In the gastrulating chick embryo, the mesoderm cells arise from the epiblast layer by ingression through the linear accumulation of cells called the primitive streak. The mesoderm cells emerge from the streak with a fibroblastic morphology and proceed to move away from the mid-line of the embryo using, as a substratum, the basement membrane of the overlying epiblast and the extracellular matrix. We have investigated the roles of fibronectin and laminin as putative substrata for mesoderm cells using complementary in vivo and in vitro methods. We have microinjected agents into the tissue space adjacent to the primitive streak of living embryos and, after further incubation, we have examined the embryos for perturbation of the mesoderm tissue. These agents were: cell-binding regions from fibronectin (RGDS) and laminin (YIGSR), antibodies to these glycoproteins, and a Fab’ fragment of the antibody to fibronectin. We find that RGDS, antibody to fibronectin, and the Fab’ fragment cause a decrease in the number of mesoderm cells spread on the basement membrane, and a perturbation of cell shape suggesting locomotory impairment. No such influence was seen with YIGSR or antibodies to laminin.

These results were extended using in vitro methods in which mesoderm cells were cultured in fibronec tin-free medium on fibronectin or laminin in the presence of various agents. These agents were: RGDS; YIGSR; antibodies to fibronectin, fibronectin receptor, laminin and vitronectin; and a Fab’ fragment of the fibronectin antiserum. We find that cell attachment and spreading on fibronectin is impaired by RGDS, antiserum to fibronectin, the Fab’ fragment of fibronectin antiserum, and antiserum to fibronectin receptor. The results suggest that although the RGDS site in fibronectin is important, it is probably not the only fibronectin cell-binding site involved in mediating the bevaviour of the mesoderm cells. Cells growing on laminin were perturbed by YIGSR, RGDS and antibodies to laminin, suggesting that mesoderm cells are able to recognise at least two sites in the laminin molecule. We conclude that the in vivo dependence of mesoderm cells on fibronectin is confirmed, but that although these cells have the ability to recognise sites in laminin as mediators of attachment and spreading, the in vivo role of this molecule in mesoderm morphogenesis is not yet certain.

Gastrulation in avian embryos results in the formation of a three-layered blastoderm from a two-layered one. This is accomplished by the ingression of cells from the upper epithelial epiblast layer through the primitive streak to give rise to mesodermal cells (Sanders, 1986; Stem and Canning, 1988; Harrisson, 1989). The primitive streak is a medial, linear accumulation of cells that are undergoing the epithelial-to-mesenchymal transformation that accompanies mesoderm differentiation (Bellairs, 1986). Having passed through the primitive streak, the mesodermal cells invade the space between the overlying epiblast and underlying endoblast layer and disperse laterally away from the mid-line. At this time the mesodermal cells are using, as a substratum for movement, the basement membrane of the epiblast, the dorsal surface of the endoblast (which has no basement membrane), and the general extracellular matrix of the tissue space, which contains high levels of hyaluronic acid at this time (Sanders, 1986).

The factors influencing the dispersal of the mesoderm cells are unclear, but may involve contact inhibition of locomotion, contact guidance by the overlying basement membrane, or simply the population pressure in the crowded mid-line (Sanders, 1989). The cells attached to the basement membrane would most likely be influenced by the fibronectin-rich content of that structure. The fibronectin is present in the lamina densa and in the so-called ‘interstitial bodies’ that are attached to the outer surface of the lamina densa and which are in direct contact with the mesoderm cells (Duband and Thiery, 1982; Sanders, 1982; Harrisson et al. 1984). The fibronectin in the lamina densa itself is not uniformly distributed, but occurs in tracts, especially rostrally, where it may (Critchley et al. 1979), or may not (Andries et al. 1985), influence the movement of the mesoderm. Such influence may then be transmitted to those mesoderm cells that are not in direct contact with the basement membrane, since behind the advancing tissue front the mesoderm consiste of a continuous sheet of fibroblast-like cells.

Despite the immunocytochemical evidence for the presence of fibronectin, and in vitro evidence for its likely effects on mesoderm cell spreading and movement (Sanders, 1980), there is no clear understanding of the role of this glycoprotein in mesoderm cell dispersal in vivo. Similar problems in the morphogenesis of the avian neural crest (Boucaut et al. 1984) and in amphibian (Boucaut et al. 1985), sea urchin (Katow, 1987) and Drosophila (Naidet et al. 1987) gastrulation have been approached by means of microinjection of antibodies to extracellular matrix glycoproteins and of cell-binding fragments from these molecules. In the present study, we have used this technique on the gastrulating chick embryo by microinjection of a cell-binding fragment from fibronectin (RGDS), and antibodies to fibronectin (anti-FN), into the tissue space adjacent to the primitive streak at the time that gastrulation is occurring. In order to obviate questions of steric hindrance and cross-linking by the antifibronectin antiserum, we have also prepared and injected a Fab’ fragment from this antibody.

Because laminin is also a significant component of the basement membrane at this early stage of development (Mitrani, 1982; Bortier et al. 1989; Zagris and Chung, 1990), we have also injected a cell-binding fragment from this molecule (YIGSR), and antibodies to laminin (antiLN).

Results were analyzed morphometrically for alterations in cell shape and confirmed that fibronectin is responsible for the attachment and spreading of the mesoderm cells, and that the RGD site makes a significant contribution to this attachment. Under these in vivo conditions, however, neither YIGSR nor anti-LN was found to disturb mesoderm cell shape or affect the number of cells attached to the basement membrane.

These results were extended by in vitro experiments in which mesoderm cell expiants were cultured on fibronectin and laminin in the presence of antibodies to these molecules, the cell-binding fragments, the Fab’ fragment, and an antibody to the fibronectin receptor. We conclude that mesoderm cells, like those of the neural crest (Dufour et al. 1988; Perris et al. 1989; Perris and Bronner-Fraser, 1989), are dependent for attachment and spreading primarily, but not exclusively, on the RGDS site in fibronectin. Cell attachment to, and spreading on, laminin was sensitive to both RGDS and YIGSR, indicating a possible role for this molecule as a mediator of mesoderm cell behaviour during gastrulation.

In vivo microinjections

Chick embryos were incubated at 37 °C until they reached stage 4 of Hamburger and Hamilton (1951). The entire embryo was prepared for culture on its vitelline membrane, ventral surface uppermost, according to the method of New (1955), using a 25 mm internal diameter glass ring for support. The embryo was cleaned of yolk using washes of Pannett, and Compton’s saline, and incubated at 37°C for 1–1.5h in order to ensure that the blastoderm was firmly adherent to the vitelline membrane.

