ArticlePDF Available

Symmetry breaking, germ layer specification and axial organisation in aggregates of mouse embryonic stem cells


Abstract and Figures

Mouse embryonic stem cells (mESCs) are clonal populations derived from preimplantation mouse embryos that can be propagated in vitro and, when placed into blastocysts, contribute to all tissues of the embryo and integrate into the normal morphogenetic processes, i.e. they are pluripotent. However, although they can be steered to differentiate in vitro into all cell types of the organism, they cannot organise themselves into structures that resemble embryos. When aggregated into embryoid bodies they develop disorganised masses of different cell types with little spatial coherence. An exception to this rule is the emergence of retinas and anterior cortex-like structures under minimal culture conditions. These structures emerge from the cultures without any axial organisation. Here, we report that small aggregates of mESCs, of about 300 cells, self-organise into polarised structures that exhibit collective behaviours reminiscent of those that cells exhibit in early mouse embryos, including symmetry breaking, axial organisation, germ layer specification and cell behaviour, as well as axis elongation. The responses are signal specific and uncouple processes that in the embryo are tightly associated, such as specification of the anteroposterior axis and anterior neural development, or endoderm specification and axial elongation. We discuss the meaning and implications of these observations and the potential uses of these structures which, because of their behaviour, we suggest to call 'gastruloids'.
Content may be subject to copyright.
Symmetry breaking, germ layer specification and axial
organisation in aggregates of mouse embryonic stem cells
Susanne C. van den Brink
*, Peter Baillie-Johnson
*, Tina Balayo
, Anna-Katerina Hadjantonakis
Sonja Nowotschin
, David A. Turner
and Alfonso Martinez Arias
Mouse embryonic stem cells (mESCs) are clonal populations derived
from preimplantation mouse embryos that can be propagated in vitro
and, when placed into blastocysts, contribute to all tissues of the
embryo and integrate into the normal morphogenetic processes,
i.e. they are pluripotent. However, although they can be steered to
differentiate in vitro into all cell types of the organism, they cannot
organise themselves into structures that resemble embryos. When
aggregated into embryoid bodies they develop disorganised masses
of different cell types with little spatial coherence. An exception to this
rule is the emergence of retinas and anterior cortex-like structures
under minimal culture conditions. These structures emerge from the
cultures without any axial organisation. Here, we report that small
aggregates of mESCs, of about 300 cells, self-organise into polarised
structures that exhibit collective behaviours reminiscent of those that
cells exhibit in early mouse embryos, including symmetry breaking,
axial organisation, germ layer specification and cell behaviour, as
well as axis elongation. The responses are signal specific and
uncouple processes that in the embryo are tightly associated, such
as specification of the anteroposterior axis and anterior neural
development, or endoderm specification and axial elongation. We
discuss the meaning and implications of these observations and the
potential uses of these structures which, because of their behaviour,
we suggest to call gastruloids.
KEY WORDS: Mouse, Gastrulation, Self-organisation, Symmetry
breaking, Polarisation, Axial elongation, Endoderm, Mesoderm,
Neural ectoderm, Pattern formation, Live cell imaging
The emergence of asymmetries within a mass of otherwise
equivalent cells is the starting event in the development and
patterning of all embryos, and results in the establishment of a
coordinate system that cells use as a reference to generate the main
axes of an organism. In animal embryos the axial organisation acts
as a reference for the process of gastrulation, a choreographed
sequence of cell movements that transforms an often hollow
epithelium into a three-layered structure endowed with a blueprint
for the organism: a head at the anterior pole and, in vertebrates, the
ectoderm, that will give rise to the nervous system on the dorsal side
and the endoderm and the mesoderm on the ventral side. The
process of gastrulation is driven by coordinated movements of
groups of cells that interpret the global coordinate system of the
embryo and give rise to the endoderm and the mesoderm (Norris
et al., 2002; Nowotschin and Hadjantonakis, 2010; Ramkumar and
Anderson, 2011; Solnica-Krezel and Sepich, 2012). Although the
outcome of gastrulation is highly conserved, the mechanics of
the process varies, even within a phylum (Keller et al., 2003). Thus,
within the chordates, anamniotes such as amphibians utilise the
convergence of a ring of cells into a small opening, or blastopore,
through which mesoderm and endoderm invaginate and perform
directional movements, whereas in amniotes, such as chicken,
mouse and the primates, the invaginating cells move within a
different geometry and configure a dynamic groove of cells called
the primitive streak (PS) that acts as the source of the endoderm
and mesoderm (reviewed by Keller et al., 2003; Solnica-Krezel and
Sepich, 2012; Tam and Loebel, 2007). In all embryos there is a
close relationship between the process of gastrulation and the
establishment of axial structures, as experiments and mutants that
disturb the establishment of the axis lead to profound alterations in
the specification and movement of the endodermal and mesodermal
precursors (Huelsken et al., 2000; Medina et al., 1997; Morkel et al.,
2003). These defects have secondary effects on the development of
the nervous system.
Studies with mouse and human embryonic stem cells (ESCs) have
shown that culture of 3D aggregates termed embryoid bodies (EBs)
leads to the formation of rudiments of tissues and organs without the
context of an embryo (Desbaillets et al., 2000; Höpfl et al., 2004).
Recently, EBs have been steered to differentiate into eye cups and
anterior neural cortical structures (Eiraku et al., 2011; Lancaster et al.,
2013; Nakano et al., 2012; Sasai, 2013; Sasai et al., 2012), a remarkable
feat as these structures emerge without a recognisable reference
coordinate system. An explanation for this observation might lie in the
intrinsic tendency of mouse ESCs (mESCs) to develop anterior neural
fates (Tropepe et al., 2001; Turneret al., 2014c; Watanabe et al., 2005;
Wataya et al., 2008). In contrast to these observations, there are no
reports of the emergence of axial structures in EBs, even though in
culture it is possible to obtain progenitor cells for mesodermal and
endodermal structures (Gadue et al., 2006; Kouskoff et al., 2005; Kubo
et al., 2004) that exhibit some of the morphogenetic properties of the
embryo (Turner et al., 2014b), and signalling can elicit a degree of
polarised gene expression in EBs (ten Berge et al., 2008). One
exception was reported in a study of P19 embryo carcinoma (EC) cells.
Under differentiation conditions, EBs made from these cells can
organise themselves into polarised and extending structures
resembling gastrulating embryos (Marikawa et al., 2009). Such
large-scale organisation has not been described in ESCs.
Here we show that small aggregates of mESCs undergo a
symmetry-breaking event in culture and that, under conditions that
Received 23 May 2014; Accepted 26 September 2014
Department of Genetics, University of Cambridge, Cambridge CB2 3EH, UK.
Developmental Biology Program, Sloan-Kettering Institute, New York, NY 10065,
*These authors contributed equally to this work
Author for correspondence (
This is an Open Access article distributed under the terms of the Creative Commons Attribution
License (, which permits unrestricted use,
distribution and reproduction in any medium provided that the original work is properly attributed.
© 2014. Published by The Company of Biologists Ltd
Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
promote the formation of mesendoderm in embryos, they exhibit
polarised expression of the endoderm marker Sox17 (Kanai-Azuma
et al., 2002) and FoxA2 (Monaghan et al., 1993; Sasaki and Hogan,
1993) and of the PS and early mesoderm marker brachyury (Bra, or
T) (Herrmann, 1991). Over time, Bra expression becomes restricted
to a small population of cells at a tip of the aggregate, which acts as a
source of cells that express Tbx6, a mesoderm gene (Chapman
et al., 1996), and these cells are extruded from the main body of
the aggregate in a process that is reminiscent of some of the
movements of gastrulation. For this reason, we call these aggregates
gastruloidsand show that, although for the most part they are
autonomous in their development, the culture conditions influence
the cell types that develop within them. We compare the behaviour
of these aggregates with that of embryos and discuss their potential
as a new experimental system with which to study mechanisms of
early mammalian development.
Symmetry breaking in differentiating EBs
The observation that P19 EC cells are able to form polarised,
elongated structures during differentiation (Marikawa et al., 2009)
prompted us to seek culture conditions in which EBs derived from
mESCs would develop similar structures. When cells were placed in
a serum and LIF hanging drop culture, cells formed aggregates and,
after removal of LIF, a small proportion changed their morphology
to an ovoid appearance, although any further suggestion of
elongation was never apparent (data not shown). In order to
stimulate the emergence of PS features, we used culture conditions
that steer the cells towards this fate in adherent culture (Gadue et al.,
2006; Turner et al., 2014b,c) and exposed EBs of different sizes to
N2B27 for 2 days followed by continuous treatment with both
activin A (Act) and CHIR99021 (Chi), a Wnt/β-catenin signalling
agonist (Act/Chi conditions) (supplementary material Fig. S1A; see
Materials and Methods for details).
The initial EBs contained 800-1000 cells and during the first
phase of aggregate formation in N2B27 we noticed that, in contrast to
cells hanging in serum and LIF where onlyone aggregatewas formed
per drop (supplementary material Fig. S1D), cells in N2B27 formed
multiple aggregates of variable sizes per drop (supplementary material
Fig. S1C). Following the change of medium into Act/Chi conditions,
the aggregates dispersed and over time we observed an increasing
number of aggregates adopting a shape that differed from their
original spherical appearance (supplementary material Fig. S1E-F);
some displayed an ovoid shape (supplementary material Fig. S1E),
Fig. 1. Comparative analysis of the effect of exposure time and signalling on aggregate formation. (A) The stimulation protocol. The vertical black dotted
lines indicate medium changes, and the red vertical line corresponds to the beginning of day 2. Aggregates were cultured in N2B27 (grey shading) continuously
), or treated with continuous (P
) or 24 h pulses (P
) of Act, Chi or Act/Chi (blue shading) before being returned to N2B27. Data for P1 not shown.
(B) Cartoon renderings (see Materials and Methods; unprocessed images are shown in supplementary material Fig. S2) of typical aggregate morphologies on day
5 following the conditions shown in A; images are not to scale. Maximum elongation was observed following pulsed treatment within the day 2-3 time frame(P
(C,C) Comparison of aggregate morphologies following a 24 h pulse on days 2-3 of Act, Chi, Act/Chi and BMP4 for three different cell types: (a) Sox17::GFP, (b)
TBX6::EYFP and (c) wild-type E14-Tg2A. (C) Aggregates were scored based on whether they were spherical (white), contained a single outgrowth (ovoid, grey),
showed overt elongation (green) or had multiple protrusions (blue). Examples of these aggregate morphologies are shown in the form of cartoon renderings,
processed as described above. (C) These data are also represented as ratios between the indicated morphologies and the proportion of aggregates with ovoid
appearance. Note how transient exposure to Chi results in a much higher ratio of elongated to ovoid morphologies. The number of Sox17::GFP, Tbx6::EYFP and
E14-Tg2A aggregates for each condition (C,C) are as follows (respectively): Act: 71, 27, 52; Chi: 60, 32, 48; Act/Chi: 69, 17, 46; BMP4: 64, 24, 38.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
resembling what has been described previously for EBs when
β-catenin is activated (see Figure 4B in ten Berge et al., 2008).
