Temporal Modulations of NODAL, BMP, and WNT Signals Guide the Spatial Patterning in Self-Organized Human Ectoderm Tissues

Abstract

•hPSCs can form concentric rings of all major cell types in ectoderm•Spatial order of ectoderm cell types is regulated by NODAL, BMP, and WNT signals•The reaction-diffusion of BMP/WNT signals can be controlled•In vitro human microtissues have great potential in studying human development Differentiation of micropatterned hPSCs in vitro has become a powerful tool for modeling the spatial pattern formation during the early stages of development. Such pattern formation is often attributed to self-organization mediated by a persistent spatial morphogen gradient. However, it is still unclear how such a gradient is regulated in ectoderm, particularly in human genetic background. Here, we successfully generated a human ectoderm model containing all major cell types in a chemically defined condition. It is demonstrated that the time-dependent reaction-diffusion of signaling molecules controls the patterning of ectoderm. Our work provides a promising controllable platform to investigate the functional role of individual gene and signal transduction in human ectoderm development. Developmental biology studies using model organisms suggest that the emergence of spatial patterning in the ectoderm is mediated by the morphogen gradients. However, it is still unclear whether the morphogen gradient is necessary and dominates the cell spatial patterning, particularly in human genetic background. Here, we demonstrate that human pluripotent stem cells can self-organize to concentric rings of all major cell types in ectoderm when cultured on micropatterned surfaces in a chemically defined condition. We reveal that modulating the dynamics of NODAL, BMP, and WNT signals is sufficient to control the spatial order of different cell types. Our mathematical model suggests that changes in wavelength and phase of signaling patterns formed via reaction-diffusion may be the mechanism by which temporal information is translated into spatial information. Together, our work demonstrates that in vitro human ectoderm microtissues have great potential in understanding the mechanisms of early-stage human development. Developmental biology studies using model organisms suggest that the emergence of spatial patterning in the ectoderm is mediated by the morphogen gradients. However, it is still unclear whether the morphogen gradient is necessary and dominates the cell spatial patterning, particularly in human genetic background. Here, we demonstrate that human pluripotent stem cells can self-organize to concentric rings of all major cell types in ectoderm when cultured on micropatterned surfaces in a chemically defined condition. We reveal that modulating the dynamics of NODAL, BMP, and WNT signals is sufficient to control the spatial order of different cell types. Our mathematical model suggests that changes in wavelength and phase of signaling patterns formed via reaction-diffusion may be the mechanism by which temporal information is translated into spatial information. Together, our work demonstrates that in vitro human ectoderm microtissues have great potential in understanding the mechanisms of early-stage human development. The development of the central nervous system begins at neural induction, during which the primitive ectoderm gives rise to neuroepithelium, neural crest, placodes, and epidermis. The regionalization of various cell types enables the precise localization of neural crest cells and placodes, which migrate from the dorsal end of the neural tube and form skull and cranial sensory organs, while epidermal cells cover the surface of embryos. A pivotal goal of neural development research is to identify the mechanism of this regionalization process.1van Ooyen A. Using theoretical models to analyse neural development.Nat. Rev. Neurosci. 2011; 12: 311-326Crossref PubMed Scopus (78) Google Scholar Numerous studies using animal models (e.g., Xenopus, chick, and zebrafish) have established that fibroblast growth factor (FGF), bone morphogenetic protein (BMP), and Wingless/int1 (WNT) signaling pathways participate in neural induction since the pioneer works of Spemann and Mangold.2Spemann H. Mangold H. Induction of embryonic primordia by implantation of organizers from a different species. 