The distribution of E-cadherin during Xenopus laevis development

Cadherins are a family of homophilic Ca²⁺-dependent cell adhesion molecules (Takeichi, 1988); they share partial identity sequence and are highly conserved through evolution (Gallin et al. 1985, 1987; Nagafuchi et al. 1987; Nose et al. 1987; Hatta et al. 1988; Miyatani et al. 1989; Detrick et al. 1990). In the same species several cadherins coexist (e.g. E-, P-, and N-cadherin in the mouse) each with a unique tissue distribution and binding specificity which is determined by restricted regions of the amino-terminal part of the molecule (Nose et al. 1990). Transfection experiments show that the expression of different types of cadherins or of different levels of the same cadherin can specify the adhesive properties of a cell, leading to the selective cell aggregation (Nagafuchi et al. 1987; Edelman et al. 1987; Hatta et al. 1988; Nose et al. 1988; Mege et al. 1988; Miyatani et al. 1989; Friedlander et al. 1989; Matsuzaki et al. 1990). A number of perturbation experiments (Gallin et al. 1986; Duband et al. 1987; Nose et a/. 1988; Detrick et al. 1990) provide strong support for the hypothesis that cadherins are essential for the control of critical events of morphogenesis (Edelman, 1985, 1986; Takeichi, 1988; Thiery, 1989).

So far three cadherins have been identified in Xenopus laevis. (1) Xenopus E-cadherin was identified in the Xenopus A6 epithelial cell line by antibody cross reactivity with the homologous molecule in dog. Several monoclonal antibodies have been prepared against this molecule but, so far, no sequence data are available (Choi and Gumbiner, 1989). Xenopus E-cadherin has a relative molecular mass of 140×10³M_r and shares several biochemical characteristics with mouse E-cadherin such as the Ca²⁺-dependent resistance to trypsin. It is expressed primarily in epithelial tissues (Choi and Gumbiner, 1989). (2) A 140×10³M_rprotein, highly homologous to mouse and chicken N-cadherin, is expressed in Xenopus mesoderm and nervous system shortly after induction (Detrick et al. 1990). The ectopic expression of this molecule in the ectoderm leads to the formation of cell boundaries and to severe morphological defects. (3) Finally a 120×10³M_r cadherin-like protein has been recently identified in Xenopus eggs and cleaving embryos. This cadherin has been proposed to play a role in the Ca²⁺ -dependent adhesion between cleavage stage blastomeres. Its possible relationship to mouse P-cadherin is not yet known (Choi et al. 1990).

In this paper we have determined the distribution of E-cadherin at all stages of Xenopus development paying particular attention to variations of its expression in regions of the embryo where the morphogenetic process is more active.

Sexually mature Xenopus laevis were obtained from the Service d’Elevage de Xenope of the Centre National de la Recherche Scientifique (Montpellier). Embryos were obtained by artificial fertilization. Embryos were stripped from females injected 12 h earlier with human chorionic gonadotropin (1000i.u.; Sigma Chemical Co., St. Louis, MO) and fertilized with the minced testis of males injected 12 h earlier with 400i.u. of human chorionic gonadotropin. The embryos were maintained in 10% filter-sterilized Holfreter’s solution at room temperature. Animals at different stages through metamorphosis were purchased from Nasco (Fort Atkinson, Wisconsin) and were either used immediately upon delivery or maintained at 24°C and fed twice weekly with Nasco Xenopus brittle. Stages of development were determined according to Nieuwkoop and Faber (1967).

The preparation and the characterization of the monoclonal antibodies to Xenopus laevis E-cadherin used in this study has been previously described in detail (Choi and Gumbiner, 1989). Most of the results were obtained using the monoclonal antibody 5D3 alone; however, no difference in staining pattern or intensity was detected using a mixture of four monoclonal antibodies (5D3, 8C2, 19A2 and 31D1) all directed against Xenopus laevis E-cadherin. In all cases we made use of diluted ascites fluids.

