/ . Embryol exp. Morph. Vol. 59, pp. 263-279, 1980
Printed in Great Britain © Company of Biologists Limited 1980
263
Coaggregation and formation of a joint
myocardial tissue by embryonic mammalian and
avian heart eel Is
ByASISH C. NAG1, MEI CHENG AND
CHRISTOPHER J. HEALY
From the Department of Biological Sciences, Oakland University,
Rochester
SUMMARY
Intercellular adhesion and tissue reconstruction from homologous dissociated embryonic
cells from two species were studied. Dissociated 12-day-old embryonic rat heart cells and
50-h-old embryonic chick heart cells were labeled with tritiated thymidine and allowed to
aggregate in Erlenmeyer flasks during rotation culture on a gyratory water-bath shaker. The
cultures were continued for 72 h. Cell aggregates were examined microscopically for evidence of contractility and subsequently processed at intervals between 1 and 72 h for transmission electron microscopic autoradiography. Rat and chick hearts used in this study
appeared comparable in their stage of development and cellular composition. With the
exception of mature blood cells and some fibroblastic non-muscle cells, all chick cardiac
muscle cells were labeled with tritiated thymidine. As the cultures continued, aggregates
increased in size by continuous accretion of cells and joining of small clusters. The cells within
these mixed aggregates exhibited synchronous contractility from 1 h until the cultures were
terminated. Most of the aggregation in both control and mixed aggregates was completed
within 24 h. Control aggregates consisted of cells from a single species, either chick or rat.
Approximately 18 % of labeled chick cardiac muscle cells established intercellular contacts
and junctions with unlabeled rat cardiac muscle cells after 6 h of culture. This increased to
72 % after 72 h in culture. The junctions observed between chick and rat cardiac muscle
cells were desmosomes and hemidesmosomes. Approximately 13 % of the cardiac muscle
cells of one species either chick or rat were found scattered within clusters which contained
.15 % of the cells from the other species after 72 h of culture. These scattered cells did not form
junctions with monospecific cell groups. The implications of these intercellular adhesions
between the same and different species are discussed, bearing two hypotheses in mind: (1)
Moscona and collaborators, (2) Burdickand Steinberg. This study suggests that cellular as
well as species identity plays an important role in the determination of intercellular adhesion
among the embryonic cells from different species.
INTRODUCTION
Cell aggregation, intercellular adhesion and tissue reconstruction from
dissociated embryonic cells derived from the same tissue but different species
1
Author's address: Department of Biological Sciences, Oakland University, Rochester,
Michigan 48063, U.S.A.
264
A. C. NAG, M. CHENG AND C. J. HEALY
have been studied by a number of investigators (Moscona, 1957a, b; 1961,
1962, 1964; Roth, 1968; Burdick and Steinberg, 1969; Burdick, 1970; Garber
& Moscona, 1964, 1972; Grady & McGuire, 1976). Moscona and his collaborators (ibid), studying aggregates containing mouse and chick embryonic
cartilage, skin, liver, neural retina and kidney cells, reported that mouse and
chick cells of the same type did not sort out from each other according to species.
These studies led to a hypothesis which advocated that, for any one embryonic
cell type, the properties responsible for cell sorting were indistinguishable among
even very distantly related warm-blooded vertebrates species. Roth (1968),
examining aggregates of embryonic chick and mouse liver, heart, and neural
retina cells, observed that the probability of intercellular adhesion between the
same type of cells derived from the same species was slightly higher than that of
the two different species. He concluded that tissue specificity plays a larger role
in aggregation than does species specificity. Later, Burdick & Steinberg (1969)
working with embryonic heart cells from mouse and chick at light microscopic
level obtained opposite results. The mouse and chick myocardial cells segregated according to species. This study led to a hypothesis that species differences
controlled selective intercellular adhesion and/or sorting out phenomena.
More recently Burdick (1972) reported that mouse and chick liver cells may
have different cell surface recognition properties. Although these cells did not
sort out from one another in bispecific aggregates, they exhibited different
sorting behaviour in tissue fusion and coaggregation experiments when mixed
with a third cell type, embryonic chick heart ventricle cells. Thus, the present
status of the intercellular adhesion phenomenon between dissociated embryonic
cells derived from the same tissue but different species remains unresolved. In
the present study the problem has been investigated with the help of the electron
microscope after labeling chick heart ventricular cells with a radioactive isotope
and subsequently mixing them with unlabeled rat ventricular cells. Our results
indicated that the labeled chick cells established intercellular contacts and
junctions with the unlabeled rat cells and formed a synchronously beating
myocardial tissue.