However, we also observed clear elongation in a few aggregates
(supplementary material Fig. S1E) and,by the fourth day in Act/Chi, a
median of 30% of aggregates exhibited a polarised, elongated
morphology. The reduction in the proportion of cells displaying
elongated aggregates at later time points reflected an increase in cells
displaying a differentiated phenotype in addition to an increase in an
apoptotic appearance (data not shown). The aggregates needed to
be in suspension for shape changes to occur (Baillie-Johnson
et al., 2014).
These results indicate that it is possible to elicit symmetry
breaking and polarisation in aggregates of ESCs.
The effect of signals and aggregate size on polarisation
The heterogeneous response ofthe EBsto ourexperimentaltreatment
could be due to several factors and we decided to focus on three that
we deemed to be most influential in the outcome of the experiment:
(1) the composition of the culture medium; (2) the timing of exposure;
and (3) the initial size of the aggregates. In these experiments we
moved from culturing the aggregates in hanging drop to 96-well
plates, as this allowed us to arrange for one, and only one, aggregate to
develop in each well, minimising the possibilityof fusions (for details
see Baillie-Johnson et al., 2014). The aggregates were first placed in
N2B27 for 2 days and assayed on the fifth day of culture.
If, after the initial 2 days in N2B27, the aggregates are left in this
medium then we observe a range of morphologies, with 20-30%
exhibiting some polarisation. When signals (Act, Chi and BMP) are
applied continuously from the third day of differentiation, the
response is signal specific: in Act/Chi many of the aggregates exhibit
a weak elongation, whereas continuous exposure to Act alone elicits
a variable number of short protrusions or invaginations per
aggregate; on its own Chi triggers a smaller number of longer and
broader protrusions (Fig. 1A,B, protocol P
). In all cases there is a
variability in the response of the aggregate to a particular culture
condition that changes with the cell line, although the structure that
emerges is specific and recognisable for each of the signals (Fig. 1B).
Using Act and Chi in combination (Act/Chi) or individually, we
next tested the effect of changing the timing of exposure to signals
(Fig. 1A,B, protocols P
). We began by restricting the exposure of
these signals from differentiation days 2-5 (Fig. 1A, protocol P
days 2-3, 3-4 and 4-5 (Fig. 1A, protocols P
and P
respectively), and returning the aggregates to N2B27 for the
duration of the timecourse. Limiting the exposure to the third day
(48-72 h) triggered the most reproducible response, with greater
than 70% of the aggregates undergoing similar morphological
changes (Fig. 1C,C). Under these conditions, Act alone produces a
number of small, broad protrusions and invaginations from a large
oval. Addition of Chi to the Act reduces the number of invaginations
and, in many instances, elicits a single elongation of 40-60 µm that
is attached to a broad mesenchymal-like structure at the distal end of
the aggregate (Fig. 1C,C). On its own a short exposure to Chi
consistently elicits a single elongation without clear protrusions or
invaginations. Exposure to the different signals limited to either the
second or the fourth day of culture produced more variable
responses, and many aggregates that did not respond (Fig. 1). These
results suggest that, during the third day of differentiation, ESCs are
in a competent state to efficiently interpret signals in the medium.
As exposure to Chi from the third day elicited a simple and
consistent response in the form of an elongation, we used this
experimental condition as the basis to analyse the effect of the initial
size of the aggregates on their polarisation (Fig. 2; supplementary
material Fig. S2A-D). Starting with different numbers of cells
revealed that only small aggregates, of 300±100 cells, exhibit the
elongation effectively. Smaller aggregates (<200 cells) either grow
or remain small and exhibit slow growth; when they grow, they tend
to produce the elongation. Larger aggregates (>600 cells) grow in
the disorganised and symmetrical manner that is regularly reported
for EBs.
These results indicate that EBs configured from mESCs are
capable of elongation in the manner that has been described for
P19 EC cells (Marikawa et al., 2009) and that this behaviour is
associated with specific culture conditions.
Spatial and temporal patterns of gene expression in
polarising EBs
The elongated aggregates resemble structures that have been
described in sea urchin and amphibian embryos (Holtfreter, 1933;
Horstadius, 1939; Ishihara et al., 1982; Keller and Danilchik, 1988)
or when animal caps from Xenopus embryos are exposed to Activin
Fig. 2. Effect of initial cell density on the
elongation of aggregates. (A) Aggregates
formed from increasing numbers of cells (200-
1600 cells) as indicated were exposed to Chi for
the duration of the experiment (at least six
aggregates per condition). (B) Aggregates with
an initial size of between 400 and 800 cells
showed elongation. (C) Aggregates with 800
cells tended to exhibit multiple elongations
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
(Green et al., 2004; Ninomiya et al., 2004; Symes and Smith, 1987).
These comparisons suggested to us that the elongated bodies might
be recapitulating some of the early events associated with
gastrulation. If this were the case, the cells involved in generating
the protrusions might represent mesendodermal tissue. To address
this and exclude the possibility that the protrusion is simply a
mechanical response to the size and shape of the aggregates without
a specific fate (i.e. that there is no correspondence between structure
and fate), we analysed the expression of genes associated with early
differentiation in culture and in embryos (Fig. 3). To begin with we
analysed the expression of Sox17 (Figs 3 and 4), a marker of
primitive and definitive endoderm (Kanai-Azuma et al., 2002), and
of Bra (Fig. 4), a gene associated with the specification of endoderm
and mesoderm in the PS (Herrmann, 1991), using fluorescent
reporter ES cell lines for both genes (Fehling et al., 2003; Niakan
et al., 2010). Aggregate formation and staining with Sox17 and Bra
antibodies confirmed that both lines are faithful reporters of the
expression of the genes (supplementary material Fig. S3) (Turner
et al., 2014b).
Following a transient exposure during day 3 to either Act or Act/Chi
we observe expression of Sox17::GFP mostly in clusters of cells
which, in the presence of Act/Chi, tend to lie within the elongating
region (Fig. 3A). Confocal optical sections through the aggregates
revealed them to be multi-layered structures (Fig. 3A,A)witha
number of aggregates displaying internal cavities and localised
indentations or pits on their surface (Fig. 3B,B;supplementary
materialMovie 1). Sox17 expression is, for the most part, restricted to
the external cells and is associated with E-cadherin (Fig. 3A), as it is in
the embryo. The amount of Sox17 expression increases with the time
of exposure and requires Act, as exposure to Chi alone reduces the
levels of expression and the numberof expressing cells (Fig. 3A). In all
cases, Sox17-expressing cells tend to invaginate, retain E-cadherin
expression and form vesicles near the surface of the aggregate
(Fig. 3B; see also Fig. 7F).
Further analysis of aggregates exposed to Act and, particularly, to
Act/Chi revealed expression of Bra (Fig. 4A,B), Sox17 (Fig. 3 and
Fig. 4B) and FoxA2 (Fig. 4B), which are all associated with the PS,
localised towards the extending tip of the aggregate; Bra and Sox17
are expressed in a mutually exclusive pattern (Fig. 4B). Whereas the
expression of Sox17 coincides with that of FoxA2 (Fig. 4B), that of
Bra correlates with a high level of β-catenin transcriptional activity
as demonstrated in a TCF/LEF::GFP (TLG) reporter cell line
(Fig. 4C). These patterns of expression are reminiscent of those in
gastrulating embryos, in which Bra and Wnt/β-catenin signalling
can be observed in the PS (Fig. 4D) (Ferrer-Vaquer et al., 2010).
Taken together, these observations suggest that the aggregates
formed from mESCs undergo morphogenetic movements that
resemble the early stages of gastrulation.
Imaging symmetry-breaking events in real time in the
In order to monitor the emergence of the polarised expression
of Sox17 and Bra, we performed live cell microscopy on E14-Tg2A,
Sox17::GFP and Bra::GFP transcriptional reporter cell lines. First,
Fig. 3. Polarisation, patterning and gene
expression in aggregates. (A,A)Two
single sections through GPI-GFP mESC
aggregates exposed to N2B27 for 5 days
with a 24 h pulse of either Act (n=10), Chi
(n=5), or Act/Chi (n=14) between 48 and
72 h and imaged by confocal microscopy
(GPI-GFP channel not shown). The
expression of the indicated markers on the
surface of the aggregates is shown in A,
with the corresponding orthogonal view
through the aggregate in A. The arrows in
Aindicate the z-section shown in A and A.
Note how the expression of Sox17 is
localised to the surface of the aggregate.
(B,B) A representative aggregate from
GPI-GFP mESCs exposed to Act between
48 and 72 h was imaged at the end of the
treatment after being fixed and stained for
E-cadherin and Sox17 (B); the boxed
region is enlarged to show E-cadherin (B).
Note the depressions that are associated
with Sox17 expression and high levels of E-
cadherin. (C) Section through an E7.5
embryo stained for Sox17 and with DAPI.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
we imaged the formation of the aggregates from a cell suspension of
mESCs in N2B27 for 48 h (Fig. 5A; supplementary material
Movie 2). Individual cells or clusters containing small numbers of
cells were found to coalesce into larger aggregates due to a
combination of the spatial constraints of the round-bottomed culture
wells and active cell movement towards the aggregate (Fig. 5A;
supplementary material Movie 2). During this time we do not
observe expression of visceral endoderm (VE) markers, such as
Gata6 (not shown) or Sox17, suggesting that the aggregates are
composed exclusively of embryonic tissues. We then focused on the
Act/Chi conditions, recording the emergence of the fluorescence
from the time of transfer from the N2B27 medium into Act/Chi
(Fig. 5B,D) and then at later stages when the aggregates were more
advanced (Fig. 5C,E; supplementary material Movies 2-4). In both
cases, we observe the development of polarised gene expression
over time, but the patterns are different for each of the genes.
In the case of Sox17 (Fig. 5B,C), after 32 h in Act/Chi we observe
an initial pattern of scatteredcells expressingthe reporter intermingled
with Bra-expressing cells (Fig. 5B,D; supplementary material
Movies 3 and 4). The videos suggest that the definitive, polarised
expression pattern is established from the aggregation of Sox17-
expressing cells on one side of the aggregate, which then proliferate
and at 96 h can be seen to be associated with the elongation. Clusters
of Sox17-expressing cells can be seen to move inside the aggregate,
which would be consistent with the invaginations described above
(Fig. 4B,C). In the case of Bra (Fig. 5D,E; supplementary material
Movies 5 and 7), Bra::GFP was initially expressed transiently across
the whole aggregate (79 h in secondary medium; Fig. 5E) before
becoming restricted to a small region (Fig. 4E, Fig. 5D,E).
Downregulation of the reporter in other regions of the aggregate
appeared to be undertaken by individual cells not within the region of
high expression. As time progressed, the aggregate increased in size
and maintained the expression of Bra::GFP within one region
(Fig. 5E; supplementary material Movie 5). These results suggest that
symmetrybreaking and polarisation of gene expression area feature of
these aggregates elicited by different signals.
Signalling and pattern formation during aggregate
In the early postimplantation epiblast, cell fate assignments are
triggered by interactions between BMP, Nodal, Wnt signalling and
their antagonists, and lead to the partitioning of the embryo into
anterior neuroectodermal and posterior mesendodermal populations
(Arnold and Robertson, 2009; Pfister et al., 2007; Tam and Loebel,
2007). To expand our studies beyond mesendoderm, we used a Sox1::
GFP reporter ES cell line to monitor neural development (Ying et al.,
2003), and a TBX6::EYFP lineto follow the emergence of mesoderm
(see Materials and Methods). We also included BMP (see also
Fig. 1C) in the repertoire of signals, as it plays a role in the earlystages
of embryonic patterning (Arnold and Robertson, 2009; Tam and
Loebel, 2007). In these experiments, aggregates were exposed to the
signals either during the third day of differentiation and then returned
to N2B27 for a further 2 days, or for the last 3 days of the experiment
(as summarised in Fig. 6).
