/. Embryo/, exp. Morph. Vol. 43, pp. 289-313, 1978
Printed in Great Britain © Company of Biologists Limited 1978
289
Localization and synthesis of alphafoetoprotein
in post-implantation mouse embryos
By M. D Z I A D E K 1 AND E. A D A M S O N 1
From the Department of Zoology,
Oxford
SUMMARY
The localization and synthesis of alphafoetoprotein (AFP) during mouse embryogenesis
were studied by immunoperoxidase and by immunoprecipitation after radioactive labelling,
using an antiserum prepared against AFP. AFP is first detectable in embryos on the 7th day
of gestation (7th day embryos). In 7th and 8th day embryos AFP is confined to visceral
(proximal) endoderm cells around the embryonic region of the egg cylinder. Visceral extraembryonic and parietal (distal) endoderm cells do not contain AFP. By the 9th day of
gestation AFP is also present in the extra-embryonic ectoderm, mesoderm and embryonic
ectoderm cells around the three cavities of the embryo. These tissues do not synthesize AFP
when cultured in isolation, but can adsorb AFP when it is added to the medium. On the
12th day of gestation AFP synthesis is confined to the endoderm layer of the visceral yolk sac.
It is concluded that the ability to synthesize AFP is a property which is restricted to the
visceral endoderm during early post-implantation development. The presence of AFP in
other tissues of the embryo appears to be due to adsorption.
INTRODUCTION
The initial allocation of cells of early mouse embryos, to the outer trophoblast layer and to the inner cell mass, depends on cell position (Tarkowski &
Wroblewska, 1967; Graham, 1971). Differences in gene activity between these
two cells can now be detected in early blastocysts (60-100 cells) by specific and
characteristic differences in synthesized proteins (Van Blerkom, Barton &
Johnson, 1976). After blastocyst implantation, a morphologically distinct
endoderm layer differentiates on the surface of the inner cell mass which faces
the blastocoelic cavity (Enders, 1971 ; Nadijcka & Hillman, 1974). This primitive
endoderm appears to give rise to the parietal (distal) endoderm lying adjacent
to the trophoblast (which together are usually called the parietal yolk sac), and
the visceral endoderm (or proximal endoderm, which is used here to mean all
endoderm cells around the egg cylinder) (Snell & Stevens, 1966; Gardner &
Papaioannou, 1975). The visceral endoderm can be further subdivided into two
regions: the visceral embryonic endoderm, overlying the embryonic region of
the egg cylinder; and the visceral extra-embryonic endoderm which overlies the
extra-embryonic region. The visceral endoderm of the yolk sac synthesizes and
secretes alphafoetoprotein (AFP) by 1 \\ days of gestation (Wilson & Zimmerman,
1
Authors' address: Department of Zoology, South Parks Road, Oxford OX1 3PS, U.K.
290
M. D Z I A D E K AND E. ADAMSON
1976). AFP is the first a-globulin to be formed by the mammalian embryo
during development, and is produced predominantly by the yolk sac and foetal
liver (Wilson & Zimmerman, 1976). AFP is not normally produced in adult
tissues, but reappears under certain conditions, such as hepatomas and
teratocarcinomas, and during liver regeneration (reviewed by Abelev, 1971).
A difference in products synthesized by the parietal and visceral endoderm
layers would provide potentially useful markers for analysis of specific gene
activation during cell differentiation in the early mouse embryo. The present
study employs the techniques of immunoperoxidase and immunoprecipitation
after radioactive labelling, in an investigation of localization and synthesis of
AFP in cells of the mouse embryo. The aim was to determine how early in
development synthesis takes place, and whether it is indeed confined to cells of
the visceral endoderm.
MATERIALS AND METHODS
(a) Preparation of anti-AFP antiserum
Embryos (from the 15th to the 20th day of gestation) from mice of several
stocks (C3H, 129J, A2G, CFLP, PO) were dissected rapidly from the uterus
into warmed phosphate buffered saline (PBS, solution A of Dulbecco & Vogt
1954) containing 1 % (w/v) sodium citrate. Amniotic fluid and foetal blood,
were collected, centrifuged to remove the blood cells, and the supernatant was
dialysed against 0-1 M - N H 4 - H C 0 3 before freeze-drying. About 20 mg of the
dried product was dissolved in sample buffer for electrophoretic separation in
Polyacrylamide slab gels (17 x 17 cm x 4-5 mm thickness) using the solutions
described by Davis (1964). After electrophoresis, two strips of the gel were
stained to locate the AFP and from this guide the AFP zone was cut out of the
unstained gel. The gel band was homogenized and the AFP extracted by three
to four successive portions of 0 1 M - N H 4 H C 0 3 with shaking at 4 °C. The
combined extracts were dialysed against 0-1 M - N H 4 H C 0 3 and freeze-dried. The
product was further purified by a second preparative Polyacrylamide electrophoretic separation on a thinner version of the same kind of gel. The gel band
containing AFP (about 1-5 mg) was excised, finely macerated in 2 ml of PBS
and was added to an equal volume of Freund's complete adjuvant (Grand
Island Biol. Co., Berkeley, Calif., U.S.A.). The mixture was emulsified in
a vortex mixer, halved and injected subcutaneously into two rabbits. One
month later the rabbits were reinjected with a solution of AFP (0-5 mg for each
rabbit) in PBS mixed with an equal volume of Freund's incomplete adjuvant.
Ten days later the animals were bled and antibodies against AFP were detected
by Ouchterlony double diffusion tests (Ouchterlony, 1958). Immunoelectrophoresis (Grabar & Williams, 1953) showed the presence of a small amount of
anti-albumin antibodies in the antisera. After absorption with adult serum
proteins the resulting antisera were specific for AFP as shown in Fig. 1. It was
shown by titration (Hudson & Hay, 1976) that 25 JLL\ of undiluted antiserum
AFP in mouse
embryos
291
Fig. 1. Immunoelectrophoresis of 5 /tl serum (centre well) from 19th day embryos.