1923.Int. J. Dev. Biol. 2001; 45: 13-38PubMed Google Scholar The currently accepted, although highly debatable, “default model” for the neural induction proposes that biomolecules (e.g., noggin, chordin) secreted by the organizer (primitive node in human) antagonize BMP signaling and prime ectoderm cells with neural fate.3Munoz-Sanjuan I. Brivanlou A.H. Neural induction, the default model and embryonic stem cells.Nat. Rev. Neurosci. 2002; 3: 271-280Crossref PubMed Scopus (449) Google Scholar, 4Harland R. Neural induction.Curr. Opin. Genet. Dev. 2000; 10: 357-362Crossref PubMed Scopus (202) Google Scholar, 5Stern C.D. Neural induction: old problem, new findings, yet more questions.Development. 2005; 132: 2007-2021Crossref PubMed Scopus (303) Google Scholar Due to the existence of a gradient of BMP antagonists, the neural plate is specified to become neuroepithelium, while the border region subjected to intermediate BMP level is specified to become neural crest and placode, flanked by epidermis, which requires the highest BMP level. Research over the past two decades has converged to the view that the neural induction process is more complex than described by this simple default model.6Stern C.D. Neural induction: 10 years on since the 'default model'.Curr. Opin. Cell Biol. 2006; 18: 692-697Crossref PubMed Scopus (100) Google Scholar,7Gerhart J. Cellular basis of morphogenetic change: looking back after 30 years of progress on developmental signaling pathways.in: Love A.C. Conceptual Change in Biology: Scientific and Philosophical Perspectives on Evolution and Development. 2015: 175-197Crossref Scopus (2) Google Scholar In particular, while the central role of BMP remains valid, new evidence suggests functional roles of WNT and FGF signals in neural induction, which may act concomitantly with, or separately from, direct BMP inhibition to regulate BMP activities.3Munoz-Sanjuan I. Brivanlou A.H. Neural induction, the default model and embryonic stem cells.Nat. Rev. Neurosci. 2002; 3: 271-280Crossref PubMed Scopus (449) Google Scholar,8Zirra A. Wiethoff S. Patani R. Neural conversion and patterning of human pluripotent stem cells: a developmental perspective.Stem Cells Int. 2016; 2016: 8291260Crossref PubMed Scopus (19) Google Scholar Moreover, their functions may not be evolutionarily conserved.8Zirra A. Wiethoff S. Patani R. Neural conversion and patterning of human pluripotent stem cells: a developmental perspective.Stem Cells Int. 2016; 2016: 8291260Crossref PubMed Scopus (19) Google Scholar For instance, WNT signaling upregulates BMP expression and promotes epidermal fate in chick embryos.9Wilson S.I. Rydstrom A. Trimborn T. Willert K. Nusse R. Jessell T.M. Edlund T. The status of Wnt signalling regulates neural and epidermal fates in the chick embryo.Nature. 2001; 411: 325-330Crossref PubMed Scopus (224) Google Scholar In contrast, WNT3a activation does not impair neural induction in mouse embryos, mouse embryonic stem cells, and human pluripotent stem cells (hPSCs).10Yoon Y. Huang T.T. Tortelote G.G. Wakamiya M. Hadjantonakis A.K. Behringer R.R. Rivera-Perez J.A. Extra-embryonic Wnt3 regulates the establishment of the primitive streak in mice.Dev. Biol. 2015; 403: 80-88Crossref PubMed Scopus (24) Google Scholar, 11Turner D.A. Hayward P.C. Baillie-Johnson P. Rue P. Broome R. Faunes F. Arias A.M. Wnt/beta-catenin and FGF signalling direct the specification and maintenance of a neuromesodermal axial progenitor in ensembles of mouse embryonic stem cells.Development. 2014; 141: 4243-4253Crossref PubMed Scopus (79) Google Scholar, 12Li W.L. Ding S. Generation of novel rat and human pluripotent stem cells by reprogramming and chemical approaches.Cell Program. Reprogram. Methods Protoc. 2010; 636: 293-300Crossref Scopus (20) Google Scholar These findings collectively demonstrate that there is still a lack of fundamental understanding of how BMP, WNT, and FGF signaling interact and guide the cell-patterning process, particularly in the context of human development. While the prevailing approach in developmental biology research is to use animal models, it is well received that neurulation events in human and mouse differ in many respects.13Copp A.J. Stanier P. Greene N.D. Neural tube defects: recent advances, unsolved questions, and controversies.Lancet Neurol. 