/. Embryo/, exp. Morplt. Vol. 51, pp. 121-135, 1979
Printed in Great Britain © Company of Biologists Limited 1979
121
An analysis of the aggregation and morphogenesis
of area opaca endoderm cells from the primitivestreak chick embryo
By NADINE MILOS, 1 SARA E. ZALIK 1
AND ROBERT PHILLIPS 1
From the Department of Zoology, University of Alberta
SUMMARY
The aggregative behaviour and subsequent morphogenesis of extra-embryonic endoderm
cells from primitive-streak chick embryos have been investigated. A relatively pure population
of area opaca endoderm cells was obtained by differential dissociation, which involves partial
separation of epiblast and endoderm cell clumps by sieving through Nitex mesh. For aggregation studies cells were cultured in rotating flasks in Leibovitz (L-15) medium, in saline or in
saline supplemented with glucose (1 mg/ml). Aggregation was monitored using the Coulter
Counter. In these three media aggregation is rapid; by lOmin an average of 61% of the
population had aggregated, to reach a plateau at 30 min when an average percent adhesion
value of 83 % was obtained. The aggregates in L-15 medium were large and compact. After
several days in culture, they cavitated and formed smooth hollow vesicles with thin walls
composed of one or a few cell layers. Aggregates formed in PCS were smaller and looser in
appearance; the addition of glucose resulted in a certain degree of compaction. Some
morphogenesis occurred under these conditions with the aggregates developing numerous
irregular cavities. These experiments suggest that some of the factors that affect cell adhesion
in early embryonic cells can be studied in vitro. The results also indicate that the ability to
cavitate is an intrinsic property of the endoderm cells of the area opaca since this occurs in
the absence of epiblast or mesoderm.
INTRODUCTION
The morphogenetic events which occur early in vertebrate embryogenesis to
transform the two-layered into the three-layered embryo are among the most
important in development. These processes, which are collectively called
gastrulation, involve, among other things, the movement of great numbers of
cells, both from the surface into the interior, and within the embryo itself.
A study of adhesion at this stage of development is biologically meaningful,
because in order for these precise cellular rearrangements to take place, cells
must carry with them the information which permits selective recognition and
adhesion to occur.
Studies using chick embryos have shown that some degree of selective
cellular affinity is present at this early stage of development. Beginning with the
1
Authors' address: Department of Zoology, University of Alberta, Edmonton, Alberta,
Canada T6G 2E9.
122
N. MILOS, S. E. ZALIK AND R. PHILLIPS
pioneering work of Zwilling (1960, 1963), embryos or parts of embryos at both
gastrular and pregastrular stages have been dissociated and the cells allowed to
reassociate into cellular aggregates (Miura & Wilt, 1970; Zalik & Sanders, 1974;
Eyal-Giladi, Kochav & Yerushalmi, 1975; Macarak, 1975; Sanders & Zalik,
1976). The different cell types that are brought together to form the aggregate
are initially intermingled. However, upon further culturing, cell sorting occurs
and in some cases a differentiated tissue pattern is formed.
It became apparent from the above results, that in order to study selective
cell adhesion at these developmental stages, one had to first investigate the
phenomenon in cell populations of a single-cell type. In this paper we describe
a procedure for the isolation of a relatively pure population of endoderm cells
from the primitive-streak chick embryo (stage 4). We have studied the kinetics
of aggregation of these cells as well as the morphogenesis of the aggregates.
These experiments have been performed under culture conditions devoid of
macromolecular constituents such as serum or embryonic extracts. Our results
demonstrate that these cells aggregate very efficiently, and that the aggregates
are capable of developing a tissue pattern which resembles, to a certain extent
the one they would eventually form in vivo. Our experiments indicate that the
ability to cavitate is an intrinsic property of the endodermal cells of the area
opaca and does not depend on the presence of other tissues. An abstract of
these results has been presented previously (Milos, Zalik & Phillips, 1977).
