Download The origin of early primitive streak - Development

87
Development 127, 87-96 (2000)
Printed in Great Britain © The Company of Biologists Limited 2000
DEV3080
Formation of the avian primitive streak from spatially restricted blastoderm:
evidence for polarized cell division in the elongating streak
Yan Wei and Takashi Mikawa*
Department of Cell Biology, Cornell University Medical College, 1300 York Avenue, New York, NY 10021, USA
*Author for correspondence (e-mail: [email protected])
Accepted 13 October; published on WWW 8 December 1999
SUMMARY
Gastrulation in the amniote begins with the formation of a
primitive streak through which precursors of definitive
mesoderm and endoderm ingress and migrate to their
embryonic destinations. This organizing center for amniote
gastrulation is induced by signal(s) from the posterior
margin of the blastodisc. The mode of action of these
inductive signal(s) remains unresolved, since various
origins and developmental pathways of the primitive streak
have been proposed. In the present study, the fate of
chicken blastodermal cells was traced for the first time in
ovo from prestreak stages XI-XII through HH stage 3,
when the primitive streak is initially established and prior
to the migration of mesoderm. Using replication-defective
retrovirus-mediated gene transfer and vital dye labeling,
precursor cells of the stage 3 primitive streak were mapped
predominantly to a specific region where the embryonic
midline crosses the posterior margin of the epiblast. No
significant contribution to the early primitive streak was
seen from the anterolateral epiblast. Instead, the precursor
cells generated daughter cells that underwent a polarized
cell division oriented perpendicular to the anteroposterior
embryonic axis. The resulting daughter cell population was
arranged in a longitudinal array extending the complete
length of the primitive streak. Furthermore, expression of
cVg1, a posterior margin-derived signal, at the anterior
marginal zone induced adjacent epiblast cells, but not those
lateral to or distant from the signal, to form an ectopic
primitive streak. The cVg1-induced epiblast cells also
exhibited polarized cell divisions during ectopic primitive
streak formation. These results suggest that blastoderm
cells located immediately anterior to the posterior marginal
zone, which secretes an inductive signal, undergo spatially
directed cytokineses during early primitive streak
formation.
INTRODUCTION
the epiblast by stage X (Eyal-Giladi and Kochav, 1976; Kochav
et al., 1980). The hypoblast underlying the epiblast begins to
grow anteriorly from the posterior margin of the area pellucida
at stage XI and covers the entire ventral surface of the area
pellucida by stage XIII. Primitive streak formation is first
detected as a cluster of cells in the posterior margin of the area
pellucida along the embryonic midline at HH stage 2
(Hamburger and Hamilton, 1951) and is completed by HH
stage 3. Mesodermal cell migration begins by HH stage 4 as
Hensen’s node and groove are established.
Recent studies have identified several transcriptional
factors and growth factor signals involved in primitive streak
formation and/or gastrulation (Stern et al., 1990; Knezevic et
al., 1995, 1997; Boncinelli and Mallamaci, 1995; Wilson et
al., 1995; Winnier et al., 1995; Seleiro et al., 1996). It has
also been shown that epiblast cells of the area pellucida are
competent to respond to inductive signaling and can be
induced to form an ectopic primitive streak (Khaner and EyalGiladi, 1986; Stern, 1990; Mitrani and Shimoni, 1990;
Mitrani et al., 1990; Ziv et al., 1992; Callebaut and Van
Nueten, 1994; Bachvarova et al., 1998). The role(s) of these
Gastrulation is the morphogenetic event through which the
three definitive germ layers, ectoderm, mesoderm and
endoderm, are established. Coordinated behaviors of
embryonic cells during gastrulation have been carefully studied
in non-amniotes, particularly frog embryos (reviewed in Keller,
1986). In the frog, animal hemisphere blastomeres involute
through the dorsal blastopore lip, undergo convergentextension and move as a sheet on the inner surface of the
blastocoel toward the animal pole of the embryo. Embryos of
higher amniotes, such as birds and mammals, do not utilize
blastopore-mediated involution as the mechanism of
gastrulation. Instead, they initially develop a primitive streak
consisting of bottle-shaped mesenchymal cells along the
posterior half of the midline of the epithelioid epiblast. Once
the primitive streak is established, mesenchymal cells begin an
anterolateral migration to differentiate into embryonic
mesoderm and endoderm.
In the chicken embryo, blastoderm cells of the area pellucida
first differentiate into a single layer of epithelioid cells called
Key words: Primitive streak, Gastrulation, Fate map, Retrovirus, DiI,
Chick, Epiblast, Cytokinesis
88
Y. Wei and T. Mikawa
factors in inducing the primitive streak is presently unclear,
partly due to uncertainty in our understanding of the origin
and development of the authentic primitive streak. The
pattern of primitive streak development might be critical for
considering the mechanisms of inductive signaling. At least
three distinct models have been proposed to describe
primitive streak formation, based on migration patterns of
epiblast cells in cultured blastodiscs (Fig. 1). The first model
hypothesizes that epiblast cells migrate from lateral regions
towards the midline when they ingress to form the primitive
streak (reviewed in Balinsky, 1975). Another model presumes
that HNK-1-positive cells, sparsely distributed within the
epiblast, individually ingress into the blastocoel, migrate to
the posterior midline and form the initial primitive streak
(reviewed in Stern, 1992). The third model predicts that
posterolateral epiblast cells sequentially migrate to the
midline and then move anteriorly along the midline (EyalGiladi et al., 1992).
In the present study, an in ovo protocol was established for
tagging chicken blastoderm cells at prestreak stages (XI-XII)
in order to trace primitive streak formation. Data are presented
that show that the majority of precursor cells giving rise to the
HH stage 3 primitive streak are restricted to a midline area
within the posterior margin of the blastodisc. During early
primitive streak formation, the precursor cells undergo
polarized cytokinesis oriented along the midline. Their
daughter cells are then longitudinally arrayed along the
primitive streak at HH stage 3.