MATERIALS AND METHODS
Twelve-day-old rat embryos (Gruneberg, 1943) and 50-h-old chick embryos
(Hamburger & Hamilton, 1951) were used. The chick heart cells were labeled
with [3H]thymidine by the administration of 10 /tCi [3H]thymidine each time,
through an opening on the shell, at 12, 24, 36 and 48 h of incubation. The
preparation of embryonic rat and chick heart cell suspensions and their aggregation by rotation has been previously described (Nag & Buszke, 1977; Nag,
1978). The ventricles were minced and incubated for 45 min in 0-25 % trypsin
(Grand Island Biological Co., Grand Island, N.Y.) in calcium- and magnesiumfree Tryode's solution. Trypsin was then discarded and the mince was washed
several times with culture medium by successive centrifugations. The culture
Coaggregation of mammalian and avion heart cells
265
medium consisted of 88 % Eagle's basal medium, 10 % fetal calf serum (GIBCO
1 % L-glutamine (GIBCO) and 1 % penicillin-streptomycin mixture (GIBCO).
The tissue pieces were dispersed into a suspension of single cells in culture
medium after washing. Clumps of cells and residual tissue were removed by
filtering cell suspensions through wire mesh (porosity 80 /mi). Cell suspensions
from chick and rat were mixed together in ratios ranging 1:1 to 4:1 (chick:rat)
at total cell concentrations of 3 x 106 cells suspended in 3 ml of medium in
25 ml Erlenmeyer flasks. Cell suspensions were gassed with 5 % CO2 + 95 %
air and incubated at 37 °C in a gyratory water bath shaker at 70 rev./min.
Cell aggregates were examined by phase-contrast microscope for evidence of
contractility I h and 72 h after initiating the cultures, and subsequently fixed in
Karnovsky's fixative (1/2 strength) for 1£ h at room temperature, post-fixed in
1 % OsO4 in 0-1 M cacodylate buffer, dehydrated in graded ethanol and propyline
oxide, and embedded in araldite. Thin sections were cut on a Porter-Blum
ultramicrotone and picked up on 200 mesh grids and attached by their edge to a
small piece of double-edged Scotch tape fixed to a microscope slide. The sections were then coated with a thin film of Ilford L-4 photographic emulsion
(Caro, 1969). Fifteen grams of this gelled emulsion was added to 15 ml of distilled water and melted at 45 °C in a water bath. A loop of nickel wire was dipped
into the emulsion and withdrawn slowly, forming a thin film in the loop. This
film gelled almost immediately. The loop was touched to the surface of the slide.
The gelled film fell on the grids and adhered to them firmly. The coated sections
were dried overnight and stored at room temperature in a black box containing
desiccant. They were exposed for 4-8 weeks, developed in Microdol-X for 3
min, rinsed briefly in distilled water, and fixed in 20 % sodium thiosulphate
solution for 10 min. The sections were then doubly stained with uranyl acetate
and lead citrate. Some of the uncoated sections from experimental and control
batches of aggregates were doubly stained as above. Since there was only a small
number of nuclei in the plane of section, serial sections were cut to trace the
nuclei in some blocks. All samples were examined and photographed with a
Philips 200 electron microscope operated at an accelerating voltage of 60 kV.
Uncoated sections were mainly used for fine structural studies. The electron
microscopic autoradiographs from mixed aggregates were analysed by counting
300-500 labeled cells of each aggregate. Five such aggregates at each time point
(6, 12, 24, 48 and 72 h) of each experiment were counted. Altogether, eight
experiments were carried out and the results were expressed as percentage of
labeled muscle cells in contact with unlabeled muscle cells.
