Self-organized morphogenesis of a human neural tube in vitro by geometric constraints

Background:

Understanding the morphogenetic events underlying de novo organ formation during human embryonic development is a topic of immense interest. Neural tube folding, the foundational event for the development of the brain and spinal cord, is one such example. Errors in neural tubulation can lead to disability and/or postnatal lethality, and hence warrants a better understanding at cell- and tissue-level. However, technical challenges hinder the accessibility of the respective cell types involved in human embryos, thereby impeding the study of this process in the native in vivo context.

To try and overcome this, various alternative systems have been developed to try and aid our understanding of organ developmental processes. Although organoids offer an alternative for recapitulating organ formation, the resultant tissue shape and cell fate patterns are anatomically incorrect with low reproducibility. Animal models are also limited by their relevance to human development. Other experimental models such as organ-on-chip, which are scalable and can be controlled to generate functional tissues, impose additional constraints preventing the self-organization observed in embryonic development. Developing an alternative experimental model which can overcome the limitations of these existing systems would therefore prove useful for understanding certain aspects of human development. Pluripotent Stem Cells (PSCs) are the ideal starting material for such a system, given they can generate all the founder lineages and any subsequent organ of interest.

In this preprint, the authors present a new experimental system to faithfully study the dynamics of organ morphogenesis in vitro. Through micropatterning of human Pluripotent Stem Cells (hPSCs), precise control over initial size, shape and cell number was achieved, leading to high reproducibility of resultant tissue patterns. As a proof of concept, the authors supplemented the system with morphogens that drive early neural development and by monitoring the system through long-term live imaging, the authors report the intricacies of neural tube folding.

Key findings:

1. Micropatterning PSC cultures yields a robust and versatile experimental system

The cell number, size, and shape of primary embryonic tissue, from which organs develop, are tightly regulated in vivo. To recapitulate this in their model system, the authors cultured PSCs in 2D on micropatterned protein islands. Matrigel was supplied to trigger 3D transition, which eventually generated a pluripotent epithelium surrounding a lumen. This 3D tissue stained positively for the core pluripotency factors: Oct4, Sox2, and Nanog. Sustained maintenance of pluripotency implies the flexibility of the experimental model, allowing the differentiation towards any organ or tissue of interest. Importantly, they found the central lumen to be physically and chemically isolated from the external environment, allowing the study of niche-mediated regulation of fate specification and tissue morphogenesis. Therefore, the engineered system possesses all the attributes to make it a reliable in vitro model to study the dynamics of organ morphogenesis.

2. Neural induction in the system replicates in vivo neurulation

The authors tested their system by focusing on neural tube formation. To do this, the system was sequentially supplemented with TGFβ inhibitor followed by BMP4, morphogens which are reported to drive early neural development [1-3]. This induced self-organized tissue folding over three days, comparable to the time period over which neural folding occurs in human embryos [4]. The resultant tissue recapitulated the anatomical organization of the embryonic neural tube.

Besides morphological similarity, the folded tissue dimensions were also found to be comparable to those of human neural plate at respective developmental stages. Furthermore, the two focal points at which tissue bending is concentrated are reminiscent of hinge points formed during neural tube development in vivo. For the bent neural tissue to be closed into a tube, non-neural ectoderm was shown to make the first contact to initiate a zippering motion. This, and the formation of an actin ring at the closing edge, have been reported to occur during in vivo neural closure [5,6]. Additional similarities to physiology include:

(i) The formation of surface ectoderm – neural tissue bilayer confirmed by immunostaining for respective canonical markers (Fig.1b,d,e)

(ii) the neural layer – ectoderm interface being enriched exclusively with fibronectin and being the origin spot for neural crest cells, and

(iii) the deposition of collagen beneath the ectoderm.

All these observations further corroborate the reliability of the experimental model.

Figure.1b: Vertical section of formed neural tube stained for N-cadherin (Neural tissue) and E-cadherin (non-neural ectoderm) to show the tissue bilayer. 1d and 1e: Additional validation for the formation of bilayer tissue interface through immunostaining for Pax6 or Otx2 (both markers for neural tissue) along with E-cadherin

3. Unraveling the tissue-wise mechanical contributions for neural folding

The roles of embryonic tissues in driving neural tubulation aren’t well known, partly owing to the contradictory results found from existing experimental models which lack multiple tissue types – notably mesoderm and non-neural ectoderm. Since this newly proposed system offers precise control over tissue composition, the authors employed it to resolve neural tube morphogenesis.

First, the authors profiled the constituent tissues during the onset of neural folding by immunostaining for lineage-specific markers. The absence of Brachyury stained cells indicates the absence of mesoderm in neural tubulation. As TGFβ inhibition, which suppresses mesoderm fates, is the obligatory first step in the experimental model to direct neurulation, the authors concluded that mesoderm is not necessary for folding morphogenesis. They then investigated the requirement of non-neural ectoderm. Strikingly, prolonged BMP4 omission, which prevents ectoderm induction, did not give rise to tubulation despite generating homogeneous neural tissue. BMP4 supplementation without neural induction did not ensure neural folding either. Therefore, the presence of both non-neural ectoderm and neural tissue is mandatory for successful neurulation.