Embryos were microinjected through the endoblast with approximately 100 nl of reagent into the space between the epiblast and endoblast (Fig. 3) using a micropipette of approximately 20–30 μm tip diameter. The injection site was kept constant in all experiments, being immediately to the right-hand side of the primitive streak, close to Hensen’s Node (Fig. 1). The endoblast layer is extremely thin at this stage of development and as a result the injection site healed over, without trace, after 1–2 h of further incubation. Successful injections were identifiable by visible fluid movement in a limited area of the tissue space during injection (van Hoof et al. 1984). Injections resulting in tears in the endoblast or excessive fluid movement, particularly across the mid-line, were discarded. Although it is possible to use finer pipettes and smaller volumes (van Hoof et al. 1984), preliminary experiments in which blue-dyed, 5.5 pm diameter, polystyrene beads (Polysciences Inc.) were injected (Fig. 4) indicated the optimum conditions for injection.

Fig. 1.

Diagrammatic representation of a chick embryo at stage 4–5. The arrow indicates the injection site at the rostral end of the primitive streak (PS). The area pellucida (AP) and area opaca (AO) are indicated.

Fig. 1.

Diagrammatic representation of a chick embryo at stage 4–5. The arrow indicates the injection site at the rostral end of the primitive streak (PS). The area pellucida (AP) and area opaca (AO) are indicated.

Fig. 2.

Western blot comparing the fibronectin content of normal serum (lanes 1, 3 and 5) with that of the serum that was run through a gelatin-Sepharose column (lanes 2, 4 and 6). Lanes 1 and 2 were made with 1 μl of serum, lanes 3 and 4 with 8 pl of serum and lanes 5 and 6 with 25μ of serum. Note the depletion of fibronectin from lanes 2, 4 and 6.

Fig. 2.

Western blot comparing the fibronectin content of normal serum (lanes 1, 3 and 5) with that of the serum that was run through a gelatin-Sepharose column (lanes 2, 4 and 6). Lanes 1 and 2 were made with 1 μl of serum, lanes 3 and 4 with 8 pl of serum and lanes 5 and 6 with 25μ of serum. Note the depletion of fibronectin from lanes 2, 4 and 6.

Fig. 3.

Light micrograph (LM) of a section through the primitive streak (ps), showing the epiblast (ep) and underlying mesoderm cells (m). The endoblast is the very thin ventral layer of cells. The arrow indicates the position of the microinjection. Scale bar, 50 μm.

Fig. 3.

Light micrograph (LM) of a section through the primitive streak (ps), showing the epiblast (ep) and underlying mesoderm cells (m). The endoblast is the very thin ventral layer of cells. The arrow indicates the position of the microinjection. Scale bar, 50 μm.

Fig. 4.

LM showing the presence of microinjected latex beads (arrows) among the mesoderm cells in the space between the epiblast and endoblast. Bar, 50 μm.

Fig. 4.

LM showing the presence of microinjected latex beads (arrows) among the mesoderm cells in the space between the epiblast and endoblast. Bar, 50 μm.

The following agents were injected: the fibronectin cell-binding fragment, RGDS (Peninsula Laboratories or Sigma Chemical Co.); the control peptide, GRGESP (Peninsula Laboratories or Bachem Inc.); the laminin cell-binding fragment, YIGSR (Peninsula Laboratories); rabbit anti-human-fibronectin antiserum (Collaborative Research Inc.); a Fab’ fragment prepared from the fibronectin antiserum; rabbit anti-mouse-laminin antiserum (Collaborative Research Inc.); Pannett and Compton’s saline. After iryection, the cultures were returned to the incubator for a further 6 h before processing.

The Fab’ fragment was prepared by precipitating 680/d of fibronectin antiserum with 350 gl saturated ammonium sulphate solution at 4°C overnight. After centrifugation, the pellet was redissolved in 680 μl 0.1 M sodium acetate buffer, pH 4.5, and 17 pl of pepsin solution (20mgml-1) was added. Digestion was carried out overnight at 37 °C. The sample was centrifuged and dialysed against one litre of PBS (pH 8.0, two changes) for 1 h at room temperature. The dialysate was treated with β-mercaptoethanol and iodoacetamide, and then dialysed exhaustively against PBS, pH 7.4, at 4°C, overnight with five changes of PBS. The final volume was 850 /d, and this was diluted 1:10 with culture medium for use. The preparation was assayed in vitro as described below.

Embryos were fixed overnight in 2.5 % glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, at 4°C. After a buffer wash, they were post-fixed in 1% buffered osmium tetroxide, pH 7.4, for lh at room temperature. After a further buffer wash, the specimens were dehydrated in a graded series of ethanol solutions and propylene oxide, and embedded in Araldite.

For light microscopy, 1 μm sections were cut and stained with Richardson’s stain (azure B and methylene blue). Results were quantified by aligning sections from the region of the injection site in the microscope field so that measurements could be made from the primitive streak to a point approximately 160 pm lateral to the streak. Three parameters were measured: (1) the distance between the dorsal-most mesoderm cells and the basal surface of the epiblast; (2) the distance between the ventral-most mesoderm cells and the dorsal surface of the endoblast; (3) the number of mesoderm cells attached to the basal surface of the epiblast. The first two of these parameters served as controls to ensure that the embryo was not being inflated by the injection process. Results were analysed statistically using a 2-way ANOVA and a 2-tailed Student’s i-test.

For scanning electron microscopy, embryos fixed as above were dehydrated with a graded series of acetone solutions and critical point dried from liquid carbon dioxide. Specimens were then split across the primitive streak at the injection site using tungsten needles, and mounted on stubs using conducting paint so that the broken edge would be visible in the microscope. Specimens were sputter-coated with gold and examined with a Philips 505 scanning electron microscope. Micrographs were made, at magnifications between ×270 and ×2000, of the exposed mesoderm cells in the regions lateral to the primitive streak at the injection site. Prints of the micrographs were made by photographically enlarging the negatives three or four times. Morphometric analysis of mesoderm cell shape was carried out on these prints using a digitizing pad interfaced to a computer and the “Bloquant’ analysis program (R and M Biometrics Inc.). By tracing the outline of exposed mesoderm cells (Figs 5, 6) it was possible to compute a ‘shape factor’ for each cell by assigning a value between 0 and 1, where 1 is a perfect circle. Results were analysed statistically using a 2-way ANOVA and a 2-tailed Student’s i-test. The magnification of the print used for tracing is not a factor to be considered in this process, and this was confirmed by measuring the shapes of cells at several different magnifications.