Electrophoresis was from left to right towards the anode and was continued for 2 h
at 8 V/cm and 4 °C in 005 M barbitone buffer, pH 8-6. The troughs were filled with
60-80 /tl of antisera from each of two rabbits (A and B) and diffusion occurred for
1-2 days before the plate was photographed using oblique illumination and a black
background. The single precipitin arc demonstrates the specificity of the antiserum
to AFP. Precipitin arcs were never produced in control experiments using adult
mouse serum (not shown).
precipitated approximately 20/tg of AFP. The concentration of antiAFP
immunoglobulin in the antiserum was approximately 2-5 mg/ml.
(b) Immunoper oxidase method
Cellular localization of AFP in early mouse embryos and isolated tissues at
later Stages was studied by the immunoperoxidase reaction on tissue sections.
(i) Tissue preparation
Embryos from various mouse strains (C3H, 129J, A2G, F 2 progeny from
C3H x 129J and CBA x C57BL) were used with no apparent differences in result.
Female mice were mated overnight and checked for vaginal plugs the next
morning. The day of the plug was designated the first day of pregnancy.
Animals were killed by cervical dislocation and embryos dissected from the
uterus. Sixth to 10th day embryos were fixed within the decidual tissue after a brief
wash in PBS. Embryos for experimental manipulation were dissected from the
decidua into a-medium+10% heat-inactivated foetal calf serum (Stanners,
Elicieri & Green, 1971). Various tissues were dissected from 14th day foetuses
and fixed separately after a brief wash in PBS (these included brain, lung, heart,
gut, kidney, gonad, amnion, yolk sac).
The Sainte-Marie technique was used for tissue fixation (Sainte-Marie, 1962),
with Engelhardt's modification employing a cold mixture of 96 % ethanol with
glacial acetic acid (99:1 v/v) for the fixative (Engelhardt, Goussev, Shipova &
Abelev, 1971). Tissue was placed in cold fixative at 4 °C for 12-24 h, then
dehydrated in two changes of pre-cooled toluene and embedded in paraffin at
292
M. D Z I A D E K AND E. ADAMSON
56 ° C. Blocks were stored at 4 °C. Serial sections were cut at 8 /im thickness
and slides dried overnight at 37 °C.
(ii) Jmmunoperoxidase reaction
Sections were dewaxed, hydrated through successive dilutions of ethanol into
PBS, and washed by gentle agitation in three successive beakers of PBS.
Sections were treated with the rabbit anti-AFP antiserum at 1/50 dilution for
30 min at room temperature; washed thoroughly with PBS; treated with goat
anti-rabbit IgG conjugated with peroxidase (Nordic Immunological Labs.,
Maidenhead, Berks., U.K.) for 30 min; and washed again. Several control
incubations were done on alternate groups of sections to test against nonspecific labelling. These were (a) first incubation with antiserum Which had been
absorbed with excess AFP, to test for specificity of the antiserum for AFP;
(b) omission of the first incubation and treatment only with goat anti-rabbit
IgG to test for cross reactivity of the second conjugate with mouse tissue; and
(c) direct peroxidase reaction on tissue sections to test for any endogenous
peroxidase activity. (Control treatments were invariably negative, and only the
first control above was used routinely.)
Peroxidase activity in the sections was located after reaction with diaminobenzidine tetrahydrochloride (75 mg DAB (Sigma Chemical Co., Surrey, U.K.)
in 100 ml 0-05 M-Tris-HCl buffer at pH 7-6, plus 3 jd H 2 0 2 (30 %)). Sections
were incubated in this reaction mixture for a standard time of 15 min at room
temperature, washed twice in 0-05 M Tris buffer, and post-fixed in 1 % osmium
tetroxide for 10 min at room temperature. Dehydration and mounting were by
standard histological methods. Peroxidase activity was observed as a dark
brown staining reaction, in contrast to low background staining.
(c) Immunoprecipitation after radioactive labelling
Synthesis and secretion of AFP by embryonic tissues was determined by the
presence of radioactively labelled AFP in the culture medium after incubation
in tritiated lysine. Tissues and embryos were radioactively labelled by incubation
in 0-5-10 ml Minimum Essential Medium (MEM) minus lysine, and containing
10 % dialysed foetal calf serum and 50 /tCi/ml [3H]lysine (1 mCi/ml, 33 Ci/mM,
Radiochemical Centre, Amersham, Bucks., U.K.). After 24 h at 37 °C the
incubates were centrifuged at low speed to obtain medium free of cells. Tissue
pellets were retained for DNA measurements. Any further debris was removed
from the medium by centrifugation at 200000 g for 30 min at 4 °C. Carrier AFP
was added (20 /ig) together with 25 jil of specific antiserum. After incubation
at 37 °C for 1 h, the medium was kept at 4 °C for 1-2 days. The immunoprecipitate was collected and washed by centrifugation and resuspension in two
successive portions of 1 ml of 1 % sodium deoxycholate (w/v), 1 % Triton
X-100 (v/v), 0-1 mM lysine in PBS, followed by one wash in 1 ml of 0-9 % NaCl
(w/v), 0-1 mM lysine. The precipitate was dissolved in 30 mM Tris-HCl, pH 6-8,
AFP in mouse embryos
293
2 % SDS (w/v), 2 % mercaptoethanol (v/v), 0-001 % bromophenol blue
(w/v), 15% glycerol (v/v) for analysis by column gels (10x0-5 cm diameter)
and/or by slab gels (17 x 17 cm x 1-5 mm thick). Both types of gel contained
7-5 % Polyacrylamide and 0-1 % SDS. The composition of the gel and solutions
was as described by Laemmli (1970). After fixing and staining in 0-25 %
Coomassie brilliant blue R in 45 % ethanol (v/v), 10 % acetic acid (v/v), and
destainingin 20 % ethanol, 10 % acetic acid, the column gels were sliced and the
radioactivity in each 1 mm slice was determined as described by Adamson (1977).
The slab gels were photographed and dried down for fluorography (Bonner &
Laskey, 1974). Tissue pellets were analysed for DNA by the fluorometric
method described by Morris & Cole (1972). The synthesis of AFP is expressed as
counts per minute (cpm) tritium-labelled AFP synthesized in 24 h per/ig DNA.
