Presumptive mesoderm cells from Xenopus laevis gastrulae attach to and migrate on substrata coated with fibronectin or laminin

During gastrulation movements in amphibian embryos the presumptive mesodermal cells migrate from the blastopore towards the animal pole region along the inner surface of the ectodermal layer. Migrating cells extend lamellipodia and filopodia, and move preferentially towards the animal pole (Nakatsuji et al. 1982). The filopodia are very frequently attached to the extracellular fibrils that form an anastomosing network on the inner surface of the ectoderm layer. These fibrils appear in late blastulae and increase in number through gastrulation in all species of urodeles and anurans studied to date (Nakatsuji & Johnson, 19836 ; Darribère et al. 1985), but they are much reduced in number or absent altogether in hybrid embryos that cease to develop at early gastrula stages (Nakatsuji & Johnson, 19846 ; Delarue et al. 1985).

If the network of fibrils is transferred from the ectodermal layer onto a coverslip, gastrula mesodermal cells readily attach to and actively migrate on such conditioned surfaces (Nakatsuji & Johnson, 1983a). The orientation of cell movement in such a situation was not random, but aligned along the animal pole-blastopore axis of the ectoderm-layer explant that conditioned the surface. Moreover, if the fibril network was artificially aligned by applying mechanical tension during the conditioning, the mesodermal cells showed aligned movement along this new axis, strongly suggesting the presence of a contact guidance system (Nakatsuji & Johnson, 1984a). During gastrulation, the ectodermal layer stretches predominantly along the blastoporeanimal pole axis to bring about epiboly (Keller, 1980). It is therefore very likely that the extracellular fibril network on the inner surface of the ectoderm layer not only provides an adequate substratum for mesodermal cell migration but also gives a degree of orientation by contact guidance. Such contact guidance and the contact inhibition of movement, which drives cells from the crowded blastopore region towards the less crowded animal pole region, are probably working together to bring about the oriented migration by the mesodermal cells during gastrulation (Nakatsuji, 1984).

Immunohistochemistry has been used to demonstrate the presence of fibronectin in these fibrils. Fluorescent staining at the light-microscope level (Boucaut & Darribère, 1983 a, b; Lee et al. 1984) and colloidal gold staining at the scanning electron microscope level (Nakatsuji et al. 1985a; Darribère et al. 1985) have demonstrated that the fibrils in anuran and urodele embryos contain fibronectin. Inhibition of gastrulation movements by injected Fab’ fragments of the anti-fibro- nectin antibody (Boucaut et al. 1984a) or synthetic peptides of the cell attachment domain of fibronectin (Boucaut et al. 1984b) have demonstrated that fibronectin is actually playing a role in cell migration. More recently, the presence of laminin in the fibril network was shown by immunofluorescent staining in newt gastrulae (Nakatsuji et al. 1985 b). Darribère et al. (1986) have observed the presence of laminin in another urodele by immunostaining, and confirmed biochemically that synthesis of laminin-like polypeptides occurs at blastula stages. The location of the fibril network suggests that it may well be a primitive form of the basement membrane. It may, therefore, contain other components of the mature basement membrane, such as type IV collagen and heparan sulphate proteoglycans.

One direct way of examining the possible role of these molecules is to study the attachment and movement of isolated mesodermal cells on substrata coated with purified components. Johnson (1985) studied the adhesion of these cells to fibronectin-coated Sepharose beads. In the present study I have used culture conditions that enable the gastrula mesodermal cells to move actively in vitro in a similar fashion and at similar rate to that occurring in vivo (Nakatsuji & Johnson, 1982). Culturedish surfaces were coated with fibronectin, laminin, type IV collagen or heparan sulphate. Adhesion and movement by dissociated mesodermal cells from Xenopus gastrulae on such surfaces were studied by use of a time-lapse video system and scanning electron microscopy.