Paraffin sections were prepared for staining using a previously published procedure (Levi et al. 1987; Gurdon et al. 1976); when this procedure was used to prepare sections of large metamorphic animals, the method was modified as described (Levi et al. 1990). Briefly, whole embryos or dissected organs were frozen in isopentane cooled in liquid nitrogen and immediately immersed in methanol at –80°C. The samples were then maintained in methanol at –80°C for periods ranging from three days .to four months depending on the size of the animal with weekly changes of cold methanol. The tissues were serially transferred to methanol equilibrated at –20°, 4°, and 20°C for at least two hour at each step. The samples were then immersed twice in xylene until completely clarified for a total period not greater than 30 min and then transferred to a solution of 50% Paraplast (Monoject Scientific, St Louis, MO) in xylene at 56°C for 20min in a vacuum oven, infiltrated three times with Paraplast at 56°C under vacuum for 45min, and embedded in Paraplast. Sections 10 μm thick were cut using a microtome (Lemarde-ley, Paris, France), floated on distilled water at 45°C, collected on washed glass slides and dried on a heating plate at 45°C for at least one hour. To determine the pattern of distribution of E-cadherin immunoreactivity in embryos between stage 2 and 20 NF, we collected the complete set of serial sections corresponding to the whole embryo and stained them at the same time.

For immunofluorescence, sections were deparaffinized in Xylene (3×2min), rehydrated and incubated sequentially with the primary monoclonal antibody (10 μg ml^-1 in PBS, 5% foetal calf serum (FCS); overnight), biotinylated goat anti-mouse IgG secondary antibodies (10 μg ml^-1 each in PBS, 5% FCS; 2 h) and FITC-conjugated streptavidin 5 μg ml^-1 in PBS, 5% FCS; 30 min). In the case of early embryos, the staining was confirmed by using phycoerythrin-conjugated streptavidin (Biomeda Corp., Foster City, CA) and a barrier filter at 550 nm to reduce the intensity of the yolk autofluorescence. The sections were observed with a Leitz epifluorescence microscope. Control sections were stained either with similar dilutions of monoclonal antibodies directed against other epitopes or omitting the primary antibody.

We could not detect E-cadherin by immunohistofluor-escence until stage 12.5. Between stage 12.5 and 13.5 there was a progressive increase in fluorescence in the non-neural ectoderm, so that by stage 13.5 an intense signal could be detected in ectodermal cell surface. At this stage the distribution of the anti-E-cadherin immunoreactivity was as follows. (1) Staining was present both in the outer and inner (sensorial) layer of the ectoderm (Fig. 1 A,C,D,G). All regions of the ectoderm were stained, with the notable exceptions of the area of the ectoderm within the developing neural plate, which was stained very weakly (Fig. 1 B,H), and a ring of cells surrounding the blastopore, which possibly corresponded to cells of the involuting marginal zone (prospective mesoderm and endoderm) prior to invagination (Fig. 1 I). (2) The mesoderm and the endoderm were not stained. (3) In tangential section through the ectoderm, E-cadherin immunostaining was predominantly associated with the cell surface in areas of cell-cell contact (Fig. 1 D).

During later development there was a progressive reduction in the size of the unstained region of the ectoderm within the neural plate so that, by stage 18, only a small ectodermal region adjacent to the neural groove was not stained (Fig. 2A). At stage 23 the epidermis of the entire embryo was strongly E-cadherin positive (Fig. 2C–E), while the endoderm (Fig. 2D), mesodermal derivatives (Fig. 2D,E) and the nervous system were negative. The cement gland and the hypophysis expressed very high levels of the molecule from the beginning of their differentiation (Fig. 2C).

