The first obvious sign of bilateral symmetry in mammalian and avian embryos is the appearance of the primitive streak in the future posterior region of a radially symmetric disc. The primitive streak marks the midline of the future embryo. The mechanisms responsible for positioning the primitive streak remain largely unknown. Here we combine experimental embryology and mathematical modelling to analyse the role of the TGFβ-related molecules BMP4 and Vg1/GDF1 in positioning the primitive streak. Bmp4 and Vg1 are first expressed throughout the embryo, and then become localised to the future anterior and posterior regions of the embryo, where they will, respectively, inhibit or induce formation of the primitive streak. We propose a model based on paracrine signalling to account for the separation of the two domains starting from a homogeneous array of cells, and thus for the topological transformation of a radially symmetric disc to a bilaterally symmetric embryo.
How do vertebrate embryos break their initial radial symmetry and establish a midline as the axis of bilateral symmetry? In amphibians and fishes, the whole embryo is initially patterned by antagonistic gradients of BMP (ventrally) and Wnt/Nodal/Activin and BMP antagonists (dorsally)1,2,3. The difference between dorsal (where gastrulation starts) and the opposite side is set up by localization of maternal determinants. However, in amniotes (birds and mammals, and presumably also reptiles) zygotic transcription starts very early, allowing embryonic regulation until quite late. For example, a chick embryo at the 20,000–50,000 cell stage can be divided into 4 or more fragments, all of which can initiate the formation of a primitive streak4,5. These observations suggest that localization of maternally produced molecules cannot be the sole determinant of bilateral symmetry or the position of the embryonic axis in amniotes. In the early chick embryo, the posterior marginal zone (adjacent to where the primitive streak will form) expresses the TGFβ superfamily member Vg1 6,7,8,9,10, which is both sufficient6,7,8,9,10 and necessary11 for primitive streak formation. The opposite (anterior) margin expresses the transcription factor Gata2, which appears to act as a weak inhibitor of primitive streak formation. Previous experiments suggested that Gata2 and Vg1 transcription is regulated independently at the opposite ends of the embryo, which led to the proposal of a Global Positioning System (GPS) to pattern the whole embryo11.
What is the molecular nature of this GPS? Gata2 knockdown causes downregulation of Bmp4 expression, consistent with an involvement of BMP in positioning the primitive streak12. This suggests that BMP signalling might constitute one of the elements in the embryo GPS. To explore this possibility, we examined the earliest expression of Bmp4 and Vg1. In situ hybridization on embryos earlier than stage X EG&K13 reveals that both Bmp4 and Vg1 are expressed ubiquitously (Supplementary Figure SF1 A-I). By stage X, the expression domains of these genes separate to opposite poles of the blastodisc (Supplementary Figure SF1 J-P). This raises the question of how this segregation takes place.
In order to understand the role of BMP4 in positioning the primitive streak, and BMP4 relation with Vg1 we analysed the effects of ectopic BMP4 in different regions of the embryo. A bead of BMP4 placed in the posterior marginal zone (Fig. 1A) causes downregulation of Vg1 (23/26, control: 0/10) (Fig. 1B,C). Vg1 downregulation was paralleled by inhibition of primitive streak formation: in 42/49 embryos incubated overnight after a posterior graft of a BMP4 bead, the primitive streak failed to form near the bead (as previously reported12), but two streaks arose from lateral positions (control: 0/32) (Fig. 1D,E). Paradoxically, grafts of a bead of BMP4 in the anterior/lateral marginal zone (Fig. 1F) caused upregulation of Vg1. In 17/34 embryos Vg1 expression was upregulated within 6 hours (control: 0/33) (Fig. 1G,H and Supplementary Figure S2). Simultaneous inducer and inhibitor effects of BMP4 on Vg1 were evident even in the same embryo (Supplementary Figure SF 2C). We grafted four BMP4 beads in the marginal zone (as shown in Fig. 1I). 9 out of 12 embryos developed multiple primitive streaks, spaced between the beads (control: 0/12) (Fig. 1J,K). The paradoxical opposite effects elicited by BMP4 on the anterior and posterior parts of the early embryo on Vg1 expression support the idea that BMP4 is part of the GPS, and is thus involved in positioning the primitive streak. If BMP4 is indeed part of the GPS system that positions Vg1, is the converse also true? To test this, we grafted a pellet of Vg1-transfected cells onto the anterior marginal zone (Fig. 1L). In 7/12 embryos, Bmp4 expression was downregulated (control: 0/12) (Fig. 1M,N).