When cells are left in N2B27, for the most part they do not undergo
any specific morphogenetic process and express Sox1::GFP
throughout the aggregate (Fig. 6 and Fig. 7A), although in 10% of
cases we observe some polarised Bra expression (not shown). The
pattern of Sox1::GFP expression does not change when inhibitors of
Nodal/Activin (SB43) or MEK (PD03) are added to the medium from
day 3 (not shown) and is consistent with the observation that mESCs
placed in N2B27 will develop, mainly, as neural precursors (Ying
et al., 2003). Also consistent with known inputs of signalling on neural
development (Andoniadou and Martinez-Barbera, 2013; Turner et al.,
2014c), Sox1::GFP expression was suppressed by exposure to Act,
although a few foci of expression remained in some aggregates (Fig. 6
Fig. 4. Polarised expression of Bra in responseto Wnt signalling. (A) Bra::GFP cells exposed to a24 h pulse of Act or Chi between 48 and 72 h and imaged at
120 h (n>3). (B) Aggregates of a Sox17::GFP cell line treated with sustained Act/Chi and stained on day 4 for either Bra (left, n>10) or FoxA2 (right, n=13).
(C,D) The β-catenin transcriptional reporter line TCF/LEF::mCherry treated with (C) Chi in culture (as in A, n=11) and compared with its expression in the PS of
an E6.5 embryo (D), both stained for Bra. These conditions not only produce elongations of the aggregates, but also result in polarised gene expression.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
and Fig. 7A). The response of genes associated with endoderm (Sox17)
and mesoderm (Tbx6) to the different signals is summarised in Fig. 6
(examples of expression patterns are shown in Fig. 7). The pattern of
responses mirrors that of embryos to the same signals (Figs 6 and 7).
For example, Act suppresses mesoderm and promotes endoderm,
whereas BMP promotes mostly mesoderm and Chi is able to elicit all
germ layers (Fig. 6). In all cases the different cell types emerge as
continuous and polarised groups of expressing cells: a pulse of Chi
leads to an increase in TBX6::EYFP expression (Fig. 7B) and
polarised β-catenin transcriptional activity (Fig. 7C) and Bra::GFP
expression (Fig. 7D). We also observe interactions between the
different signals; thus, BMP appears to quench the effects of Chi on
Sox1 expression, and Act suppresses the effects of BMP on Tbx6
expression (not shown). Prolonged exposure to a signal or signal
combination tends to increase the response in terms of expression but
has a negative effect on polarisation of the expression (Fig. 6 and data
not shown).
The exposure to a pulse of Chi led to an elongation that,
surprisingly, exhibits Sox1::GFP expression in the elongating cells
(Fig. 7A), with Bra expression restricted to the tip of the elongate in a
small region that does not express Sox1::GFP (Fig. 7A, Chi pulse
insets; Fig. 7D). The elongated region exhibits a complex structure,
with most of the cells expressing Sox2 and, often, Sox17 in vesicles
that form near the surface and have lower levels of Sox2 (Fig. 7F);
Sox1 expression is non-overlapping with that of Sox17. In the non-
elongated region we observe low levels of Sox2 expression. This
arrangement is reminiscent of the situation in the embryo, where the
endoderm, which expresses Sox17 and Sox2 (Wood and Episkopou,
1999), lies underneath the developing nervous system (Fig. 7G). In
addition, the aggregates express TBX6, which is usually associated
with mesoderm formation (Chapman et al., 1996), and appear to
recapitulate events associated with axial extension (see Turner et al.,
2014a). The aggregates lack a notochord, which, in the embryo, lies
between the nervous system and the gut. An important feature of the
development of these aggregates is the timing of the events, which is
reliable and reproducible from experiment to experiment: the Sox17
expression precedes and initiallyoverlaps with Bra expression and the
extrusion of cells, and Tbx6 expression follows a few hours later (see
Fig. 8) (Turner et al., 2014a).
These results complement the morphological changes described
above and are consistent with what is known about the early events
in the embryo, namely the existence of a pre-proneural basal state in
the epiblast with the mesendoderm being specified by BMP, Nodal/
Act and Wnt signalling (Turner et al., 2014c). N2B27 appears to be
a transitional medium in which cells can adopt a primary neural fate
(Turner et al., 2014c). Furthermore, in the aggregates, as in the
embryo, Act initiates endoderm development (Sox17) and BMP
initiates mesoderm (Tbx6) development.
Cell movement in polarised aggregates
When Sox17::GFP aggregates were cultured in Act/Chi for 120 h, we
observed cells that were being extruded from a region adjacent to the
primary focal point of reporter expression (Fig. 8A; supplementary
material Movie 6). As time progressed, the frequency of this event
Fig. 5. Emergence of polarised gene expression in aggregates. (A) Stills from live cell imaging of mESCs in suspension in N2B27 showing aggregate
formation within the first 8 h. (B,C) Emergence and progression of Sox17::GFP following addition of secondary Act/Chi medium. (D) Early stages of Sox17::GFP
and Bra expression. Initially, Bra and Sox17::GFP are heterogeneously expressed before polarisation occurs. (E) Live imaging of Bra::GFP mESCs following
addition of Act/Chi. Every cell initially expresses Bra before downregulation in regions that will not form the elongation. A, B, C and E are from supplementary
material Movies 1-4, respectively. Data are representative from at least two experiments.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
increased and many more cells were seen to emerge from the same
point. Close observation revealed two types of movement: Sox17-
expressing cells appeared to move inside the aggregate, close to its
wall, whereas others, not expressing Sox17, moved towards the
outside (supplementary material Movie 6). Cell movements can
also be observed after exposure to Chi and in other cell lines such as
Bra::GFP (Fig. 7D and Fig. 8B; supplementary material Movie 7) and
TBX6::EYFP (Fig. 7B and Fig. 8C; supplementary material Movies 8
and 9). Movies show that the extruded cells stem from the region of
Bra::GFPexpression(Fig. 7B) and that they express Tbx6 (Fig. 8C,C;
supplementary material Movies 8 and 9), suggesting that they are
mesodermal. The extruded cells produce floating trails in the medium
or attach to the main body of the aggregate (Fig. 7B and Fig. 8C)
and, when they attach, they maintain expression of Tbx6, suggesting
that continuing Tbx6 expression requires some substrate that can only
be provided by other cells. It appears as if the cells prefer to attach to
the anteriorsection of the aggregate, suggesting that there are
differences between the two regions.
We do not observe filopodia or lamellipodia in the cells leaving
the aggregate but observe the emergence of blebs (Fig. 8D,D;
supplementary material Movie 10), which have been associated
with cell movements during gastrulation in zebrafish (Paluch and
Raz, 2013).
We have shown that under defined culture conditions aggregates of
mESCs undergo processes that resemble the collective behaviours of
cells in early mouse embryos: symmetry breaking, axial organisation,
germ layer specification, gastrulation and axis elongation. This is
surprising in light of the fact that EBs are commonly used in
differentiation experiments and yet, with two exceptions, there have
been no reports of similar behaviours. One of the exceptions is an
account of elongation and polarised gene expression in aggregates of
P19 EC cells exposed to serum (Marikawa et al., 2009). The second is
a report of weak polarisation of Bra expression in EBs of mESCs
exposed to agonists of Wnt signalling (ten Berge et al., 2008), which
resembles the early stages of what we report here. It is possible that
similar partial polarisation events occur at low frequency in EBs but
are generally overlooked. We believe that the consistency and
magnitude of the behaviour that we observe in our aggregates are
founded, principally, in two aspects of our experimental protocol: the
sequence of culture conditions that we use and the initial number of
cells in the aggregate.
In our adherent cultures we had noticed that exposure of
differentiating ESCs to N2B27 for 2 days results in a homogeneous
response to external signals (Turner et al., 2014c). We reasoned that
this treatment allows all cells in the culture to enter a state resembling
the postimplantation epiblast, where they become competent to
respondto signals (Sterneckert et al., 2010; Turner et al., 2014c), and it
is for this reason that we used this protocol as the basis for our
experiments. The second element in our protocol that differs from
standard procedures concerns the number of cells in the initial
aggregate, which appears to be a critical variable in the experiments.
Aggregates above or below 300±100 cells (average aggregate
diameter of 100 µm) will either not develop or do so into
amorphous masses of cells, characteristic of the EB protocols in
current use. This size of 300±100 cells is reminiscent of that of early
postimplantation embryos and perhaps defines an optimal length scale
for a unique outcome of the biochemical reactions that mediate
symmetry breaking and polarisation. Experiments searching for
conditions that mimic the emergence of the postimplantation epiblast
from mESCs also find that the number of starting cells is a critical
parameter of the process (Bedzhov and Zernicka-Goetz, 2014). A
surprising conclusion from these observations that will need to be
pursued is that early patterning events do not scale easily and this is in
agreement with recent observations on the emergence of germ layers
in micropatterns of human ESCs (Warmflash et al., 2014).
Our observations raise many questions about symmetry breaking
in early embryos, the ability of cell ensembles to respond to signals
and the different behaviour of mESCs in adherent and three-
dimensional cultures. For reasons of space, here we shall focus on
two specific issues concerning how the system that we have
Fig. 6. Qualitative summary of the
tissue-specific response of
aggregates to different signalling
environments. Reporter lines were
used for neural (Sox1::GFP), endoderm
(Sox17::GFP), mesendoderm
(Bra::GFP) and paraxial mesoderm
(TBX6::EYFP) as well as for Wnt
signalling (TCF/LEF::mCherry).
Aggregates from the different lines were
treated as indicated and the results in
terms of expression levels within the
population as a whole are summarised
by the colour intensity within each
square. Representative examples are
shown in Fig. 7. ND, not determined.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
described informs our understanding of the mechanisms underlying
early mammalian development; other issues will be discussed
elsewhere (see Turner et al., 2014a).
Symmetry breaking and axis specification
Our results reveal that, under appropriate culture conditions,
aggregates of mESCs have an intrinsic ability for symmetry
breaking and stable polarisation of gene expression. This pattern
resembles events in the embryo at E6.0 with some, perhaps
informative, differences.
In the embryo, the initial localisation of the PS canbe identified as a
focus of Bra expression in the proximal posterior region of the embryo
(Wilkinson et al., 1990) and its specification follows a sequence of
events associated with the localisation of ligands for BMP, Nodal and
Wnt signalling to the same region (reviewed by Arnold and
Robertson, 2009; Pfister et al., 2007; Rossant and Tam, 2009; Tam
and Loebel, 2007). This process requires first the specification and
localisation of the anterior visceral endoderm (AVE) to the
prospective anterior region of the conceptus, where it acts as a
source of antagonists of Wnt, BMP and Nodal signalling (Arkell and
Tam, 2012). It is thought that the action of the AVE positions or
restricts the PS to the opposite end of the epiblast (Perea-Gomez et al.,
2004; Perea-Gómez, 2014; Rivera-Pérez and Magnuson, 2005). Our
results raise questions about the actual role of the AVE, since they
show that a stable axis, as reflected by localised expression of Bra,
Sox17 and FoxA2, can beinitiated without external influences. In our
experiments the signals are ubiquitous and so the symmetry-breaking
event must be intrinsic to the aggregates, raising the possibility that a
similar spontaneous event takes place in the embryo. This conclusion
is at odds with the large body of experimentalevidence suggesting that
the anteroposterior axis requires a sequence of interactions between
extraembryonic and embryonic tissues (Rossant and Tam, 2009).