2013; 12: 799-810Abstract Full Text Full Text PDF PubMed Scopus (289) Google Scholar Recent works have explicitly demonstrated the potential of using hPSCs cultured in two-dimensional (2D) and 3D microenvironments to model key features of gastrulation.14Warmflash A. Sorre B. Etoc F. Siggia E.D. Brivanlou A.H. A method to recapitulate early embryonic spatial patterning in human embryonic stem cells.Nat. Methods. 2014; 11: 847-854Crossref PubMed Google Scholar, 15Heemskerk I. Burt K. Miller M. Chhabra S. Guerra M.C. Liu L. Warmflash A. Rapid changes in morphogen concentration control self-organized patterning in human embryonic stem cells.eLife. 2019; 8https://doi.org/10.7554/eLife.40526Crossref PubMed Scopus (43) Google Scholar, 16Shao Y. Taniguchi K. Townshend R.F. Miki T. Gumucio D.L. Fu J. A pluripotent stem cell-based model for post-implantation human amniotic sac development.Nat. Commun. 2017; 8: 208Crossref PubMed Scopus (106) Google Scholar We previously reported an in vitro neuroectoderm model by culturing hPSCs on micropatterned surfaces.17Xue X. Sun Y. Resto-Irizarry A.M. Yuan Y. Aw Yong K.M. Zheng Y. Weng S. Shao Y. Chai Y. Studer L. Fu J. Mechanics-guided embryonic patterning of neuroectoderm tissue from human pluripotent stem cells.Nat. Mater. 2018; 17: 633-641Crossref PubMed Scopus (85) Google Scholar In that model, hPSCs near the border region of the micropattern spontaneously differentiated into neural plate border cells, whereas cells in the center of the pattern differentiated into neuroepithelial cells under the dual Smad inhibition condition. We further revealed that such self-organization was dictated by biomechanical forces arising from geometrical confinement. However, this model lacks epidermis and placodal cells, and how BMP and WNT signaling pathways work in conjunction with biomechanical forces to regulate cellular spatial patterning is still unclear. Here, we report a complete in vitro ectoderm model containing all the major cell types (neuroepithelium, neural crest, placodes, and epidermis) by fine-tuning biochemical signals under chemically defined culture conditions. Moreover, we found that changing the sequence of NODAL, BMP, and WNT signaling activation could modulate the spatial order of the rings consisting of neural crest and epidermis. A mathematical model based on reaction-diffusion was developed to demonstrate how such temporal information was translated to spatial patterning. Together, our work suggests that the time-dependent reaction-diffusion of signaling molecules dictates the patterning of ectoderm, in contrast to a conventional morphogen gradient-mediated patterning model. The in vitro human ectoderm microtissues provide an engineered and controllable experimental platform from which to investigate the functional role of individual genes and molecules in ectoderm development in humans. We previously demonstrated that dual Smad inhibition and a transient WNT activation led to the regionalization of Pax6+ neuroepithelial cells and Pax3+ neural plate border cells on micropatterned surfaces.17Xue X. Sun Y. Resto-Irizarry A.M. Yuan Y. Aw Yong K.M. Zheng Y. Weng S. Shao Y. Chai Y. Studer L. Fu J. Mechanics-guided embryonic patterning of neuroectoderm tissue from human pluripotent stem cells.Nat. Mater. 2018; 17: 633-641Crossref PubMed Scopus (85) Google Scholar To achieve complete ectoderm with epidermal, neural crest, and placodal cells, it is necessary to further activate BMP and WNT signals.18Gomez G.A. Prasad M.S. Sandhu N. Shelar P.B. Leung A.W. Garcia-Castro M.I. Human neural crest induction by temporal modulation of WNT activation.Dev. Biol. 2019; 449: 99-106Crossref PubMed Scopus (16) Google Scholar, 19Dincer Z. Piao J. Niu L. Ganat Y. Kriks S. Zimmer B. Shi S.H. Tabar V. Studer L. Specification of functional cranial placode derivatives from human pluripotent stem cells.Cell Rep. 2013; 5: 1387-1402Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar, 20Byrne C. Tainsky M. Fuchs E. Programming gene expression in developing epidermis.Development. 