MATERIALS AND METHODS
Virus
A replication-defective variant of spleen necrosis retrovirus
(Dougherty and Temin, 1986) encoding bacterial β-galactosidase (βgal) was generated from a packaging cell line, D17.2G, as described
(Mikawa et al., 1992a). The virus particles were concentrated to titers
of ~107-108 virions/ml (Mima et al., 1995) just before injection.
Propagation of the β-gal viral vector in vitro and proof of helper
virus-free stocks have been presented elsewhere (Mikawa et al.,
1992b; Mikawa and Fischman, 1992).
In vivo labeling
Fertilized chicken eggs were obtained from an outbred flock (Truslow
Farms, MD). A small window of ~20 mm diameter was opened in the
shell at the lateral side of the unincubated egg and the underlying shell
membrane removed. Embryos were illuminated by cool light through
a blue filter and monitored through a color videocamera (Hitachi
VKC150) connected to a trinocular stereomicroscope. The monitored
image of embryos were examined to identify Koller’s sickle
characteristic of the posterior margin of stages XI-XIII prestreak
embryos (Eyal-Giladi and Kochav, 1976). Embryos lacking an
obvious Koller’s sickle were not used.
DiI (1,1′-dioctadecyl 3,3,3′,3′-tetramethyl indocarbocyanine
perchlorate: Molecular Probes, Inc.), at 0.25 % in dimethyl
formamide or concentrated viral suspensions containing 100 µg/ml
polybrene, were pressure injected in ovo (Mikawa and Gourdie,
1996) into one of five selected sites (Fig. 1) under the control of a
Picospritzer II (General Valve Co., NJ). To map these sites within
the blastodisc, the embryonic midline was first drawn on the monitor
as the line intersecting the midpoint of Koller’s sickle. A complete
circle was then superimposed from the arc. Site 1 was demarcated
as the point where the midline meets Koller’s sickle, site 2 was
designated as the region anterior to site 1 by 1/5 the diameter of the
circle, site 3 was defined as the spot immediately inside the circle
along a diagonal 135° from the apex of the AP axis, site 4 was the
midpoint between sites 2 and 3, and site 5 was defined as the region
posterior to the midpoint of Koller’s sickle. Other general
procedures for microsurgery, injection of virus and incubation of
infected embryos have been described previously (Mikawa et al.,
1992a,b; Itoh et al., 1996).
Histology
All embryos were fixed by injecting 2% paraformaldehyde in PBS
underneath the embryonic discs. The fixed discs were then removed
from the eggs and processed for morphological examination. Virally
tagged embryos were stained with X-gal for detection of β-gal in
whole mount as described (Mikawa et al., 1992a). DiI in some
labeled embryos was photoconverted, according to Ruiz i Altaba
et al. (1993). Embryos exhibiting β-gal-positive cells or
photoconverted DiI precipitates were embedded in Epon or paraffin,
serially sectioned at 4-7 µm thickness and examined by bright-field,
phase or Hoffman Modulation Contrast optics. Actin and
chromosomes of stages XI, XII, 2+ and 3 were stained with
fluorescein-conjugated phalloidin and DAPI (4′, 6-diamidino-2phnylindole, Molecular Probe), respectively.
BrdU labeling
S-phase nuclei during primitive streak formation were labeled with
a thymidine analogue, BrdU (5′-bromo-2′deoxyuridine, Amersham).
At various times of incubation, 500 µg of BrdU per egg were
introduced onto embryos through a small window in the egg
shell and shell membrane. Eggs were reincubated for further
development until stages 3− to 3. All embryos were then fixed
and removed from the eggs as above. After removal of
unincorporated BrdU by washing with PBS, embryos were
immunolabeled in whole mount with anti-BrdU antibodies
(Amersham) followed by detection with the ABC kit (Vector lab)
with a peroxidase-based substrate. BrdU-negative nuclei were
counterstained with DAPI (1 µg/ml).
Ectopic induction of primitive streak
Ectopic primitive streaks were induced by implanting COS cells
expressing a chick Vg1 (cVg1, gifts from Drs Claudio Stern and Jane
Dodd, Columbia College of Physicians and Surgeons, NY), according
to Shah et al. (1997). After 7-9 hours of incubation, the embryos were
fixed with 2% paraformaldehyde in PBS and those exhibiting an
ectopic primitive streak were further processed for DAPI staining or
in situ hybridization.
In situ hybridization
Fixed embryos were hybridized with 1 µg/ml of DIG-labelled
antisense or sense RNA probe for chicken Brachyury (provided by Dr
Raymond B. Runyan, University of Arizona), according to Henrique
et al. (1995) with the slight modification of substituting proteinase K
digestion with three 30 minute incubations with RIPA-buffer (50 mM
NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM
EDTA, 50 mM Tris-HCl, pH 8.0). Immunodetection of the RNA
probes and color development were carried out according to Henrique
et al. (1995).
Photography and digital imaging
All images of embryos and histological sections were photographed
using Kodak Ektachrome 400 film or captured directly by a digital
camera, DKC-5000 (SONY, Japan). Color slides were scanned into
Adobe Photoshop 5.0 (Adobe Systems, Inc., Mountain View, CA)
using a Nikon film scanner LS-1000 (Nikon, Japan). All images
were adjusted for brightness and contrast, and then cropped. To
normalize distribution of β-gal-positive cells, digital images from
six embryos were superimposed after each was adjusted to 16%
opacity.
The origin of early primitive streak
RESULTS
Tagging of blastodiscs in ovo
To identify the origin of the early primitive streak in ovo, five
sites of unincubated, prestreak-stage embryos were tagged and
the fate of labeled cells was examined at HH stage 3 prior to
anterolateral migration of mesoderm (Fig. 1B). These sites
were specifically selected to test the three models proposed for
the genesis of the primitive streak in the in vitro fate map
studies (Fig. 1A). The predicted pattern of labeled cell
distribution in each model is summarized in Fig. 1C.