(d) Experimental manipulations
For analysis of AFP synthesis by isolated tissues, dissections were made by
two techniques.
(i) Manual dissections using tungsten needles to cut 7th day and 8th day
egg cylinders into separate portions.
(ii) Enzymic separation of tissue layers using a 2-5 % pancreatin 0-5 %
trypsin solution (Levak-Svajger, Svajger & Skreb, 1969). Egg cylinders were
incubated in the enzyme solution for 1 h at 4 °C, and washed several times in
culture medium with added serum. Tissue layers were separated by pipetting
the embryo in a glass pipette with boie diameter slightly less than the diameter
of the egg cylinder. To separate the mesoderm from the endoderm layer of
visceral yolk sacs, the dissected yolk sacs were incubated in the enzyme solution
for 2 h at 4 °C, and the layers were teased apart and washed thoroughly.
(e) Tissue culture
Isolated embryonic tissues were cultured in a-medium plus 10 % foetal calf
serum on bacteriological dishes (Sterilin) at 37 °C in 5 % C 0 2 in air. Cultures
were maintained for up to 72 h.
RESULTS
The results are divided into three sections: the first being a study of the
localization of AFP at different stages of mouse embryogenesis; the second
a study of AFP localization and synthesis by isolated tissue layers; and the third,
a direct analysis of AFP synthesis by labelled amino acid incorporation.
(a) Localization of AFP in mouse embryos
6th to 7th day embryos
AFP was first observed in five out of nine 7th day embryos. None of the ten
6th day embryos analysed were labelled for AFP in any cells. AFP in 7th day
embryos was confined to some of the squamous cells forming the visceral
294
M. D Z I A D E K AND E. ADAMSON
2A
K-
40 \
^ «j
^a^-»^
* •<
t
£^t
_%
^ , *3
AFP in mouse embryos
295
embryonic endoderm (Fig. 2). Serial sections showed that positive cells were
situated predominantly on one side of the egg cylinder, but did not form a continuous patch, since negative cells were present among the positive. All other
cells of the egg cylinder, parietal endoderm and trophoblast were negative.
Early 8th day embryos
Sixteen embryos on the 8th day of gestation were studied. These varied in stage
of development, ranging from a pre-primitive streak stage to a stage of exocoelomic cavity formation. At all these stages, all cells of the visceral embryonic
endoderm were labelled for AFP (Figs. 3, 4). A transition from presence to
absence of cellular AFP was observed between the visceral embryonic endoderm
and visceral extra-embryonic endoderm. This was a consistent observation.
Visceral endoderm cells around the egg cylinder grade from columnar,
vesiculated cells characteristic of the visceral extra-embryonic endoderm, into
flat squamous cells of the embryonic region. The transition from AFP-positive
to negative cells was not directly related to a particular cell morphology, since
it occurred in the zone where both types of cells were present (Fig. 3). In the
stages prior to formation of the exocoelomic cavity, no AFP activity was
observed in the ectoderm, mesoderm or parietal endoderm. At a slightly later
stage when the exocoelom had started to form, AFP was observed at the
luminal surface of embryonic ectoderm cells bordering the pro-amniotic
cavity (Fig. 4).
Late 8th to early 9th day embryos
Three cavities are present in the egg cylinder by 8£ days of gestation; the
amniotic, exocoelomic and ectoplacental cavities (see Snell & Stevens (1966)
for detailed description of morphogenesis in mouse embryos). Ninth day
embryos retained this same basic structure but had grown in size. Six embryos
were studied in detail. AFP labelling in the visceral endoderm was now confined
to cells around the midgirth of the embryo (Fig. 5). The squamous visceral
embryonic endoderm cells around the embryonic region of the embryo were no
longer highly labelled, and the most heavily labelled endoderm cells were
situated in a zone where a more columnar morphology was starting. The future
visceral yolk sac is thought to form from the endoderm plus mesoderm layer
Fig. 2. Immunoperoxidase reaction on sections of a 7th day mouse embryo. (A) AFP
activity is confined to some cells of the visceral embryonic endoderm layer (arrows).
(B) Control section treated with antiserum absorbed with AFP shows no labelling
in this region.
Fig. 3. Sections of a primitive streak stage embryo on the 8th day of gestation,
showing (A) localization of AFP in the visceral embryonic endoderm, and (B) no
activity on a control section. Eect, embryonic ectoderm; VEend, visceral embryonic
endoderm; Ex. ect. extra-embryonic ectoderm; VEx. end, visceral extra-embryonic
endoderm; mes, mesoderm; PE, parietal endoderm.
296
M. DZIADEK AND E. ADAMSON
4A
* **
;
*•
ÉJ.
Y
** »*^
» ~* -
s.
„'«£
""^C
« , - *-*mm "' - - v*
'a
100 fxm
Fig. 4. Sagittal sections through an 8th day embryo at the stage of formation of the
exocoelom. (A) Visceral embryonic endoderm cells are labelled for AFP, and ectoderm cells around the proamniotic cavity show labelling on the luminal border.
(B) Control sections treated with absorbed antiserum show no labelling. PA, proamniotic cavity; EC, exocoelom; Eect, embryonic ectoderm; VE end, visceral
embryonic endoderm; Ex ect, extra-embryonic ectoderm; VEx end, visceral extraembryonic endoderm; PE, parietal endoderm; mes, mesoderm.
TWI
297
AFP in mouse embryos
5 A*-**
* »•** k.'/.t.
5B
VE end.
V Ex end
f
I
%**.
Ex ect
EP
•**
* mes
!
100 ^m
Fig. 5. Sagittal sections through a 9th day embryo showing (A) AFP in visceral endoderm cells at the midgirth of the embryo (arrows); on the periphery of embryonic
ectoderm cells bordering the amniotic cavity; in mesoderm cells around the exocoelomic cavity and on the surface of the allantois; and in extra-embryonic ectoderm
cells around the ectoplacental cavity. (B) No labelling on control section treated with
absorbed antiserum. A, amniotic cavity; EC, exocoelomic cavity; EP, ectoplacental
cavity; am, amnion; ch, chorion; all, allantois; E ect, embryonic ectoderm; VE end,
visceral embryonic endoderm; Ex ect, extra-embryonic ectoderm; Vex end, visceral
extra-embryonic endoderm; mes, mesoderm; PE, parietal endoderm.