Taken together, the above experiments suggest that BMP4 and Vg1 can inhibit each other’s expression when misexpressed in each other’s domain, but overexpression of BMP4 anteriorly paradoxically induces Vg1. What mechanisms could account for this? Opposite effects of BMP4 on Vg1 in different regions of the embryo can hardly be explained by assuming any prior difference between cells in anterior and posterior regions. In order to get insights on how a homogeneous field of cells can give rise to a distinct pattern which results in an antero-posterior symmetry we formulated a mathematical model of BMP4 and Vg1 interactions.
Mathematical modelling has been widely used to explore self-regulated pattern formation in biological systems. The most commonly used models are based on reaction-diffusion (RD) mechanisms14,15 based on long-range diffusion of morphogens that can generate patterns at long range, and have recently been used to understand how patterns such as the formation of structures like rugae in the hard palate16 and digit patterning17 occur in mouse. The standard RD approach postulates that the spatial distribution of molecular signals (morphogens) is determined by direct interactions among them, and by their diffusion across a given domain14,15. However, RD systems are not appropriate to model interactions in the early chick blastoderm, because a) the embryo at this stage is a very large flat disk (about 3 mm diameter), just one cell thick, suspended between two large volumes of fluid (albumen dorsally, yolk ventrally), with virtually no extracellular space to establish a stable gradient based on diffusion; b) the source of Vg1 is comparatively far away from the opposite pole of the embryo; c) free, extracellular diffusion cannot provide an efficient physical mechanism for anterior and posterior embryonic regions to interact via diffusive chemical signals. A more parsimonious explanation could involve paracrine (local) signalling between nearby cells18. Both BMP4 and Vg1 are secreted signalling proteins that interact with specific membrane receptors, located in the same cell or in nearby cells. Therefore, they do not need to diffuse across particularly large distances within the embryo to be fully functional. We propose a model based on the idea that BMP4 and Vg1 interact by a short-range paracrine activity, whereby the signalling process is maintained by signal renewal triggered by signal-receptor interactions, rather than following from collision-like chemical reactions outside the cell (see scheme in Supplementary Material SM1 for details). An algorithm that only requires two transcription factors, labelled FB and FV, mediates feedback interactions between BMP4 and Vg1 in neighbouring cells. Pairwise interactions between BMP4, Vg1, FB and FV are represented by means of Hill-type equations (see SM1), which have been used in a variety of genetic systems19,20, because they can describe activation and inhibition mechanisms in a straightforward manner.
With these elements we implemented an agent-based model in which the same functional relations between BMP4 and Vg1 operate in each individual cell of the marginal zone (see SM2, with diagram in A5). Starting from a ubiquitous, uniform expression pattern in a homogeneous field of cells proposed to be identical, the model can generate a coherent collective behaviour, leading to segregation of BMP4 and Vg1 to opposite poles of the embryo (Fig. 2A). Importantly, no initial bias is necessary to induce the breaking of radial symmetry of the embryo. In this respect, all cells are proposed to behave according to the same interactions as described in SM2, and simulations were performed starting from initially homogeneous values. Therefore, the resulting macroscopic pattern is an emergent property of the model.
Next, we tested if the model can reproduce the paradoxical effects of ectopic expression of BMP4 in anterior and posterior regions of the embryo. We postulate that initially, interactions between BMP4 and Vg1 (described in SM2) take place continuously in every cell in the epiblast, irrespective of their location in the embryo. Indeed, the model reproduces our experimental results: in normal embryos, the domains of expression of BMP4 and Vg1 segregate to opposite sides of the blastoderm. In experimental embryos, the model reproduces the findings that overexpression of BMP4 posteriorly inhibits Vg1 expression (Fig. 2B), whereas anterior misexpression of BMP4 paradoxically induces Vg1 near the site of overexpression (Fig. 2C). The model also predicts the previously reported observation8 that ectopic expression of Vg1 anteriorly induces Vg1 expression, and also that it should inhibit Bmp4 expression there (Fig. 2D).