One way to reconcile our observations with those of the genetic
analysis of early development would be to entertain the possibility
that the function of the AVE is not to break the symmetry of the
embryo but rather to ensure that an event that can happen
spontaneously has a reproducible outcome, i.e. the AVE ensures the
maintenance of a regionprimed for anterior neural development at the
opposite pole to that of the PS and, more importantly, endows this
region with an anterior neural fate potential (Albazerchi and Stern,
Fig. 7. Gene and tissue-specific response of aggregates to different signalling environments. (A-E) Representative examples from the summary in Fig. 6.
Sox1::GFP (A, n=16, 14, 16, 24, 16 per labelled condition, respectively), TBX6::EYFP (B, n=11 or 10 per labelle d condition, respectively), TCF/LEF::mCherry (C,
n=4 per condition), Bra::GFP (D, n=9, 11 or 11 per labelled condition, respectively) and Sox17::GFP (E, n>3) mESCs were treated as indicated. The boxed region
in the Chi pulse image (A) is enlarged section to the right to show a region within the tip of the aggregate that is negative for Sox1::GFP. Compare with the
expression pattern of Bra and Sox17::GFPand Wnt activity from Figs 3 and 4. Note that the expression of the reporters is associated with specific morphogenetic
events; two examples of each are given. The colour coding of each treatment label corresponds to that used in Fig. 6. (F,G) Aggregates of Sox1::GFP mESCs
following a pulse of Chi on day 3 were stained for GFP (Sox1), Sox2 and Sox17. The boxed region is magnified to the right and also shows orthogonal views.
Sox1 at the tip of the aggregate is co-expressed with high levels of Sox2, whereas Sox2 levels decrease in regions high for the endoderm marker Sox17.
(G) Section through a 12-somite stage embryo stained for Sox2 and with DAPI. Note how Sox2 is expressed in the neural tissue (n) and in the gut ( g), similar to the
expression pattern seen in F.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
2007; Bertocchini and Stern, 2002). This suggests that it is possible to
uncouple symmetry breaking and anterior neural specification. The
latter requires suppression of Nodal, BMP and Wnt signalling
(Andoniadou and Martinez-Barbera, 2013; Stern, 2005) and this, in
terms of patterning anterior and posterior domains, can only be
achieved by a localised source which, in the embryo, is provided by
the AVE. Consistent with this, aggregates maintained in N2B27, or in
N2B27 in the presence of BMP inhibitors, for the most part remain
symmetrical and express a neural fate, probably mimicking the
specification of anterior neural fate under these conditions (Eiraku
et al., 2011). In the future it will be interesting to provide localised
inhibition of BMP and Nodal in aggregates exposed to Act, BMP and
Chi to try to obtain both anterior neural and mesendodermal fates in
the same group of cells. However, this only provides a partial
explanation, as embryos lacking a VE develop symmetrically (Perea-
Gómez et al., 1999; Waldrip et al., 1998), suggesting that our
experimental conditions might be generating a situation that does not
occur in the embryo. One explanation is that, in addition to
maintaining a proneuroectodermal region, a key function of the
AVE is to biasa spontaneous symmetry-breaking event that is intrinsic
to the epiblast. Our experiments might be creating these imbalances
by an excess of specific signals in the medium, a hypothesis that will
be of interest to test in further experiments.
The events that lead to symmetry breaking remain out of the scope
of this work; however, our observations provide some hints as to
their constraints. There is clearly a defined length scale to the
process, as only aggregates of a certain size undergo the unique
event. The symmetry-breaking event must contain an activating and
an inhibitory component that are linked, i.e. once the process has
started it can inhibit itself within a certain length scale to make the
process unique (Meinhardt, 2012). In support of this suggestion,
there is evidence for the potential to generate multiple PSs or axes
from a single embryo in mouse (Merrill et al., 2004; Perea-Gomez
et al., 2002) and chicken (Bertocchini and Stern, 2002; Bertocchini
et al., 2004) but only one emerges in the embryo. The possibility that
limitation of signalling range plays a role can be gauged in our
experiments, which show that persistent signalling can give rise to
multiple patterning foci.
Finally, and in the context of symmetry breaking, our experimental
system underpins a well-known connection between Wnt signalling
and axial elongation (Martin and Kimelman, 2009; Petersen and
Reddien, 2009) and provides an opportunity to probe into its
Gastrulation in culture?
In the mouse embryo, one of the consequences of the formation of
an anteroposterior axis is the localisation of the start of gastrulation
to the posterior proximal region, a process that will generate the
primordia for the endoderm and the mesoderm as well as reveal the
axial organisation of the embryo (Nowotschin and Hadjantonakis,
2010; Ramkumar and Anderson, 2011; Tam and Gad, 2004). The
start of this process is manifest in the localisation of the expression
of BMP, Nodal and Wnt3 to this region and, more significantly,
of Bra to the emergent PS (Herrmann, 1991; Pfister et al., 2007).
A central feature of this structure is an epithelial-to-mesenchymal
transition (EMT), which, under the control of specific signals, leads
to a germ layer-specific behaviour: in the endoderm, cells re-
epithelialise (Burtscher and Lickert, 2009; Kwon et al., 2008; Lewis
and Tam, 2006), whereas in the mesoderm they become highly
mesenchymal (Nakaya and Sheng, 2008). As a result of these
movements the three germ layers are distributed relative to each
other. We observe related behaviours when our aggregates are
exposed to Act, BMP and, in particular, Wnt/β-catenin. On its own,
Act treatment elicits the expression of the endodermal marker Sox17
Fig. 8. Gastrulation-like movements in aggregates. (A-C) Cell extrusion and intra-aggregate movement in (A,A) Sox17::GFP, (B,B) Bra::GFP and (C-C)
TBX6::EYFP. Aggregates were treated as indicated. Images correspond to the indicated time points from the associated movies (see supplementary material
Movies 5-8). Arrows in Cindicate a single extrudedcell from a second aggregate between102 and 103 h (see supplementary material Movie 8); curvedarrows in A
indicate the directionof movement of the cells leaving the aggregate. (D,D) Live imaging of GPI-GFP mESCs following treatment with Act/Chi, showing membrane
blebbing at the elongated region of the aggregate. Two different zplanes are shown. Arrows indicate the blebbing region (see supplementary material Movie 10).
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
in a group of cells that express E-cadherin and form coherent
epithelial groups on the outer edges of the aggregate, as they do in
the embryo. By contrast, BMP and especially Wnt/β-catenin favour
the extrusion of cells from a domain that expresses Bra and TBX6, a
gene associated with paraxial mesoderm (Chapman et al., 1996).
These observations suggest that our culture system recapitulates
some of the features of gastrulation, even though the behaviour of
the mesodermal-like cells is the reverse of that in the embryo, where
cells move inward rather than outward. It is likely that this
topological switch reflects the architecture of the aggregates.
At the end of gastrulation, amniote embryos undergo a process of
axial extension that generates the spinal cord and the paraxial
mesoderm from a population of stem-like cells located in the distal
end of the embryo (Kondoh and Takemoto, 2012; Wilson et al.,
2009). This process relies on Wnt signalling and a localised source
of Bra expression at the tip of the extension (reviewed by Wilson
et al., 2009). We observe that transient exposure of the aggregates to
Chi is able to elicit this structure (see also Turner et al., 2014a).
Furthermore, in some of these aggregates we observe endoderm
embedded in this tissue near its surface, a situation that, once again,
mimics the embryo (Fig. 9).
We note that, independentof the geometry, the sequence and timing
of the events that we observe are reproducible and can be related
to events in the embryo (Fig. 9). Thus, day 3 seems to be similar to
E5.5-6.5 as it is at this time that we observe the cell extrusion and
intrusion, with the elongation starting sometime during day 4, which
would thus be homologous to E7.5 in the embryo. Altogether, these
observations further emphasize the similarity between the processes
that we have uncovered here and the events in the embryo. The
movements are related to those of cells in gastrulating embryos and for
this reason we term these aggregates gastruloids.
Final consideration and prospective uses of gastruloids
A somewhat surprising aspect of the gastruloids is that they allow
the uncoupling of processes that in the embryo are tightly linked, such
as specification of the anteroposterior axis and anterior neural
development or endoderm specification and axial elongation. This
could be construed to mean that, on the whole, the aggregates are not
reflecting the situation in vivo. This contention could be underlined by
the differential relative topology of the movements of endodermal and
mesodermal cells in the aggregates. However, we believe that
gastruloids reflect embryonic events and that they do so in the same
manner as the emergence of eye cups (Nakano et al., 2012) and
cerebroids (Lancasteret al., 2013) and offer an additional experimental
system with which to explore the mechanisms of self-organisation
processes in cellular ensembles. Furthermore, we believe that the
deconstruction of developmental events achieved here and its
comparison with the events in embryos will allow a detailed
mechanistic analysis of processes that, like gastrulation and axial
extension, have significant mechanical and geometrical inputs that
make them difficult to study in vivo. Naturally, the conclusions from
this work will ultimately have to be tested in embryos, but this should
not deter their use for analytical purposes. We have begun to do this by
probinginto the mechanisms of axial elongation (Turner et al., 2014a).
Tissue culture, FACS, immunofluorescence and confocal
Routine tissue culture, FACS analysis, immunofluorescence and confocal
microscopy were performed as described previously (Faunes et al., 2013;
Kalmar et al., 2009; Turner et al., 2014a,b,c). Primary antibodies used for
immunofluorescence were: goat anti-Bra (Santa Cruz Biotechnology, sc-
17743; 1:200), rat anti-E-Cadherin (Takara, M108; 1:200), goat anti-Sox17
(R&D Systems, AF1924; 1:500) and goat anti-FoxA2 (Santa Cruz
Biotechnology, sc-6554; 1:500). Alexa-conjugated secondary antibodies
were from Invitrogen and were used at 1:500 dilution. Hoechst 3342
(Invitrogen) stained the nuclei and was used at 1:1000 dilution.
Cell lines
The cell lines used are E14-Tg2A, Sox1::GFP (Ying et al., 2003), Sox17::
GFP (Niakan et al., 2010), Bra::GFP (Fehling et al., 2003), the Wnt/
β-catenin transcriptional reporter TCF/LEF::mCherry (Faunes et al., 2013;
Ferrer-Vaquer et al., 2010), TBX6::EYFP (this is a knock-in into the Tbx6
locus; A.-K.H. and S.N.) and CAG::GPI-GFP (referred to hereafter as
GPI-GFP) (Rhee et al., 2006).
Aggregate culture and imaging
A detailed protocol for the growth of the aggregates, with trouble-shooting,
is provided elsewhere (Baillie-Johnson et al., 2014). Images in Fig. 1 were
generated by manipulating the brightness and contrast of pictures of the
aggregates in addition to edge detection; the outlines were enhanced
manually through tracing. The original unprocessed images of the
aggregates are provided in supplementary material Fig. S1G,H. N2B27
(NDiff) was sourced from StemCells (USA) and tissue culture slides for
monolayer imaging were obtained from Ibidi (Germany). All experimental
conditions were repeated at least twice.