MATERIALS AND METHODS
Fertile white Leghorn eggs from the University of Alberta Experimental Farm
were incubated for 20 h at 39 °C. Embryos were dissected out into Pannett and
Compton saline (Pannett & Compton, 1924) (PCS) with 15 ITIM Hepes buffer
(Sigma), (pH 7-5). They were then removed from the yolk and vitelline membrane and swirled through numerous changes of PCS to remove adhering yolk
until the inner portion of the area opaca was visibly loosened from the overlying ectoderm. Blastoderms at stage 4 were then selected from the population
[full streak with no visible head process (Hamburger & Hamilton, 1951)] and
stored overnight at R.T. in Leibovitz (L-15) medium with glutamine (Gibco),
pH 7-5, supplemented with 20 /*g/ml gentamycin (Schering Corp.). All operations were carried out under sterile conditions.
Cell type separation
The next day the area pellucida was dissected out with fine scissors from the
centre of each blastoderm. Area opaca tissue from 100 to 120 embryos was then
pooled, washed three times in PCS and resuspended in 4 ml PCS. Tissues were
pipetted gently about 20 times using pipettes of decreasing bore size (approximate inner diameters, 5 and 2 mm) in order to gently dislodge the cells from
one another. The resulting suspension was filtered twice through Nitex mesh
Aggregation of chick endoderm cells
123
(44 /im - B. & S. H. Thompson & Co. Ltd, Montreal, Canada). Sheets of area
opaca ectoderm were collected on the mesh and could be shaken off into PCS
while endoderm cell clumps passed through it.
Endoderm cell dissociation
Contaminating yolk has been a major problem in this study. Since Bellairs
(1963) has reported that in the embryo white yolk is intercalated among the
endoderm cells of the area opaca, the presence of a certain amount of yolk in
our preparations is not surprising. However, intracellular yolk is also released
from cells damaged during dissociation. Our dissociation technique has therefore been designed to result in the lowest contamination with free yolk which
is compatible with a reasonably high yield of single cells. The suspension
containing the clumps of area opaca endoderm cells was divided between four
test tubes into 1 ml aliquots which were spun in a clinical centrifuge for 5 min
at 19 g with rapid acceleration. Each pellet was then resuspended in approximately 500 /d of cold CaMg-free PCS (pH 7-8). Pairs of resuspended pellets
were combined into two tubes, each containing 1 ml of cell suspension. These
were incubated on ice for 10 min with occasional shaking and then pipetted
three times with a flamed Pasteur pipette. Cells were then centrifuged as above,
and pellets were resuspended by shaking and combined in 1 ml cold CaMg-free
PCS. After centrifugation the pellets were again resuspended in cold CaMg-free
PCS (usually 100 /d for each ten embryos). This cell suspension was held on
ice until used. This method of dissociation gave a cell suspension free of slime.
The gentle washes aided in removing a large proportion of the free yolk and
resulted in loose pellets which could be resuspended relatively easily with
minimal cell damage. Trypan blue was excluded by 96% of the cells.
Aggregation assays
One hundred microlitre aliquots of cell suspension were dispensed into
siliconized 10 ml Erlenmeyer flasks with flat bottoms containing 2 ml of either
L-15 medium, PCS or PCS supplemented with 1 mg/ml glucose (G-PCS). The
pH of all media was adjusted to 7-5. After inoculation, 100 ji\ of cell suspension
containing approximately 2 x 104 cells was removed for initial counting
purposes. The flasks were incubated in a gyrotory shaker (New Brunswick
Scientific Co., Inc., New Brunswick, New Jersey) at 37 °C and 80 rev./min.
Aggregation was monitored using the Coulter Counter TAII (Coulter
Electronics, Hialeah, Fla.) fitted with a 400 jLtm aperture tube (size calibration
control set at 118-2). Based on cell sizing 85 % of the dissociated cells fall within
the diameter range measured by channels 5-7 (channel 5 accumulates counts
due to spherical particle diameters of 16-20-2 /im; channel 6 counts particles of
20-2-25-4 /<m; and channel 7 counts particles of 25-4-32 ^m). Counts accumulated in channels 8-16 throughout the course of the experiment will not be
presented here.