The ear placodes develop as a thickening of the sensorial layer of the ectoderm between stage 21 and 22. Shortly after their formation the level of E-cadherin expression in the placode epithelium started to diminish (Fig. 3A). During placode depression (Fig, 3B) and closure of the ear vesicle (Fig. 3C) all the cells of the placode, with the exception of those still in contact with the inner layer of the ectoderm, did not express high levels of the molecule. After expansion of the ear vesicle (Fig. 3D), E-cadherin was not detectable in the auditory organ. A similar rapid down modulation of E-cadherin expression could be observed during the formation of the olfactory organ so that invaginating cells of the nasal pit destined to form the sensory part of the organ expressed much lower levels of E-cadherin compared to the surrounding epidermis (Fig. 3E). It is interesting to note that during placode formation the invaginating epithelial cells both in the otic and olfactory placode express high levels of the neural cell adhesion molecule N-CAM (Levi et al. 1987), which is not expressed in the epidermis. Therefore a sharp border is formed between N-CAM- and E-cadherin-positive territories.

The region of the ectoderm corresponding to the presumptive embryonic epidermis started to express E-cadherin since stage 13.5 NF. All epithelial cells of the embryonic, pre-larval and larval skin, including the outer epithelial layer and the inner sensorial layer of the embryonic (Fig. 4A) and pre-larval (Fig. 4B) skin and the large unicellular glands of the larval skin (Fig. 4C), expressed high levels of the molecule. During metamorphosis a complete restructuring of the skin takes place under the control of thyroid hormones. This process occurs differently in various areas of the body. In large areas of the body skin, there is a strong proliferation in the stratum germinativum which leads to the formation of a new epidermis beneath the remnants of the larval epidermis. This new epidermis consists initially of a monostratified epithelium and of proliferating groups of cells, which will develop into glands. Other areas of the skin, such as the tail skin, the skin of the body wall overlying the developing limbs and several stripes of skin on the body degenerate, gradually becoming disorganized and then undergoing a process of cornification. During this whole process, all cells of the epidermis continued to be strongly immunopositive for E-cadherin, while no reactivity was detected in the underlying dermis (Fig. 4D,E,F,H; Fig. 5A–D). Surprisingly, this was true also for areas of degenerating skin both in the body (Fig. 4G; Fig. 5A–C) and in the tail (Fig. 5D), even when the epidermis started to be completely disorganized (Fig. 5B,C). The draining ducts of both granular and mucus glands in the postmetamorphic skin were strongly positive while the glandular epithelium was very weakly stained (Fig. 4 E,F). In the adult skin, strong E-cadherin immunoreactivity persisted in the stratum germinativum of the epidermis and in the exocrine ducts and diminished gradually in more external layer of the epidermis (Fig. 4I).

E-cadherin was not detectable by immunohistofluorescence in any endodermal derivative until a well-differentiated epithelium had formed. Because the development of the digestive tract occurs more rapidly in rostral than in caudal regions, the molecules started to accumulate at different stages of development in different portions of the digestive tract. In a stage 41 embryo, for example, the tall columnar epithelium of the midgut had just formed when E-cadherin staining appeared as faint, radially oriented lines corresponding to the cell membranes (Fig. 6A), while the stomach and the duodenum were already well differentiated and anti-E-cadherin antibodies stained all epithelial cells brightly (Fig. 6B). The epithelium of the larval intestine was strongly labeled by anti-E-cadherin antibodies (Fig. 6C). This staining persisted through metamorphosis even when’ the epithelium started to become vacuolated (Fig. 6D) and disorganized (Fig. 6E), and was clearly present also in the reorganized postmetamorphic intestinal epithelium (Fig. 6F).

The monostratified epithelium of the gall bladder as well as the pancreatic acini and the hepatic, bile and pancreatic ducts were strongly positive throughout development while parenchymal cells of the liver showed only a faint immunoreactivity. This was most evident in sections where the relative intensity of immunoreactivity of these structures could be compared (Fig. 6A,B; Fig. 7A,B). Other positive epithelia included the gills, developing lungs, the Wolffian duct and some tubules in the mesonephros; however, the pronephros and most tubules of the early mesonephros were negative.