298
M. D Z I A D E K AND E. ADAMSON
around the exocoelom. At this stage in development AFP positive visceral
endoderm cells extended less than half way up this primitive yolk sac region
(Fig. 5).
Most of the AFP labelling was now in non-endodermal tissues. Embryonic
ectoderm cells adjacent to the amniotic cavity had high levels of AFP, localized
at the peripheral border. The amnion showed strong intracellular labelling.
Extra-embryonic ectodermal cells around the ectoplacental cavity showed light
labelling for AFP (Fig. 5). Mesodermal tissue of the embryonic region, which
forms a continuous layer between the endoderm and ectoderm, showed no AFP
activity. However, the single layer of mesoderm lining the entire exocoelom was
very heavily labelled. This mesoderm is thought to contribute to the amnion
when associated with embryonic ectoderm, to the yolk sac in association with
extra-embryonic endoderm and to the chorion with extra-embryonic ectoderm.
In 9th day embryos the allantois is present as a bud of mesoderm cells extending
into the exocoelom (Fig. 5). Only the cells on the outer surface of this structure
were labelled for AFP.
Late 9th day embryos
Embryos had grown very much in size by this stage. The span of AFP positive
cells in the visceral endoderm had increased, but the degree of labelling within
these cells appeared to be reduced in the five embryos studied. Very few visceral
endoderm cells forming the early yolk sac were AFP positive, whereas the
mesoderm layer was heavily labelled. Thickenings were now observed in this
mesoderm, but as in the allantois, only the outer cells adjacent to the exocoelom
were positively labelled (Fig. 6).
The amnion and embryonic ectodermal tissue showed the highest degree of
labelling. Ectodermal cells at the very base of the embryonic region showed very
strong intracellular labelling, compared to the still peripheral staining pattern
FIGURES 6-8
Fig. 6. Section through a mid 9th day embryo, where the chorion has moved closer to
the roof of the ectoplacental cavity. Localization of AFP in tissues is identical to an
early 9th day embryo (Fig. 5). Thickenings in the mesoderm of the presumptive yolk
sac are labelled only in the outer layer of cells (arrows), as in the allantois.
Fig. 7. Sections of a late 9th day embryo, where the ectoplacental cavity has almost
disappeared. (A) Extra-embryonic ectoderm cells no longer contain AFP (cf. Fig. 6),
although mesoderm cells lining the exocoelomic cavity are highly labelled. (B)
Embryonic cells at the base of the embryo show very strong AFP activity (arrow).
Visceral endoderm cells of the presumptive yolk sac are lightly labelled for AFP.
Fig. 8. Section through visceral and parietal yolk sacs of a 10th day embryo. AFP is
found in the endodermal, mesodermal and haematopoietic components of the
visceral yolk sac. Parietal endoderm is not labelled. A, amniotic cavity; EC, exocoelomic cavity; EP, ectoplacental cavity; am, amnion; eh, chorion; All, allantois;
Ex ect, extra-embryonic ectoderm; mes, mesoderm; PE, parietal endoderm; E eet,
embryonic ectoderm; V end, visceral endoderm; BI, blood islands.
AFP in mouse embryos
299
300
M. D Z I A D E K AND E. ADAMSON
in the rest of the ectoderm (Figs. 6, 7B). The chorion now lies adjacent to cells
at the roof of the ectoplacental cavity, and fusion takes place a short time later.
In early 9th day embryos extra-embryonic ectoderm cells lining the ectoplacental
cavity still contained AFP (Fig. 6), whereas at a later stage they were no longer
labelled (Fig. 7A).
10th day embryos
Organogenesis of the embryo has accelerated by the 10th day of gestation,
when a number of somites have formed, the neural tube has closed, and the
heart is pulsating regularly. AFP was distributed in most tissues, although the
ectodermal surface of the embryo was the most heavily labelled. The amnion
surrounding the embryo was also quite heavily labelled. The visceral yolk sac
is now an extensive structure, encasing the embryo and amnion. The outer layer
of columnar, highly vacuolated endoderm cells was labelled throughout, but not
as heavily as the underlying mesoderm layer (Fig. 8). Haematopoietic cells in
the blood islands forming in the yolk sac were also AFP positive.
14th day embryos
The visceral yolk sac of 14th day embryos contained AFP in only the endoderm
layer (Fig. 9), and all endoderm cells were labelled. The amnion was also highly
labelled. Foetal liver showed AFP activity within small groups of cells scattered
throughout this organ (Fig. 10).
Other tissues of the embryo also contained AFP. Kidney and gonadal tissue
showed AFP labelling in the mesenchymal cells, but not in the tubular epithelium.
Similarly, gut tissue was positive in the outer mesodermal/mesenchymal layer,
but the inner epithelial lining was negative. Brain tissue was lightly labelled in
most regions, although negative areas were observed in symmetrical position in
both halves of the brain. Heart tissue showed endogenous peroxidase activity
in blood cells within the ventricles, and high levels of AFP in muscle cells lining
these cavities. Lung tissue was generally negative, but had a positive outer layer
of AFP.
(b) AFP synthesis by isolated tissues
AFP is a protein which is secreted into the embryonic fluid after its production,
and therefore becomes available for uptake by cells of the embryo which do not
synthesize this product. The immunoperoxidase technique cannot distinguish
FIGURES 9 AND
10
Fig. 9. Sections through the visceral yolk sac of a 14th day embryo. (A) The highly
vesiculated columnar endoderm cells contain high levels of AFP while the mesoderm
component is now unlabelled (cf. Fig. 8). (B) Control section treated with absorbed
antiserum is not labelled, end, endoderm; mes, mesoderm.
Fig. 10. Sections through the liver of a 14th day embryo. (A) Hepatocytes containing
AFP are scattered throughout the liver. (B) Control section is completely negative.