The model invokes intracellular factors downstream of BMP4 and Vg1, FB and FV, respectively. The transcription factor Gata2 is a candidate for FB. Gata2 is expressed throughout the embryo before stage X, but eventually co-localizes in the future anterior region with Bmp4 11. The model reproduces these changes in Gata2 expression (Fig. 3A). The model also predicts that Gata2 knockdown should result in Bmp4 downregulation (Fig. 3B) and that anterior/lateral overexpression of BMP4 should increase Gata2 mRNA levels (Fig. 3C). The first prediction agrees with published results11, so we tested the second by grafting a BMP4 bead in the anterior/lateral marginal zone. In 13/19 embryos Gata2 expression was upregulated (control 0/14) (Fig. 3D). In the posterior region, Pitx2, a transcription factor that regulates Vg1 expression21, could be a possible candidate for FV. Pitx2 is slightly upregulated after BMP4 misexpression in the anterior/lateral marginal zone, with expression extending from the posterior region towards the bead (Figure SF3A,B, 6/18 embryos, control: 0/21, figure SF3C). However, upregulation of Pitx2 after BMP4 misexpression is only seen in a proportion of the embryos and is weaker than that of Vg1 in the same experimental conditions, keeping open the possibility that other molecules could fulfil the role of FV (see ref.21 for genes expressed in the posterior region of the early embryo).
In this paper we described a self-organizing process to account for the breaking of radial symmetry and the establishment of bilateral symmetry in the avian embryo. The model (SM2) is based on a paracrine mode of action of BMP4 and Vg1, whereby a set of complex interactions with two intracellular factors, FB/Gata2 and FV gives rise to a spatial pattern of expression that defines the anterior and posterior poles of the blastodisc and thereby anticipates the position of the primitive streak (see SM2). A role of BMP4 and Vg1/Nodal in early embryo polarity has been described in frog and fish22,23. However, to the best of our knowledge, this is the first description of dynamic interactions between BMP4 and Vg1 driving symmetry breaking that eventually results in primitive streak formation in amniotes. Interestingly, the model also suggests an explanation for the formation of twins (see SM2).
Could a similar mechanism work in mammals? In mouse, asymmetric Nodal activity drives movement of the distal VE towards the future anterior region, thus establishing the position of primitive streak formation24. Bmp4 is expressed in the distal ring of extraembryonic epiblast in the early mouse embryo, and Bmp4 downregulation prevents gastrulation and mesoderm formation25. The latter effect is due to Bmp4 influence on Nodal antagonists in the VE26, which supports an inhibitory role in primitive streak formation. Onset of an ectopic primitive streak-like structure in the amnion, with Nodal upregulation, occurs in the Bmp-effector Smad5 knockout, suggesting the presence of a Bmp/Nodal antagonism27. Whether or not interactions between Bmp4 and Nodal determine the position of the primitive streak in the mouse embryo remains unknown.
Paracrine mechanisms could explain the emergence of self-organized spatial patterns in other multicellular patterning systems, such as small aggregates of mouse Embryonic Stem Cells28 or micro-patterned cultures of human Embryonic Stem Cells29,30. For example, in the latter case, cells confined to a disk shape self-organize into patterned concentric areas, reminiscent to a certain extent of the three concentric areas that define the early chick embryo (AO, MZ and AP). This suggests the potential deployment of similar mechanisms in the patterning of a group of cells arranged in a blastodisc shape. The model proposed here could help to design experiments to test whether similar mechanisms could operate under these conditions.
Embryos and manipulation
Fertile hens’ eggs were obtained from Granja Gibert (Spain) (Brown Bovan Gold) and staged in Roman numerals for pre-primitive streak stages13 and in Arabic numerals31 starting from stage 2, when the primitive streak appears. Embryos were cultured in modified New culture32,33. Pre-stage X embryos were collected using a manual retrieval method as previously described34. Cut-in-half experiment on stage X embryos was carried out as previously described10. No live vertebrates were used for the experiments.
In situ hybridization
To misexpress Vg1, we used a Dorsalin-cVg1 expression construct7. We transplanted COS cells transfected with the construct of interest, and pellets of 1000 cells were generated from hanging drops and grafted into host embryos as previously described7,8,12,43. For misexpression via BMP4-conjugated to heparin beads (SIGMA), recombinant BMP4 (RD systems) was used at 15 μg/ml. Control beads were incubated in PBS.
Birsoy, B., Kofron, M., Schaible, K., Wylie, C. & Heasman, J. Vg 1 is an essential signaling molecule in Xenopus development. Development 133, 15–20, https://doi.org/10.1242/dev.02144 (2006).
Lu, F. I., Thisse, C. & Thisse, B. Identification and mechanism of regulation of the zebrafish dorsal determinant. Proceedings of the National Academy of Sciences of the United States of America 108, 15876–15880, https://doi.org/10.1073/pnas.1106801108 (2011).
Fauny, J. D., Thisse, B. & Thisse, C. The entire zebrafish blastula-gastrula margin acts as an organizer dependent on the ratio of Nodal to BMP activity. Development 136, 3811–3819, https://doi.org/10.1242/dev.039693 (2009).