We thank K. Niakan for the Sox17::GFP cell line, E. Davies for sharing data and
J. Brickman, J. Briscoe, S. Muñoz-Descalzo, J. Nichols, A. Perea-Gomez,
C. Schro
̈eter, T. Rodriguez and C. Stern for discussions and constructive criticisms.
Competing interests
The authors declare no competing financial interests.
Author contributions
A.M.A. conceived the project, and S.C.B., P.B.J., T.B., D.A.T., S.N. and A.-K.H.
carried out the experiments. A.M.A. and D.T. wrote the paper.
This work is fun ded by a European Research Council (E RC) Advanced Investigator
Award to A.M.A. (D.A.T. and T.B.) wit h the contribution of a Proje ct Grant from the
Wellcome Trust to A.M.A., an Engineering and Physical Sc iences Research
Fig. 9. Comparison of events in embryos and aggregates. (Top) Timeline of
embryogenesis, with the illustrated stages acting as landmarks. (Bottom) A
representation of the behaviour of aggregates exposed to different signalling
environments over the indicated periods of differentiation, as inferred from our
experiments labelled here as a, b and c. We propose that the third day of
differentiation of the aggregates is equivalent to the E5.5-6.0 postimplantation
epiblast. DD, day of aggregate differentiation. The dark blue shading indicates
anterior Sox1 expression.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
Council (EPSRC) Studentship to P.B.-J. and E rasmus, Stichting dr. Hendrik
Mullers Vaderlandsch Fonds and Fundatie van de Vrijvrouwe van Renswoude
te s-Gravenhage to S.C.B. A.-K.H. was funded by a grant from the Nation al
Institutes of Health (NIH) [RO1-HD052115] and S.N. by a Muscular Dystrophy
Association Development Grant [186552]. Deposited in PMC for immediate
Supplementary material
Supplementary material available online at
Albazerchi, A. and Stern, C. D. (2007). A role for the hypoblast (AVE) in the
initiation of neural induction, independent of its ability to position the primitive
streak. Dev. Biol. 301, 489-503.
Andoniadou, C. L. and Martinez-Barbera, J. P. (2013). Developmental
mechanisms directing early anterior forebrain specification in vertebrates. Cell.
Mol. Life Sci. 70, 3739-3752.
Arkell, R. M. and Tam, P. P. L. (2012). Initiating head development in mouse
embryos: integrating signalling and transcriptional activity. Open Biol. 2, 120030.
Arnold, S. J. and Robertson, E. J. (2009). Making a commitment: cell lineage
allocation and axis patterning in the early mouse embryo. Nat. Rev. Mol. Cell Biol.
10, 91-103.
Baillie-Johnson, P., van den Brink, S. C., Balayo, T., Turner, D. A. and Martinez
Arias, A. (2014). Generation of aggregates of mouseES cells that show symmetry
breaking, polarisation and emergent collective behaviour in vitro. BioRxiv. doi:
Bedzhov, I. and Zernicka-Goetz, M. (2014). Self-organizing properties of mouse
pluripotent cells initiate morphogenesis upon implantation. Cell 156, 1032-1044.
Bertocchini, F. and Stern, C. D. (2002). The hypoblast of the chick embryo
positions the primitive streak by antagonizing nodal signaling. Dev. Cell 3,
Bertocchini, F., Skromne, I., Wolpert, L. and Stern, C. D. (2004). Determination of
embryonic polarity in a regulative system: evidence for endogenous inhibitors
acting sequentially during primitive streak formation in the chick embryo.
Development 131, 3381-3390.
Burtscher, I. and Lickert, H. (2009). Foxa2 regulates polarity and epithelialization
in the endoderm germ layer of the mouse embryo. Development 136, 1029-1038.
Chapman, D. L., Agulnik, I., Hancock, S., Silver, L. M. and Papaioannou, V. E.
(1996). Tbx6, a mouse T-Box gene implicated in paraxial mesoderm formation at
gastrulation. Dev. Biol. 180, 534-542.
Desbaillets, I., Ziegler, U., Groscurth, P. and Gassmann, M. (2000). Embryoid
bodies: an in vitro model of mouse embryogenesis. Exp. Physiol. 85, 645-651.
Eiraku, M., Takata, N., Ishibashi, H., Kawada, M., Sakakura, E., Okuda, S.,
Sekiguchi, K., Adachi, T. and Sasai, Y. (2011). Self-organizing optic-cup
morphogenesis in three-dimensional culture. Nature 472, 51-56.
Faunes, F., Hayward, P., Descalzo, S. M., Chatterjee, S. S., Balayo, T., Trott, J.,
Christoforou, A., Ferrer-Vaquer, A., Hadjantonakis, A.-K., Dasgupta, R. et al.
(2013). A membrane-associated β-catenin/Oct4 complex correlates with ground-
state pluripotency in mouse embryonic stem cells. Development 140, 1171-1183.
Fehling, H. J., Lacaud, G., Kubo, A., Kennedy, M., Robertson, S., Keller, G. and
Kouskoff, V. (2003). Tracking mesoderm induction and its specification to the
hemangioblast during embryonic stem cell differentiation. Development 130,
Ferrer-Vaquer, A., Piliszek, A., Tian, G., Aho, R. J., Dufort, D. and
Hadjantonakis, A.-K. (2010). A sensitive and bright single-cell resolution live
imaging reporter of Wnt/ß-catenin signaling in the mouse. BMC Dev. Biol. 10, 121.
Gadue, P., Huber, T. L., Paddison, P. J. and Keller, G. M. (2006). Wnt and TGF-
beta signaling are required for the induction of an in vitro model of primitive streak
formation using embryonic stem cells. Proc. Natl. Acad. Sci. USA 103,
Green, J. B. A., Dominguez, I. and Davidson, L. A. (2004). Self-organization of
vertebrate mesoderm based on simple boundary conditions. Dev. Dyn. 231,
Herrmann, B. G. (1991). Expression pattern of the Brachyury gene in whole-mount
TWis/TWis mutant embryos. Development 113, 913-917.
Holtfreter, J. (1933). Die totale Exogastrulation, eine Selbstablo
̈sung des
Ektoderms vom Entomesoderm. Wilhelm Roux Arch. Entwicklungsmech. Org.
129, 669-793.
̈pfl, G., Gassmann, M. and Desbaillets, I. (2004). Differentiating embryonic
stem cells into embryoid bodies. In Germ Cell Protocols, Volume 2: Molecular
Embryo Analysis, Live Imaging, Transgenesis, and Cloning (Methods in
Molecular Biology) Vol. 254 (ed. H. Schatten), pp. 79-98. New York: Humana
Horstadius, S. (1939). The mechanisms of sea urchin development, studied by
operative methods. Biol. Rev. 14, 132-179.
Huelsken, J., Vogel, R., Brinkmann, V., Erdmann, B., Birchmeier, C. and
Birchmeier, W. (2000). Requirement for beta-catenin in anterior-posterior axis
formation in mice. J. Cell Biol. 148, 567-578.
Ishihara, K., Tonegawa, Y., Suyemitsu, T. and Kubo, H. (1982). The blastocoelic
fluid of sea urchin embryo induces exogastrulation. J. Exp. Zool. 220, 227-233.
Kalmar, T., Lim, C., Hayward, P., Muñoz-Descalzo, S., Nichols, J., Garcia-
Ojalvo, J. and Martinez Arias, A. (2009). Regulated fluctuations in nanog
expression mediate cell fate decisions in embryonic stem cells. PLoS Biol. 7,
Kanai-Azuma, M., Kanai, Y., Gad, J. M., Tajima, Y., Taya, C., Kurohmaru, M.,
Sanai, Y., Yonekawa, H., Yazaki, K., Tam, P. P. L. et al. (2002). Depletion of
definitive gut endoderm in Sox17-null mutant mice. Development 129,
Keller, R. and Danilchik, M. (1988). Regional expression, pattern and timing of
convergence and extension during gastrulation of Xenopus laevis. Development
103, 193-209.
Keller, R., Davidson, L. A. and Shook, D. R. (2003). How we are shaped: the
biomechanics of gastrulation. Differentiation 71, 171-205.
Kondoh, H. and Takemoto, T. (2012). Axial stem cells deriving both posterior
neural and mesodermal tissues during gastrulation. Curr. Opin. Genet. Dev. 22,
Kouskoff, V., Lacaud, G., Schwantz, S., Fehling, H. J. and Keller, G. (2005).
Sequential development of hematopoietic and cardiac mesoderm during
embryonic stem cell differentiation. Proc. Natl. Acad. Sci. USA 102, 13170-13175.
Kubo, A., Shinozaki, K., Shannon, J. M., Kouskoff, V., Kennedy, M., Woo, S.,
Fehling, H. J. and Keller, G. (2004). Development of definitive endoderm from
embryonic stem cells in culture. Development 131, 1651-1662.
Kwon, G. S., Viotti, M. and Hadjantonakis, A.-K. (2008). The endoderm of the
mouse embryo arises by dynamic widespread intercalation of embryonic and
extraembryonic lineages. Dev. Cell 15, 509-520.
Lancaster, M. A., Renner, M., Martin, C.-A., Wenzel, D., Bicknell, L. S., Hurles,
M. E., Homfray, T., Penninger, J. M., Jackson, A. P. and Knoblich, J. A. (2013).
Cerebral organoids model human brain development and microcephaly. Nature
501, 373-379.
Lewis, S. L. and Tam, P. P. L. (2006). Definitive endoderm of the mouse embryo:
formation, cell fates, and morphogenetic function. Dev. Dyn. 235, 2315-2329.
Marikawa, Y., Tamashiro, D. A. A., Fujita, T. C. and Alarcón, V. B. (2009).
Aggregated P19 mouse embryonal carcinoma cells as a simple in vitro model to
study the molecular regulations of mesoderm formation and axial elongation
morphogenesis. Genesis 47, 93-106.
Martin, B. L. and Kimelman, D. (2009). Wnt signaling and the evolution of
embryonic posterior development. Curr. Biol. 19, R215-R219.
Medina, A., Wendler, S. R. and Steinbeisser, H. (1997). Cortical rotation is
required for the correct spatial expression of nr3, sia and gsc in Xenopus embryos.
Int. J. Dev. Biol. 41, 741-745.
Meinhardt, H. (2012). Turings theory of morphogenesis of 1952 and the
subsequent discovery of the crucial role of local self-enhancement and long-
range inhibition. Interface Focus 2, 407-416.
Merrill, B. J., Pasolli, H. A., Polak, L., Rendl, M., Garcı
a, M. J., Anderson,
K. V. and Fuchs, E. (2004). Tcf3: a transcriptional regulator of axis induction in the
early embryo. Development 131, 263-274.
Monaghan, A. P., Kaestner, K. H., Grau, E. and Schu
̈tz, G. (1993).
Postimplantation expression patterns indicate a role for the mouse forkhead/
HNF-3 alpha, beta and gamma genes in determination of the definitive endoderm,
chordamesoderm and neuroectoderm. Development 119, 567-578.
Morkel, M., Huelsken, J., Wakamiya, M., Ding, J., van de Wetering, M., Clevers,
H., Taketo, M. M., Behringer, R. R., Shen, M. M. and Birchmeier, W. (2003).