AFP in mouse embryos
301
9A
EMB 43
302
M. D Z I A D E K AND E. ADAMSON
Fig. 11. Immunoperoxidase reaction for AFP on sections through isolated parietal
yolk sac(PYS) tissues. (A) PYS from 8th day embryo, without culture. (B) 8th day P YS
after 48 h in culture. (C) PYS endoderm layers from 10th day embryo, without culture.
(D) 10th day PYS endoderm after 48 h in culture. Parietal endoderm cells do not contain AFP, either before or after culture. Blood cells which infiltrate the PYS show
endogenous peroxidase activity (A). Extensive basement membrane was produced
during culture (arrows).
between cellular localization of AFP due to synthesis or due to adsorption from
an external source. To determine which cell types of the embryo synthesize AFP,
isolated tissues were cultured in AFP-free medium. Cellular localization of AFP
after culture must then be due to synthesis by that tissue.
(i) AFP in isolated endodermal tissues after culture
Parietal endoderm was manually dissected from ten 7th day embryos, twelve
8th day embryos and eight 10th day embryos, and was either washed and fixed
immediately, or cultured in suspension for 48 h. The 7th and 8th day parietal
endoderm samples were contaminated with trophoblast and adhering blood
cells, but these blood cells decreased substantially in number during culture.
Parietal endoderm showed no AFP activity at any of the three developmental
stages, whether immediately after dissection or after 48 h in culture (Fig. 11).
Any blood cells present were stained because of endogenous peroxidase activity
303
AFP in mouse embryos
Extra-embryonic
region
7th-day
egg cylinder
Embryonic
region
Enzyme
separation
\
Endoderm
Ectoderm •
.48 h in vitro
100 urn
D
Fig. 12. Experimental procedure for the separation and culture of visceral endoderm
and ectoderm tissue layers in 7th day egg cylinders. Immunoperoxidase reaction
for AFP showed activity in only endoderm tissue (A, C). Control sections (B, D)
treated with absorbed antiserum were unlabelled.
304
M. D Z I A D E K A N D E. A D A M S O N
Extra-embryonic
region
Ectoderm + mesoderm
Endoderm
I
48 h in vitro
B
D
^v
.«•**
S. ^
, r-
•-
Î**/ -'
,4^- •
fy 100/zm
Fig. 13. Experimental procedure for the separation and culture of visceral endoderm
and ectoderm plus mesoderm components of 8th day egg cylinders. Immunoperoxidase reaction for AFP showed activity in only endoderm tissue (A, C). Control
sections (B, D) treated with absorbed antiserum were unlabelled.
AFP in mouse embryos
305
Extra-embryonic
region
Enzyme separation
Extra-embryonic
ectoderm
Mesoderm
Embryonic ectoderm
30 h in vitro
B
**u'
AFP added
to culture medium
I
I
Fig. 14. Experimental procedure for the isolation and culture of extraembryonic
ectoderm, mesoderm and embryonic ectoderm from 8th day embryos. Immunoperoxidase reaction for AFP showed no activity in these tissues (A, B, C) until AFP
was added to the culture medium (D, E, F).
306
M. D Z I A D E K AND E. ADAMSON
in these cells. Parietal endoderm formed extensive basement membrane under
in vitro conditions (Fig. 11).
Visceral endoderm from ten 7th day and twelve 8th day egg cylinders was
separated enzymically from the underlying tissue. The separated components
were cultured in suspension for 48 h (Figs. 12,13). Isolated endoderm from both
7th day and 8th day embryos formed vesicles very rapidly, which were strongly
positive for AFP after 48 h culture in all cases (Figs. 12, 13). Isolated ectoderm
from 7th day embryos remained as a solid structure during culture in which mesodermal elements could not be identified. No cells in these endoderm-free
aggregates contained AFP (Fig. 12). Isolated ectoderm plus mesoderm tissues
from 8th day embryos were also entirely unlabelled. for AFP in ten out of twelve
cases (Fig. 13). In the other two cases, cells containing AFP were present in
a single-layered group at one pole of the aggregate. These had a columnar,
vesiculated appearance characteristic of visceral endoderm cells, and unlike
the normal morphology of mesoderm or ectoderm cells. It was concluded that
enzymic separation of the endoderm layer from these two egg cylinders was
incomplete.
The presence of large quantities of AFP in visceral endoderm cells after
culture indicated that these cells are able to continue AFP synthesis under the
in vitro conditions used in these experiments. The culture conditions therefore
did not prevent AFP synthesis, nor did they initiate synthesis in cells which
were not labelled in vivo, since parietal endoderm cells did not contain AFP
after culture. In vitro synthesis of AFP therefore corresponded to that in vivo
for at least these two endoderm derivatives, and the same test was applied to
other embryonic tissues which labelled for AFP in the whole embryo.
FIGURE 15
Immunoprecipitation of newly synthesized AFP. Tissues were incubated in medium
containing 50 /*Ci/ml-[3H]lysine for 24 h before the secreted proteins were immunoprecipitated with anti-mouse AFP antiserum (see the Materials and Methods section).
The washed immunoprecipitates were then analysed using SDS-containing Polyacrylamide gels (7-5 %, Laemmli, 1970).
(A) 1 mm slices from a stained column gel (sketched at the top) were counted for
tritium-labelled proteins. The tissue samples were 10th day visceral yolk sac x — x ;
10th day decidua O—O, part of the resulting immunoprecipitates were used in (B)
and (C). (B) Photograph of a stained slab gel and its fluorograph (C) to show that
washed immunoprecipitates still contain traces of contaminating proteins but that
the predominating radioactive protein is AFP. The stained bands were identified
by marker proteins run on adjacent slots. The positions of AFP and heavy chain
of IgG are shown in (A) and (B). On this fluorograph, AFP appears as a doublet
in track 1, but this may be an artefact since it was seen only in some gels. In contrast, radioactive AFP analysed by slicing gels was always a single peak. Track 1,
9th day whole embryos; track 2,9th day decidual tissue; track 3, 10th day decidual
tissue; track 4, 10th day visceral yolk sac.