Lutz, H. Sur la production expérimentale de la polyembryonie et de la monstruosité double chez les oiseaux. Arch. Anat. Microsc. Morphol. Exp 38, 79–144 (1949).
Spratt, N. T. & Haas, H. Integrative mechanisms in development of the early chick blastoderm. I. Regulative potentiality of separated part. J. Exp. Zool. 145, 97–137 (1960).
Seleiro, E. A., Connolly, D. J. & Cooke, J. Early developmental expression and experimental axis determination by the chicken Vg1 gene. Current biology: CB 6, 1476–1486 (1996).
Shah, S. B. et al. Misexpression of chick Vg1 in the marginal zone induces primitive streak formation. Development 124, 5127–5138 (1997).
Skromne, I. & Stern, C. D. A hierarchy of gene expression accompanying induction of the primitive streak by Vg1 in the chick embryo. Mech Dev 114, 115–118 (2002).
Bertocchini, F., Skromne, I., Wolpert, L. & Stern, C. D. 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, https://doi.org/10.1242/dev.01178 (2004).
Bertocchini, F. & Stern, C. D. The hypoblast of the chick embryo positions the primitive streak by antagonizing nodal signaling. Developmental cell 3, 735–744 (2002).
Bertocchini, F. & Stern, C. D. Gata2 provides an early anterior bias and uncovers a global positioning system for polarity in the amniote embryo. Development 139, 4232–4238, https://doi.org/10.1242/dev.081901 (2012).
Streit, A. et al. Chordin regulates primitive streak development and the stability of induced neural cells, but is not sufficient for neural induction in the chick embryo. Development 125, 507–519 (1998).
Eyal-Giladi, H. & Kochav, S. From cleavage to primitive streak formation: a complementary normal table and a new look at the first stages of the development of the chick. I. General morphology. Developmental biology 49, 321–337 (1976).
Turing, A. M. The chemical basis of morphogenesis. Phil Trans R Soc Lon B: Biol Sci 641, 37–72 (1952).
Gierer, A. & Meinhardt, H. A theory of biological pattern formation. Kybernetik 12, 30–39 (1972).
Economou, A. D. et al. Periodic stripe formation by a Turing mechanism operating at growth zones in the mammalian palate. Nature genetics 44, 348–351, https://doi.org/10.1038/ng.1090 (2012).
Raspopovic, J., Marcon, L., Russo, L. & Sharpe, J. Modeling digits. Digit patterning is controlled by a Bmp-Sox9-Wnt Turing network modulated by morphogen gradients. Science 345, 566–570, https://doi.org/10.1126/science.1252960 (2014).
van Boxtel, A. L. et al. A Temporal Window for Signal Activation Dictates the Dimensions of a Nodal Signaling Domain. Developmental cell 35, 175–185, https://doi.org/10.1016/j.devcel.2015.09.014 (2015).
Burrill, D. R. & Silver, P. A. Making cellular memories. Cell 140, 13–18, https://doi.org/10.1016/j.cell.2009.12.034 (2010).
Schultz, D., Walczak, A. M., Onuchic, J. N. & Wolynes, P. G. Extinction and resurrection in gene networks. Proceedings of the National Academy of Sciences of the United States of America 105, 19165–19170, https://doi.org/10.1073/pnas.0810366105 (2008).
Torlopp, A. et al. The transcription factor Pitx2 positions the embryonic axis and regulates twinning. eLife 3, e03743, https://doi.org/10.7554/eLife.03743 (2014).
Reversade, B. & De Robertis, E. M. Regulation of ADMP and BMP2/4/7 at opposite embryonic poles generates a self-regulating morphogenetic field. Cell 123, 1147–1160, https://doi.org/10.1016/j.cell.2005.08.047 (2005).
Xu, P. F., Houssin, N., Ferri-Lagneau, K. F., Thisse, B. & Thisse, C. Construction of a vertebrate embryo from two opposing morphogen gradients. Science 344, 87–89, https://doi.org/10.1126/science.1248252 (2014).
Yamamoto, M. et al. Nodal antagonists regulate formation of the anteroposterior axis of the mouse embryo. Nature 428, 387–392, https://doi.org/10.1038/nature02418 (2004).
Winnier, G., Blessing, M., Labosky, P. A. & Hogan, B. L. Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes & development 9, 2105–2116 (1995).
Soares, M. L., Torres-Padilla, M. E. & Zernicka-Goetz, M. Bone morphogenetic protein 4 signaling regulates development of the anterior visceral endoderm in the mouse embryo. Development, growth & differentiation 50, 615–621, https://doi.org/10.1111/j.1440-169X.2008.01059.x (2008).