Figs. 5 and 6.

Scanning electron micrograph (SEM) showing mesoderm cells (Fig. 5). The cell indicated by the star is represented in Fig. 6 by a computer tracing that was used with the ‘Bioquant’ program for assessment of cell shape. Bar, 5 μm.

Figs. 5 and 6.

Scanning electron micrograph (SEM) showing mesoderm cells (Fig. 5). The cell indicated by the star is represented in Fig. 6 by a computer tracing that was used with the ‘Bioquant’ program for assessment of cell shape. Bar, 5 μm.

In vitro experiments

Embryos at stage 5 of Hamburger and Hamilton (1951) were removed from their yolk and vitelline membrane and washed with Tyrode’s saline. Using fine tungsten needles, the endoblast was removed and mesoderm tissue was dissected away without the aid of enzymatic digestion. The tissue was cultured on glass coverslips in drops of 199 culture medium (GIBCO) containing 5 % fibronectin-free fetal bovine serum and l0 μgml-1 gentamycin. Fibronectin-free serum was prepared by passing the serum through a gelatin-Sepharose column (Pharmacia). Western blotting was used to confirm that the serum was fibronectin-free (Fig. 2), by running the serum in volumes of 1 id, 8 pl and 25 μl against anti-human fibronectin antiserum (Collaborative Research Inc; 20 pl in 7 μl of 5 % skimmed milk). The result was visualized using 125I-labeled Protein A in 5 % skimmed milk.

Glass coverslips were used uncoated, or coated with either fibronectin or laminin. Coating was carried out by exposing the coverslips to 50 μg ml-1 fibronectin or laminin in culture medium for 45 min at room temperature, followed by washing with medium.

As required, the culture medium was supplemented with RGDS, GRGESP and YIGSR at a concentrations ranging from 10−2M to 10−3M; rabbit antisera to human fibronectin (anti-FN), mouse laminin (anti-LN), and bovine vitronectin (anti-VN); rabbit IgG, or rabbit antiserum to human fibronectin receptor (the last three from Calbiochem Corporation) at final dilutions ranging from 1:10 to 1:100. The Fab’ preparation was diluted 1:10 with culture medium before use. Cultures were placed in an atmosphere of 5% CO2 at 37 °C for 24 h, after which cells were observed to define their morphology, as either fibroblastic, epithelial or rounded.

In vivo microinjections

Light microscopy

Embryos were injected with: RGDS, GRGESP, YIGSR (all at 10”2M), anti-FN, or anti-LN (the last two at a dilution of 1:40 in saline). In a fixed field (see Materials and methods), measurements were made of the gap between the mesoderm cells and the epiblast; the gap between the mesoderm cells and the endoblast; and the number of cells attached to the basement membrane. Results (Tables 1 and 2) showed that in no case was there a significant effect on the distance between the mesoderm tissue and the basement membrane of the epiblast, indicating that the microinjection process did not hyperinflate the embryo. In only two pairings did any difference appear in the distance between the mesoderm tissue and the endoblast after 6 h incubation. This was presumably a reflection of the very thin character of the endoblast, making it extremely sensitive to any added extracellular volume, van Hoof et al. (1984) have shown that after microinjection the occasional increase in the volume of this tissue space is usually restored to normal after 1 h of incubation. With respect to the numbers of mesoderm cells attached to the basement membrane, however, significant differences were found in comparisons of GRGESP vs RGDS; GRGESP us anti-FN; anti-FN us YIGSR; YIGSR us RGDS; and anti-LN vs anti-FN. The ANOVA discriminated between GRGESP and anti-FN, but not between RGDS and GRGESP, indicating that the antibody was more effective than the peptide. Similarly, anti-LN vs anti-FN was significantly different, while anti-LN vs RGDS was not, again illustrating that the anti-FN was more effective that the RGDS peptide. The Fab’ fragment from the fibronectin antiserum behaved in a similar way to the antibody and the cell-binding peptide. Untreated embryos showed a value of 13.31±2.56 (n=ll) for the number of mesoderm cells attached to the basement membrane. This was not significantly different from the GRGESP, YIGSR or anti-LN values. Saline injections had minimal effect.

Table 1.

Light microscope analysis of the distance between the mesoderm layer and the epiblast or endoblast, and the number of mesoderm cells attached to the basement membrane after various treatments (±S.D.)

Light microscope analysis of the distance between the mesoderm layer and the epiblast or endoblast, and the number of mesoderm cells attached to the basement membrane after various treatments (±S.D.)
Light microscope analysis of the distance between the mesoderm layer and the epiblast or endoblast, and the number of mesoderm cells attached to the basement membrane after various treatments (±S.D.)
Table 2.

Statistical analysis of Table 1, showing significant differences with the 2-way ANOVA**, and the 2-tailed Student’s t-test* (P<0.05)

Statistical analysis of Table 1, showing significant differences with the 2-way ANOVA**, and the 2-tailed Student’s t-test* (P<0.05)
Statistical analysis of Table 1, showing significant differences with the 2-way ANOVA**, and the 2-tailed Student’s t-test* (P<0.05)

Scanning electron microscop

Embryos injected with GRGESP, and then examined for mesoderm cell morphology (Fig. 7), in general showed cells with irregular outline, attached to both the epiblast basement membrane and to the endoderm by means of filopodia and lamellipodia (Fig. 8) similar to the case in untreated embryos (England and Wakely, 1977). By contrast, RGDS-injected (Fig. 9), anti-FN-injected and Fab’ fragment-injected cells tended to show rounder profiles with fewer fine processes, indicating lower levels of adhesion and spreading. This was confirmed statistically using the ‘shape-factor’ analysis (Fig. 11; Tables 3 and 4), which indicated that there were significant differences in shape between cells in embryos treated with RGDS and GRGESP, Fab’ and GRGESP, and most particularly between those treated with anti-FN and GRGESP. Similar analysis of embryos injected with YIGSR and anti-LN, using GRGESP as a control peptide (Fig. 10), indicated that no differences could be detected between pairs of these treatments (Fig. 12; Tables 3 and 4). On the other hand, significant differences were found between the following pairings: GRGESP vs anti-LN, YIGSR vs GRGESP, and anti-FN vs anti-LN.

