lithium-treated embryos suggest that the onset of gastrulation is not delayed, although gastrulation is abnormal. We have shown (Osborn & Stanisstreet, in preparation) that exogastrulae (El in Table 1) have about half as many cells as control neurulae, and that by the onset of gastrulation, lithium-treated embryos have significantly fewer cells than gastrulae. Thus it is possible that the abnormal gastrulation of lithium-treated embryos could be due to the presence of fewer but larger cells in these embryos. However, the results of a series of experiments by Cooke (1972, 1973a, 1973b) suggest that inhibition of cell division by Mitomycin-C does not prevent differentiation or post-gastrula morphogenesis.
The present results suggest that cell differentiation of ectoderm, judged by the appearance and growth of cilia, is delayed by some hours in lithium-induced exogastrulae. This finding is confirmed in principle by transmission electron microscope studies (Osborn, Smith & Stanisstreet, in preparation) which suggest that the formation of mucous-secreting cells, a characteristic of differentiated ectoderm (Billett & Gould, 1971), is delayed in exogastrulae. The reason for this delay in cell differentiation is not clear.
Ectoderm isolated from gastrulae forms ciliated cells (Grunz 1973, Grunz et al. 1976) and recently it has been shown that the animal blastomeres of the 8-cell stage of Triturus alpestris are determined to form ciliated epidermis (Grunz et al. 1976). Thus the differentiation of ciliated epidermis does not appear to rely on inductive interactions which could be altered by lithium treatment.
It could be that ciliated cells appear later in cultured isolated ectoderm than in ectoderm in situ: ectoderm taken from gastrulae of Xenopus laevis and cultured at 15 °C for 15 h did not show cilia (Billett, 1968). If the differentiation of isolated ectoderm is delayed, lithium treatment might mimick the isolation effect in some way. However, no absolute comparisons of the time of appearance of cilia and mucous vesicles in isolated ectoderm and whole embryos using the same species at the same temperature are available.
Treatment of isolated ectoderm with a vegetalizing agent prevents the formation of ciliated cells (Grunz et al. 1976). However, presumably in this system ectoderm cells are ‘vegetalized’ into endoderm and mesoderm cells which do not form cilia. It is conceivable that lithium, which is also a ‘vegetalizing’ agent, could act in a similar but less extreme manner to delay cilia formation in exogastrulae produced by a short pulse of lithium treatment. The mechanism of the vegetalizing action of the lithium ion is not known, but biochemical comparisons between lithium-treated and control embryos have shown several gross differences at the transcriptional and translational level (reviewed by Stanis-street & Osborn, 1976), and it is likely that the delay in cell differentiation of ectoderm in lithium-induced exogastrulae is due to some lesion in the mechanisms of gene expression.
ACKNOWLEDGEMENTS
We wish to thank Mr C. J. Veltkamp for his expert help with the scanning electron microscopy, and Dr C. F. H. Vickers of the Department of Dermatology who allowed us to use the critical-point drying apparatus. J. C. O. thanks the Wellcome Trust for a Research Scholarship.