124
N. MILOS, S. E. ZALIK AND R. PHILLIPS
Particle counts were performed by adding an aliquot of cell suspension to
100 ml of cold 3 % glycerine in PCS with gentle mixing. The glycerine helped
keep the cells and aggregates in suspension for the time of the count (64 sec).
Care was taken to ensure that no more than 1250 counts per 2 ml sample were
accumulated in channels 2-16. This keeps coincidence loss below 5 % (Coulter
Counter TAII Reference Manual). The 100 ji\ aliquot of cell suspension that
was removed at the time of inoculation was diluted and counted to give the
initial count for all flasks (To); previous experiments had established that
initial inoculation differences varied by only 2-4 %. At the desired time intervals
(Tt) each flask was shaken gently to distribute cells and aggregates and a
200 ju\ aliquot was removed, diluted and counted. An increased sample volume
at Tt was possible because of the decrease in the total number of particles that
occurs during aggregation. The volume of sample removed at To and Tt was
taken into account when the percent adhesion was calculated. Since some free
extracellular yolk (10-25%) and a few cell clumps of equivalent volume were
present which would be counted in channels 5-7, the percent adhesion was
calculated in terms of particle number:
,, .
no. of particles in channels 5-7, Ta - no. of particles in channels 5-7, T,
n,
%
adhesion =
r^-A—;
, „ „
— x 100
no. of particles in channels r5-7, To
Some aggregates were harvested at 30 min of culture, fixed in Bouin's fixative,
lightly stained in alcoholic Eosin and photographed. When aggregates were
maintained in culture for longer periods of time the rotation was increased to
100 rev./min to inhibit further cell accretion. In this case gentamycin (20 /tg/ral)
was added to the medium. The aggregates were harvested at 24, 48 and 72 h
and examined under the dissecting microscope. They were then fixed, embedded
and stained for histological examination.
RESULTS
The area opaca endoderm cell suspensions obtained with our dissociation
technique consist of approximately 80% single cells, 10% pairs and 10%
small cell clumps (three to four cells). The cells are large, ranging in diameter
from 11 to 40 (im (Fig. 1) with an average diameter of 24/on. They contain
vast quantities of intracellular yolk which often obscures the nucleus. Peripherally, one or more hyaline lobopodia may be observed at the cell surface (inset
to Fig. 1). Such large protrusions have previously been described on dissociated
area pellucida cells of the chick (Overton, 1962) and on embryonic cells isolated
from amphibians (Holtfreter, 1943; Satoh, Kageyama & Sirakami, 1976;
Fraser & Zalik, 1977) and teleosts (Trinkaus, 1963).
Aggregation of chick endoderm cells
125
40 i-
•3 30
20
§ 10
10
15
20
25
30
Diameter (/urn of extraembryonic endodermal cells
35
40
Fig. 1. Size range of single, dissociated area opaca endodermal cells. The diameters of
138 single, dissociated cells suspended in PCS were measured under the microscope.
The cell diameters were grouped into 5 /<m ranges. Inset: a single area opaca endoderm cell. Note the large lobopodium (arrow) and the great number of yolk platelets
inside the cell. The bar is 20 //m.
Aggregation in L-15 medium
Endoderm cells of the area opaca were aggregated in L-15 culture medium
at 37 °C and 80 rev./min and particle counts were performed at 10, 20, 30 and
60 min of culture. Although representative experiments are shown in Fig. 2
this pattern of aggregation has been consistently observed in experiments
performed over a 1 year period with embryos from several different flocks. The
maximum rate of aggregation as measured by the decrease in particle counts
occurred during the first 10 min of culture (Fig. 2). At 20 min the change in the
percent adhesion value was smaller, and by 30 min of aggregation the curve
had reached a plateau. Aggregation kinetics were measured at two different
initial inoculation sizes. At an average inoculation size of 9-3 x 104 particles/ml
an average of 49 % aggregation was observed by 10 min, 64% by 20 min, and
69 % by 30 min. Doubling the inoculation size to an average of 2 x 105 particles/
ml resulted in an average percent adhesion value of 63% at 10 min, 79% at
20 min, and 86% at 30 min. The differences in the average percent adhesion
values between the two different inoculations at 10, 20, and 30 min were thus
14, 16 and 17 %, respectively. By 60 min an average of 80% of the particles had
aggregated at the low inoculation size as opposed to 90 % at the high inoculation
size. Doubling the inoculation size had a significant effect on the kinetics of
adhesion at 10, 30 and 60 min (P < 0-01-0-05 when compared with an F test).