Table 3.

The mean and standard deviation values for shape factors, obtained after various treatments, analyzed with the scanning electron microscope

The mean and standard deviation values for shape factors, obtained after various treatments, analyzed with the scanning electron microscope
The mean and standard deviation values for shape factors, obtained after various treatments, analyzed with the scanning electron microscope
Table 4.

Statistical analysis of shape factors from Table 3 showing significant differences using the 2-way ANOVA**, and 2-tailed Student’s t-test*

Statistical analysis of shape factors from Table 3 showing significant differences using the 2-way ANOVA**, and 2-tailed Student’s t-test*
Statistical analysis of shape factors from Table 3 showing significant differences using the 2-way ANOVA**, and 2-tailed Student’s t-test*
Fig. 7.

SEM through the injection site showing mesoderm cells (m) after microinjection of GRGESP. ep, epiblast, en, endoblast. Bar, 5 pm.

Fig. 7.

SEM through the injection site showing mesoderm cells (m) after microinjection of GRGESP. ep, epiblast, en, endoblast. Bar, 5 pm.

Fig. 8.

SEM showing mesoderm cells (m) after microinjection of GRGESP. The cells are attached to the basal surface of the epiblast (ep) by means of broad lamellipodia (arrow). Bar, 1 μm.

Fig. 8.

SEM showing mesoderm cells (m) after microinjection of GRGESP. The cells are attached to the basal surface of the epiblast (ep) by means of broad lamellipodia (arrow). Bar, 1 μm.

Fig. 9.

SEM showing mesoderm cells after microinjection of RGDS. The cells tend to appear more rounded that those in Fig. 7. Bar, 5 μm.

Fig. 9.

SEM showing mesoderm cells after microinjection of RGDS. The cells tend to appear more rounded that those in Fig. 7. Bar, 5 μm.

Fig. 10.

SEM showing mesoderm cells after microinjection of anti-laminin antiserum. The cells tend to retain the irregular shape and the lamellipodia seen in Fig. 7. Bar, 2 μm.

Fig. 10.

SEM showing mesoderm cells after microinjection of anti-laminin antiserum. The cells tend to retain the irregular shape and the lamellipodia seen in Fig. 7. Bar, 2 μm.

Fig. 11.

Histogram showing the distribution of mesoderm cell shapes after microinjection with GRGESP, RGDS and antifibronectin antiserum. Note the shift to the right in the values for RGDS and anti-FN, indicating a tendency to more round cells.

Fig. 11.

Histogram showing the distribution of mesoderm cell shapes after microinjection with GRGESP, RGDS and antifibronectin antiserum. Note the shift to the right in the values for RGDS and anti-FN, indicating a tendency to more round cells.

Fig. 12.

Histogram showing the distribution of mesoderm cell shapes after microinjection with anti-laminin antiserum, YIGSR and GRGESP. All treatments show a similar distribution of cell shapes.

Fig. 12.

Histogram showing the distribution of mesoderm cell shapes after microinjection with anti-laminin antiserum, YIGSR and GRGESP. All treatments show a similar distribution of cell shapes.

In vitro experiments

When mesoderm tissue was explanted onto uncoated glass coverslips in medium with 5 % fibronectin-free serum, the cells emigrated from the tissue mass showing an exclusively fibroblastic morphology (Fig. 13), as they do in medium containing normal serum (Sanders, 1980). These cells would not attach to substrata in the absence of serum, or in the presence of RGDS (Fig. 14). However, removal of vitronectin from the fetal bovine serum by addition of antiserum to bovine vitronectin (anti-VN) at concentrations ranging from 1:10 to 1:100 had no effect on the attachment of these cells to uncoated glass during two days of culture. All of the following experiments were performed in the presence of fibronectin-free serum. The results are summarized in Table 5.

Table 5.

In vitro cell morphologies after various treatments

In vitro cell morphologies after various treatments
In vitro cell morphologies after various treatments
Fig. 13.

Phase-contrast micrograph showing mesoderm cells 16 h after explantation onto an uncoated glass substratum in medium containing fibronectin-free serum. The cells at the edge of the explant show a fibroblast-like morphology. Phase-bright vacuoles in the cytoplasm of the cells are yolk droplets. Bar (Fig. 13–16), 100 μm.

Fig. 13.

Phase-contrast micrograph showing mesoderm cells 16 h after explantation onto an uncoated glass substratum in medium containing fibronectin-free serum. The cells at the edge of the explant show a fibroblast-like morphology. Phase-bright vacuoles in the cytoplasm of the cells are yolk droplets. Bar (Fig. 13–16), 100 μm.

Fig. 14.

Mesoderm cells on an uncoated glass substratum for 16 h in the presence of RGDS. The cells are unable to attach or spread.

Fig. 14.

Mesoderm cells on an uncoated glass substratum for 16 h in the presence of RGDS. The cells are unable to attach or spread.

Fibronectin substratum

Mesoderm cells growing on a fibronectin substratum take on an exclusively epithelioid morphology that is distinctly different from their exclusively fibroblastic appearance on glass (Fig. 15; Sanders, 1980). We have used these two morphologies as criteria for assessing the effects of different antibodies and peptides on these cells in vitro. Cells on fibronectin-coated coverslips were subjected at the time of explantation to medium containing RGDS at a range of concentrations: at 10−2M there was no attachment of the cells and therefore no outgrowth; at 10−3M the expiants seemed to be unaffected by the treatment; at 5×10−3M the outgrowing cells showed mixed morphologies (Fig. 16), consisting of: small epithelial sheets, individual fibroblast-like cells, and some rounded cells. All subsequent experiments were performed with RGDS at a concentration of 5×10−3M. A similar range of cell morphologies was obtained if the cells were explanted in normal fibronectin-free medium and changed to RGDS-containing medium after 24 h. Conversely, when cells were grown initially in RGDS-containing medium and after 24 h changed to normal fibronectin-free medium, the exclusively epithelioid morphology was restored (Fig. 17). Mesoderm cells exposed to medium containing GRGESP (Fig. 18) or rabbit IgG, in place of RGDS, also showed epithelioid characteristics.

Fig. 15.

Mesoderm cells 16 h after explantation onto a fibronectin-coated substratum. The tissue has formed a coherent sheet of epithelial cells.

Fig. 15.