In the previous fate map studies on cultured embryos,
Koller’s sickle representing the future posterior of the embryo
has been a reliable landmark for detecting the embryonic
anteroposterior (AP) axis, while the boundary between the area
pellucida and area opaca has been used to detect the embryonic
and extraembryonic boundaries. We first examined whether
these landmarks, which are clearly identifiable in vitro, were
detectable in ovo within the prestreak embryo. Fortunately, an
arc- or crescent-shaped cluster of cells characteristic of
Koller’s sickle was evident in ovo (Fig. 2A). To confirm that
the arc-shaped cluster of cells is Koller’s sickle demarcating
the posterior margin of epiblast, the midpoint of the arc was
labeled with DiI (Fig. 2A). Immediately after DiI labeling,
embryos were photographed (Fig. 2A), fixed and removed
from the egg (Fig. 2B). The DiI-labeled region was identified
as the site where the embryonic AP axis intersected the
midpoint of Koller’s sickle (Fig. 2B). Histological inspection
of these embryos further proved that DiI was localized in
Koller’s sickle, a cluster of cells protruding toward the
underlying hypoblast (Fig. 2C,D). Sagittal sections of such
prestreak embryos revealed that the anterior portion of epiblast
was composed of a simple cuboidal epithelium, while the more
posterior region was more stratified. This gradient in epithelial
aspect of the epiblast was used to verify the AP axis in the
subsequent studies. Koller’s sickle was also used as a landmark
for tagging individual sites of prestreak blastodiscs in ovo.
Migration pattern of labeled cells during primitive
streak formation
The fate of tagged cells was examined at HH stage 3 (Fig. 3).
In the first set of experiments, blastoderm cells were tagged
with DiI (Fig. 3A). DiI-positive (DiI+) cells from the posterior
end of the midline (site 1) were predominantly localized to the
primitive streak (Fig. 3A). The results indicated that DiIlabeled embryos developed normally and the migration pattern
of the embryonic cells could be examined in ovo. The results
also suggested that cells of the HH stage 3 primitive streak
predominantly arose from a small region within the posterior
margin of the embryonic midline. However, since the DiI
signal fades after cycles of several cell division, there might
have been a population of cells that lost DiI but had migrated
into the primitive streak. In addition, the DiI signal seen in HH
stage 3 embryos might have arisen from cells labeled at the
time of injection and from a second population that later passed
through the DiI-deposit, since in several instances DiI crystals
remained undissolved at the injection site.
To clarify those uncertainties of DiI-based cell tagging and
tracing, cells of prestreak blastodiscs were genetically tagged
with retrovirally encoded β-galactosidase (Fig. 3B). Marking
of site 1 (n=82) gave rise to a β-galactosidase-positive (β-gal+)
89
primitive streak in 75.1% of the tagged embryos (Fig. 3B). In
the remainder of the embryos, β-gal+ cells were localized to
areas adjacent to the primitive streak, either anteriorly (3.9%),
laterally (20.7%) or posteriorly (1.3%). Of the embryos
with primitive streak labeling, 35% exhibited β-gal+ cells
exclusively in the primitive streak (Fig. 3C), while 65%
contained β-gal+ cells both in the primitive streak and in
an area juxtaposed to the primitive streak. Histological
examination of these embryos (Fig. 3C) confirmed that tagging
of site 1 gave rise to the primitive streak which was richly
populated with β-gal+ cells. Tagging of other regions rarely
resulted in significant numbers of DiI+ or β-gal+ cells within
the primitive streak (Fig. 4A-H). Instead, cells from site 2
(n=8) were usually found in the area anterior to the primitive
streak (Fig. 4A,E) and those from site 3 (n=22) were mainly
distributed in the area lateral to the primitive streak (Fig. 4B,F).
Marking of cells at site 4 (n=24) and site 5 (n=8) gave rise to
β-gal+ cells in the area anterolateral to the primitive streak
(Fig. 4C,G) and in the extraembryonic region posterior to the
primitive streak (Fig. 4D,H), respectively.
To normalize for variation in cell migration and targeting
accuracy, the distribution patterns of tagged cells from 6
embryos were superimposed (Fig. 5). The data clearly
demonstrated that in ovo the majority of cells forming the HH
stage 3 primitive streak were derived from the posterior region
of the blastodisc, particularly from the midline portion (site 1).
Little contribution was evident from other areas surrounding
site 1. Neither migration of precursor cells from the lateral
epiblast and subsequent ingression along the midline, nor
ingression of precursor cells sparsely distributed in the epiblast
and followed by accumulation to the posterior region was
detected as the main route for generating the HH stage 3
primitive streak. These results suggest that the majority of
precursor cells giving rise to the HH stage 3 primitive streak
are already localized in the posterior-midline region of stages
XI-XII embryos.
Developmental process of HH stage 3 primitive streak
Although these results map the precursors of the HH stage 3
primitive streak predominantly to site 1 in the prestreak-stage
embryo, it still does not explain how cells restricted to such a
small area can generate the primitive streak, which expands to
~1 mm in length. Three potential hypotheses were tested (Fig.
6). The first, based on the model proposed by Eyal-Giladi et
al. (1992), assumes that cells localized to the midline region of
the precursor area first migrate cranially forming the most
anterior domain of the primitive streak, followed by laterally
localized precursors, which give rise to the more posterior
sections of the streak (Fig. 6A). The second hypothesis predicts
that daughter cells from different regions of the precursor area
randomly migrate along the midline and intermingled within
the resulting primitive streak (Fig. 6B). The third hypothesis
assumes that individual cells within the precursor area migrate
and/or proliferate unidirectionally toward the anterior tip of the
primitive streak (Fig. 6C).