1994; 120: 2369-2383Crossref PubMed Google Scholar Thus, we sought to investigate whether adding BMP4 and a β-catenin stabilizer, CHIR, to fully defined E6 medium supplemented with transforming growth factor β (TGF-β) inhibitor SB431542 (SB) (10 μM) is sufficient to induce those cell types (referred to henceforth as the “B.C.SB protocol,” Figure S1A). As shown in Figure 1, by continuously activating BMP and WNT signals, micropatterned (diameter 600 μm) H9 human embryonic stem cells (hESCs) self-organized into structures with four distinct zones. We found that cells located in both the center and outermost rings of the pattern were CDX2+, indicating a trophoblast fate, consistent with previous findings that BMP4 activation in the absence of FGF signaling can induce trophoblast differentiation.21Xu R.H. Chen X. Li D.S. Li R. Addicks G.C. Glennon C. Zwaka T.P. Thomson J.A. BMP4 initiates human embryonic stem cell differentiation to trophoblast.Nat. Biotechnol. 2002; 20: 1261-1264Crossref PubMed Scopus (823) Google Scholar,22Amita M. Adachi K. Alexenko A.P. Sinha S. Schust D.J. Schulz L.C. Roberts R.M. Ezashi T. Complete and unidirectional conversion of human embryonic stem cells to trophoblast by BMP4.Proc. Natl. Acad. Sci. U S A. 2013; 110: E1212-E1221Crossref PubMed Scopus (119) Google Scholar In addition, we found that the layer next to the outer ring of trophoblasts expressed SOX10, PAX3, and SOX9, markers for neural crest cells.23Basch M.L. Bronner-Fraser M. Garcia-Castro M.I. Specification of the neural crest occurs during gastrulation and requires Pax7.Nature. 2006; 441: 218-222Crossref PubMed Scopus (259) Google Scholar Cells expressing TP63 and E-cadherin (E-cad), epidermis markers,24Pattison J.M. Melo S.P. Piekos S.N. Torkelson J.L. Bashkirova E. Mumbach M.R. Rajasingh C. Zhen H.H. Li L.J. Liaw E. et al.Retinoic acid and BMP4 cooperate with p63 to alter chromatin dynamics during surface epithelial commitment.Nat. Genet. 2018; 50: 1658-1665Crossref PubMed Scopus (29) Google Scholar,25Haremaki T. Metzger J.J. Rito T. Ozair M.Z. Etoc F. Brivanlou A.H. Self-organizing neuruloids model developmental aspects of Huntington’s disease in the ectodermal compartment.Nat. Biotechnol. 2019; 37: 1198-1208Crossref PubMed Scopus (41) Google Scholar were found between the SOX10+ PAX3+ SOX9+ cells and center CDX2+ cells. Using an independent control human induced pluripotent stem cell (hiPSC) line, we observed similar spatial patterning of these cell types (Figure S2). As CDX2 signals in the center and edge of the micropatterns needed to be collected at different focal distances, we used confocal microscopy to further characterize the 3D cytoarchitecture in micropatterns. Confocal images of nuclei staining suggested that micropatterned cells formed a dome-shaped monolayer with colony thickness less than 22 μm (Figure S3). Importantly, we did not find cells co-expressing TP63 and SOX10, suggesting the separation of cells with neural crest and epidermal fates. In addition, CDX2 confocal images confirmed that these trophoblast-like cells occupied the center and the outermost region of the colony (Figure S4). We further examined the existence of placodal cells and found that AP2+ SIX1+ placodal cells19Dincer Z. Piao J. Niu L. Ganat Y. Kriks S. Zimmer B. Shi S.H. Tabar V. Studer L. Specification of functional cranial placode derivatives from human pluripotent stem cells.Cell Rep. 2013; 5: 1387-1402Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar,26Tchieu J. Zimmer B. Fattahi F. Amin S. Zeltner N. Chen S. Studer L. A modular platform for differentiation of human PSCs into all major ectodermal lineages.Cell Stem Cell. 2017; 21: 399-410.e7Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar were mainly located in the ring containing SOX10+ neural crest cells (Figure S5), suggesting that under this condition, the neural crest cells and placodal cells were not yet separated. To examine whether the localization of the putative neural crest/placodes and epidermis is a result of cell sorting, we immunostained TP63 and SOX10 at days 5, 6, and 7 of the differentiation (Figure S6) and found that both TP63+ and SOX10+ cells emerged at day 6 with the same localization. These results are consistent with our previous findings in neuroectoderm patterning and suggest that cell migration or cell