RESULTS
Muscle cells were the most numerous cell type in the hearts of both species
at the stage of development employed in the present study. In addition, there
were non-muscle cells in the cell suspensions. With the exception of mature
266
f
A. C. NAG, M. CHENG AND C. J. HEALY
Coaggregation of mammalian and avian heart cells
267
blood cells and some fibroblastic non-muscle cells, all chick cardiac muscle cells
were labeled with [3H]thymidine (Fig. 1). Within one hour of culture, most of
the cells formed small clusters of two to five cells in the chick-rat aggregates as
well as control aggregates containing cells of one species. The experimental
and control aggregates did not exhibit differences in aggregation after trypsinization. Although it was reported (Curtis, 1967) that different cell types may
have different recovery times after trypsinization, our studies (Nag, 1978;
unpublished data) along with others (Shimada, Moscona & Fischman, 1974),
indicate that the cardiac muscle cells of chick and rat do not show any differences in the recovery time which might affect intercellular adhesion through the
formation of junctions. Labeled chick cardiac muscle cells established desmosomes with the unlabeled rat cardiac muscle cells as early as one hour of culture
(Figs. 2, 3). The cells within these mixed aggregates exhibited synchronous
contractility from one hour to the end of the culture period. Most of the aggregations were completed within 24 h. Bispecific aggregates with 1:1 (chick-rat)
cell ratio, after 1 to 72 h of culture, showed cardiac muscle cells of both species
interspersed throughout the aggregates (Fig. 4).
This observation did not conform with that of the light microscopic study of
Burdick & Steinberg (1969), who reported that mouse and chick heart cells were
extensively sorted out after 2 to 2-5 days of culture. Our electron microscopic
studies revealed the presence of desmosomes and hemidesmosomes between the
closely adherent labeled chick and unlabeled rat cardiac muscle cells at all stages
of aggregation studied (Figs 5, 6, 7). Although cardiac muscle cells were readily
distinguished from the non-muscle cells with the help of myonbrils, there was
sometimes difficulty in identifying muscle cells with myofibrils during lowpower electron microscopy, which was circumvented by the presence of pockets
of abundant glycogen particles, another characteristic feature of the cardiac
muscle cells (Figs 8, 9). In spite of tracing the nuclei on the serial sections, some
nuclei were not found due to technical difficulties. Consequently, approximately 1-5 % cells were not identified, owing to the absence of nuclei in the
cells. The measurements of intercellular contact of the labeled chick cardiac
muscle cells with the unlabeled rat cardiac muscle cells indicated that approximately 18 % of labeled chick cells established intercellular contacts after 6 h of
culture. This percentage increased from 18 to 72 % after 72 h in culture (Table
1, Fig. 10). The interspersed chick and rat cardiac muscle cells underwent
differentiation of myofibrils during the culture period. The dimensions of the
intercellular space in some of the cell adhesion regions between chick and rat
Fig. 1. Autoradiograph of the freshly dissociated labeled chick cardiac muscle cells.
Note the incorporation of tritiated thymidine into the nuclei as exhibited by specific localization of silver grains (Sg) on the nuclei. Mb, myofibril; Mf, myofilaments;
N, nucleus, x 5750.
268
A. C. NAG, M. CHENG AND C. J. HEALY
Fig. 2. A portion of an aggregate of heart cells after 1-hour in rotation culture.
Note labeled chick cardiac muscle cell established a desmosomal junction (arrows)
with the unlabeled rat cardiac muscle cell. Mf, myofilaments; N, nucleus, x 15500.
Fig. 3. A higher magnification electron micrograph of a portion of Fig. 2, showing
particularly the well-developed desmosome (arrows), x 41250.
Coaggregation of mammalian and avion heart cells
269
Fig. 4. Three hour aggregate showing the interspersed labeled and unlabeled heart
cells. The labeled chick cell contained vacuolated (Vc) disentegrated myofilaments
(Dm) which are commonly visible in freshly dissociated cardiac muscle cells as well
as muscle cells of the early aggregates. Fb, fibroblastic cell; G, glycogen particles;
Mb, myofibril; N, nucleus, x 19315.
cells ranged between 2 and 10 nm. Intercellular spaces were not discernible in
close membrane contact regions.
In bispecific aggregates (1:1, chick:rat) approximately 13 % of the cardiac
muscle cells of one species were found scattered within clusters comprising
15 % of other species (chick or rat) at the end of the culture period (Table 1).