Mesoderm cells 16 h after explantation onto a fibronectin-coated substratum. The tissue has formed a coherent sheet of epithelial cells.

Fig. 16.

Mesoderm cells on a fibronectin-coated substratum for 16 h in the presence of RGDS. A variety of morphologies are present, including small sheets of epithelial cells and individual fibroblast-like cells.

Fig. 16.

Mesoderm cells on a fibronectin-coated substratum for 16 h in the presence of RGDS. A variety of morphologies are present, including small sheets of epithelial cells and individual fibroblast-like cells.

Fig. 17.

Mesoderm cells on fibronectin treated with RGDS as in Fig. 16, but then transferred to medium without RGDS for a further 24 h. The cells have resumed their epithelial morphology. Bar (Figs 17–23), 100 μm.

Fig. 17.

Mesoderm cells on fibronectin treated with RGDS as in Fig. 16, but then transferred to medium without RGDS for a further 24 h. The cells have resumed their epithelial morphology. Bar (Figs 17–23), 100 μm.

Fig. 18.

Mesoderm cells on fibronectin for 16 h in the presence of GRGESP. The cells have retained an epithelial morphology’.

Fig. 18.

Mesoderm cells on fibronectin for 16 h in the presence of GRGESP. The cells have retained an epithelial morphology’.

In the presence of antibody to fibronectin, mesoderm cells failed to spread either on fibronectin or on uncoated glass (see also Sanders, 1980). The Fab’ fragment of the antibody produced a similar though less pronounced effect on adhesion and spreading, which was reversible. Under these conditions, cells spread primarily with a fibroblastic morphology. With antibody to fibronectin receptor, however, over a range of dilutions from 1:10 to 1:100, the result was similar to that in the presence of RGDS, showing a mixture of cell morphologies including small epithelial sheets and individual fibroblastic cells (Fig-19). As with uncoated glass the removal of vitronectin from the fetal bovine serum by addition of antiserum to bovine vitronectin (anti-VN) had no effect on the attachment or spreading of the mesoderm cells on fibronectin over a 2-day period.

Laminin substratum

In contrast to the situation on fibronectin, mesoderm tissue was able to spread on a laminin substratum with a fibroblastic morphology (Fig. 20). When YIGSR was added to the medium over the same concentration range as that used for RGDS, cell spreading on laminin was completely abolished (Fig. 21). The same result was obtained for cells grown on uncoated glass. As in the case of fibronectin, cells treated for 24 h with YIGSR recovered their normal appearance when transferred to normal medium (Fig. 22).

Fig. 19.

Mesoderm cells on fibronectin for 16 h in the presence of an antiserum to the fibronectin receptor. The cells show a variety of morphologies similar to those seen in the presence of RGDS CFig. 16).

Fig. 19.

Mesoderm cells on fibronectin for 16 h in the presence of an antiserum to the fibronectin receptor. The cells show a variety of morphologies similar to those seen in the presence of RGDS CFig. 16).

Fig. 20.

Mesoderm cells after 16 h on a laminin-coated substratum. The cells are largely of a fibroblastic morphology, in contrast to their shape on fibronectin (Fig. 15).

Fig. 20.

Mesoderm cells after 16 h on a laminin-coated substratum. The cells are largely of a fibroblastic morphology, in contrast to their shape on fibronectin (Fig. 15).

Fig. 21.

Mesoderm cells on laminin after 16 h in the presence of YIGSR. The cells are unable to attach and spread.

Fig. 21.

Mesoderm cells on laminin after 16 h in the presence of YIGSR. The cells are unable to attach and spread.

Fig. 22.

Mesoderm cells on laminin, treated with YIGSR as in Fig. 21, but transferred to medium without YIGSR for a further 24h. The cells regain the ability to attach and spread on the substratum.

Fig. 22.

Mesoderm cells on laminin, treated with YIGSR as in Fig. 21, but transferred to medium without YIGSR for a further 24h. The cells regain the ability to attach and spread on the substratum.

Cells explanted on laminin and treated with medium containing RGDS (Fig. 23) or anti-laminin antiserum behaved in the same way as those treated with YIGSR, being unable to spread. Cells treated with rabbit IgG were unaffected.

Fig. 23.

Mesoderm cells on laminin after 16 h in the presence of RGDS. The cells are unable to attach and spread, but will recover if transferred to medium without RGDS.

Fig. 23.

Mesoderm cells on laminin after 16 h in the presence of RGDS. The cells are unable to attach and spread, but will recover if transferred to medium without RGDS.

The demonstrated presence of fibronectin in the basement membrane of the gastrulating chick embryo, with regional variations in distribution (Critchley et al. 1979; Duband and Thiery, 1982; Sanders, 1982), has long been held to be significant for the attachment, spreading and outward migration of the mesoderm cells from the primitive streak following ingression. There has been, however, no evidence for its precise roles in any of these processes. Some evidence suggests that, in contrast to neural crest cells, mesoderm cells in the area vasculosa are themselves engaged in the synthesis and deposition of fibronectin identical in its splicing to that found in the immediate environment and which is synthesized by agacent cells (ffrench-Constant and Hynes, 1988). This would suggest that in this location at least, the mesoderm cells are not using exogenous fibronectin as a guidance cue for their movements. In this case, presumably, the fibronectin is merely providing an hospitable substratum for adhesion and spreading. Cell surface receptors for fibronectin (presumably the and other integrins; Akiyama et al. 1990) have been shown to be present on cells of the early chick embryo, including mesoderm cells, using antibodies to the 140 x 103Mr fibronectin receptor complex (Duband et al. 1986), and the CSATor JG22 antibodies (Krotoski et al. 1986). These may also be shown to perturb neural crest cell migration when microinjected in vivo (Bronner-Fraser, 1985, 1986).

In vivo microinjections

In the present study, we have shown that in vivo microinjection of a cell-binding site from fibronectin, and antibodies to fibronectin, can prevent the attachment of mesoderm cells to the basement membrane and alter their shape in a manner that suggests a disturbance in their migratory capacity. We obtained this in vivo result by injection of the RGDS cell-binding peptide, which in other cell types is capable of inhibiting both adhesion and motility in response to fibronectin (Straus et al. 1989; Aznavoorian et al. 1990). However, this is not the only cellbinding region in fibronectin (Yamada, 1989). Other cell and matrix-binding sites in the molecule may be required, perhaps to interact synergistically with the RGD site, for the attachment of avian neural crest cells to substrata, at least in vitro (Dufour et al. 1988; Perris et al. 1989).