Beta-catenin regulates Cripto- and Wnt3-dependent gene expression programsin
mouse axis and mesoderm formation. Development 130, 6283-6294.
Nakano, T., Ando, S., Takata, N., Kawada, M., Muguruma, K., Sekiguchi, K.,
Saito, K., Yonemura, S., Eiraku, M. and Sasai, Y. (2012). Self-formation of optic
cups and storable stratified neural retina from human ESCs. Cell Stem Cell 10,
Nakaya, Y. and Sheng, G. (2008). Epithelial to mesenchymal transition during
gastrulation: an embryological view. Dev. Growth Diff. 50, 755-766.
Niakan, K. K., Ji, H., Maehr, R., Vokes, S. A., Rodolfa, K. T., Sherwood, R. I.,
Yamaki, M., Dimos, J. T., Chen, A. E., Melton, D. A. et al. (2010). Sox17
promotes differentiation in mouse embryonic stem cells by directly regulating
extraembryonic gene expression and indirectly antagonizing self-renewal. Genes
Dev. 24, 312-326.
Ninomiya, H., Elinson, R. P. and Winklbauer, R. (2004). Antero-posterior tissue
polarity links mesoderm convergent extension to axial patterning. Nature 430,
Norris, D. P., Brennan, J., Bikoff, E. K. and Robertson, E. J. (2002). The Foxh1-
dependent autoregulatory enhancer controls the level of Nodal signals in the
mouse embryo. Development 129, 3455-3468.
Nowotschin, S. and Hadjantonakis, A.-K. (2010). Cellular dynamics in the early
mouse embryo: from axis formation to gastrulation. Curr. Opin. Genet. Dev. 20,
Paluch, E. K. and Raz, E. (2013). The role and regulation of blebs in cell migration.
Curr. Opin. Cell Biol. 25, 582-590.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
Perea-Gómez, A. (2014). Formation of the anterior-posterior axis in mammals. In
Principles of Developmental Genetics (ed. L. Wolpert and C. Tickle). Amsterdam:
Elsevier (in press).
Perea-Gómez, A., Shawlot, W., Sasaki, H., Behringer, R. R. and Ang, S. (1999).
HNF3beta and Lim1 interact in the visceral endoderm to regulate primitive streak
formation and anterior-posterior polarity in the mouse embryo. Development 126,
Perea-Gomez, A., Vella, F. D. J., Shawlot, W., Oulad-Abdelghani, M., Chazaud,
C., Meno, C., Pfister, V., Chen, L., Robertson, E., Hamada, H. et al. (2002).
Nodal antagonists in the anterior visceral endoderm prevent the formation of
multiple primitive streaks. Dev. Cell 3, 745-756.
Perea-Gomez, A., Camus, A., Moreau, A., Grieve, K., Moneron, G., Dubois, A.,
Cibert, C. and Collignon, J. (2004). Initiation of gastrulation in the mouse embryo
is preceded by an apparent shift in the orientation of the anterior-posterior axis.
Curr. Biol. 14, 197-207.
Petersen, C. P. and Reddien, P. W. (2009). Wnt signaling and the polarity of the
primary body axis. Cell 139, 1056-1068.
Pfister, S., Steiner, K. A. and Tam, P. P. L. (2007). Gene expression pattern and
progression of embryogenesis in the immediate post-implantation period of
mouse development. Gene Expr. Patterns 7, 558-573.
Ramkumar, N. and Anderson, K. V. (2011). SnapShot: mouse primitive streak. Cell
146, 488-488.e2.
Rhee, J. M., Pirity, M. K., Lackan, C. S., Long, J. Z., Kondoh, G., Takeda, J. and
Hadjantonakis, A.-K. (2006). In vivo imaging and differential localization of lipid-
modified GFP-variant fusions in embryonic stem cells and mice. Genesis 44,
Rivera-Pérez, J. A. and Magnuson, T. (2005). Primitive streak formation in mice is
preceded by localized activation of Brachyury and Wnt3. Dev. Biol. 288, 363-371.
Rossant, J. and Tam, P. P. L. (2009). Blastocyst lineage formation, early embryonic
asymmetries and axis patterning in the mouse. Development 136, 701-713.
Sasai, Y. (2013). Next-generation regenerative medicine: organogenesis from stem
cells in 3D culture. Cell Stem Cell 12, 520-530.
Sasai, Y., Eiraku, M. and Suga, H. (2012). In vitro organogenesis in three
dimensions: self-organising stem cells. Development 139, 4111-4121.
Sasaki, H. and Hogan, B. L. (1993). Differential expression of multiple fork head
related genes during gastrulation and axial pattern formation in the mouse
embryo. Development 118, 47-59.
Solnica-Krezel, L. and Sepich, D. S. (2012). Gastrulation: making and shaping
germ layers. Annu. Rev. Cell Dev. Biol. 28, 687-717.
Stern, C. D. (2005). Neural induction: old problem, new findin gs, yet more questions.
Development 132, 2007-2021.
Sterneckert, J., Stehling, M., Bernemann, C., Araúzo-Bravo, M. J., Greber, B.,
Gentile, L., Ortmeier, C., Sinn, M., Wu, G., Ruau, D. et al. (2010). Neural
induction intermediates exhibit distinct roles of Fgf signaling. Stem Cells 28,
Symes, K. and Smith, J. C. (1987). Gastrulation movements provide an early
marker of mesoderm induction in Xenopus laevis. Development 101, 339-349.
Tam, P. P. L. and Gad, J. M. (2004). Gastrulation in the mouse embryo. In
Gastrulation:From Cells to Embryo (ed. C. D. Stern), p. 731. New York: Cold
Spring Harbor Laboratory Press.
Tam, P. P. L. and Loebel, D. A. F. (2007). Gene function in mouse embryogenesis:
get set for gastrulation. Nat. Rev. Genet. 8, 368-381.
ten Berge, D., Koole, W., Fuerer, C., Fish, M., Eroglu, E. and Nusse, R. (2008).
Wnt signaling mediates self-organization and axis formation in embryoid bodies.
Cell Stem Cell 3, 508-518.
Tropepe, V., Hitoshi, S., Sirard, C., Mak, T. W., Rossant, J. and van der Kooy, D.
(2001). Direct neural fate specification from embryonic stem cells: a primitive
mammalian neural stem cell stage acquired througha default mechanism. Neuron
30, 65-78.
Turner, D. A., Hayward, P., Baillie-Johnson, P., Rué, P., Broome, R., Faunes, F.
and Martinez Arias, A. (2014a). Wnt/ß-catenin and FGF signaling direct the
specification and maintenance of a neuromesodermal axial progenitor in
ensembles of mouse embryonic stem cells. Development 141, 4243-4253.
Turner, D. A., Rué, P., Mackenzie, J. P., Davies, E. and Martinez Arias, A.
(2014b). Brachyury cooperates with Wnt/ß-Catenin signalling to elicit primitive-
streak-like behaviour in differentiating mouse embryonic stem cells. BMC Biol. 12,
Turner, D. A., Trott, J., Hayward, P., Rué, P. and Martinez Arias, A. (2014c). An
interplay between extracellular signalling and the dynamics of the exit from
pluripotency drives cell fate decisions in mouse ES cells. Biol. Open 3, 614-626.
Waldrip, W. R., Bikoff, E. K., Hoodless, P. A., Wrana, J. L. and Robertson, E. J.
(1998). Smad2 signaling in extraembryonic tissues determines anterior-posterior
polarity of the early mouse embryo. Cell 92, 797-808.
Warmflash, A., Sorre, B., Etoc, F., Siggia, E. D. and Brivanlou, A. H. (2014).
A method to recapitulate early embryonic spatial patterning in human embryonic
stem cells. Nat. Chem. Biol. 11, 847-854.
Watanabe, K., Kamiya, D., Nishiyama, A., Katayama, T., Nozaki, S., Kawasaki,
H., Watanabe, Y., Mizuseki, K. and Sasai, Y. (2005). Directed differentiation of
telencephalic precursors from embryonic stem cells. Nat. Neurosci. 8, 288-296.
Wataya, T., Ando, S., Muguruma, K., Ikeda, H., Watanabe, K., Eiraku, M.,
Kawada, M., Takahashi, J., Hashimoto, N. and Sasai, Y. (2008). Minimization of
exogenous signals in ES cell culture induces rostral hypothalamic differentiation.
Proc. Natl. Acad. Sci. USA 105, 11796-11801.
Wilkinson, D. G., Bhatt, S. and Herrmann, B. G. (1990). Expression pattern of the
mouse T gene and its role in mesoderm formation. Nature 343, 657-659.
Wilson, V., Olivera-Martinez, I. and Storey, K. G. (2009). Stem cells, signals and
vertebrate body axis extension. Development 136, 1591-1604.
Wood, H. B. and Episkopou, V. (1999). Comparative expression of the mouse
Sox1, Sox2 and Sox3 genes from pre-gastrulation to early somite stages. Mech.
Dev. 86, 197-201.
Ying, Q.-L., Stavridis, M., Griffiths, D., Li, M. and Smith, A. (2003). Conversion of
embryonic stem cells into neuroectodermal precursors in adherent monoculture.
Nat. Biotechnol. 21, 183-186.
RESEARCH ARTICLE Development (2014) 141, 4231-4242 doi:10.1242/dev.113001
Supplementary Figure Legends
Development | Supplementary Material
Figure S1. Elongation of aggregates from mouse ESCs. (A) Schematic representation of differentiation time-
course; HD: aggregates forming in hanging-drops; Act/Chi: Medium condition of 100 ng/ml Act + 3 μM Chi.
(B) Hanging drops of ~200 mESCs per 40 μl drop. (C) mESCs in HD phase grown in Serum with LIF or (D)
N2B27 only. (E) Proportion of cells displaying elongated morphologies following HD phase of N2B27. Each
point represents the median from two independent experiments and (F-Fʹʺ) Representative images of elongating
cells from (E) after 1 (F), 3 (Fʹ), 6 (Fʺ) and 11 (Fʹʺ) days Act/Chi. Arrows in (C) and (D) indicate individual
aggregates. (G) Original, unprocessed images used to generate the cartoon renderings of the typical aggregate
morphologies in Fig. 1B. Maximum elongation was observed following pulsed treatment within the day 2-3
time-frame. (H) Original, unprocessed images of aggregates used to generate the examples of morphologies
used in Fig. 1C. See materials and methods section for explanation of rendering process. Images not to scale.
Scale bar: 100 μm.
Development | Supplementary Material
Development | Supplementary Material
Development | Supplementary Material
Development | Supplementary Material
Development | Supplementary Material
Figure S2. A comparative analysis of exposure time, signalling and initial cell density on aggregate formation.
(A-D) Aggregates formed from 200 (A), 400 (B), 800 (C) and 1600 (D) cells and imaged daily from day 3 to 6.
Development | Supplementary Material
Six representative examples per condition are shown. Scale bar: 200 μm. (E) Aggregates formed using different
initial plating densities (200, 400, 800 and 1600 cells per well) were cultured in N2B27 with a 24 h pulse on day
2-3 of Chi (Eʹ), Act (Eʺ) or Act/Chi (Eʹʺ). Representative aggregates are shown for each condition plating
density; the same aggregate was followed throughout the time-course. See Fig. 2A (protocol P3) for stimulation
protocol. Aggregates in (E) not to scale.