To test these possibilities, a subset of cells within the
precursor region was labeled with β-gal and the distribution of
β-gal+ cells in the resulting primitive streak was examined at
HH stage 3. When a large number of precursor cells were
tagged, there was no detectable pattern of β-gal+ cell
distribution in the primitive streak (Fig. 7A,B), as seen in the
90
Y. Wei and T. Mikawa
consistent only with the third model (Fig. 6C) in which
individual precursor cells generate daughter cells arranged in
a longitudinal array along the embryonic AP axis.
Fig. 1. (A) Three distinct models proposed for genesis of the early
primitive streak. Blue lines indicate migration or movement of
primitive streak precursors. AP, anteroposterior axis. (B) Relative
positions of five sites tagged with DiI or β-gal virus at pre-streak
stages in vivo. (C) Predicted contribution of labeled cells (blue-filled
circles) to the primitive streak by marking sites 1-5.
above experiment (Fig. 3). However, when the retroviral titer
was reduced, a subdomain within the precursor area was
tagged, giving rise to β-gal+ cells, which were arranged in a
longitudinal strand that often extended the entire length of the
HH stage 3 primitive streak (Fig. 7C-F). Tagging of smaller
regions resulted in a narrower β-gal+ band. DiI-labeling gave
rise to similar results (not shown). Histological sections
through embryos bearing a single β-gal+ stripe revealed that
β-gal+ cells formed arrays in either the dorsal epithelioid or
middle mesenchymal layers (Fig. 7G,H). Importantly, when a
few cells within the precursor area were tagged (Fig. 8A-C),
β-gal+ daughter cells occupied only a portion of the entire
longitudinal array.
Such a unique arrangement of daughter cells was observed
only when parental cells were tagged within the precursor area,
which extended anteriorly along the embryonic midline to 200300 µm from the Koller’s sickle (Fig. 8C). These data are
Oriented cell division in the developing primitive
streak
This unique arrangement of daughter cells in a longitudinal
array could be generated by elongation of individual cells
and/or by oriented cell division along the midline. Cell shape
and the orientation of cell division were both examined in the
developing primitive streak and the non-gastrulating epiblast.
Cell shapes were outlined by staining the actin-containing
adherens belt with phalloidin, while the direction of cell
division was predicted by the orientation of DAPI-stained
metaphase chromosome plates. In the dorsal view of embryos
stained with phalloidin, no elongation of cell shape along the
AP axis was apparent in any area of the blastodisc. Instead,
primitive streak cells were smaller (8.14±0.33 µm, n=96) than
those in non-primitive streak epiblast layer (10.07±0.36 µm,
n=96) (Fig. 9A,B).
But an alignment of metaphase chromosome plates was
clearly evident in the developing primitive streak, especially
when compared to other regions of the DAPI-stained embryos
(Fig. 9C,D). The majority of primitive streak cells in
metaphase exhibited metaphase plates aligned perpendicular to
the AP axis (Fig. 9C). This oriented plane of the metaphase
chromosome plates in the streak was distinct from the random
orientation seen in non-primitive streak epiblast areas (Fig.
9D). Quantitative data (Fig. 9E) showed that in the primitive
streak, about 50% of metaphase cells set up a plane of cell
division perpendicularly to the AP axis, and an additional 30%
of cells exhibited a plane tilted by more than 45° to the AP
axis. Such polarized cell division was not evident once
primitive streak elongation completed at stage 3+-4. These
results suggest that, as the early primitive streak develops
anteriorly along the midline, primitive streak cells undergo the
oriented cytokinesis along the AP axis.
The increase in cell number along the AP axis and the
consequence that this would have on the linear dimension of
the early streak was established by counting the number of cells
and measuring their diameter (Table 1). Based on the results
of DiI and viral tagging, the anterior boundary of the precursor
area extended 200-300 µm from the Koller’s sickle, while the
Fig. 2. (A) Dorsal view of a
prestreak embryo immediately
after DiI labeling (arrow) in
ovo. Inset, higher
magnification of DiI-labeled
area. Note: DiI was introduced
to the midpoint of the white
arc (Koller’s sickle). (B) DiIlabeled embryos, fixed,
isolated, photoconverted and
viewed after removal of yolk.
Inset: higher magnification of
DiI-labeled (arrow) area.
(C) Higher magnification of
DiI-labeled area in D.
(D) Montage of sagittal
sections along the midline of
DiI-labeled (arrow) prestreak embryo. a-p, anteroposterior axis; d-v, dorsal-ventral axis; e, epiblast; h, hypoblast, kc, Koller’s sickle; y, yolk.
The origin of early primitive streak
Fig. 3. Distribution of DiI+ (A) and β-gal+ (B,C) cells in HH stage 3
embryos which were marked at site 1. (A,B) Whole-mount dorsal
view; (C) a transverse section. ps, primitive streak; e, epiblast; h,
hypoblast.
91
primitive streak length of 1.27 mm was obtained by direct
measurement of HH stage 3 embryos under a dissecting
microscope. These values were then divided by the average
diameter of precursor and primitive streak cells, 8.8 µm and
8.1 µm, respectively, giving rise to estimated cell numbers of
23-34 cells in the precursor area along the AP axis and 127
cells in the single column of the primitive streak. Thus, a 6- to
7-fold increase in cell numbers along the AP axis was
estimated during primitive streak extension between prestreakstage XII and HH stage 3 (Table 1). This estimate is consistent
with the data obtained independently by counting cells of
isolated, trypsinized tissue fragments, in which 2.3×103 cells
in the precursor area (n=4) and 1.4×104 cells in the early streak
(n=3) were scored on average. These results suggest that
precursor cells underwent three cell cycles on average during
this 12 hour interval. This is consistent with the mitotic
index (~2-4 hours, for cells of chicken gastrulae) which
was independently obtained in earlier studies based on
incorporation of a thymidine analogue (Sanders et al., 1993).
However, since only a small number of cells were in the
metaphase at the any given time point during primitive streak
formation (Fig. 9C,D), it was uncertain whether primitive cells
divided asynchronously and quickly completed M-phase or
only a subpopulation of primitive streak entered an active cell
cycle. These possibilities were tested by monitoring DNA
synthesis of cells in developing primitive streak (Fig. 10).