Our in vivo experiment does not directly address the question of the role of other active cell-binding sites in fibronectin, but we do show that the polyclonal antibody to fibronectin is significantly more effective as a perturbant than the RGDS peptide. This does suggest that other sites, blocked by the antibody, may be acting together with the RGDS site to effect normal cell attachment. A similar conclusion is reached by the ability of the more sensitive i-test to distinguish between the results of injecting GRGESP vs. anti-FN, whereas injections of GRGESP vs RGDS could only be distinguished using the less sensitive ANOVA. The Fab’ fragment from the fibronectin antiserum produced a less pronounced effect than the whole antibody, perhaps indicating some cross-linking activity of the intact antibody. This may also be reflected by the in vitro results indicating that the Fab’ fragment allows some cell outgrowth, while the intact antibody does not.

We are confident that the RGDS peptide is most likely interfering with cell attachment to either fibronectin or laminin, but not types I or II collagen (Dedhar et al. 1987; Carson et al. 1988) or tenascin (Bourdon and Ruoslahti, 1989), since these three extracellular matrix components do not show widespread distribution in the embryo at this early stage of development (Sanders, 1989). The possible perturbation by RGDS of mesoderm cell interaction with type IV or VI collagens in the basement membrane (Carson et al. 1988; Aumailley et al. 1989) is difficult to assess owing to lack of information as to the distribution of these collagen components at this developmental stage.

The results of the in vivo microinjections of RGDS are in agreement with those on migrating cells of the avian neural crest (Boucaut et al. 1984), Wolffian duct (Jacob et al. 1989), precardiac mesoderm (Linask and Lash, 1988) and epiblast (Lash et al. 1990). These in vivo studies all show that blockade of the RGDS binding-site on fibronectin is sufficient to inhibit cell spreading and movement. The question of the possible role of other fibronectin cellbinding sites on morphogenesis may at present only be realistically addressed in vitro.

Despite the fact that laminin is present in the basement membrane of the epiblast (Mitrará, 1982; Bortier et al. 1989), and possibly even among the mesoderm cells themselves (Zagris and Chung, 1990), we were unable to demonstrate any effect on the mesoderm cells of injecting either anti-LN or YIGSR. While anti-FN significantly reduced the number of cells attached to the basement membrane, anti-LN did not. The use of YIGSR to inhibit mesoderm cell invasion of reconstituted basement membrane in a model system (Sanders, 1991) has also failed to demonstrate any measurable effect. We cannot rule out the possibility that the effects of YIGSR and anti-LN are too subtle to be picked up by the in vivo microinjection technique, and indeed, the present in vitro results (see below) suggest that this might be the case.

Because laminin also contains an RGDS cell-binding region (Grant et al. 1989; Aumailley et al. 1990), the interpretation of the results of RGDS injections is made more difficult. However, the paucity of laminin in comparison with fibronectin makes it unlikely that the effects of RGDS are due to laminin alone. Also, this peptide has been shown to be only weakly active in inhibiting functional interactions of cells with laminin or type I collagen in comparison with fibronectin (Yamada and Kennedy, 1987).

In vitro experiments

Mesoderm cells grown on uncoated glass condition the substratum with exudate containing fibronectin, so that under these circumstances the fibroblast-like cell spreading is abolished by the presence of anti-FN in the medium (Sanders, 1980) and markedly reduced in the presence of the Fab’ fragment. It is demonstrated here that medium containing RGDS or anti-fibronectin receptor antiserum has the same effect. The same result was obtained using medium containing YIGSR and anti-LN, suggesting that under in vitro conditions laminin secreted by the mesoderm cells may be a significant factor in their spreading behaviour. This was surprising, and yet mesoderm cells can be shown by immunofluorescence to synthesize laminin in vitro (Sanders, unpublished results), and they appear to have laminin in their immediate environment in vivo (Zagris and Chung, 1990).

Mesoderm cells explanted onto fibronectin, which would normally show an epithelial morphology, do not spread in the presence of anti-FN, as shown previously (Sanders, 1980). The response to the presence of RGDS and antifibronectin receptor antiserum, by contrast, was a reduction in the proportion of spread cells, with the cultures consisting of cells showing a variety of morphologies: epithelial, fibroblastic and unspread. This clearly indicates the importance of the RGDS cell-binding site in the fibronectin molecule, but the failure of RGDS to abolish attachment supports the view that this site is not the only one required for attachment and spreading of embryonic cells (Dufour et al. 1988). The observation that mesoderm cells spread on fibronectin in vitro as an epithelium, but in vivo as a mesenchyme, indicates that other factors are also in play; recent work suggests that these may be soluble growth factors, which are also able to influence the phenotype of the mesoderm cells (Sanders and Prasad, 1991).

Mesoderm cells growing on laminin show a fibroblastic morphology, in contrast to their epithelial morphology on fibronectin. This attachment was sensitive to the presence of both YIGSR and RGDS, in addition to anti-LN, suggesting that not only are these cells capable of using receptors for laminin for their attachment, but that more than one cell-binding region in the laminin molecule is significant. This result is not consistent with that found in similar experiments using neural crest cells (Perris et al. 1989), where neither YIGSR nor RGDS disrupted attachment, spreading or migration on a laminin/nidogen substratum. While there is no reason to expect that mesoderm cells should behave in the same way as neural crest cells, Bilozur and Hay (1987) have reported that neural crest cells migrating in a reconstituted basement membrane matrix are influenced by the addition of YIGSR, as we have shown here for mesoderm cells on a planar laminin substratum. Inhibitory effects of YIGSR and anti-LN were not discernable in the in vivo microinjection experiments, however. This may of course be due to the fact that the in vivo technique is less sensitive than the in vitro one, but it may also indicate that the situation in vivo is complicated by the presence of other factors, such as the predominance of fibronectin, that mask effects of YIGSR and anti-LN. It is also possible that these other factors render laminin binding in vivo by the mesoderm cells physiologically irrelevant for their attachment, spreading and movement.

The in vitro attachment and spreading of the mesoderm cells was clearly not influenced by the presence in the fetal bovine serum of vitronectin, since the addition of antibovine vitronectin antiserum was without effect on either uncoated glass or on fibronectin substrata. It is probable that, since the vitronectin molecule has a collagen-binding site (Izumi et al. 1988), much of the vitronectin was removed from the serum at the time that fibronectin was removed. However, the results may also support the idea that cell attachment mediated by these two adhesion glycoproteins occurs by different mechanisms (Dejana et al. 1988; Ylanne, 1990).