AFP in mouse embryos
307
AFP
Heavy
chain of
IgG
Jri
gsa a
J
200
100
\*Q4-l»-o+.o+*ytM-o~,
o-*^la*a*o^u*ateA>lÉ^o<igha
4
6
Distance migrated (cm)
B
•
- * •
*% X
AFP—>-
Heavy chain —>of IgG
„A>;
1
2
1
2
3
+
308
M. D Z I A D E K AND E. ADAMSON
(ii) AFP in isolated non-endodermal tissues after culture
Extra-embryonic ectoderm, embryonic ectoderm and mesoderm tissues were
isolated from 20 8th day embryos in the following way (Fig. 14). Two transverse
cuts were made through the egg cylinders to give three portions; a complete
extra-embryonic fragment; a complete embryonic fragment; and a mid-region
composed of a mixture of tissues, which was discarded. The embryonic and
extra-embryonic tissues were incubated in a pancreatin/trypsin solution for
1 h, after which the ectoderm and mesoderm layers were easily isolated. Tissues
were washed in medium containing added serum for 2 h to remove all enzyme
activity and then transferred to fresh medium for a further 30 h culture period.
Extra-embryonic ectoderm tissue remained as a solid aggregate in culture,
whereas embryonic ectoderm and mesoderm adhered readily to the culture dish
and grew out as a monolayer. No ectodermal or mesodermal tissues contained
AFP after the culture period (Fig. 14), and therefore do not synthesize AFP
when isolated and grown in vitro. To determine whether these tissues can adsorb
AFP, isolated fragments were cultured for 24 h in standard medium, which was
then replaced with medium containing approximately 6-7 mg/ml of AFP.
Tissues were incubated in this medium for a further 8 h before fixation. All three
tissue types showed AFP activity in their cells after such treatment (Fig. 14).
(c) AFP synthesis, measured by immunoprecipitation
The synthetic products of whole early embryos, and isolated tissues from later
stages were analysed to determine whether the presence of AFP detected within
these embryos and tissues by immunoperoxidase was due to actual synthesis by
that embryo or tissue.
(i) 7th to 10th day embryos
Embryos from each day of development were dissected from the uterus and
incubated in [3H]lysine as described in 'Materials and Methods'. Newly
synthesized AFP which was secreted into the culture medium was determined
by the amount of radioactivity in protein of appropriate mobility, after electrophoresis of immunoprecipitates.
The earliest stage at which AFP could be detected by this method was the
8th day whole embryo, at a low level. By the 9th day of gestation, synthesis was
readily detectable at 708 cpm/^g DNA/24 h. Radioactive AFP in immunoprecipitates was measured as counts per minute under the peak (Fig. 15 A).
Fig. 15 B shows that AFP was the only radioactive protein present in the stained
immunoprecipitate (see Figure legend). This demonstrates the specificity of the
antiserum and the purity of the resulting antibody-antigen insoluble complex
after a rigorous washing procedure. Recoveries were assumed to be similar in
all incubates and averaged 80-100 %.
Embryos on the 10th day of development were dissected into parietal yolk
AFP in mouse embryos
309
sac, visceral yolk sac, amnion, decidual tissue and embryo proper. The visceral
yolk sac was the only portion found to synthesize and secrete AFP, at 2318
cpm//4g DNA/24 h.
(ii) Synthesis by tissues of later embryos
On the 12th day of gestation, the visceral yolk sac was larger than on the
10th day, and could be separated enzymically into the component mesoderm and
endoderm layers. These tissue layers were incubated separately in [3H]lysine.
No newly synthesized AFP could be detected in the culture medium from the
mesoderm incubate, while the endoderm layer had secreted 1861 cpm AFP//6g
DNA/24 h. The apparent reduction in the amount of AFP synthesis detected
between 10th day and 12th day yolk sacs is likely to be a result of exposure of
the latter to proteolytic enzymes during tissue separation.
The liver and yolk sac from later stage embryos up to the 18th day of gestation
were found to be secreting newly synthesized AFP at high rates. No quantitative
analysis was done.
DISCUSSION
Initiation of AFP synthesis in the mouse embryo
The first cells of the mouse embryo in which AFP can be detected by the
immunoperoxidase technique are visceral embryonic endoderm cells of the
7th day egg cylinder. Our observations confirm the suggestion that this might be
the case (Engelhardt, Poltoranina & Yazova, 1973). It follows that AFP is likely to
be synthesized by the first cells in which it is found, since no other parts of the
embryo or uterine decidua contain AFP for secretion at this stage. Adsorption
from the maternal serum is also unlikely, since extremely low levels of AFP are
normally present in adult sera, and furthermore, cellular localization of AFP
also appears under in vitro conditions in the absence of maternal serum, in the
endoderm layer formed on the surface of inner cell masses isolated from blastocysts (M. Dziadek, unpublished observations). No actual synthesis of AFP by
7th day embryos could be detected by the immunoprecipitation technique, but
8th day embryos secreted detectable amounts of newly synthesized AFP. Since
cellular localization of AFP is confined to the visceral embryonic endoderm of
8th day embryos, it can be safely concluded that these cells initiate AFP synthesis
on the 7th day of gestation, but secretion (if any at this early stage) is at too low
a level to be detected by the immunoprecipitation technique.
AFP synthesis is not initiated synchronously in all cells of the visceral
embryonic endoderm of the 7th day egg cylinder, nor it is initiated in one small
patch. All visceral embryonic endoderm cells contain AFP by the 8th day of
gestation, which may be due to successive initiation in all cells of this region, or
proliferation and movement of the initially small clones of AFP positive cells,
However, it is not clear whether all AFP positive visceral embryonic endoderm
310
M. D Z I A D E K AND E. ADAMSON
cells synthesize this protein, or whether some accumulate the synthetic products
of their neighbours.
AFP positive cells do not appear in the visceral extra-embryonic endoderm
of 8th or 9th day egg cylinders. It seems that synthetic activity in the visceral
endoderm of 8th day egg cylinders is related in some way to the type of underlying tissue, since visceral endoderm cells positioned over the embryonic
ectoderm of the egg cylinder synthesize AFP, whereas cells positioned over the
extra-embryonic ectoderm do not. Tissue relationships are altered as a result of
morphogenetic movements after mesoderm formation and migration, and in
the 9th day embryo such a precise correlation of AFP synthesis with a particular
type of underlying tissue is not evident. Studies concerning the tissue interactions involved in the initiation and maintenance of AFP synthesis through
subsequent development will be presented in a separate report.