Pereira, P. N. et al. Antagonism of Nodal signaling by BMP/Smad5 prevents ectopic primitive streak formation in the mouse amnion. Development 139, 3343–3354, https://doi.org/10.1242/dev.075465 (2012).
van den Brink, S. C. et al. Symmetry breaking, germ layer specification and axial organisation in aggregates of mouse embryonic stem cells. Development 141, 4231–4242, https://doi.org/10.1242/dev.113001 (2014).
Warmflash, A., Sorre, B., Etoc, F., Siggia, E. D. & Brivanlou, A. H. A method to recapitulate early embryonic spatial patterning in human embryonic stem cells. Nature methods 11, 847–854, https://doi.org/10.1038/nmeth.3016 (2014).
Etoc, F. et al. A Balance between Secreted Inhibitors and Edge Sensing Controls Gastruloid Self-Organization. Developmental cell 39, 302–315, https://doi.org/10.1016/j.devcel.2016.09.016 (2016).
Hamburger, Va. H. H. L. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92 (1951).
New, D. A. T. A new technique for the cultivation of the chick embryo in vitro. J. Embryol. Exp. Morphol. 3, 326–331 (1955).
Stern, C. D. & Ireland, G. W. An integrated experimental study of endoderm formation in avian embryos. Anatomy and embryology 163, 245–263 (1981).
Lee, H. C. et al. Cleavage events and sperm dynamics in chick intrauterine embryos. PloS one 8, e80631, https://doi.org/10.1371/journal.pone.0080631 (2013).
Stern, C. D. Detection of multiple gene products simultaneously by in situ hybridization and immunohistochemistry in whole mounts of avian embryos. Current topics in developmental biology 36, 223–243 (1998).
Liem, K. F. Jr, Tremml, G., Roelink, H. & Jessell, T. M. Dorsal differentiation of neural plate cells induced by BMP-mediated signals from epidermal ectoderm. Cell 82, 969–979 (1995).
Kispert, A., Ortner, H., Cooke, J. & Herrmann, B. G. The chick Brachyury gene: developmental expression pattern and response to axial induction by localized activin. Developmental biology 168, 406–415, https://doi.org/10.1006/dbio.1995.1090 (1995).
Kispert, A., Koschorz, B. & Herrmann, B. G. The T protein encoded by Brachyury is a tissue-specific transcription factor. The EMBO journal 14, 4763–4772 (1995).
Knezevic, V., De Santo, R. & Mackem, S. Two novel chick T-box genes related to mouse Brachyury are expressed in different, non-overlapping mesodermal domains during gastrulation. Development 124, 411–419 (1997).
Sheng, G. & Stern, C. D. Gata2 and Gata3: novel markers for early embryonic polarity and for non-neural ectoderm in the chick embryo. Mech Dev 87, 213–216 (1999).
Logan, M., Pagan-Westphal, S. M., Smith, D. M., Paganessi, L. & Tabin, C. J. The transcription factor Pitx2 mediates situs-specific morphogenesis in response to left-right asymmetric signals. Cell 94, 307–317 (1998).
Zhu, L. et al. Cerberus regulates left-right asymmetry of the embryonic head and heart. Current biology: CB 9, 931–938 (1999).
Skromne, I. & Stern, C. D. Interactions between Wnt and Vg1 signalling pathways initiate primitive streak formation in the chick embryo. Development 128, 2915–2927 (2001).
The authors are grateful to A. Martinez-Arias and M. Ros for helpful comments on the manuscript, and to Marisa Junco for excellent technical assistance. This work was funded by grants from MINECO to C.F.A. and M.A.H. (MTM2014-53156-P), from NIH (RO1-GM56656), the BBSRC (BB/C512753/1), the European Research Council (“GEMELLI”) and a Wellcome Trust investigator award to C.D.S., and from the Spanish Ministry of Science and Innovation to F.B. (BFU2010-19656). F.B. is a Ramon y Cajal fellow.
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
About this article
Cite this article
Arias, C.F., Herrero, M.A., Stern, C.D. et al. A molecular mechanism of symmetry breaking in the early chick embryo. Sci Rep 7, 15776 (2017). https://doi.org/10.1038/s41598-017-15883-8
This article is cited by
Expression patterns of signalling molecules and transcription factors in the early rabbit embryo and their significance for modelling amniote axis formation
Development Genes and Evolution (2021)
From genes to environment in shaping of an embryo: understanding embryonic-extraembryonic interactions at the BSDB autumn meeting in Oxford
Development Genes and Evolution (2019)