Aggregates with 4:1 (chick:rat) mixed cell ratio exhibited the expected
crowding of labeled chick cells along with scanty non-labeled rat cells. Although
18
EMB 59
270
A. C. NAG, M. CHENG AND C. J. HEALY
Fig. 5. Six hour aggregate showing intercellular contact and desmosomes (arrows)
between the labeled chick and unlabeled rat cardiac muscle cells. Mf, myofilaments;
IS, nucleus, x 19315.
labeled chick cells were in close apposition with one another due to their large
numbers, scanty scattered non-labeled rat cells were present among them
(Fig. 11). These scattered rat cells were in close apposition to chick cells. Our
analysis indicated that approximately 2 % of the labeled cells established
contact with the unlabeled cells after 6 h of culture. This rose to 26 % after
24 h, and leveled off until the termination of culture (Table 2, Fig. 12). Monospecific associations of rat cells were rare in these bispecific aggregates (4:1,
chick:rat cells).
Coaggregation of mammalian and avian heart cells
271
Fig. 6. Closely adherent labeled chick and unlabeled rat muscle cells after 9-hours
in rotation culture contained hemidesmosomes (large arrows) and desmosomes
(small arrows) between them. Mb, myofibril; Mf, myofilaments; N, nucleus,
x 12990.
Fig. 7. A higher magnification micrograph of a portion of Fig. 6 showing the well
differentiated hemidesmosomes (large arrows), x 32450.
18-2
272
A. C. NAG, M. CHENG AND C. J. HEALY
Fig. 8. Twelve hour aggregate showing interspersed labeled and unlabeled chick and
rat cardiac muscle cells respectively. Note the pockets of glycogen particles (G) in
the muscle cells, x 49 500.
Fig. 9. A higher power electron micrograph of a portion of Fig. 8 showing the distribution of glycogen particles in a labeled chick cell, x 20295.
20 ( + 0-5)
3-8 (±1-2)
70 (±2-5)
80 (±1-8)
80 (±20)
21 (±0-5)
4-3 (±1-7)
8-2 (±3-0)
6-5 ( + 20)
7-7 (±1-5)
18 (±41)
53 (±8-3)
72 ( + 10-1)
72 (±10-3)
72 (±120)
478 (±35)
502 (±42)
464 (±27)
388 ( + 31)
497 ( + 25)
6
12
24
48
72
37-5
20-3
7-2
5-5
6-7
(±7-2)
(+6-4)
(+31)
(±1-7)
(+1-5)
Labeled
cells (%)
A
* Mean of 40 aggregates at each time point with standard deviations (±). (See Materials and Methods for details).
t Cells adhered via desmosomes or hemidesmosomes.
t No junctions between the cells. Standard deviations in parenthesis.
unlabeled
cells (%)
labeled
cells (%)
Adhered
cellsf (%)
No. of labeled/
unlabeled cells*
A
Period in
culture (h)
A
5-6 (±20)
80 (±2-2)
5-6 (±1-7)
39-8) +7-5)
19-5 (±5-3)
Non-labeled
cells (%)
Table 1. Quantitative analysis of the adherent and non-adherent cardiac muscle cells in aggregates (1:1 cell ratio; chick:rat)
Adherent labeled to labeled;
unlabeled to unlabeled cardiac
muscle cells
Adherent labeled to unlabeled cardiac
cardiac muscle cellsf
a
rS
op
Oq
i
274
A. C. NAG, M. CHENG AND C. J. HEALY
80
70
60
50
40
30
20
10
10
12
24
48
Hours in culture
72
Fig. 10. Quantitation of labeled/unlabeled adhered cells at different times of culture (1:1 cell ratio; chick:rat). Standard deviations (±) are represented by bars on
the columns.
DISCUSSION
In the present study cardiac muscle cells did not show active segregation or
sorting-out phenomena according to species, as reported by Burdick and Steinberg (1969). Intercellular contact or adhesion through intercellular junctions
cannot be determined with certainty with the light microscope. Since we
observed close intercellular contacts through close membrane appositions and
the adhesion through intercellular junctions between the embryonic rat and
chick cardiac muscle cells, the concept of true segregation or sorting out does
not hold good for the embryonic myocardial cells of the two species we have
studied, i.e. rat and chick. Our findings partially support the hypothesis of
Moscona and his collaborators (1957a, b, 1961, 1964, 1973) that bispecific
combinations of homologous cells do not sort out sharply according to species.