4. ROCK/Shroom3 signaling axis mediated apical constriction drives neural tubulation

The authors were next interested in how their model system could offer insight into Neural Tube Defects (NTDs). At the molecular level, the authors focused on Shroom3/Rho-kinase (ROCK) signaling, given its role in human NTDs (7,8). Accordingly, ROCK inhibition (ROCKi) through the addition of Y-27632 abrogated the curvature necessary for neural folding (Fig.2b). The cell shape change known to drive neural tissue bending observed in control samples is apical surface constriction. Thus, the authors measured the apical surface area of both individual neural cells and of the entire neural tissue in control samples and drug-treated ones. On both counts, ROCKi samples had larger values compared to the control. Additional evidence to support the loss of apical constriction included the disruption of Shroom3 and F-actin localization upon Y-27632 treatment. No change was observed between control and ROCKi samples with regards to F-actin localization at the basal surface of neural cells or either surface of non-neural tissue. These findings indicate that the newly described system can facilitate the study of developmental disorders.

Figure.2b: Vertical sections of immunostained neural tubes showing bent tissue and hinge formation in control sample (top) and the lack of both in ROCK-inhibitor (Y27632) treated sample (bottom)

5. Shape follows size during neural tubulation

In vivo, the neural plate varies in width along the anterior-posterior axis, with the latter region being the narrowest. To see if and how tissue size influences its folding shape, the authors induced neurulation in PSCs cultured on micropatterns of different widths. The narrow patterns (<150um width) produced tissue folds reminiscent of posterior regions with their single central hinge and u-shape, whereas the wider patterns generated two lateral hinges, a trait typical of the anterior region (Fig.3). When measured, neural tissue apical constriction did not vary with micropattern size, thereby ruling out the possibility of a difference in cell behaviors to drive observed morphological changes. Another key experimental outcome was the constant width of surface ectoderm across all micropatterns, compared to neural plate size which scaled linearly. Finally, through in silico modeling, the authors found that the combination of tissue size, apical contractility, and basal adhesion between constituent tissues, account for hinge number and the fold shape.

Figure.3: Vertical sections of neural tubes formed on micropatterns of different sizes. Smaller micropatterns generate tissue folds with only one hinge point and larger micropatterns display two hinges. White arrows point the hinges.

Conclusion

In this preprint, the authors present a versatile and highly reproducible experimental model. Through their proof-of-concept experiments they have more than validated the reliability and amenability of the system.

Why I chose this preprint

It presents a detailed and user-friendly protocol to replicate the self-organization of early embryonic tissues in vitro without compromising the dimensionality. Besides reproducibility, the protocol offers precise control over tissue size, shape and composition, the key parameters elusive in current experimental models. The construction of the system is such that it facilitates the investigation of a wide variety of research questions pertaining to embryogenesis, specifically the events occurring late in the timeline such as gastrulation and organogenesis. Moreover, the proposed experimental model could help us investigate developmental disorders by understanding the contributions at both cellular and tissue level in organ shaping.

References

1. L. Li et al., “Ectodermal progenitors derived from epiblast stem cells by inhibition of Nodal signaling” J. Mol. Cell Biol. 7, 455–465 (2015)
2. T. Haremaki et al.,” Self-organizing neuruloids model developmental aspects of Huntington’s disease in the ectodermal compartment” Nat. Biotechnol. 37, 1198–1208 (2019)
3. G. Britton, I. Heemskerk, R. Hodge, A. A. Qutub, A. Warmflash, “A novel self-organizing embryonic stem cell system reveals signaling logic underlying the patterning of human ectoderm” Development. 146, dev179093 (2019)
4. O. R., M. F., “Emergence of intraembryonic hematopoietic precursors in the pre-liver human embryo” Developmental Stages in Human Embryos. (Contrib. Embryol., Carnegie Inst. Wash., 1987)
5. H. J. Ray, L. A. Niswander, “Dynamic behaviors of the non-neural ectoderm during mammalian cranial neural tube closure” Dev. Biol. 416, 279–285 (2016)
6. A. Rolo et al.,” Regulation of cell protrusions by small GTPases during fusion of the neural folds” Elife. 5, e13273 (2016)

Questions to the author

1. Is it possible to produce micropatterns of complex geometry? If yes, have you tried trapezoid patterns to generate both anterior and posterior neural folds in the same sample?
2. Neural folding was observed in 3D cultures but not in 2D, with the key difference between the absence of lumen formation in the latter. In this regard, what is the plausible role of central lumen in directing neural tubulation?
3. Did you observe a group of Pax3+ve cells flanking the ends of neural plate (NP), which are collectively referred to as Neural Plate Border (NPB)? If not, can the group cells of neural fold that are between the hinge and surface ectoderm be the NPB in your system?
4. Did the folded region of neural tissue between the hinge and epidermis have the same dimensions across patterns, specifically when there are 2 hinges? This question is inspired from a recent preprint [9] which showed that NP-NPB fate boundary width is independent of micropattern size.

 

Tags: micropatterning, morphogenesis, neural tubulation, pluripotent stem cells

Posted on: 22 September 2021

doi: https://doi.org/10.1242/prelights.30682

Read preprint (No Ratings Yet)