The addition of 25 units/ml DNase (DN-100, Sigma) did not affect the early
kinetics (Steinberg, 1963).
9
EMB 51
126
N. MILOS, S. E. ZALIK AND R. PHILLIPS
100
0
10
20
30
Time (min)
50
•60
Fig. 2. Aggregation kinetics of area opaca endoderm cells in L-l 5 culture medium.
Average percent adhesion values of three separate experiments performed at two
different average inoculations have been plotted. The maximum and minimum
values are indicated above and below each average percent adhesion value
(-O- 9-3 x 104, channels 5-7, particles/ml; - • - 2 x 105, channels 5-7, particles/ml).
Visible aggregates appear in each flask as early as 5 min of culture. By 30 min
they have grown to a relatively large size (Fig. 3), ranging in diameter from
about 800 to 1500 /tm. Externally, they look fairly compact but exhibit numerous,
irregular protrusions, each of which is a cell clump. This latter observation
suggests that aggregate to aggregate accretion generates much of these large
cell masses. Internally, such aggregates are composed of an array of cells with
some interspersed free yolk.
Two types of aggregate develop in L-l5 medium upon further culturing at
lOOrev./min. The first type is round and smooth in shape (Fig. 4a), and
consists of cells with lightly stained nuclei which contain one or two nucleoli.
The cytoplasm of the cells is vacuolated and also stains lightly (Fig. 4 b). The
morphological feature which distinguishes this aggregate is the presence of
small round spaces arranged at random in the midst of the main mass of endoderm cells (Fig. 4a, b).
The second type of aggregate that has developed by 24 h of culture is similar
in cellular morphology to the type described above except that it contains a
single, large cavity in addition to the numerous small spaces. The development
of this smooth-walled cavity imparts a doughnut-like appearance to crosssections of such aggregates (Fig. 5).
Sometimes one or several dark clumps of darkly staining cells are visible
among the main mass of endoderm cells (Figs. 4a, b and 5). The cytoplasm of
these cells is not vacuolated and stains relatively evenly. They may correspond
Aggregation of chick endoderm cells
127
Fig. 3. Early area opaca endoderm cell aggregate. Area opaca endoderm cell aggregates were harvested from L-15 medium at 30 min of aggregation, fixed in Bouin's
fixative, stained in alcoholic Eosin and transferred to glycerine for photography.
Note the compact appearance of this aggregate and the surface protrusions. The
bar is 500 /<m.
to area opaca ectoderm or mesoderm cells, which would be starting to invade the
area opaca posteriorly at stage 4 (Romanoff, 1960). These cells may have passed
through the Nitex mesh and may have been incorporated into the initial aggregates, either as clumps of undissociated cells or as single cells which subsequently may have sorted out. Muira & Wilt (1970) in their study of blood
island formation in area opaca aggregates have reported the presence of small
groups of cells which did not differentiate into blood. The groups of cells that
they report may be similar to those described here. The development of blood
has occurred only rarely in our aggregates.
Aggregates harvested at 48 h of culture are similar in appearance to those
obtained at 24 h, although their cavities are usually larger (Fig. 6 a). Occasionally, cells at the periphery of some aggregates organize themselves to form
extremely thin walls one to three cell layers thick. In these cases the cells lose
much of their vacuolation and may become elongated in shape (Fig. 6 b). It is
interesting to note that once cavitation occurs the aggregates start to float in
the medium.
Between 48 and 72 h of culture these aggregates develop irregular contours.