In our studies of bispecific aggregates with 1:1 cell ratio, we observed that
approximately 72 % of the labeled chick cells established intercellular contacts
with unlabeled rat cells after 72 h of culture. Among the rest of the cells, approximately 15 % of the cells of one species came in close contact with one another
and often showed intercellular junctions. The rest (approximately 13 %) of the
Coaggregation of mammalian and avion heart cells
275
11
Fig. 11. Autoradiograph of part of a 24-hour mixed aggregate (4:1; chick: rat cells),
showing the presence of a non-labeled rat cardiac muscle cell (Nc) among the
labeled chick cardiac muscle cells. Close intercellular contacts (small arrows) between
chick and rat cells and a hemidesmosome (Hd) on the rat cell are shown. A labeled
degenerated (Dc) cell is present in the field. Fm, free myofilaments; G, pocket of
glycogen particles; Mb, myofibril; N, nucleus, x 11250.
cells were scattered among the monospecific myocardial cells of the other species
(rat or chick). These scattered cells did not exhibit intercellular junctions with
adjacent cell groups. Since approximately 1-5 % of cells were not identified
owing to the absence of nuclei in the plane of sections, the above quantification
503
481
493
389
421
(±47)
(±42)
(±38)
(±37)
(±27)
labeled cells
57-4 (±12-0)
18-5 (±40)
0-6 (±015)
0
0
labeled to
unlabeled cells
2 (±0-5)
13 (±5-3)
26 (±7-2)
26 (±6-3)
26 ( ± 7 0 )
Labeled to
labeled cells
260 (±6-3)
640 (±8-4)
73-4 (±101)
740 (±7-2)
740 ( ± 8 0 )
14-6 (±4-4)
4-3 (±1-2)
0
0
0
unlabeled cells
Non-adherent cardiac muscle cellsj (%)
* Mean of 40 aggregates at each time point with standard deviations (±). (See Materials and Methods for details).
t Cells adhered via desmosomes or hemidesmosomes.
t No junctions between the cells. Standard deviations in parenthesis.
6
12
24
48
72
Period in No. of labeled/
culture (h) unlabeled cells*
A
Adherent cardiac muscle cellsf
Table 2. Quantitative analysis of the adherent and non-adherent cardiac muscle cells in aggregates (4:1 cell ratio; chick:rat)
wX
r
O
O
O
a
x
o
o
p
to
Coaggregation of mammalian and avian heart cells
277
40
I **
J
30
*8
.2 3
20
10
12
12
24
48
Hours in culture
72
Fig. 12. Quantitation of adhered labeled/unlabeled cardiac muscle cells in aggregates containing 4:1 mixed cell ratio (chick:rat). Bars represent standard deviations
should be considered slightly above or below the actual number of adherent
cells. The monospecific association of a small number of cells observed in the
present studies may be accounted for by the effect of random cell association
which is not expected to show that all chick and rat cardiac muscle cells will
adhere with one another in a 1:1 ratio. These observations raise a question as to
the problem of sorting out of a small number of cells in these bispecific homologous cell aggregates. These small monospecific myocardial cell groups from
each species did not impair or interfere with the coaggregation of most of the
bispecific homologous cells, unlike the heterologous cells of the bispecific
aggregates, where heterologous cells were sorted out into separate groupings
and reconstructed different tissues, each consisting only of cells from the corresponding species (Moscona, 1973). The small monospecific myocardial cell
groups in our study did not show the pattern of sorting out reported by Burdick
& Steinberg (1969), who reported that mouse cells were sorted out at the periphery and the chick cells were located internally. Moreover, our findings on the
bispecific aggregates containing 4:1 (chick:rat) mixed cell ratio demonstrate
that the embryonic rat and chick cardiac muscle cells establish intercellular
junctions between them, irrespective of cell numbers from each species in the
culture.
The present findings provide evidence that the embryonic rat and chick
cardiac muscle cells can coaggregate and form a bispecific tissue which consists
of mosaics of cells and cell groups from the two species. Such bispecific myocardial tissue is capable of beating synchronously. It is evident from our studies
that homologous cells from different species can recognize their cellular identity
and form joint tissues. Although monospecific association of a small number of
cells was found in the present study, one possible interpretation for their
278
A. C. NAG, M. CHENG AND C. J. HEALY
presence was random collision between cells during rotation which resulted in
the aggregation of a small number of monospecific cells as discussed above. The
present study suggests that the cellular as well as the species identity play an
important role in the determination of intercellular adhesion among the
embryonic homologous cells from different species.