We thank the Medical Research Council of Canada for an Operating Grant to E.J.S. in support of this work. We are also grateful to Dr W. J. Schneider for the Western blotting, Dr W. J. Gallin for preparation of the Fab’ fragment and Esther Cheung for technical assistance.

Akiyama
,
S. K.
,
Nagata
,
K.
and
Yamada
,
K. M.
(
1990
).
Cell surface receptors for extracellular matrix components
.
Biochim. biophys. Acta
1031
,
91
110
.
Andries
,
L.
,
Vanroelen
,
C.
,
van Hoof
,
J.
and
Vakaet
,
L.
(
1985
).
Inhibition of cell spreading on the band of extracellular fibres in early chick and quail embryos
.
J. Cell Sci.
74
,
37
50
.
Aumailley
,
M.
,
Gerl
,
M.
,
Sonnenberg
,
A.
,
Deutzmann
,
R.
and
Timpl
,
R.
(
1990
).
Identification of the arg-gly-asp sequence in laminin A chain as a latent cell-binding site being exposed in fragment Pl
.
FEBS Lett.
262
,
82
86
..
Aumailley
,
M.
,
Mann
,
K.
,
von der Mark
,
H.
and
Timpl
,
R.
(
1989
).
Cell attachment properties of collagen type VI and arg-gly-asp dependent binding to its ^(VT) and cr3VI chains
.
Expl Cell Res.
181
,
463
474
.
Aznavoorian
,
S.
,
Stracke
,
M. L.
,
Krutzsch
,
H.
,
Schiffmann
,
E.
and
Liotta
,
L. A.
(
1990
).
Signal transduction for chemotaxis and haptotaxis by matrix molecules in tumor cells
.
J. Cell Biol.
110
,
1427
1438
Bellairs
,
R.
(
1986
).
The primitive streak
.
Anat. Embryol.
174
,
1
14
.
Bilozur
,
M. E
and
Hay
,
E. D.
(
1987
).
Neural crest migration in 3D extracellular matrix utilizes laminin, fibronectin and collagen
.
Devi Biol.
125
,
19
33
.
Borher
,
H.
,
De Bruyne
,
G.
,
Espeel
,
M.
and
Vakaet
,
L.
(
1989
).
Immunohistochemistry of laminin in early chicken and quail blastoderms
.
Anat. Embryol.
180
,
65
69
.
Boucaut
,
J. C.
,
Darribère
,
T.
,
Li
,
S. D.
,
Boulekbache
,
H.
,
Yamada
,
K. M.
and
Thiery
,
J. P.
(
1985
).
Evidence for the role of fibronectin in amphibian gastrulation
.
J. Embryol. exp. Morph. Suppl.
89
,
211
227
.
Boucaut
,
J. C.
,
Darribère
,
T.
,
Poole
,
T J.
,
Aoyama
,
H.
,
Yamada
,
K. M.
and
Thiery
,
J. P.
(
1984
).
Biologically active synthetic peptides as probes of embryonic development: a competitive peptide inhibitor of fibronectin function inhibits gastrulation in amphibian embryos and neural crest cell migration in avian embryos
.
J. Cell Biol.
99
,
1822
1830
.
Bourdon
,
M. A.
and
Ruoslahti
,
E.
(
1989
)
Tenascin mediates cell attachment through an RGD-dependent receptor
.
J. Cell Biol.
108
,
1149
1155
.
Bronner-Fraser
,
M.
(
1985
).
Alterations in neural crest migration by a monoclonal antibody that affects cell adhesion
.
J. Cell Biol.
101
,
610
617
Bronner-Fraser
,
M.
(
1986
).
An antibody to a receptor for fibronectin and laminin perturbs cranial neural crest development m vivo. Devi Biol.
117
,
528
536
.
Carson
,
D. D.
,
Tang
,
J. P.
and
Gay
,
S.
(
1988
).
Collagens support embryo attachment and outgrowth in vitro: effects of the arg-gly-asp sequence
.
Devi Biol.
127
,
368
375
.
Critchley
,
D R.
,
England
,
M. A.
,
Wakely
,
J.
and
Hynes
,
R. O
(
1979
).
Distribution of fibronectin in the ectoderm of gastrulating chick embryos
.
Nature
280
,
498
500
.
Dedhar
,
S.
,
Ruoslahti
,
E.
and
Pierschbacher
,
M. D.
(
1987
).
A cell surface receptor complex collagen type I recognises the arg-gly-asp sequence
J. Cell Biol.
104
,
585
593
.
Dejana
,
E.
,
Colella
,
S.
,
Conforti
,
G.
,
Abbadini
,
M.
,
Gaboli
,
M.
and
Marchisio
,
P. C
(
1988
).
Fibronectin and vitronectin regulate the organization of their respective Arg-Gly-Asp adhesion receptors in cultured human endothelial celia
.
J. Cell Biol.
107
,
1215
1223
.
Duband
,
J.-L.
,
Rocher
,
S.
,
Chen
,
W-T.
,
Yamada
,
K. M.
and
Thiery
,
J. P.
(
1986
).
Cell adhesion and migration in the early vertebrate embryo: location and possible role of the putative fibronectin receptor complex J
.
Cell Biol.
102
,
160
178
.
Duband
,
J.-L.
and
Thiery
,
J.-P.
(
1982
).
Appearance and distribution of fibronectin during chick embryo gastrulation and neurulation
.
Devi Biol.
94
,
337
350
.
Dufour
,
S.
,
Duband
,
J.-L.
,
Humphries
,
M. J.
,
Obara
,
M.
,
Yamada
,
K. M.
and
Thiery
,
J. P.
(
1988
).
Attachment, spreading and locomotion of avian neural crest cells are mediated by multiple adhesion sites on fibronectin molecules
.
EMBO J.
7
,
2661
2671
.
England
,
M. A.
and
Wakely
,
J.
(
1977
).
Scanning electron microscopy of the development of the mesoderm layer in chick embryos
.
Anat. Embryol.
150
,
291
300
.
ffrench-Conotant
,
C.
and Hynes, R. 0
. (
1988
).
Patterns of fibronectin gene expression and splicing during cell migration in chicken embryos
.
Development
104
,
369
382
.
Grant
,
D. S.
,
Tashiro
,
K-I.
,
Segui-Real
,
B.
,
Yamada
,
Y.
,
Martin
,
G. R.
and
Kleinman
,
H. K.
(
1989
).
