Choreographed morphogenetic events underlie early foregut development in the mouse embryo

Why I chose this preprint

The anterior embryo is one of the best-mapped regions in developmental biology at the molecular level — but the large-scale morphogenetic movements that actually build its geometry, and especially how multiple tissues coordinate to produce them, remain largely overlooked. This preprint addresses this gap directly: the long-term 3D live-imaging dataset presented here is an amazing resource for studying how these dramatic, coordinated shape changes are orchestrated. The authors took a good first step in identifying and quantitatively describing the stereotypical sequence of foregut involution and began testing several candidate processes that might contribute to it. Still, many of the driving mechanisms remain speculative. I hope highlighting this work brings it wider visibility and draws more attention to these relatively underexplored areas in embryonic morphogenesis.

Background

In the mouse embryo, organogenesis of the brain, heart, and gut is initiated at the anterior end during embryonic days 7.5 to 8.5 (E7.5–E8.5): the cranial neural plate elevates and generates the forebrain; the cardiac crescent coalesces to form the linear heart tube; and the anterior endoderm folds inward to form the foregut pocket.

This preprint focuses on foregut involution. At E7.5, the anterior surface endoderm is a salt-and-pepper mosaic of 1) embryonic visceral endoderm (emVE) cells derived from the primitive endoderm and 2) definitive endoderm (DE) cells derived from the epiblast, which ingress through the primitive streak (PS) during gastrulation and intercalate into the emVE. This squamous endodermal sheet rapidly involutes and extends inward to give rise to the anterior gastrointestinal organs.

Key Findings

1. Long-term live imaging of E7.5 mouse embryos reveals a stereotyped sequence of foregut involution

Using lightsheet and spinning-disc microscopy, the authors live-imaged the full process of foregut involution from E7.5 to ~E8 and described its morphological sequence.

Quantitatively, spherical-harmonics-based analysis confirmed that the involution process is highly consistent between embryos.

Fig. 1 Long-term live imaging of E7.5 mouse embryos using light-sheet microscopy captured a stereotypical sequence of involution. Figure adapted from the preprint where it is available under a CC-BY-NC 4.0 International license.

2. Actomyosin contractility appears not required for initiating involution

The authors reported intense cortical actin condensation along the central midline of the embryo, especially in the forming pocket.

However, immunostaining did not show high levels of phosphorylated myosin light chain accompanying this actin enrichment. In addition, treatment with the myosin inhibitor blebbistatin or the ROCK inhibitor Rockout did not prevent the initiation of folding.

3. Foregut involution is spatiotemporally correlated with a peak of cell death in the embryonic visceral endoderm (emVE) lineage, yet is not dependent on it

Tracking in the live-imaging data showed a peak of cell death in the anterior endoderm coinciding with foregut involution. More detailed inspection showed that dying cells assemble a contractile actomyosin ring at their apex, and apoptotic bodies are subsequently extruded bidirectionally toward both the apical and basal sides, with a basal bias.

TUNEL assays and cleaved-caspase-3 (cCasp3) immunostaining in emVE-labelled embryos confirmed that these cell deaths are largely restricted to emVE lineage cells.

Treatment with a caspase inhibitor or emVE-specific overexpression of the anti-apoptotic factor BCL-2 effectively blocked emVE cell death and resulted in the formation of multilayered emVE cell aggregates across the anterior endoderm; however, foregut involution was not greatly disrupted.

Fig. 2 Blocking emVE-specific cell deaths resulted in formation of emVE cell aggregates but did not significantly disrupt foregut morphogenesis. Figure adapted from the preprint where it is available under a CC-BY-NC 4.0 International license.

4. Transcriptional profiles show clear distinctions between emVE/DE lineages and anterior/posterior positions

The authors performed bulk RNA-seq on embryos carrying both emVE (Afp-kGFP) and pan-endoderm (Foxa2-eGFP) markers, which allowed identification and FACS sorting of emVE and DE cells.

PCA separated emVE and DE cells into distinct clusters and further resolved anterior and posterior populations within both lineages.

Among the differentially expressed genes, the authors highlighted upregulation of morphogenesis- and cell-adhesion-related genes in anterior DE, and downregulation of cell-cycle genes in anterior emVE.

5. Detailed characterization of invagination defects following VE-specific Bmp2 deletion

Bmp2 knockout in VE cells has been reported to result in anterior disorganization. Bulk RNA-seq confirmed Bmp2 expression in emVE, as well as the presence of BMP receptors and canonical SMADs, with Acvr1/Bmpr2 and Smad5 showing positional and stage-dependent correlations, respectively, potentially linking them to the anterior-specific function of BMP2.

The authors live-imaged Bmp2 VE-KO embryos, showing that the majority managed to form a shallow foregut pocket or indentation, but at an aberrant position and with delayed timing. Subsequent ventral folding lost central symmetry and was accompanied by aberrant folding and buckling of the neural headfolds.

The authors noted that the position of the neural headfolds still correlated with the position of lateral endodermal involutions, and that failed central foregut pocket formation was also concurrent with perturbed positioning of the heart field. They therefore proposed that these structures are coupled and act as a single mechanical unit.

In addition, notochord morphogenesis was severely perturbed in Bmp2 VE-KO embryos, with the notochord becoming wider and showing compromised continuity. Spatial restriction of emVE apoptosis was also lost in Bmp2 VE-KO embryos, with cell death occurring broadly across the anterior endoderm rather than being restricted to the central proximal region.

Fig. 3 Bmp2 VE-KO embryos showed aberrant positioning and progression of involution, accompanied by disruption of notochord morphology and widespread cell deaths across endoderm. Figure adapted from the preprint where it is available under a CC-BY-NC 4.0 International license.

Questions to the authors

1. I find it rather surprising that compromising actomyosin contractility does not prevent the initiation of folding. Do you have any hypotheses about what the key driver of foregut invagination might be? For example, do the deformation patterns visible in your live-imaging data provide any clues about the tissue-level forces or strains involved?
2. The tight correlation between the neural headfolds and foregut is very interesting. Do you have any indications of how the endoderm and ectoderm might be mechanically coupled during foregut involution — for example, through ECM distribution/architecture or coordinated versus relative tissue movements? Based on the morphogenetic dynamics you observe, which tissue do you think is most likely to provide the primary driving force?
3. Avian and reptile embryos also undergo large-scale anterior folding that shapes the forebrain, heart, and foregut at roughly equivalent developmental stages. However, these embryos begin from a flat, multilayer disc shape, in contrast to the cup-shaped mouse embryo. Do you think the underlying mechanisms coordinating these morphogenetic movements are likely to be conserved across species, or might the different initial geometries require fundamentally different mechanical strategies?