In adult Xenopus, E-cadherin staining was found in epithelial tissues lining the surfaces of the body. A very intense staining was detected in the lungs (Fig. 8A), skin (Fig. 51), kidney tubules (Fig. 8B), intestinal epithelium (Fig. 8C) and oviduct epithelium. No E-cadherin staining was detectable in all other tissues examined (brain, nerve, skeletal muscle, heart and aorta). In the adult liver we could detect only a faint staining on the hepatocytes while the epithelia of the hepatic and bile ducts were intensely labeled (Fig. 8D).

In this study, we examined the distribution of E-cadherin, a Ca²⁺-dependent cell adhesion molecule during the development of Xenopus laevis. Our major findings are as follows, (a) E-cadherin is not present in the first stages of Xenopus development (as previously observed by Choi and Gumbiner, 1989). (b) It is first detectable by immunohistochemistry in the ectoderm of stage 13.5 gastrulae only in areas of prospective epidermal differentiation, (c) During placode development, the intensity of E-cadherin staining diminished very rapidly after placode thickening, (d) Endodermal cells express E-cadherin at detectable levels only when a differentiated columnar epithelium forms in the digestive tract, (e) During metamorphosis high levels of E-cadherin are present in many epithelia; low levels are present in the liver. E-cadherin persist in the skin and in the gut epithelia even when they start to become disorganized and degenerate during the metamorphic climax, (f) In the adult, E-cadherin is present in the skin, the gut, the lungs, the glands and the pancreas; hepatocytes are weakly stained and most other tissues are negative.

E-cadherin/Uvomorulin was first identified in cleavage stage mouse embryos and plays a critical role in blastomere compaction in the 8- and 16-cell stage morulas (Hyafin et al. 1980; Ogou et al. 1983; Damsky er al. 1983; Shirayoshi et al. 1983; Vestweber and Kemler, 1984; Johnson et al. 1986; Vestweber et al. 1987). During later development of the mouse embryo, after being expressed by all cells of the embryo, E-cadherin disappears from the mesoderm, but persists in the ectoderm and in most regions of the endoderm (Damjanov et al. 1986); the molecule is then present in the liver and most epithelia of the adult (Peyrieras et al. 1983; Vestweber and -Kemler, 1984; Takeichi, 1988). However, no systematic study of the distribution of E-cadherin during mouse development has yet appeared. L-CAM is a calcium-dependent cell adhesion molecule in the chicken (Gallin et al. 1983, 1985, 1987). It has been repeatedly proposed that L-CAM is the avian analogue of E-cadherin. However, this possibility has been recently questioned on the basis of sequence data (Nose et al. 1990). The distribution of L-CAM during the embryonic development of the chicken has been determined in detail: like E-cadherin, L-CAM is present on most cells of the early embryo and disappears then from the mesoderm and neural structures to persist on most limiting epithelia (Thiery et al. 1984).

Xenopus E-cadherin has been identified on the basis of its cross-reactivity with a polyclonal antibody directed against mammalian E-cadherin. It shares several biochemical properties peculiar to mouse E-cadherin (Choi and Gumbiner, 1989).

The distribution of E-cadherin immunoreactivity in Xenopus laevis differs from those of E-cadherin in the mouse and L-CAM in the chicken. First, both mouse-E-cadherin and L-CAM are already present in the very first stages of embryonic development and have therefore been called primary CAMs (Edelman, 1988a,b). In Xenopus, E-cadherin appears only in specific regions of the ectoderm when a well-differentiated epithelium is formed; in the endoderm at variance with other species no E-cadherin is present until the gut epithelium starts to form; finally, in several developing epithelia such as the placodes, the pronephros and the liver, the levels of E-cadherin expression in Xenopus are much lower than those observed in other species where E-cadherin is a prevalent molecule.