Not all visceral embryonic endoderm cells maintain a high level of AFP
synthesis after the 9th day of development. AFP localization becomes limited
to cells at the midgirth of the embryo, whereas cells around the lower embryonic
region no longer label strongly. The number of visceral endodermal cells
containing AFP increases quite markedly from the late 9th day to the 10th day
of gestation. In the late 9th day embryo only a relatively small proportion of the
visceral endoderm cells of the presumptive yolk sac region contain AFP,
compared to the entire visceral endoderm of the 10th day yolk sac. Until the
factors controlling AFP synthesis in visceral endoderm cells are understood it
is not possible to use AFP synthesis as a marker to follow the cell lineage of
visceral yolk sac formation. For instance, it is not known whether the AFPpositive visceral embryonic endoderm cells in the 8th and 9th day embryos form
a continuous lineage and contribute to the entire yolk sac endoderm, or whether
AFP negative visceral extra-embryonic endoderm cells are subsequently induced
to synthesize AFP and also contribute to the 10th day visceral yolk sac.
Adsorption v. synthesis by non-endoderm tissues
By the stage when three cavities have formed in the egg-cylinder AFP is no
longer confined to visceral embryonic endoderm cells but is also present in
extra-embryonic ectoderm, mesoderm and embryonic ectoderm cells which line
the ectoplacental, exocoelomic and amniotic cavities respectively. When these
tissues were cultured in isolation no AFP synthesis could be detected. Isolated
tissues in cell culture might fail to produce AFP for at least two reasons. Firstly,
because they do not synthesize AFP in the intact embryo, and secondly because
normal cellular relationships have been upset after tissue dissociation. We
think that the first explanation is correct, because tissue isolation does not
inhibit AFP synthesis in other tissues. Furthermore, it appears that AFP
positive cells in the ectoderm and mesoderm tissues of the intact embryo always
have a peripheral position, with part of their surface exposed to an embryonic
cavity. Peripheral labelling suggests AFP adsorption, and we have shown that
AFP in mouse embryos
311
these tissues can adsorb AFP in cell culture, and that newly synthesized AFP
is secreted by 8th and 9th day embryos, and hence AFP is available for uptake
by cells at these early stages. The mechanism by which AFP is transported to the
cavities of embryo after secretion by the visceral embryonic endoderm has yet
to be determined.
Adsorption of AFP from foetal serum probably also accounts for the presence
of this protein in the wide variety of tissues of 14th day embryos other than the
yolk sac and liver. Adsorption of AFP into particular cell layers of tissues, and
the changing capacity for adsorption, as seen in the mesoderm of the visceral
yolk sac between the 10th and 14th days of gestation, and in the extra-embryonic
ectoderm prior to fusion of the chorion with the ectoplacental cone, reflects
differences in the distribution of AFP, or some specificity in the process of
adsorption. This raises the question of what role AFP plays during embryogenesis. The exact function of AFP is still unclear, although it has been shown to
have some role in the suppression of immunological responses, (Ogra, Murgita
& Tcmasi, 1974; Murgita & Tomasi, 1975), and has high affinity binding
properties for oestrogen in the rat and mouse (Uriel, De Nechaud & Dupiers,
1972; Uriel, Bouillon & Dupiers, 1975; Uriel, Bouillon, Aussei & Dupiers,
1976; Nunez, Engelmann, Benassay & Jayle, 1971; Aussei, Uriel & MercierBodard, 1973). AFP may play a role in the sexual differentiation of the
rat brain by functioning as an oestradio-binding protein (Attardi & Ruoslahti,
1976). AFP may act as a more general carrier protein, and it has been proposed that it could act as a foetal albumin, due to similarities in the physicochemical and binding properties of these two proteins (Belanger et al. 1975).
Which of these functions are relevant to the presence of AFP in early mammalian development has yet to be determined.
AFP as a visceral endoderm marker
The present study establishes that AFP synthesis in the early postimplantation
mouse embryo is confined initially to visceral embryonic endoderm cells, and
later to the entire visceral endoderm layer of the yolk sac. Further experimentation (M. Dziadek, in preparation) has shown that visceral extra-embryonic
endoderm cells also have the capacity for AFP synthesis. On this evidence it
can be concluded that AFP synthesis is a specific property of visceral endoderm
cells in the early embryo (prior to formation of the foetal liver). Parietal endoderm
is unable to synthesize AFP, shown by both the immunoperoxidase and the
radioimmunoprecipitation techniques, but has been shown to synthesize
plasminogen activator, which is not a product of the visceral endoderm
(Strickland, Reich & Sherman, 1976). It is now possible to use AFP and
plasminogen activator as biochemical markers for the differentiation of primitive
endoderm into the visceral and parietal elements, and a closer analysis of the
mechanisms involved in the differentiation of these endodermal derivatives is
in progress. These markers should also prove useful for the identification of cell
types in differentiating teratocarcinoma cell lines.
312
M. D Z I A D E K A N D E. A D A M S O N
The authors gratefully acknowledge Dr C. F. Graham for valuable discussion, Drs R. L.
Gardner, J. W. McAvoy, V. E. Papioannou and P. Thorogood for critical reading of the
manuscript, and J. Haywood for photography. M. Dziadek is supported by a Flinders
University of South Australia Overseas Scholarship, and E. Adamson by the Medical
Research Council.
REFERENCES
ABELEV, G. I. (1971). Alphafetoprotein in ontogenesis and its association with malignant
tumors. Adv. Cancer Res. 14, 295-358.
ADAMSON, E. D. (1977). Acetylcholinesterase in mouse brain, erythrocytes and muscle.
/ . Neurochem. 28, 605-615.
ATTARDI, B. & RUOSLAHTI, E. (1976). Foetoneonatal oestrodiol-binding protein in mouse
brain cytosol is a-foetoprotein. Nature, Lond. 263, 685-687.
AUSSEL, C , URIEL, J. & MERCIER-BODARD, C. (1973). Rat alpha-fetoprotein: isolation,
characterization and estrogen-binding properties. Biochimie 55, 1431-1437.