Figure S3. Sox17::GFP faithfully reports the localisation of Sox17. Sox17::GFP cells exposed to continuous
Act/Chi after 2 days N2B27 were fixed and stained for Sox17. Confocal imaging confirms the faithful reporting
of Sox17 gene expression by the Sox17::GFP fluorescent reporter. Scale bar: 50 μm. Hoechst was used to stain
the nuclei.
Development | Supplementary Material
Supplementary Movie Legends
Movie 1. Sections through the aggregate shown in Fig. 3B and B’. Hoechst, E-Cadherin and Sox17 are
labelled in blue, red and white, respectively.
Development | Supplementary Material
Movie 2. Aggregation of ES cells within the first 48 h of suspension culture. A suspension of mouse ES cells in
N2B27 were plated in 96-well plates as described (see Materials and Methods; and Baillie-Johnson et al., 2014)
and imaged for 48 h. Over time, individual or small clusters of ES cells begin to aggregate at the bottom of the
well (refer to Fig. 5A).
Movie 3. Emergence Sox17::GFP expression following addition of Act and Chi. Aggregated Sox17::GFP ES
cell were imaged for ~73 h in N2B27 supplemented with Act and Chi. Note how the expression of Sox17::GFP
is heterogeneously expressed before a regionalised expression pattern is adopted (see Fig. 5B).
Development | Supplementary Material
Movie 4. Generation of polarised expression of Sox17::GFP. Polarised expression of Sox17::GFP becomes
more pronounced as the aggregate increases in size throughout time (see Fig. 5C).
Movie 5. Formation of polarised Bra::GFP expression in aggregated ES cells. Aggregates of Bra::GFP ES cells
in N2B27 were transferred to Act and Chi and filmed for 120 h. Initially, Bra::GFP is expressed throughout the
whole aggregate before the down-regulation in regions that will not form the extensions (see Fig. 5E).
Development | Supplementary Material
Development | Supplementary Material
Movies 6-9. Gastrulation-like movements in aggregates. Cell extrusion and intra-aggregate movement in
Sox17::GFP (Movie 6), Bra::GFP (Movie 7) and Tbx6::EYFP (Movies 8,9). Aggregates were treated as in Fig.
8 indicated and the stills correspond to the indicated time points from the associated movies (see Movies 6-9).
Movie 10. Polarised cell blebbing in an aggregate. Aggregates of GPI-GFP ES cells were transferred to Act and
Chi conditions on day 3 and were imaged over 2 h with confocal microscopy.
Development | Supplementary Material

Supplementary resource (1)

... Amongst other properties, these models represent a useful minimal system to investigate the communication between embryonic and extra-embryonic tissue. However, the aggregation of mouse ESCs alone can also be induced to generate distinct embryonic-like structures referred to as gastruloids [8][9][10] . They can organize polarities and elongate an anterior-posterior axis in the absence of any extra-embryonic tissue and in a highly reproducible manner [8][9][10][11][12] , which in many respects mimics tail bud elongation, thus revealing the previously underappreciated ability of embryonic cells to form complex embryonic-like structures with minimal pre-existing positional cues 13 . ...
... However, the aggregation of mouse ESCs alone can also be induced to generate distinct embryonic-like structures referred to as gastruloids [8][9][10] . They can organize polarities and elongate an anterior-posterior axis in the absence of any extra-embryonic tissue and in a highly reproducible manner [8][9][10][11][12] , which in many respects mimics tail bud elongation, thus revealing the previously underappreciated ability of embryonic cells to form complex embryonic-like structures with minimal pre-existing positional cues 13 . This ability which we refer to as selforganisation can be triggered by various phenomenon; for example a recent study 14 identified an early radial symmetry breaking determining a binary response to WNT. ...
... Gastruloids develop axially organised structures composed of cells derived from all three germ layers 10,16,18 . Following a pulse of exogeneous WNT activation (Chiron, CHIR99021), they break symmetry and elongate an antero-posterior axis [8][9][10] without the influence of extra-embryonic tissue or from the implantation into the uterine wall. This self-organization capacity may involve distinct cell signalling, mechanical constraints between inner and outer cells, or a differential accessibility to Chiron, as well as heterogeneity in the starting cellular state. ...
Full-text available
Gastruloids have recently emerged as an efficient four-dimensional model for studying some aspects of post-implantation embryonic patterning. They undergo gastrulation-like processes leading to the self-organization into highly reproducible biological objects. Here, we sought to uncover the molecular and cellular mechanism underlying this remarkable property. We report that self-organization competence is associated with a cell-specific coordination of a Cadherin switch. We find that N-Cadherin hinders gastruloids morphogenetic competence, for its inactivation leads to the formation of trunk-like structures in absence of extra-cellular matrix analogues. In contrast, E-Cadherin repression by Snai1 is critical for self-organization: Snai1 establishes a cell-specific repressive pace by triggering the repression of a pluripotency-associated transcription program and its chromatin landscape, thus allowing a proper transition from E- to N-Cadherin to occur. Altogether, this work establishes a molecular mechanism that integrates the exit from pluripotency and the pace of cell differentiation, leading to the observed self-organizing potential of gastruloids.
... For example, symmetry breaking is achievable in EBs in vitro, where the local activation of the WNT signaling pathway induces AP polarity and the formation of a primitive streak-like region, resulting in differentiation into a mesendodermal compartment (in contrast, neurectodermal differentiation is achieved by WNT signaling inhibition) 68 . Furthermore, the activation of WNT signaling at an appropriate time frame (at the initial stages of culture, which is usually within 74 h post-aggregation) induces the formation of a progressive elongating domain similar to that of the in vivo tail bud of the embryo, as well as the specification of the endoderm, which all lead to simulating aspects of the gastrulation period of the embryo 69 . Additionally, the precise timing regulated by WNT/β-catenin and Nodal signaling induces symmetry breaking in gastruloids even without the presence of extraembryonic tissue or localized signaling, allowing further polarization to take place, which is evidenced by distinct T/ Brachyury expression 70 . ...
... Early works to recapitulate aspects of gastrulation were conducted using mouse stem cells. The first absolute 3D gastruloid was created by van den Brink and colleagues in 2014 69 . Consisting of only 300 mouse ES cells, the EB was cultured in vitro to create a 3D gastruloid, where hallmarks of early mouse embryonic development, such as symmetry breaking and axial organization, were observable under the activation of WNT/ β-catenin signaling, which further caused germ layer specification and minor axial elongation upon the addition of specific morphogens such as CHIR99021 and Activin A 77 . ...
... Mouse ESC-based gastruloids induced without surrounding extraembryonic tissues showed localized T/ Brachyury expression with polarity and extension to one side, suggesting the notion of AP axis development. Thus, this discovery suggested that extraembryonic tissues are not necessarily required for the embryo to undergo self-patterning 69 . ...
Full-text available
Recent discoveries in stem cell and developmental biology have introduced a new era marked by the generation of in vitro models that recapitulate early mammalian development, providing unprecedented opportunities for extensive research in embryogenesis. Here, we present an overview of current techniques that model early mammalian embryogenesis, specifically noting models created from stem cells derived from two significant species: Homo sapiens , for its high relevance, and Mus musculus , a historically common and technically advanced model organism. We aim to provide a holistic understanding of these in vitro models by tracing the historical background of the progress made in stem cell biology and discussing the fundamental underlying principles. At each developmental stage, we present corresponding in vitro models that recapitulate the in vivo embryo and further discuss how these models may be used to model diseases. Through a discussion of these models as well as their potential applications and future challenges, we hope to demonstrate how these innovative advances in stem cell research may be further developed to actualize a model to be used in clinical practice.
... Gastruloids are three-dimensional aggregates of ESCs that have the inherent capacity to self-organise into polarised structures that resemble gastrulating embryos [7,8]. Gastruloids derived from mouse ESCs (mESCs) demonstrated collective behaviours reminiscent of the posterior cells in the early mouse embryo, such as symmetry breaking, axial organisation, germ-layer specification, and axial elongation [9]. ...
... This demonstrates that despite the complexity of the embryonic signalling environment, relatively homogeneous pluripotent cells have an intrinsic ability to self-organise and recapitulate many aspects of gastrulation when given a single signalling cue. It is important to note however, that several factors, including the formulation of the culture medium, exposure timing, and initial aggregate size, can influence the cellular composition of gastruloids experiments [7,8]. ...
Full-text available
Gastrulation represents a pivotal phase of development and aberrations during this period can have major consequences, from minor anatomical deviations to severe congenital defects. Animal models are used to study gastrulation, however, there is considerable morphological and molecular diversity of gastrula across mammalian species. Here, we provide an overview of the latest research on interspecies developmental control across mammals. This includes single-cell atlases of several mammalian gastrula which have enabled comparisons of the temporal and molecular dynamics of differentiation. These studies highlight conserved cell differentiation regulators and both absolute and relative differences in differentiation dynamics between species. Recent advances in in vitro culture techniques have facilitated the derivation, maintenance and differentiation of cell lines from a range of species and the creation of multi-species models of gastrulation. Gastruloids are three-dimensional aggregates capable of self-organising and recapitulating aspects of gastrulation. Such models enable species comparisons outside the confines of the embryo. We highlight recent in vitro evidence that differentiation processes such as somitogenesis and neuronal maturation scale with known in vivo differences in developmental tempo across species. This scaling is likely due to intrinsic differences in cell biochemistry. We also highlight several studies which provide examples of cell differentiation dynamics being influenced by extrinsic factors, including culture conditions, chimeric co-culture, and xenotransplantation. These collective studies underscore the complexity of gastrulation across species, highlighting the necessity of additional datasets and studies to decipher the intricate balance between intrinsic cellular programs and extrinsic signals in shaping embryogenesis.
... Gastruloids are often defined as three-dimensional aggregates of stem cells that faithfully recapitulate the spatial and genetic composition of the gastrulating embryo. These aggregates exhibit collective behaviors akin to those observed during early embryonic development, such as symmetry breaking and axis elongation [1][2][3][4][5][6]. However, there are no universal criteria for defining an 'optimal' gastruloid. ...
Full-text available
The young field of gastruloids brings promise to modeling and understanding early embryonic development. However, being a complex model, gastruloids are prone to variability at different levels. In this perspective, we define the different levels of gastruloid variability, and parameters over which it can be measured. We discuss potential sources for variability, and then propose methods to better control and reduce it. We provide an example from definitive endoderm progression in gastruloids, where we harness gastruloid-to-gastruloid variation in early parameters to identify key driving factors for endoderm morphology. We then devise interventions that steer morphological outcome. A better control over the developmental progression of gastruloids will enhance their utility in both basic research and biomedical applications.
... mental systems to address this question. 3D spherical aggregates of mouse or human ESCs self-organize into gastruloids, which are embryonic organoids that undergo symmetry breaking and exhibit axial organization and gene expression patterns that mirror events in the embryo [6][7][8][9][10]. Depending on the differentiation protocol, i.e. depending on the culture conditions and the signaling molecules present in the medium, gastruloids have the capacity to develop advanced structures strikingly similar to organs such as (1) somites and neural-tube [11][12][13][14] or (2) gut and heart [11,15,16]. ...