BrdU-positive nuclei were found
throughout the primitive streak
(Fig. 10A,B). Furthermore, the
majority of primitive streak cells
were labeled with BrdU within a
few hours of incubation (Fig. 10C),
indicating that the majority of cells
in the developing primitive streak
undergo an active, asynchronized
cell cycle.
Vg1 induce adjacent epiblast
cells to form an ectopic
primitive streak
These results show that epiblast
Fig. 4. Distribution of DiI+ (A-D) and
β-gal+ (E-H) cells in HH stage 3
embryos which were marked at site 2
(A,E), site 3 (B,F), site 4 (C,G) and site
5 (D,H).
Fig. 5. Probability in distribution of daughter cells derived from the tagged sites of prestreak embryos at HH stage 3. Digital images of six
embryos from each labeling group were superimposed. Each number corresponds to the tagged site indicated in the diagram of prestreak
embryos. Bar cord grades the presence of tagged cells in individual area; deeper blue represents the highest probability.
92
Y. Wei and T. Mikawa
Table 1. Primitive streak extension between stages XII and 3
Factor
1,270±84 µm (n=8)
200 µm⬉lpa⬉300 µm (n=18)
Cell diameter (Ø)
Primitive streak cells (Øps)
Precursor area cells (Øpa)
8.14±0.33 µm (n=96)
8.8±0.36 µm (n=92)
Estimated cell numbers along the AP axis (l/Ø)
Primitive streak (l/Ø)ps
Precursor area (l/Ø)pa
Fig. 6. Three distinct models proposed for the early primitive streak
formation. See text.
cells immediately anterior to the posterior marginal zone are
the main precursors of stage 3 primitive streak. Therefore, it is
likely that a posterior marginal zone-derived signal induces
primitive streak by recruiting adjacent epiblast cells, rather
than attracting either a lateral or anterior population to migrate
towards the midline. To test this possibility, COS cells
expressing cVg1 (cVg1/COS), one of the primitive-streakinducing factors, were placed at the marginal zone in the
anterior portion of prestreak-stage embryos. Within 12-14
Score
Length along the AP axis (l)
Primitive streak (lps)
Precursor area (lpa)
(l/Ø)ps=123.7±29.0
22.7⬉(l/Ø)pa⬉34.1
The maximal ps/pa ratio in
Length = lps/minlpa
Cell number = (l/Ø)ps/min (l/Ø)pa
6.35
5.45
Required cell cycle (log2ps/pa) in increase of
Length = log2lps/minlpa
Cell number = log2(l/Ø)ps/min (l/Ø)pa
2.67
2.45
(lps) of HH stage 3 embryos was directly measured under a dissect
microscope and presented as a mean value with 95% confidence interval,
while (lpa) was estimated based on the results of Fig. 8D. (Ø) was obtained by
analysing Phalloidin-stained embryos as in Fig. 9.
hours after implantation, a primitive-streak-like cluster of cells
formed at the site where cVg1/COS was implanted (Fig.
Fig. 7. Development of the early primitive streak by unidirectional growth of precursors along the AP axis. The precursor region was labeled
with β-gal virus either broadly (A,B) or discretely (C-F). Distribution of β-gal+ cell arrays (arrows) within the primitive streak was examined
by transverse histological sections of discretely labeled embryos (G,H). Diagrams illustrate relative positions of labeled cells within the
primitive streak.
Fig. 8. β-gal+ arrays oriented along the AP axis (A-C). A few cells in the precursor region were labeled with β-gal virus. (D) Frequency of
primitive streak labeling (n=16) depends on distance (d in left diagram) of a tagging site from Koller’s sickle along the embryonic midline
(arrow). Bar graph represents % of embryos exhibiting tagged cells only in primitive streak (filled bar), in both primitive streak and anterior
epiblast (shaded bar), or only in anterior epiblast (open bar).
The origin of early primitive streak
93
Fig. 9. Oriented metaphase
chromosome plates in the early
primitive streak. HH stages 3−
embryos were stained with
fluorescent phalloidin for actin
(A,B) and with DAPI for nuclei
(C,D). The developing primitive
streak (A,C) and non-primitive
streak epiblast (B,D) were
examined. (E) Angle of
metaphase chromosome plates to
the AP axis (AP) were scored in
the primitive streak (filled bar)
and non-primitive streak epiblast
(shaded bar). Primitive streak
nuclei were examined in both the
upper epithelioid layer and the
middle mesenchymal layer. n:
total number of counted
metaphase nuclei in the primitive
streak (ps) and non-primitive
streak epiblast (epi).
11A,B) as shown by Shah et al. (1997). In situ hybridization
of a primitive streak marker, Brachyury, identified that the
cVg1-induced cluster of cells was indeed an ectopic primitive
streak (Fig. 11C). To determine the source of cells forming
cVg1-induced, ectopic primitive streak, epiblast cells at
various positions relative to cVg1/COS were labeled with DiI
and DiO (Fig. 11D-G). Marking of cells lateral or distant from
cVg1/COS gave rise to an unlabeled primitive streak (Fig.
11E). By contrast, labeling of epiblast cells immediately
adjacent to cVg1/COS frequently generated a DiI+ or DiO+
primitive streak (Fig. 11F,G). The plane of cell division was
then examined in cVg1-induced primitive streak. Metaphase
chromosome plates in cVg1-induced primitive streak were
aligned perpendicularly to the ectopic streak axis (Fig. 11H),
Fig. 10. BrdU incorporation during primitive streak formation.
(A) Dorsal view of HH stage 3 primitive streak incubated with BrdU
for 2 hours before fixation, stained with BrdU antibody (left panel)
and DAPI (right panel). Dark-brown (left panel) and blue (right
panel) spots are BrdU-positive nuclei, while white spots (right panel)
are BrdU-negative nuclei. (B) Time course of BrdU incorporation
during primitive streak formation. The ratio of BrdU-positive nuclei
to total nuclei counted (n) is presented with standard deviation (bar).
similarly to those seen in authentic primitive streaks (Fig. 9).