BELANGER, L., HAMEL, D., LACHANCE, L., DUFOUR, D., TREMBLAY, M. & GAGNON, P. M.
(1975). Hormonal regulation of a-foetoprotein. Nature, Lond. 256, 657-659.
BONNER, W. M. & LASKEY, R. A. (1974). A film detection method for tritium-labelled proteins
and nucleic acids in Polyacrylamide gels. Eur. J. Biochim. 46, 83-88.
DAVIS, B. J. (1974). Disc electrophoresis. II. Methods and applications to human serum
proteins. Ann. NY. Acad. Sei. 121, 404-427.
DULBECCO, R. & VOGT, M. (1954). Plaque formation and isolation of pure lines with poliomyelitis viruses. / . exp. Med. 99, 167-182.
ENDERS, A. C. (1971). The fine structure of the blastocyst. In Biology of the Laboratory
Mouse, 2nd ed. (ed. E. L. Green), pp. 205-245. New York: McGraw-Hill.
ENGELHARDT, N . V., GOUSSEV, A. I., SHIPOVA, L. J. & ABELEV, G. I. (1971). Immuno-
fluorescent study of a-fetoprotein in liver and liver tumors. I. Technique of a-fetoprotein
localization in tissue sections. Int. J. Cancer 7, 198-206.
ENGELHARDT, N . V., POLTORANINA, V. S. & YAZOVA, A. K. (1973). Localization of alpha-
fetoprotein in transplantable murine teratocarcinomas. Int. J. Cancer 11, 448-459.
GARDNER, R. L. & PAPAIOANNOU, V. E. (1975). Differentiation in the trophectoderm and
inner cell mass. In The Early Development of Mammals (2nd Symposium of the British
Society for Developmental Biology) (ed. M. Balls & A. E. Wild), pp. 107-132. Cambridge
University Press.
GRABAR, P. & WILLTAMS, C. A. (1953). Méthode permettant l'étude conjugée des propriétés
electrophorétiques et immunochimique d'un mélange de protéines. Application au sérum
sanguin. Biochim. biophys. Acta 10, 193-194.
GRAHAM, C. F. (1971). The design of the mouse blastocyst. In Control Mechanisms of Growth
and Differentiation (Symp. Soc. Exp. Biol. XXV), pp. 371-378. Cambridge University
Press.
HUDSON, L. & HAY, F. C. (1976). Practical Immunology, pp. 101-102, Oxford: Blackwell
Scientific Publications.
LAEMMLr, U. K. (1970). Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature, Lond. 227, 680-685.
LEVAK-SVAJGER, B., SVAJGER, A. & SKREB, N. (1969). Separation of germ layers in presomite
rat embryos. Experientia 25, 1311-1312.
MORRIS, G. E. & COLE, R. J. (1972). Cell fusion and differentiation in cultured chick muscle
cells. Expl Cell Res. 75, 191-199.
MURGITA, R. A. & TOMASI, T. B. (Jr). (1975). Suppression of the immune response by
a-fetoprotein. I. The effect of mouse a-fetoprotein on the primary and secondary antibody
response. J. exp. Med. 141, 269-286, 440-452.
NADIJCKA, M. & HILLMAN, N . (1974). Ultrastructural studies of the mouse blastocyst
substages. / . Embryol. exp. Morph. 32, 675-695.
NUNEZ, E., ENGELMANN, F., BENASSAY, G. & JAYLE, M. F. (1971). Identification et puri-
fication préliminaire de la foetoprotéine liant les oestrogènes dans la sérum de rats
nouveau-nés. C. r. hebd. Séanc. Acad. Sei., Paris 273, 831-834.
AFP in mouse embryos
313
OGRA, S. S., MURGITA, R. & TOMASI, T. B. (1974). Immunosuppressive activity of mouse
amniotic fluid. Immun. Commun. 3, 497-508.
OucHTERLONY, O. (1958). Diffusion in gel methods for immunological analysis. Prog. Allergy
5, 1-78.
SAINTE-MARIE, G. (1962). A paraffin embedding technique for studies employing immunofluorescence. / . Histochem. Cytochem. 10, 250-256.
SNELL, G. D. & STEVENS, L. C. (1966). Early embryology. In Biology of the Laboratory
Mouse, 2nd ed. (ed. E. L. Green), pp. 205-245. New York: McGraw-Hill.
STANNERS, C. P., ELICEIRI, G. L. & GREEN, H. (1971). Two types of ribosome in mouse-
hamster hybrid cells. Nature, Lond. 230. 52-54.
STRICKLAND, S., REICH, E. & SHERMAN, M. I. (1976). Plasminogen activator in early embryogenesis. Cell 9, 213-240.
TARKOWSKI, A. K. & WROBLEWSKA, J. (1967). Development of blastomeres of mouse eggs
isolated at the four and eight-cell stage. J. Embryol. exp. Morph. 18, 155-180.
URIEL, J., BOUILLON, D. & DUPIERS, M. (1975). Affinity chromatography of human, rat and
mouse a-fetoprotein on estradiol-sepharose adsorbents. FEBS Letters, Amsterdam 53,
305-308.
URIEL, J., BOUILLON, D., AUSSEL, C. & DUPIERS, M. (1976). Alpha-fetoprotein: the major
high affinity estrogen binder in rat uterine cytosols. Proc. natn Acad. Sei., U.S.A. 73,
1452-1456.
URIEL, J., D E NECHAUD, B. & DUPIERS, M. (1972). Estrogen-binding properties of rat, mouse
and man fetospecific serum proteins. Demonstration by immunoautoradiographic methods.
Biochem. biophys. Res. Commun. 46, 1175-1180.
VAN BLERKOM, J., BARTON, S. C. & JOHNSON, M. H. (1976). Molecular differentiation in the
preimplantation mouse embryo. Nature, Lond. 259, 319-321.
WILSON, J. R. & ZIMMERMAN, E. F. (1976). Yolk sac: site of developmental microheterogeneity of mouse a-fetoprotein. Devi Biol. 54, 187-200.
(Received 1 August 1977, revised 6 September 1977)