Internally, they are either uncavitated or contain several irregular cavities,
suggesting that the main cavity has either collapsed or has been replaced. Zalik
& Sanders (1974) successfully harvested cavity-containing aggregates formed
from dissociated cells of unincubated blastodiscs up to 7 days of culture. This
may have been due to the presence of foetal calf serum in the medium. In the
present experiments, unsupplemented L-15 was used.
Aggregation in PCS and G-PCS
Area opaca endoderm cells contain a massive amount of yolk which represents an available source of nutrients (Bellairs, 1958). We were therefore
9-2
128
N. MILOS, S. E. ZALIK AND R. P H I L L I P S
Fig. 4. For legend see facing page.
Aggregation of chick endoderm cells
129
Fig. 5. Cavitated area opaca endoderm cell aggregate (24 h). Experimental procedure
is the same as Fig. 4. A large, smooth-walled central cavity (c) has formed in this
aggregate which has an extremely smooth outer wall. The cell mass is vacuolated and
several clumps of contaminating cells are present (arrow). The large dark spheres
are yolk platelets. The bar is 200 /mi.
interested to see if both the early and late (tissue construction) phases of
aggregation could occur under conditions where the cells were more dependent
upon stored reserves. Although representative experiments are presented, these
patterns of aggregation have been observed in these media in experiments
performed over a 1 year period with embryos from several different flocks.
The early aggregation kinetics in PCS are similar to those observed in L-15.
At an average initial inoculation of 2 x 105 particles/ml, average percent
adhesion values of 58, 73, 81, and 89 % were obtained at 10, 20 30 and 60 min
of culture, respectively. The addition of glucose to the saline resulted in early
aggregation kinetics that were also similar to those measured in L-15. When the
percent adhesion values at 10, 20, 30 and 60 min in the three media were compared using an F test they were not found to be significantly different.
Fig. 4. Area opaca endoderm cell aggregate (24 h). Area opaca endoderm cell aggregates were harvested from L-15 medium at 24 h of culture, sectioned at 6/im and
stained with Haematoxylin and Eosin. (a) Note the smooth outer surface and the
small holes (arrows) interspersed throughout this aggregate. On the left is a small,
round clump of cells believed to be contaminating area opaca ectoderm or mesoderm. The bar is 100 /JLTCX. (b) A higher magnification of the area enclosed in the box.
The vacuolation of the area opaca endoderm cells is evident and contrasts with the
evenly staining cytoplasm of the cell clump. A mitotic figure is visible in the midst
of the clump. The bar is 25 /«n.
130
N. MILOS, S. E. ZALIK AND R. P H I L L I P S
Fig. 6. Cavitated area opaca endoderm cell aggregate (48 h). Experimental procedure
is the same as Fig. 4 except that the aggregates were harvested at 48 h of culture.
(a) This cell aggregate consists of extremely vacuolated cells and its wall is only
several cells thick in places (arrows). The large, central cavity (c) contains precipitate-like material as well as what appear to be pycnotic cell clumps (*). The bar
is 100 //m. (b) In this aggregate a section of the wall became very thin and appears
at this point to consist of only two cell processes (arrow). Some cells lining the
cavity (c) are elongated and have lost their vacuolation (arrows). Some vacuoles
are still visible in the outer cells. The bar is 50 jum.
Aggregates harvested from PCS at 30 min of culture and observed under the
dissecting microscope differ from those obtained in L-15. The aggregates
formed in saline are usually somewhat smaller in size, varying in diameter from
about 700 to 1000 ^m. Also, they are generally more irregular in appearance
and are composed of fewer clumps of cells which are also more irregularly
arranged (Fig. 7). The looseness exhibited by these aggregates compared with
those formed in L-15 is apparent when they are removed from the aggregation
flasks with wide-mouth pipettes. This procedure breaks PCS aggregates apart
while L-15 aggregates can be harvested intact. When glucose is added to the
Aggregation of chick endoderm cells
131
Fig. 7. Early area opaca endoderm cell aggregate harvested from PCS. Experimental
procedure is the same as Fig. 3 except that PCS was used as the culture medium.