The present data are also different from those obtained using anti –chicken-L-CAM antibodies on Xenopus embryos (Levi et al. 1987); beside staining most epithelia, these antibodies stained diffusely the early embryo and brightly some structures (such as the otic vesicle) that are not stained by anti-E-cadherin antibodies. One possible explanation for the difference is that anti-L-CAM antibodies might have recognized more than one cadherin in Xenopus and that their staining patterns represented the sum of the distributional maps of two or more cadherins. Another possible explanation for the different distributions observed is that anti-L-CAM and anti-Xenopus-E-cadherin recognize different cadherins. This would imply that more than one cadherin coexist on the same epithelial cell; such a situation occurs in chicken development where both L-CAM and A-CAM (probably the chicken equivalent of N-cadherin) are occasionally coexpressed by the same cells (e.g. the endoderm, the otic and lens placode) (Thiery et al. 1982; Duband et al. 1988). Indeed a second Xenopus cadherin has been identified recently (Choi et al. 1990); this molecule shares with E-cadherin and all other known cadherins a common cytoplasmic domain, it is present in the eggs and cleaving embryos and it has a relative molecular mass of 120000, the same as that recognized in Xenopus by anti-L-CAM antibodies (Levi et al. 1987).

Given the high homology between different cadherins and the possibility that not al] cadherins have yet been described, it is not possible to compare unequivocally the distribution of cadherins in different species until complete sequence data are available. In this respect, the situation of Xenopus laevis may be particularly difficult because it has a pseudotetrapioid genome where polymorphic alleles of (he same gene might have diverged during evolution; this has been suggested already for Xenopus N-cadherin, which is present in two closely related variants (Detrick et al. 1990). Furthermore, it has been reported (Detrick et al. 1990) that antibodies raised against fusion proteins encoding Xenopus N-cadherin recognize several bands in Western blots of Xenopus brain tissue.

The early distribution of E-cadherin in Xenopus resembles closely that of a number of epidermal antigens (Itoh et al. 1988; Jones, 1985). Several studies suggest that the animal pole region of the embryo at cleavage stage differentiates autonomously into epidermis (Slack, 1984, 1985; Jones and Woodland, 1986, 1987); however, when the ectoderm is cultured in the presence of mesoderm, a process of induction leads to suppression of expression of epidermal markers and enhances the expression of neural genes such as the neural cell adhesion molecule N-CAM. Although expressed in most epidermal regions of the late gastrula, E-cadherin never accumulated in regions overlying the neural plate. Therefore, it is conceivable that, during neural induction, E-cadherin expression is inhibited in the neural plate by (he same set of molecular signals that activate the expression of N-CAM.

A further point of interest of our data is given by the persistence of E-cadherin in degenerating epithelia during metamorphosis. Indeed we could detect a clear immunoreactivity even when the epithelia started to become disorganized before falling apart. This is somehow surprising as it is supposed that E-cadherin is responsible for the stabilization of cell-to-cell contacts in the epithelia and one might expect its down regulation before epithelia disorganization. One must however, remember that the adhesive function of E-cadherin can be modulated in several ways; for example, it is known that in the absence of calcium the adhesive properties of the molecule are lost, and several observations indicate that a complex interaction with cytoskeletal elements is needed to assure cadherin function (Nagafuchi and Takeichi, 1988; Ozawa et al. 1989; Friedlander et al. 1989; Jaffe et al. 1990); the sole presence of the protein is not sufficient to assess its functionality.

This work was in part supported by the Centre National de la Recherche Scientifique, a grant of the Fondation pour la Recherche sur les Myopathies (contract AFM, 1990), and by the National Institute of Health grant No. GM37432. This work was done under the tenure of an Established Investigatorship from the American Heart Association to Dr Barry Gumbiner. Dr Giovanni Levi was a recipient of a grant from the Associazione Italiana per la Promozione delle Ricerche Neurologiche, ARIN.