The results indicate that only epiblast cells adjacent to an
inductive signal undergo the primitive streak formation and
these exhibit a polarized cell division.
DISCUSSION
Using both DiI and retroviral cell tagging procedures, the
present study visualizes for the first time the in ovo
development pattern of chicken blastoderm cells during
formation of the early primitive streak. The data map early
primitive streak precursors to a small area where the midline
crosses the posterior margin (Koller’s sickle) of the prestreak
blastodisc. Vg1-induced primitive streak formation strongly
suggests that individual blastoderm cells adjacent to an
inductive signal proliferate along the AP axis and exhibit
polarized cell division during HH stage 3 primitive streak
formation.
This study examined the fate of cells from five selected sites
of prestreak embryos. DiI and retroviral tagging only of site 1
produced labeled cells throughout the primitive streak. A DiI+
or β-gal+ primitive streak was never generated by labelling the
other four sites. Our results suggest that neither epiblast cell
migration from lateral to the midline nor random ingression of
epiblast cells followed by subsequent accumulation to the
posterior region is the primary pathway for generating the HH
stage 3 primitive streak in ovo. Rather, these results define a
model in which primitive streak precursor cells are already
localized to the posterior-midline region of stages XI-XII
blastodiscs. This conclusion is further supported by our present
results that exogenous Vg1 induces an ectopic primitive streak
by recruiting cells adjacent to the signal but not those lateral
or distal from it. These results also corroborate previous
observations that the epiblast portion of the posterior marginal
zone contributes to the primitive streak (Stern, 1990; EyalGiladi et al., 1992; Bachvarova, 1998).
These in ovo results differ from previously proposed models
for the location and migration pathways of precursors
94
Y. Wei and T. Mikawa
Fig. 11. Induced primitive streak exhibits same
characteristics as authentic primitive streak. Dorsal
views of embryos into which cVg1-negative (A) and
cVg1-positive (B,C) COS cells (arrows) were
implanted. Authentic (ps) and induced (ips) primitive
streaks were identified by in situ hybridization for
Brachyury (C). The origin of Vg1-induced primitive
streak was determined by labeling epiblast cells with
DiI (e1) and DiA (e2 and g) (D). The anteroposterior
polarity was illustrated by a and p. The localization of
tagged cells from each site was indicated in fluorescent
images (E,G) of resulting embryos. (F) Bright-field
image of G. Implanted COS cells are indicated by
white circles, while authentic (ps) and induced (ips)
primitive streaks were outlined by dotted lines.
(H) Angle of metaphase chromosome plates were
scored in Vg1-induced streak as in Fig. 9.
generating the early primitive streak (reviewed in
Stern, 1992; Eyal-Giladi et al., 1992). Although
the reason for this variation is not fully understood,
significantly different experimental approaches
were used to analyze cell movement in these
studies. Model-b (Fig. 1A) is based on the
expression pattern of HNK-1, a sulfated form of
glucuronic acid, and the fate of an internalized
HNK1-antibody complex (Stern, 1992). Model-c (Fig. 1A) is
based on the fate of a tissue fragment implanted into an isolated
blastodisc (Eyal-Giladi et al., 1992) and on the distribution of
diffusible vital dyes in ovo (Gräper, 1929). By contrast, the
present study directly traced the fate of blastoderm cells using
two non-invasive, non-diffusible cell tagging methods, DiI
labeling and retroviral-mediated gene transfer. Importantly,
these two independent in ovo approaches led to the same
conclusion.
The posterior-midline epiblast, including the primitive
precursor area, is stratified and not a simple epithelium. Our
injection of higher viral titer into the precursor area, probably
tagging both the surface and deeper layer of epiblast cells,
results in β-gal+ cells distributed to both the surface and deeper
layer of the primitive streak (Fig. 3C). In contrast, injection of
lower viral titers to the area tagged a population of cells in the
surface layer (Fig. 7G) or in the middle layer (Fig. 7H).
Consistent with our observations, it has been shown that cells
of both the surface and deeper layer of the posterior-midline
epiblast generate a subpopulation of primitive streak cells
(Bachvarova et al., 1998). The origin of these posterior-midline
cells of the stage XI-XII embryos remains to be identified.
This study examined the location and developmental pattern
of precursor cells that form the HH stage 3 primitive streak. At
later stages, the primitive streak is composed of cells derived
Fig. 12. Schematic illustration of the directed
cell division underlying the early primitive
streak development along the AP axis. Blue
spot in the left diagram marks the primitive
streak precursor area at prestreak stages (XXII), while blue bar in the right diagram
represents HH stage 3 primitive streak (ps). In
between, polarized cell division of precursor
cells are illustrated. Blue arrows, inductive
signal; AP, anteroposterior axis.
from the lateral epiblast. For example, at HH stage 4, cells that
originally formed the HH stage 3 primitive streak migrate
laterally and anteriorly and some of them differentiate into
lateral mesoderm (Eyal-Giladi et al., 1992; Garcia-Martinez
and Schoenwolf, 1993), while the HH stage 4 primitive streak
consists of cells that were originally localized in the
posterolateral region of the epiblast (Stern et al., 1995; Hatada
and Stern, 1994). Thus, the origin and migration pattern of cells
generating the early primitive streak are different from those
constituting the later stage primitive streak.