The more irregular contours of this aggregate are evident when compared with
Fig. 3. The bar is 500//m.
saline, the aggregates become more compact and in many cases indistinguishable
from those formed in L-15 (Fig. 3). A stimulatory effect of glucose on adhesion
has also been reported by Umbreit & Roseman (1975) for neural retina cells.
By 24 h, the majority of aggregates cultured in PCS have irregular external
contours. In G-PCS there are always more smooth aggregates which are at
intermediate stages of cavitation and contain many small spaces. A single,
central smooth-walled cavity has never been observed in aggregates cultured
in PCS or G-PCS for 24 h although such a cavity is often present in aggregates
cultured in L-15 for this period of time.
After 48 h of culture all PCS and G-PCS aggregates have irregular external
contours. Internally they contain numerous, irregular cavities (Fig. 8a). These
cavities are interconnected by vacuolated endoderm cells which form cellular
partitions one or several layers thick. Where the partitions are only one cell
thick, the cells are often elongated in shape, appearing stretched in the direction
of the partition (Fig. 86).
DISCUSSION
There are several reasons for fractionating the embryo into subpopulations
of cells. As stated in the Introduction, these early embryos consist of cells of
differing fates which may or may not resemble each other in their surface
characteristics. In addition, cells of differing sizes and presumably differing
densities are present at this stage of development. Larger or more dense cells
could in principle aggregate faster under these experimental conditions than
smaller less dense cells because they can be more easily brought into proximity
with one another. Thus, the adhesive behaviour of a mixed population of cells
may be the result of physical parameters possessed by one type of cell which
are unrelated to cell surface properties (see also Whur, Koppel, Urquhart &
Williams, 1977). Also, any dissociation procedure used on such a mixed cell
132
N. MILOS, S. E. ZALIK AND R. P H I L L I P S
(b)
Fig. 8. Twenty-four hour area opaca endoderm cell aggregate harvested from G-PCS.
Experimental procedure is the same as Fig. 4 except that G-PCS was the culture
medium, (a) Note the irregular exterior profile of this aggregate. Only partial cavitation has occurred and the walls of the cavities (c) are irregular in shape. Several
clumps of contaminating cells are present (arrows). The bar is 100/<m. (b) Another
section through this aggregate which illustrates very thin cellular partitions (arrows).
A free pycnotic cell (*) is visible to the right of the central partition. The bar is 50/<m.
Aggregation of chick endoderm cells
133
population may select for certain cell types in preference to others. Eyal-Giladi
et ah (1975) have previously demonstrated that the epiblast and hypoblast of
the area pellucida of the chick embryo incubated to stage XIII of Eyal-Giladi &
Kochav (1976) are differentially dissociable.
We have used the greater cohesiveness of the area opaca ectoderm compared
to the endoderm at stage 4 (Hamburger & Hamilton, 1951) in designing our
dissociation technique. Bellairs (1963) has described terminal bars in the
ectoderm of the area opaca. These junctions may confer greater cohesion on
the ectoderm and allow it to be separated as sheets from the endoderm cells
which do not form an epithelium and are irregularly and loosely arranged. The
endoderm cells are also separated by varying quantities of yolk spheres and yolk
membranes; these additional features may facilitate their dissociation.
The area opaca endoderm cells have been identified on the basis of our own
examination of histological sections from whole embryos and pieces of tissue
dissected from living embryos, from the morphology of the aggregates (see also
Sanders & Zalik, 1976) as well as the morphological descriptions of Bellairs
(1963). While the cells that we obtain are of varying sizes (Fig. 1), the majority
are large, and contain abundant yolk. Most of them, with the exception of the
primordial germ cells, are destined to form the inner lining of the yolk sac,
although they may not all be at the same stage of differentiation (Bennett, 1973).