sorting is not the major driving force for such pattern formation. We next sought to investigate whether the initial seeding density could affect the cytoarchitecture of the patterned cells by plating cells under three different seeding densities (40,000 cells cm−2, 80,000 cells cm−2, and 160,000 cells cm−2). Interestingly, we found a gradual shift of both the TP63+ and SOX10+ cells toward the border of the circular patterns with increasing seeding density (Figures 2A and 2B ). The average intensity peaked at 81, 152, and 260 μm from the pattern center, respectively for TP63 and 158, 184, and 259 μm, respectively for SOX10 when the seeding density increased from 40,000 to 160,000 cells cm−2 (Figure 2C). We also examined the effect of pattern size on the spatial patterning of the ectoderm microtissue in vitro (Figure 3). To this end, we generated circular adhesive islands with diameters of 200, 600, and 1,000 μm. In contrast to our previous finding, which showed that the neuroepithelium-neural plate border two-layer patterning is insensitive to pattern size, in this condition patterns with a diameter of 200 μm failed to produce intact rings of TP63+ cells. For 1,000-μm patterns, TP63+ cells or SOX10+ cells did not form an intact ring either, and the boundaries between TP63+ cells and SOX10+ cells were not as distinct as in 600-μm patterns, as reflected by the overlapping peak of intensities (324 μm versus 342 μm). Noticeably, in all these conditions we did not observe any cell co-expressing TP63 and SOX10 by examining the merged images, and the relative localization of TP63+ and SOX10+ rings remained unchanged. We previously described a mechanics-driven model to explain the formation of the two-layer neuroectoderm microtissues based on the distinct mechanical status of cells at the center and periphery of the patterns.17Xue X. Sun Y. Resto-Irizarry A.M. Yuan Y. Aw Yong K.M. Zheng Y. Weng S. Shao Y. Chai Y. Studer L. Fu J. Mechanics-guided embryonic patterning of neuroectoderm tissue from human pluripotent stem cells.Nat. Mater. 2018; 17: 633-641Crossref PubMed Scopus (85) Google Scholar However, the observed patterns here were significantly different because of the existence of apparent periodic structures (two CDX2+ layers) and the sensitivity to pattern size. Thus, a simple mechanical or chemical gradient was not sufficient to explain the formation of such complex structures. Moreover, the observed relative localization of epidermis and neural crest/placodes is opposite to the ectoderm development in vivo, suggesting that such pattern formation is not solely autonomous and is strongly affected by external chemical signals. Thus, we aimed to develop a theoretical model to explain these observations and predict under what conditions the order of epidermis and neural crest/placodes can be reversed to mimic ectoderm development in vivo. To simplify the model, we first hypothesized that BMP and WNT signals do not strongly crosstalk with each other and that the cell fate depends on the synergistic effects of BMP and WNT activities. BMP4-Noggin and WNT-Dickkopf (DKK) are two well-established activator-inhibitor pairs,27Tewary M. Ostblom J. Prochazka L. Zulueta-Coarasa T. Shakiba N. Fernandez-Gonzalez R. Zandstra P.W. A stepwise model of Reaction-Diffusion and Positional-Information governs self-organized human peri-gastrulation-like patterning.Development. 2017; 144: 4298-4312Crossref PubMed Scopus (72) Google Scholar,28Sick S. Reinker S. Timmer J. Schlake T. WNT and DKK determine hair follicle spacing through a reaction-diffusion mechanism.Science. 2006; 314: 1447-1450Crossref PubMed Scopus (394) Google Scholar and cell-secreted proteins from the micropatterned cells can rapidly diffuse to the culture medium.24Pattison J.M. Melo S.P. Piekos S.N. Torkelson J.L. Bashkirova E. Mumbach M.R. Rajasingh C. Zhen H.H. Li L.J. Liaw E. et al.Retinoic acid and BMP4 cooperate with p63 to alter chromatin dynamics during surface epithelial commitment.Nat. Genet. 