Interestingly, the β-gal-labelled cells were arrayed
longitudinally in the early primitive streak. Potential
mechanisms by which progenitor cells at the posterior margin
give rise to the extended streak include directed cell
intercalation, directed cell migration, cell shape change and
oriented cell division. Mitotic activity has been shown to be
associated with primitive streak development (Sanders et al.,
1993), and β-gal expression in cells forming the early primitive
streak requires at least one cell cycle for integration and
expression. This is clearly different from the non-mitotic
convergent and extension mechanism underlying frog
gastrulation (reviewed in Keller, 1986). Furthermore, cells of
the developing avian primitive streak do not exhibit elongation
perpendicular to the AP axis, the cell shape change predicted
to occur with lateral-to-medial cell sliding during
The origin of early primitive streak
convergent/extension (Keller, 1986). Our results do not support
oriented cell intercalation as a major mechanism accounting
for primitive streak development. Indeed, β-gal+ cells are
arranged in a longitudinal array without significant interruption
by β-gal-negative cells. A number of studies have
demonstrated that migration of mesodermal cells is directed by
a fibronectin cue (Harrison et al., 1993). However, fibronectin
is absent from the primitive streak (Duband and Thiery, 1982).
The hypothesis that the hypoblast underlying the epiblast may
provide a cue for direction of primitive streak development is
controversial (Azar and Eyal-Giladi, 1981; Khaner, 1995).
Polarized cell division plays a role in shaping the neural
plate (Hamburger, 1948; Martin, 1967; Sausedo et al., 1997;
Concha and Adams, 1998). Here we demonstrate that cells in
the developing primitive streak have metaphase chromosome
plates with a plane perpendicular to the AP axis. Cells in an
ectopic primitive streak induced by Vg1 exhibit similar
polarity in their cell division. The orientation of metaphase
chromosome plates together with the observed longitudinal
arrays of β-gal+ cells in the primitive streak suggest that
directional cell division plays a role in primitive streak
elongation along the midline. This idea is consistent with the
model illustrated in Fig. 12, where a local signal(s), such as
epithelial scatter factor (Stern et al., 1990), activin (Mitrani et
al., 1990; Ziv et al., 1992) and Vg1 (Seleiro et al., 1996), in a
confined area of the posterior marginal zone, induces a
blastoderm cell subpopulation that differentiates into the early
primitive streak. During this process, primitive streak cells
undergo cytokinesis with the plane perpendicular to the AP
axis. Future studies should address mechanisms that orient the
plane of cytokinesis in streak-forming cells. It remains to be
established if oriented cell division is causal for primitive
streak elongation or an associated phenomenon.
The authors thank Drs I. Skromne, C. D. Stern and J. Dodd for
cVg1 cDNA and their guidance of ectopic primitive streak induction
with Vg1. Our thanks extend to Drs P. Wilson, D. A. Fischman, J.
Hyer and K. Kelly for their comments and discussions on this study
and Ms L. Cohen-Gould, L. Miroff and L.-L. Ong for their technical
assistance. This work was supported in part by grants from the NIH
and the Mathers’ Foundation. T. M. is an Irma T. Hirschl Scholar.
REFERENCES
Azar, Y. and Eyal-Giladi, H. (1981). Interaction of epiblast and hypoblast in
the formation of the primitive streak and the embryonic axis in chick, as
revealed by hypoblast-rotation experiments. J. Embryol. Exp. Morph 61,
133-144.
Bachvarova, R. F., Skromne, I. and Stern, C. D. (1998). Induction of
primitive streak and Hensen’s node by the posterior marginal zone in the
early chick embryo. Development 125, 3521-3534.
Balinsky, B. I. (1975). Introduction to Embryology. Philadelphia: Saunders.
Boncinelli, E. and Mallamaci, A. (1995). Homeobox genes in vertebrate
gastrulation. Curr. Opin. Genet. Dev. 5, 619-627.
Callebaut, M. and Van Nueten, E. (1994). Rauber’s (Koller’s) Sickle: The
early gastrulation organizer of the avian blastoderm. Eur. J. Morph. 32, 3548.
Cheng, N. N., Kirby, C. M. and Kemphues, K. J. (1995). Control of cleavage
spindle orientation in C. elegans: the role of the genes par-2 and par-3.
Genetics 139, 549-559.
Concha, M. L. and Adams, R. J. (1998). Oriented cell divisions and cellular
morphogenesis in the zebrafish gastrula and neurula: a time-lapse analysis.
Development 125, 983-994.
Dougherty, J. P. and Temin, H. M. (1986). High mutation rate of a spleen
necrosis virus-based retrovirus vector. Mol. Cell. Biol. 168, 4387-4395.
95
Duband, J. L. and Thiery, J. P. (1982). Appearance and distribution of
fibronectin during chick embryo gastrulation and neurulation. Dev. Biol. 94,
337-350.
Eyal-Giladi, H. and 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). Dev. Biol. 49, 321-337.
Eyal-Giladi, H., Debby, A. and Harel, N. (1992). The posterior section of
the chick’s area pellucida and its involvement in hypoblast and primitive
streak formation. Development 116, 819-830.
Garcia-Martinez, V. and Schoenwolf, G. C. (1993). Primitive-streak origin
of the cardiovascular system in avian embryos. Dev. Biol. 159, 706-719.
Gräper, L. (1929). Die primitiventwicklung des huhnchens nach
stereokinematographischen untersuchungen, kontrolliert durch vitale
farbmarkierung und verglichen mit der entwicklung anderer wirbeltiere. W.
Roux Arch. EntwMech Org. 116, 382-429.
Hamberger, V. (1948). The mitotic patterns in the spinal cord of the chick
embryo and their relation to histogenetic processes. J. Comp. Neurol. 88,
221-284.
Hamburger, V. and Hamilton, H. L. (1951). A series of normal stages in the
development of the chick embryo. J. Morph. 88, 49-67.
Harrisson, F., Nassauw, L. V., Hoof, J. V. and Foidart, J. M. (1993).
Microinjection of antifibronectin antibodies in the chicken blastoderm:
inhibition of mesoblast cell migration but not of cell ingression at the
primitive streak. Anat. Rec. 236, 685-696.
Hatada, Y. and Stern, C. D. (1994). A fate map of the epiblast of the early
chick embryo. Development 120, 2879-2889.
Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J. and Ish-Horowicz,
D. (1995). Expression of Delta homologue in prospective neurons in the
chick. Nature 375, 787-790.