In view of the ease with which the area opaca endoderm cells are dissociated
compared with the rest of the embryo (Milos, Phillips & Zalik, in preparation),
their efficient aggregation kinetics are perhaps surprising. The kinetics that we
observe may reflect either the negligible loss or rapid regeneration of surface
receptors involved in making adhesions. In addition the large size and density
of the cells may be a factor affecting aggregation. It is also possible that in vivo
the cells may be prevented from making strong adhesions because of intercalated yolk and yolk membranes (Bellairs, 1963).
The rapid aggregation kinetics do, however, preclude an analysis of aggregation at the level of the whole population. Theoretically, the disappearance of
single cells and the appearance of clumps of varying sizes and their accretion
with one another can be monitored with this aperture tube up to a particle
volume of 2-1 x 106 ^m3. However, even by 10 min of aggregation clumps have
formed which are beyond the range of the aperture tube. Thus, aggregation
under these conditions is too fast to permit a total channel analysis. Total
aggregation kinetics, especially cell to cell adhesion, have not been studied until
this is done (see also Whur et ah, 1977). Such a study may be possible in this
system if we decrease the size of the initial inoculum.
The aggregation kinetics and aggregate morphogenesis that occur in L-15
medium have arbitrarily been chosen as the base line pattern. The experiments
in PCS and G-PCS show that although almost as efficient single particle disappearance occurs in these two media as in L-15, aggregate growth and morphogenesis are inhibited. Thus, there are additional metabolic requirements for
134
N. MILOS, S. E. ZALIK AND R. PHILLIPS
compaction and morphogenesis which cannot be met by PCS or G-PCS. More
experiments are needed to determine the nature of these parameters. That
cavitation in L-15 medium is not an adaptation of these aggregates to our
culture conditions is supported by our findings that aggregates constructed of
area pellucida cells at the same developmental stage are compact and do not
cavitate.
In vivo the endoderm cells of the area opaca arrange themselves into a single
layer of columnar cells which forms the absorptive layer of the yolk sac.
Bennett (1973) has suggested that this morphological differentiation depends on
the presence of mesoderm. In many of the aggregates cultured in L-15 the walls
of the aggregates become very thin (see Fig. 6 b) and acquire an epithelial appearance. It is thus possible that cavitation and the formation of an epithelium are
intrinsic properties of the endoderm cells of the area opaca which do not depend
either on the presence of other cell types or on normal epiboly with the vitelline
membrane as substrate. The mesoderm may thus play only a permissive role.
The spreading tendency of amphibian endoderm and the cavitation that occurs
within aggregates of these cells have been described by Holtfreter (1944).
Whether specialized junctions appear in chick endoderm aggregates which
create a microenvironment as is the case in the mammalian embryo (Ducibella,
Albertini, Anderson & Biggers, 1975) remains to be established. Our observation
that in some cases cavities (see Fig. 6 b) contain a small amount of cellular debris
is similar to that of Martin, Wiley & Damjanov (1977) who observed debris in
developing cavities of teratocarcinoma embryoid bodies. The suggestion by
these authors that localized death may contribute to cavity formation warrants
further study.
In conclusion, we have shown that a relatively large population of endoderm
cells can be isolated and made to aggregate under simple culture conditions.
Most importantly, the aggregates are capable of constructing a cohesive tissue
pattern. This suggests that some of the parameters that affect the adhesion of
these cells in vitro may be relevant to the morphogenesis that occurs in vivo.
This work was supported by the National Cancer Institute and the National Research
Council of Canada. We thank Jack Scott for help with the photography and Eva Dimitrov
and Vi Scott for help with the embryos.
REFERENCES
R. (1958). The conversion of yolk into cytoplasm in the chick blastoderm as shown
by electron microscopy. /. Embryol. exp. Morph. 6, 149-161.
BELLAIRS, R. (1963). Differentiation of the yolk sac of the chick studied by electron microscopy. J. Embryol. exp. Morph. 11, 201-225.
BENNETT, N. (1973). Study of yolk-sac endoderm organogenesis in the chick using a specific
enzyme (cysteine lyase) as a marker of cell differentiation. /. Embryol. exp. Morph. 29,
159-174.