Two different laminin domains mediate the differentiation of human endothelial cells into capillary-like structures in vitro.
Cell
58
,
933
943
.
Hamburger
,
V.
and
Hamilton
,
H. L.
(
1951
).
A series of normal stages in the development of the chick embryo
.
J. Morph.
88
,
49
92
.
Harrisson
,
F.
(
1989
).
The extracellular matrix and cell surface, mediators of cell interactions in chicken gastrulation
.
Int. J. devl Biol.
33
,
417
438
.
Harrisson
,
F.
,
Vanroelen
,
C
,
Foidart
,
J-M.
and
Vakaet
,
L.
(
1984
).
Expression of different regional patterns of fibronectin immunoreactivity during mesoblast formation in the chick blastoderm
.
Devi Biol.
101
,
373
381
Izumi
,
M.
,
Shima-Oka
,
T.
,
Morishita
,
N.
,
Ii
,
I.
and
Hayashi
,
M.
(
1988
).
Identification of the collagen-binding domain of vitronectin using monoclonal antibodies
.
Cell Struct. Funct.
13
,
217
225
.
Jacob
,
M.
,
Christ
,
B.
,
Jacob
,
H. J.
,
Flamme
,
I.
,
Britsch
,
S.
and
Poelmann
,
R. E.
(
1989
).
The role of fibronectin and laminin in the migration of the Wolffian duct of avian embryos
.
Cell Differ. Dev.
27
Suppl
.
S79
.
Katow
,
H.
(
1987
).
Inhibition of cell surface binding of fibronectin and fibronectin-promoted cell migration by synthetic peptides in sea urchin primary mesenchyme cells in vitro.
Dev. Growth Differ.
29
,
579
589
.
Krotoski
,
D. M.
,
Domingo
,
C.
and
Bronner-Fraser
,
M.
(
1986
).
Distribution of a putative cell surface receptor for fibronectin and laminin in the avian embryo
.
J. Cell Biol.
103
,
1061
1071
.
Lash
,
J. W.
,
Gosfield
,
E.
,
Ostrovsky
,
D.
and
Bellairs
,
R.
(
1990
).
Migration of chick blastoderm under the vitelline membrane: the role of fibronectin
.
Devi Biol.
139
,
407
416
.
Linask
,
K. K.
and
Lash
,
J. W.
(
1988
).
A role for fibronectin in the migration of avian precardiac cells. I. Does-dependent effects of fibronectin antibody
.
Devi Biol.
129
,
315
323
.
Mitrani
,
E.
(
1982
).
Primitive streak-forming cells of the chick invaginate through a basement membrane
.
Wilhelm Roux Arch. EntwMech. Org.
191
,
320
324
.
Naidet
,
C.
,
SémEriva
,
M.
,
Yamada
,
K. M.
and
Thiery
,
J. P.
(
1987
).
Peptides containing the cell-attachment recognition signal Arg-Gly- Asp prevent gastrulation in Drosophila embryos
.
Nature
325
,
348
350
New
,
D. A. T.
(
1955
)
A new technique for the cultivation of the chick embryo in vitro.
J. Embryol. exp. Morph.
3
,
320
331
.
Perris
,
R.
and
Bronner-Fraser
,
M.
(
1989
).
Recent advances in defining the role of the extracellular matrix in neural crest development
.
Comm, devl Neurobiol.
1
,
61
83
.
Perris
,
R.
,
Paulsson
,
M.
and
Bronner-Fraser
,
M.
(
1989
).
Molecular mechanisms of avian neural crest cell migration on fibronectin and laminin
.
Devl Biol.
136
,
222
238
.
Sanders
,
E. J.
(
1980
).
The effect of fibronectin and substratum-attached material on the spreading of chick embryo mesoderm cells in vitro.
J. Cell Sci.
44
,
225
242
.
Sanders
,
E. J.
(
1982
).
Ultrastructural immunocytochemical localization of fibronectin in the early chick embryo
.
J. Embryol exp. Morph.
71
,
155
170
.
Sanders
,
E. J.
(
1986
).
Mesoderm migration in the early chick embryo
. In
Developmental Biology. A Comprehensive Synthesis,
vol.
2
, (ed.
L.
Browder
), pp.
449
480
. New York: Plenum.
Sanders
,
E. J.
(
1989
).
The Cell Surface m Embryogenesis and Carcinogenesis. Caldwell, New Jersey, The Telford Press
.
Sanders
,
E. J.
(
1991
).
Embryonic cell invasiveness: an in vitro study of chick gastrulation
.
J. Cell Sci.
98
,
403
407
.
Sanders
,
E. J.
and
Prasad
,
S.
(
1991
).
Possible roles for TGF/51 in the gastrulating chick embryo
.
J. Cell Sci.
99 (in press
).
Stern
,
C. D.
and
Canning
,
D. R.
(
1988
).
Gastrulation in birds: a model system for the study of animal morphogenesis
.
Experientia
44
,
651
657
.
Straus
,
A. H.
,
Carter
,
W. G.
,
Wayner
,
E. A.
and
Hakomori
,
S.
(
1989
).
Mechanism of fibronectin-mediated cell migration: dependence or independence of cell migration susceptibility on RGDS-directed receptor (integrin)
.
Expl Cell Res.
183
,
126
139
.
van Hoof
,
J.
,
Harrisson
,
F.
and
Vakaet
,
L.
(
1984
).
Microinjection in the chick embryo Expl Cell Res
155
,
278
282
.
Yamada
,
K. M.
(
1989
).
Fibronectin domains and receptors In Fibronectin
(ed.
D. F.
Mosher
), pp.
47
121
.
Academic Press
,
San Diego
.
Yamada
,
K. M.
and
Kennedy
,
D. W.
(
1987
).
Peptide inhibitors of fibronectin, laminin, and other adhesion molecules: unique and shared features
.
J. Cell. Physiol.
130
,
21
28
.
Ylanne
,
J.
(
1990
).
RGD peptides may only temporarily inhibit cell adhesion to fibronectin
.
FEBS Lett.
267
,
43
45
.
Zagrib
,
N.
and
Chung
,
A. E.
(
1990
).
Distribution and functional role of laminin during induction of the embryonic axis in the chick embryo
.
Differentiation
43
,
81
86
.