Itoh, N., Mima, T. and Mikawa, T. (1996). Loss of fibroblast growth factor
receptors is necessary for terminal differentiation of embryonic limb muscle.
Development 122, 291-300.
Keller, R. E. (1986). The cellular basis of amphibian gastrulation. In
Developmental Biology: A Comprehensive Synthesis (ed. L. Browder), Vol.
2, pp. 241-327. New York: Plenum Press.
Khaner, O. (1995). The rotated hypoblast of the chicken embryo does not
initiate an ectopic axis in the epiblast. Proc. Natl. Acad. Sci. USA 92, 1073310737.
Khaner, O. and Eyal-Giladi, H. (1986). The embryo-forming potency of the
posterior marginal zone in stages X through XII of the chick. Dev. Biol. 115,
275-281.
Knezevic, V., Ranson, M. and Mackem, S. (1995). The organizer-associated
chick homeobox gene, Gnot1, is expressed before gastrulation and regulated
synergistically by activin and retinoic acid. Dev. Biol. 171, 458-470.
Knezevic, V., DeSanto, R. and Mackem, S. (1997). Two novel chick T-box
genes related to mouse Brachyury are expressed in different, nonoverlapping mesodermal domains during gastrulation. Development 124,
411-419.
Kochav, S., Ginsburg, M. and Eyal-Giladi, H. (1980). From cleavage to
primitive streak formation: a complementary normal table and a new look
at the first stages of the development of the chick (II). Dev. Biol. 79, 296308.
Martin, A. H. (1967). Significance of mitotic spindle fibre orientation in the
neural tube. Nature 216, 1133-1134
Mikawa, T., Borisov, A., Brown, A. M. C. and Fischman, D. A. (1992a).
Clonal analysis of cardiac morphogenesis in the chicken embryo using a
replication-defective retrovirus: I. formation of the ventricular myocardium.
Dev. Dyn. 193, 11-23.
Mikawa, T., Cohen-Gould, L. and Fischman, D. A. (1992b). Clonal analysis
of cardiac morphogenesis in the chicken embryo using a replicationdefective retrovirusIII: Polyclonal origin of adjacent ventricular myovytes.
Dev. Dyn. 195, 133-141.
Mikawa, T. and Fischman, D. A. (1992). Retroviral analysis of cardiac
morphogenesis: discontinuous formation of coronary vessels. Proc. Natl.
Acad. Sci USA 89, 9504-9508.
Mikawa, T. and Gourdie, R. G. (1996). Pericardial mesoderm generates a
population of coronary smooth muscle cells migrating into the heart along
with ingrowth of the epicardial organ. Dev. Biol. 173, 221-232.
Mima, T., Ueno, H., Fischman, D. A., Williams, L. T. and Mikawa, T.
(1995). FGF-receptor is reguired for in vivo cardiac myocyte proliferation
at early embryonic stages of heart development. Proc. Natl. Acad. Sci. USA
92, 467-471.
Mitrani, E. and Shimoni, Y. (1990). Induction by soluble factors of organized
axial structures in chick epiblasts. Science 247, 1092-1094.
96
Y. Wei and T. Mikawa
Mitrani, E., Ziv, T., Thomsen, G., Shimoni, Y., Melton, D. A. and Bril, A.
(1990). Activin can induce the formation of axial structures and is expressed
in the hypoblast of the chick. Cell 63, 495-501.
Ruiz i Altaba, A., Warga, R. M. and Stern, C. D. (1993). Fate maps and cell
lineage analysis. In Essential Developmental Biology (ed. D. D. Stern and
P. W. H. Holland). pp81-95. Oxford: IRL Press.
Sanders, E. J., Varedi, M. and French, A. S. (1993). Cell proliferation in the
gastrulating chick embryos: a study using BrdU incorporation and PCNA
localization. Development 118, 389-399.
Sausedo, R. A., Smith, J. L. and Schoenwolf, G. C. (1997). Role of
nonrandomly oriented cell division in shaping and bending of the neural
plate. J. Comp. Neurol. 381, 473-488.
Seleiro, A. P. E., Connolly, D. J. and Cooke, J. (1996). Early developmental
expression and experimental axis determination by the chicken Vg1 gene.
Curr. Biol. 6, 1476-1486.
Shah, S. B., Skromne, I., Hume, C. R., Kessier, D. S., Lee, K. J., Stern,
C. D. and Dodd, J. (1997). Misexpression of chick Vg1 in the
marginal zone induces primitive streak formation. Development 124,
5127-5138.
Stern, C. D. (1990). The marginal zone and its contribution to the hypoblast
and primitive streak of the chick embryo. Development 109, 667-682.
Stern, C. D. (1992). Mesoderm induction and development of the embryonic
axis in amniotes. Trends Genet. 8, 158-163.
Stern, C. D., Ireland, G. W., Herrick, S. E., Gherardi, E., Gray, J.,
Perryman, M. and Stoker, M. (1990). Epithelial scatter factor and
development of the chick embryonic axis. Development 110,1271-1284.
Stern, C. D., Yu, R. T., Kakizuka, A., Kintner, C. R., Mathews, L. S., Vale,
W. W., Evans, R. M. and Umesono, K. (1995). Activin and its receptors
during gastrulation and the later phases of mesoderm development in the
chick embryo. Dev. Biol. 172, 192-205.
Wilson, V., Manson, L., Skarnes, W. C. And Beddington, R. S. (1995). The
T gene is necessary for normal mesodermal morphogenetic cell movements
during gastrulation. Development 121, 877-886.
Winnier, G., Blessing, M., Labosky, P. A. and Hogan, B. L. M. (1995). Bone
Morphogenetic protein-4 is required for mesoderm formation and patterning
in the mouse. Genes Dev. 9, 2105-2116.
Ziv, T., Shimoni, Y. and Mitrani, E. (1992). Activin can generate ectopic
axial structures in chick blastoderm explants. Development 115, 689-694.