DUCIBELLA, T., ALBERTINI, D. F., ANDERSON, E., & BIGGERS, J. D. (1975). The preimplantation mammalian embryo: characterization of intercellular junctions and their appearance
during development. Devi Biol. 45, 231-250.
BELLAIRS,
Aggregation of chick endoderm cells
135
H. & KOCHAV, S. (1976). From cleavage to primitive streak formation: a
complementary normal table and a new look at the first stages of the development of
the chick. I General morphology. Devi Biol. 49, 321-337.
EYAL-GILADI, H., KOCHAV, S. & YERUSHALMI, S. (1975). The sorting-out of thymidinelabelled chick hypoblast cells in mixed epiblast-hypoblast aggregates. Differentiation 4,
57-60.
FRASER, B. R. & ZALIK, S. E. (1977). Lectin-mediated agglutination of amphibian embryonic
cells. /. Cell Sci. 27, 227-243.
HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development of
the chick embryo. /. Morph. 88, 49-92.
HOLTFRETER, H. (1943). A study of the mechanics of gastrulation. Part I. /. exp. Zool. 94,
261-318.
HOLTFRETER, H. (1944). A study of the mechanics of gastrulation. Part U. /. exp. Zool. 95,
171-212.
MACARAK, E. J. (1975). An ultrastructural study of adhesive junctions in reaggregates of
unincubated chick embryos. / . Morph. 147, 41-49.
MARTIN, G. R., WILEY, L. M. & DAMJANOV, I. (1977). The development of cystic embryoid
bodies in vitro from clonal teratocarcinoma stem cells. Devi Biol. 61, 230-244.
MILOS, N., ZALIK, S. & PHILLIPS, R. (1977). The aggregation of extraembryonic endodermal
cells from early chick embryos. J. Cell Biol. 75, 36a.
MIURA, Y. & WILT, F. H. (1970). The formation of blood islands in dissociated-reaggregated
chick embryo yolk sac cells. Expl Cell Res. 59, 217-226.
OVERTON, J. (1962). Desmosome development in normal and reassociating cells in the early
chick blastoderm. Devi Biol. 4, 532-548.
PANNETT, C. A. & COMPTON, A. (1924). The cultivation of tissues in saline embryonic juice.
Lancet 205, 381-384.
ROMANOFF, A. L. (1960). The Avian Embryo. Structural and Functional Development. New
York: Macmillan.
SANDERS, E. J. & ZALIK, S. E. (1976). Aggregation of cells from early chick blastoderms.
Differentiation 6, 1-11.
SATOH, N., KAGEYAMA, T. & SIRAKAMI, K. I. (1976). Mobility of dissociated embryonic cells
in Xenopus laevis: its significance to morphogenetic movements. Devel., Growth and Diff.
18, 55-67.
STEINBERG, M. S. (1963). 'ECM': its nature, origin and function in cell aggregation. Expl
Cell Res. 30, 257-279.
TRINKAUS, J. P. (1963). The cellular basis of Fundulus epiboly. Adhesivity of blastula and
gastrula cells in culture. Devi Biol. 7, 513-532.
UMBREIT, J. & ROSEMAN, S. (1975). A requirement for reversible binding between aggregating
embryonic cells before stable adhesion. /. biol. Chem. 250, 9360-9368.
WHUR, P., KOPPEL, H., URQUHART, C. & WILLIAMS, D. C. (1977). Quantitative electronic
analysis of normal and transformed BHK21 fibroblast aggregation. /. Cell Sci. 23, 193209.
ZALIK, S. E. & SANDERS, E. J. (1974). Selective cellular affinities in the unincubated chick
blastoderm. Differentiation 2, 25-28.
ZWILLING, E. (1960). Some aspects of differentiation: disaggregation and reaggregation of
early chick embryos. Natn. Cancer Jnst. Monogr. 2, 19-39.
ZWILLING, E. (1963). Survival and non-sorting of nodal cells following dissociation and
reaggregation of definitive streak chick embryos. Devi Biol. 7, 642-652.
EYAL-GILADI,
(Received 14 June 1978, revised 13 December 1978)