Xenopus laevis eggs were obtained from matings artificially induced by injecting human chorionic gonadotropin into males and females (Gurdon, 1967), and staged according to Nieuwkoop & Faber (1967). Bufo bufo japonicus embryos were collected from natural spawning. Jelly coats were dissolved with 1% sodium thioglycolate in 10% modified Steams’ solution (MSS) (Nakatsuji & Johnson, 1982). The vitelline membrane was removed with two pairs of fine forceps. The dorsal part of early to mid-gastrulae (stage 10 · 5 – 11) was dissected out with hair loops in MSS. Mesodermal and ectodermal cell layers were separated and dissociated into single cells by incubating the cell masses in a 0 · 05 M-sodium citrate solution in Ca²⁺/Mg²⁺-free MSS for 20 min and mild pipetting. Dissociated cells were transferred into a culture medium, which contained 0 · 1 mM-Ca²⁺ and 0 · 5 % bovine serum albumin, and had a pH value of 8 · 0 (Nakatsuji & Johnson, 1982).

A circular area (diameter, 15mm) of a plastic tissue culture dish (diameter, 35mm; Falcon) was covered with solutions containing human plasma fibronectin (5 – 10 μ g; Bethesda Research Laboratories, Gaithersburg, Maryland, USA), laminin isolated from mouse EHS sarcoma (5 – 17 μ g; BRL), collagen type IV isolated from EHS sarcoma or bovine lens capsules (10—100 μ g; BRL or Nitta Gelatin, Yao, Japan) or heparan sulphate isolated from bovine kidney (10 – 100 μ g; Seikagaku Kogyo, Tokyo, Japan). Collagen dissolved in an acetic acid solution was applied to the dish surface directly or after mixing with the culture medium, which resulted in the formation of fibrous aggregates. Dishes were either incubated at 37°C in a humidified CO₂ incubator overnight, or dried at room temperature. In both cases they were rinsed with the culture medium before seeding with the dissociated cells.

Previous studies (Nakatsuji & Johnson, 1982, 1983a) have shown that the gastrula mesodermal cells attach to an appropriate surface (e.g. a conditioned surface) within 1 h, but do not attach to untreated tissue culture plastic or glass surfaces during the 4-h culture period used in this study. Therefore, the coated area of dish surfaces was examined during a period of 1 – 2 h after seeding with the dissociated cells. Cell morphology was recorded by taking still photographs, and cell behaviour was monitored by time-lapse video recording (Hamamatsu C1965 camera and a timelapse controller of JVC-Sankei, model SIV-J). Unattached cells were spherical and moved with the slightest agitation of the dish. On the other hand, attached cells extended filopodia and lamellipodia, which caused distortion of the cell body from the unattached spherical shape, thus making them easily recognizable under a microscope or in photographic prints. The percentage of attached cells was counted from photographic prints. The mean velocity of cell movement was calculated by marking the approximate centre of the cell body at 10-min intervals, using a video image analyser (For.A, Tokyo, model FVW-300).

Mesodermal cells attached to coated glass coverslips were fixed with a 2 · 5 % glutaraldehyde in 0 · 05 M-phosphate buffer (pH 7 · 3) for 1 day at room temperature. They were post-fixed with a 1 % OSO₄ solution in the same buffer for 1 h at room temperature. Next, cells were dehydrated through a graded ethanol series and critical-point dried using liquid CO₂. The samples were mounted on aluminium stubs, sputter-coated with platinum/palladium, and examined with a Hitachi S-800 scanning electron microscope.

Table 1 shows the percentage of attached cells on various substrata and the mean rate of cell movement on the fibronectin- and laminin-coated surfaces. On the untreated culture dish surface (control), only 5% of the mesodermal cells attached after 2h of culture in medium containing 0 · 5% bovine serum albumin (Fig. 2D). Fibronectin was a good (≈ 80% attached) substratum for mesodermal cell adhesion (Figs 1A, 3A,B). The attached cells showed rapid, active migration (Table 1). An increase in the amount of fibronectin from 5 to 10 μ g did not cause any substantial increase in cell attachment or motility. On the other hand, ectodermal cells showed no substantial attachment to the fibronectin-coated substratum (Fig. 1B).