Full-text available
During early development of multi-cellular animals, cells self-organize to set up the body axes, such as the primary head-to-tail axis, based on which the later body plan is defined. Several signaling pathways are known to control body axis formation. Here, we show, however, that tissue mechanics plays an important role during this process. We focus on the emergence of a primary axis in initially spherical aggregates of mouse embryonic stem cells, which mirrors events in the early mouse embryo. These aggregates break rotational symmetry to establish an axial organization with domains of different expression profiles, e.g. of the transcription factor T/Bra and the adhesion molecule E-cadherin. Combining quantitative microscopy and physical modeling, we identify large-scale tissue flows with a recirculation component and demonstrate that they significantly contribute to symmetry breaking. We show that the recirculating flows are explained by a difference in tissue surface tension across domains, akin to Marangoni flows, which we further confirm by aggregate fusion experiments. Our work highlights that body axis formation is not only driven by biochemical processes, but that it can also be amplified by tissue flows. We expect this type of amplification to operate in many other organoid and in-vivo systems.
Full-text available
Combining high‐throughput generation and high‐content imaging of embryo models will enable large‐scale screening assays in the fields of (embryo) toxicity, drug development, embryogenesis, and reproductive medicine. This study shows the continuous culture and in situ (i.e., in microwell) imaging‐based readout of a 3D stem cell‐based model of peri‐implantation epiblast (Epi)/extraembryonic endoderm (XEn) development with an expanded pro‐amniotic cavity (PAC) (E3.5 E5.5), namely XEn/EPiCs. Automated image analysis and supervised machine learning permit the identification of embryonic morphogenesis, tissue compartmentalization, cell differentiation, and consecutive classification. Screens with signaling pathway modulators at different time windows provide spatiotemporal information on their phenotypic effect on developmental processes leading to the formation of XEn/EPiCs. Exposure of the biological model in the microwell platform to pathway modulators at two time windows, namely 0–72 h and 48–120 h, show that Wnt and Fgf/MAPK pathway modulators affect Epi differentiation and its polarization, while modulation of BMP and Tgfβ/Nodal pathway affects XEn specification and epithelialization. Further, their collective role is identified in the timing of the formation and expansion of PAC. The newly developed, scalable culture and analysis platform, thereby, provides a unique opportunity to quantitatively and systematically study effects of pathway modulators on early embryonic development.
Full-text available
The anterior-posterior axis of the mouse embryo is defined before formation of the primitive streak, and axis specification and subsequent anterior development involves signaling from both embryonic ectoderm and visceral endoderm. Τhe Wnt signaling pathway is essential for various developmental processes, but a role in anterior-posterior axis formation in the mouse has not been previously established. β-Catenin is a central player in the Wnt pathway and in cadherin-mediated cell adhesion. We generated β-catenin–deficient mouse embryos and observed a defect in anterior-posterior axis formation at embryonic day 5.5, as visualized by the absence of Hex and Hesx1 and the mislocation of cerberus-like and Lim1 expression. Subsequently, no mesoderm and head structures are generated. Intercellular adhesion is maintained since plakoglobin substitutes for β-catenin. Our data demonstrate that β-catenin function is essential in anterior-posterior axis formation in the mouse, and experiments with chimeric embryos show that this function is required in the embryonic ectoderm.
Full-text available
The first inductive interaction in amphibian develop-ment is mesoderm induction, in which an equatorial mesodermal rudiment is induced from the animal hemisphere under the influence of a signal from vegetal pole blastomeres. We have recently dis-covered that the Xenopus XTC cell line secretes a factor which has the properties we would expect of a mesoderm-inducing factor. In this paper, we show that an early response to this factor by isolated Xenopus animal pole regions is a change in shape, involving elongation and constriction. We show by several criteria, including general appearance, timing, rate of elongation and the nonrequirement for cell division that these movements resemble the events of gastrulation. We also demonstrate that the move-ments provide an early, simple and reliable indicator of mesoderm induction and are of use in providing a 'model system' for the study of mesoderm induction and gastrulation. For example, we show that the timing of gastrulation movements does not depend upon the time of receipt of a mesoderm-induction signal, but on an intrinsic gastrulation 'clock' which is present even in those animal pole cells that would not normally require it.
Full-text available
Embryonic Stem cells derived from the epiblast tissue of the mammalian blastocyst retain the capability to differentiate into any adult cell type and are able to self-renew indefinitely under appropriate culture conditions. Despite the large amount of knowledge that we have accumulated to date about the regulation and control of self-renewal, efficient directed differentiation into specific tissues remains elusive. In this work, we have analysed in a systematic manner the interaction between the dynamics of loss of pluripotency and Activin/Nodal, BMP4 and Wnt signalling in fate assignment during the early stages of differentiation of mouse ES cells in culture. During the initial period of differentiation, cells exit from pluripotency and enter an Epi-like state. Following this transient stage, and under the influence of Activin/Nodal and BMP signalling, cells face a fate choice between differentiating into neuroectoderm and contributing to Primitive Streak fates. We find that Wnt signalling does not suppress neural development as previously thought and that it aids both fates in a context dependent manner. Our results suggest that as cells exit pluripotency they are endowed with a primary neuroectodermal fate and that the potency to become endomesodermal rises with time. We suggest that this situation translates into a "race for fates" in which the neuroectodermal fate has an advantage.
Full-text available
The formation of the primitive streak is the first visible sign of gastrulation, the process by which the three germ layers are formed from a single epithelium during early development. Embryonic stem cells (ESCs) provide a good system for understanding the molecular and cellular events associated with these processes. Previous work, both in embryos and in culture, has shown how converging signals from both nodal/TGFβR and Wnt/β-catenin signalling pathways specify cells to adopt a primitive-streak-like fate and direct them to undertake an epithelial-to-mesenchymal transition (EMT). However, many of these approaches have relied on genetic analyses without taking into account the temporal progression of events within single cells. In addition, it is still unclear to what extent events in the embryo are able to be reproduced in culture. Here, we combine flow cytometry and a quantitative live single-cell imaging approach to demonstrate how the controlled differentiation of mouse ESCs towards a primitive streak fate in culture results in cells displaying many of the characteristics observed during early mouse development including transient brachyury expression, EMT and increased motility. We also find that the EMT initiates the process, and this is both fuelled and terminated by the action of brachyury, whose expression is dependent on the EMT and β-catenin activity. As a consequence of our analysis, we propose that a major output of brachyury expression is in controlling the velocity of the cells that are transiting out of the primitive streak.
Full-text available
The development of the central nervous system is known to result from two sequential events. First an inductive event of the mesoderm on the overlying ectoderm that generates a neural plate which, after rolling into a neural tube, acts as the main source of neural progenitors. Second, the axial regionalization of the neural plate that will result in the specification of neurons with different anteroposterior identities. While this view applies with ease to amphibians and fish, it is more difficult to confirm in amniote embryos where a specialized population of cells emerges at the end of gastrulation that, under the influence of Wnt and FGF signalling, expands and generates the spinal cord and the paraxial mesoderm. This population is known as the long-term Neuromesodermal precursor (NMp). Here we show that controlled increases of Wnt/ß-catenin and FGF signalling during adherent culture differentiation of mouse Embryonic Stem (ES) cells generates a population with many of the properties of the NMp. A single cell analysis of gene expression within this population reveals signatures characteristic of stem cell populations. Furthermore, when this activation is triggered in three-dimensional aggregates of ES cells, the population self-organizes macroscopically and undergoes growth and axial elongation that mimics some of the features of the embryonic spinal cord and paraxial mesoderm. We use both adherent and three dimensional cultures of ES cells to probe the establishment and maintenance of NMps and their differentiation.
Full-text available
Dissociated mouse embryonic stem (ES) cells were cultured to form aggregates in small volumes of basal medium in U-bottomed, non tissue-culture-treated 96-well plates and subsequently maintained in suspension culture. After growth for 48 hours, the aggregates are competent to respond to ubiquitous experimental signals which result in their symmetry-breaking and generation of defined polarised structures by 96 hours. It is envisaged that this system can be applied both to the study of early developmental events and more broadly to the processes of self-organisation and cellular decision-making. It may also provide a suitable niche for the generation of cell types present in the embryo but unobtainable from conventional adherent culture.
Full-text available
Background The formation of the primitive streak is the first visible sign of gastrulation, the process by which the three germ layers are formed from a single epithelium during early development. Embryonic stem cells (ESCs) provide a good system for understanding the molecular and cellular events associated with these processes. Previous work, both in embryos and in culture, has shown how converging signals from both nodal/TGFβR and Wnt/β-catenin signalling pathways specify cells to adopt a primitive-streak-like fate and direct them to undertake an epithelial-to-mesenchymal transition (EMT). However, many of these approaches have relied on genetic analyses without taking into account the temporal progression of events within single cells. In addition, it is still unclear to what extent events in the embryo are able to be reproduced in culture. Results Here, we combine flow cytometry and a quantitative live single-cell imaging approach to demonstrate how the controlled differentiation of mouse ESCs towards a primitive streak fate in culture results in cells displaying many of the characteristics observed during early mouse development including transient brachyury expression, EMT and increased motility. We also find that the EMT initiates the process, and this is both fuelled and terminated by the action of brachyury, whose expression is dependent on the EMT and β-catenin activity. Conclusions As a consequence of our analysis, we propose that a major output of brachyury expression is in controlling the velocity of the cells that are transiting out of the primitive streak. Electronic supplementary material The online version of this article (doi:10.1186/s12915-014-0063-7) contains supplementary material, which is available to authorized users.
Reciprocal interactions between extra-embryonic and embryonic tissues are required to establish the anterior-posterior axis of the mouse embryo. A clear morphological manifestation of anterior-posterior asymmetry in the embryo is the onset of gastrulation in the posterior region. Induction of the primitive streak depends on Nodal and Wnt signaling, and is modulated by the Anterior Visceral Endoderm. Before gastrulation, migration of the Distal Visceral Endoderm is the symmetry-breaking event, which defines the anterior pole of the mouse conceptus. Proximal-distal regionalization of the visceral endoderm precedes migration of the Distal Visceral Endoderm. The Distal Visceral Endoderm, which constitutes a heterogeneous population of cells, is induced by Nodal and restricted by Bmp signaling. The movement of the Distal Visceral Endoderm is driven by active cell migration and cell intercalation, accompanied by a global movement of visceral endoderm (VE) cells. The direction of the movement is controlled by Nodal and Wnt signaling. The observations in the mouse model may not fully apply to other mammalian species. The anterior visceral endoderm (AVE) has an equivalent in the rabbit and other vertebrate species. However, the equivalent of the extra-embryonic ectoderm is unclear.
Embryos allocate cells to the three germ layers in a spatially ordered sequence. Human embryonic stem cells (hESCs) can generate the three germ layers in culture; however, differentiation is typically heterogeneous and spatially disordered. We show that geometric confinement is sufficient to trigger self-organized patterning in hESCs. In response to BMP4, colonies reproducibly differentiated to an outer trophectoderm-like ring, an inner ectodermal circle and a ring of mesendoderm expressing primitive-streak markers in between. Fates were defined relative to the boundary with a fixed length scale: small colonies corresponded to the outer layers of larger ones. Inhibitory signals limited the range of BMP4 signaling to the colony edge and induced a gradient of Activin-Nodal signaling that patterned mesendodermal fates. These results demonstrate that the intrinsic tendency of stem cells to make patterns can be harnessed by controlling colony geometries and provide a quantitative assay for studying paracrine signaling in early development.