The authors express their appreciation for the helpful criticism by Drs Radovan Zak and
Saradindu Dutta, the University of Chicago, and Wayne State University, respectively. The
authors acknowledge the research support of grants from the Michigan Heart Association
Grant-in-Aid, and NIH BSRG No. 34173.
REFERENCES
BURDICK:, M.. L. (1970). Cell sorting out according to species in aggregates contains mouse
and chick embryonic mesoblast cells. /. exp. Zool. 175, 357-367.
BURDICK, M. L. (1972). Differences in the morphogenetic properties of mouse and chick
embryonic liver cells. / . exp. Zool. 180, 117-126.
BURDICK, M. L. & M. S. Steinberg (1969). Embryonic cell adhesiveness: Do species differences exist among warm blooded vertebrates? Proc. natn. Acad. Sci. U.S.A. 63,1169-1173.
CARO, L. G. (1969). A common source of difficulty in high resolution radioautography.
J.CellBiol. 41, 918-919.
CURTIS, A. S. G. (1967). The Cell Surface; Its Molecular Role in Morphogenesis, London:
Logos Press Ltd. pp. 129-31.
GARBER, B. & A. A. MOSCONA (1964). Aggregation in vivo of dissociated cells. I. Reconstruction of skin in the chorioallantoic membrane from suspensions of embryonic chick
and mouse skin cells. / . exp. Zool. 55, .179-202.
GARBER, B. & A. A. MOSCONA (1972). Reconstruction of brain tissue from cell suspensions.
I. Aggregation patterns of cells dissociated from different regions of the developing brain.
II. Specific enhancement of aggregation of embryonic cerebral cells by supernatant from
homologous cell cultures. Devi Biol. 27, 217-234.
GRADY, S. R. & E. J. MCGUIRE (1976). Intercellular adhesive selectivity. HE. Species selectivity of embryonic liver intercellular adhesion. /. Cell Biol. 71, 96-106.
GRUNEBERG, J. (1943). The development of some external features in mouse embryos.
/. Heredity 34, 89-92.
HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development
of the chick embryo. /. Morph. 88, 49-92.
MOSCONA, A. (1957a). The development in vitro of chimeric aggregates of dissociated embryonic chick and mouse cells. Proc. natn. Acad. Sci. U.S.A. 43, 184-194.
MOSCONA, A. (19576). Formation of lentoids by dissociated retinal cells of the chick embryo.
Science 125, 598-602.
MOSCONA, A. (1961). Rotation-mediated histogenetic aggregation of dissociated cells: A
quantifiable approach to cell interactions in vitro. Expl Cell Res. 22, 455-475.
MOSCONA, A. A. 1(962). Analysis of cell recombinations in experimental synthesis of tissues
in vitro. J. cell. comp. Physiol. Suppl. I. 60, 65-71.
MOSCONA, A. A. (1964). Studies on stability of phenotypic traits in embryonic integumental
tissues and cells. In: The Epidermis (ed. W. Montagna), pp. 83-104. New York: Academic
Press.
MOSCONA, A. A. (1973). Cell Aggregation. In Cell Biology in Medicine (ed. Bitter, E. E.),
pp. 571-591. New York: John Wiley & Sons Inc.
NAG A. C. (1978). Reconstruction of mammalian heart tissue from embryonic heart cell
suspension with reference to the aggregation of adult heart cells. Cytobiologie 23,199-223.
NAG, A. C. & D. S. BUSZKE (1977). Reconstruction of mammalian heart tissue from isolated
embryonic heart cells. 35th Ann. Proc. Elec. Micro. Soc. Amer. (ed. G. W. Bailey), pp.
580-581.
Coaggregation of mammalian and heart cells
ROTH, S. (1968). Studies on intercellular adhesive selectivity. Devi Biol. 18, 602-631.
SHIMADA, S., A. A. MOSCONA & D. A. FISCHMAN (1974). Scanning electron microscopy
279
of cell
aggregation: cardiac and mixed retina-cardiac cell suspensions. Devi Biol. 36, 428-446.
(Received 7 January 1980, revised 22 March 1980)