2018; 50: 1658-1665Crossref PubMed Scopus (29) Google Scholar For such systems, the classic Turing's reaction-diffusion model29Turing A.M. The chemical basis of morphogenesis.Bull. Math. Biol. 1953; 52 (discussion 19-52): 153-197Crossref Scopus (297) Google Scholar predicts the presence of periodic activities of BMP and WNT signals (Figure 4A). We found that to achieve the four-layer patterns we observed in Figures 1 and S2, one would expect periodic high and low BMP and WNT activities with a similar wavenumber (reflected by the number of regions within each pattern) but a phase shift (i.e., the sequence of high or low level of signals) within the pattern (Figures 4A and 4B). We next sought to understand which factors can be used to tune the phase and wavenumber in this system. We adopted an axisymmetric 2D form of the reaction-diffusion model and specifically studied the effect of individual parameters on the growth of Turing instability. A brief description of the mathematical model and model calculations are given in Experimental Procedures. We first found that with constant diffusivities, the phase of the patterns can be reversed by changing the initial conditions of the partial differential equations. Our simulation results in Figure 4C show that the phase is sensitive to a subtle change in initial conditions. Furthermore, we demonstrated that the wavenumbers in the circular domain could be varied by changing the parameter value of a only. As shown in Figure 4D, the value of a=0.3078 leads to a simple low-high pattern (two regions), a value of a=0.6156 leads to three regions, while a=1.5390 yields four regions. Noticeably, the parameter a represents the coefficient of the production rate of the activator (e.g., BMP). Our theoretical analysis strongly suggests that the number of the regions and order of cell types in each region can be tuned by changing the initial condition and the level of BMP/WNT activators. Based on the theoretical model we proposed, to derive intact ectoderm microtissues with proper regionalization of neuroepithelium, neural crest/placodes, and epidermis, a possible way is to reduce the number of layers in the pattern from three to two and synchronize the phase of BMP and WNT patterns (Figure 4E). This might be achieved by lowering the activities of exogenous activators and changing the initial conditions as suggested by the model. To test this possibility, we first modified the B.C.SB protocol (Figure S1B) to delay the starting time for BMP and/or WNT activation from day 3 to day 5, which effectively reduced the initial signaling activations (i.e., a smaller a value in our model) and also changed the initial conditions. We found that when both BMP and WNT signal activation was delayed, we could obtain a three-layer structure with Pax6+ neuroepithelial cells in the center of the pattern, flanked by SOX10+ PAX3+ SOX9+ neural crest cells. A significant amount of TP63+ E-cad+ epidermal cells were located on the border of the pattern (Figure 5 for H9 cells and Figure S7 for control hiPSCs), and only a very small number of CDX2+ cells were found in the outermost region of the patterns. This agrees with the model prediction that both BMP and WNT activities need to be tuned down. In contrast, solely delaying the activation of WNT signals led to a three-layer structure with neural crest cells located on the border of the pattern and TP63+ cells located on the inner side of the SOX10+ cells. The number of Pax6+ cells also significantly decreased in the pattern center (Figures 5 and S7). Solely delaying the activation of BMP signals, however, completely disrupted the ring-shape pattern formation and largely inhibited neuroepithelial cell differentiation (Figures 5 and S7). We next examined whether modulating the concentration of BMP and WNT activators could also affect the relative localization of different cell types. By changing the dosage of BMP4 added to the culture medium using the B.C.SB protocol, we found that the exogenous BMP4 concentration strongly influenced the neural crest and epidermal differentiation (Figure 6). When BMP4 was not added, epidermis differentiation was completed inhibited and neural crest differentiation was also significantly reduced, resulting in incomplete ring structures. A low concentration of BMP4 (1 ng·mL−1, compared with 5 ng·mL−1 for the optimal condition shown in Figure 1) also failed to induce an intact ring of TP63+ cells while strongly promoting neural crest differentiation, as the full width at half maximum (FWHM) for spatial distributions of SOX10 intensity increased from 84 μm for 5 ng·mL−1 to 280 μm for 1 ng·mL−1 (Figure 6C). A high concentration (20 ng·mL−1) of BMP4 prohibited neural crest differentiation and strongly promoted epidermis differentiation, and both cell types spread randomly on the patterns. In contrast to BMP4, the concentration of CHIR has only marginal effects on ring formation (Figure 7). Without WNT activation, we could still observe both TP63+ and SOX10+ rings of cells located in the same order as shown in Figure 1A. However, a high concentration of CHIR (1,200 nM) led to widespread morphology of the cells and inhibited epidermal differentiation. A much higher concentration of CHIR (3 μM) completely inhibited both epidermal and neural crest differentiation (Figure 7). Although changing the concentrations of BMP or WNT activator can significantly influence the differentiation potential, we did not observe the change of relative localization between epidermis and neural crest cells. Together, these results showed that the absolute concentrations of BMP and WNT mainly affect the width of each ring of different cell type without changing their relative position.Figure 7Role of Exogenous WNT Activation in Cell-Fate Decision and Pattern Formation in Ectoderm MicrotissuesShow full caption(A) Representative immunofluorescence images showing cell nuclei (DAPI), neural crest marker (SOX10), and epidermis marker (TP63). Cells were differentiated using the B.C.SB protocol with modified CHIR concentration as indicated. Scale bar, 100 μm.(B) Corresponding colorimetric maps showing the average fluorescence intensity of SOX10 and TP63 staining.(C) Plots showing the quantitative average intensity in relation to the distance to the center of the pattern. In these experiments, H9 hESCs were induced using the B.C.SB protocol in E6 medium from day 3 to day 8.View Large Image Figure ViewerDownload (PPT) (A) Representative immunofluorescence images showing cell nuclei (DAPI), neural crest marker (SOX10), and epidermis marker (TP63). Cells were differentiated using the B.C.SB protocol with modified CHIR concentration as indicated. Scale bar, 100 μm. (B) Corresponding colorimetric maps showing the average fluorescence intensity of SOX10 and TP63 staining. (C) Plots showing the quantitative average intensity in relation to the distance to the center of the pattern. In these experiments, H9 hESCs were induced using the B.C.SB protocol in E6 medium from day 3 to day 8. By delaying the starting point of BMP and WNT activation, we successfully synchronized the phases of BMP and WNT patterns. The hPSCs generated ectoderm-like structures with neuroepithelial cells in the pattern centers, although the neural crest and epidermis layers were not completely separated (Figures 5 and S7, top panels). In addition, changing the concentration of exogenous BMP4 and WNT activator could regulate the width of each ring without affecting their relative locations (Figures 6 and 7). Thus, we asked whether a combination of these two approaches, i.e., simultaneously changing the BMP4 initial concentration and delaying its activation start time, could lead to a more accurate ectoderm in vitro model. We further modified the B.C.SB protocol (Figure S1B) to reduce the concentration of BMP4 from 5 ng mL−1 to 1 ng mL−1 or increase to 40 ng mL−1. We found that when BMP4 concentration was reduced, hESCs formed a three-layer structure with Pax6+ neuroepithelial cells in the center of the pattern, flanked by a layer of SOX10+ SOX9+ PAX3+ neural crest cells and then another layer of TP63+ E-cad+ epidermis. Only a negligible amount of CDX2+ cells was located on the border of the pattern (Figures 8A and 8B , top panels). In contrast, higher concentration of BMP4 led to a three-layer structure with a significant amount of CDX2+ cells located on the border of the pattern, PAX6+ neuroepithelium in the center, and a mixed population of neural crest and epidermis cells between trophoblast and neuroepithelium (Figures 8A and 8B, bottom panels). Together, these results suggested that the time-dependent reaction-diffusion of signaling molecules dictates the patterning of ectoderm (Figure 8C). The in vitro complete human ectoderm microtissues we generated may