Mesodermal cells also attached to the laminin-coated substrata (Figs 1C, 3C,D). The percentage of attached cells was slightly lower than that on the fibronectin- coated surface. There was a marked difference in the shape of attached cells; the cell body became somewhat flattened on the fibronectin-coated surface (Figs 1A, 3A), while it remained more rounded on the laminin-coated surface (Figs 1C, 3C). In the latter case, filopodia and lamellipodia seemed to be the only parts of cell surface that attached to the substratum. It was clear that the mesodermal cells were less strongly attached to the laminin- than to the fibronectin-coated surface, because the attached cells could be more easily detached when the laminin-coated dish was shaken. Attached mesodermal cells showed active migration on the laminin-coated surface, although their mean rate of movement was lower than that on fibronectin-coated surfaces (Table 1). The ectoderm cells did not attach to the laminin-coated surface (Fig- 1D).

Coating with collagen type IV or heparan sulphate did not increase the number of attached cells above the control level (Table 1, Fig. 2A-C), even with a relatively large amount (100 μ g). Ectodermal cells did not attach to such surfaces either. In the case of type IV collagen, no substantial attachment was observed when either an acidic solution or a solution adjusted to neutral pH, with the resulting aggregate formation (Fig. 2B), was used for the coating.

Fig. 4 shows two examples that indicate possible guiding effects by fibronectin- or laminin-containing substrata on the migration of mesodermal cells. Fig. 4A shows accumulation of the mesodermal cells inside the fibronectin-coated area after one day of cell movement, with a strikingly clear border between coated and non-coated surfaces. Fig. 4B shows mesodermal cells attached to a meshwork formed by the aggregated laminin molecules. Part of the meshwork was accidentally aligned, and the mesodermal cells showed elongation and movement along the aligned fibrils by the mechanism of contact guidance.

The present study shows that mesodermal cells from X. laevis gastrulae attach to and migrate actively on surfaces coated with fibronectin or laminin. There are some differences between the modes of attachment to fibronectin- and laminin-coated substrata. Mesodermal cells adhere more strongly and move faster on the fibronectin-, than the laminin-coated surfaces. Furthermore, when there is heterogeneity in the distribution of these molecules, the cells can be guided by the mechanism of haptotaxis or contact guidance.

There is accumulating evidence that fibronectin is not only present on the extracellular fibrils on the inner surface of the gastrula ectoderm layer, but that it also actually plays a role in directed cell migration (Boucaut et al. 1984a,b). Strong attachment and active migration by mesodermal cells on a fibronectin-coated surface support these observations. There have been reports of laminin being present in place of fibronectin (Nakatsuji et al. 1985 b; Darribère et al. 1986). The present study shows that laminin can be an adequate substratum for attachment and migration by mesodermal cells, although whether it is actually playing a role in cell migration during gastrulation must be confirmed by other studies.

Type IV collagen and heparan sulphate, other components of the mature basement membrane, did not serve as good substrata for cell adhesion. There are no reports concerning the presence or absence of these molecules in amphibian gastrulae. There are several possibilities. The first is that there are no such molecules in gastrulae and they are not playing any role in gastrulation. The second is that these molecules are present in amphibian gastrulae but are not involved in cell adhesion. For example, type IV collagen is a good candidate for assembling fibrils, adsorbing fibronectin and laminin molecules on its surface, but by itself does not form an adhesion site for mesodermal cells. The third possibility is that they are present and playing roles in gastrulation, but that these molecules isolated from mouse tumour and bovine kidney are different from their amphibian counterparts, and thus do not work properly. More evidence is needed to decide which is the case.

I thank Izumi Fuketa for excellent technical assistance. A batch of Bufo eggs was collected by Yuzo Kadokawa. Type IV collagen was a gift from Nitta Gelatin Co., Ltd.