A chemo-mechanical model of endoderm movements driving elongation of the amniote hindgut

Abstract

While mechanical and biochemical descriptions of development are each essential, integration of upstream morphogenic cues with downstream tissue mechanics remains understudied in many contexts during vertebrate morphogenesis. A posterior gradient of Fibroblast Growth Factor (FGF) ligands generates a contractile force gradient in the definitive endoderm, driving collective cell movements to form the hindgut. Here, we developed a two-dimensional chemo-mechanical model to investigate how mechanical properties of the endoderm and transport properties of FGF coordinately regulate this process. We began by formulating a 2-D reaction-diffusion-advection model that describes the formation of an FGF protein gradient due to posterior displacement of cells transcribing unstable Fgf8 mRNA during axis elongation, coupled with translation, diffusion, and degradation of FGF protein. This was used together with experimental measurements of FGF activity in the chick endoderm to inform a continuum model of definitive endoderm as an active viscous fluid that generates contractile stresses in proportion to FGF concentration. The model replicated key aspects of hindgut morphogenesis, confirms that heterogeneous - but isotropic - contraction is sufficient to generate large anisotropic cell movements, and provides new insight into how chemo-mechanical coupling across the mesoderm and endoderm coordinates hindgut elongation with outgrowth of the tailbud.

Application of tissue-scale tension to avian epitheliain vivoto study multiscale mechanics and inter-germ layer coupling

“…Embryos were harvested onto filter paper rings and transferred to plates containing EC culture medium (Chapman et al, 2001). Endoderm-specific electroporation was carried out as previously described (Nerurkar et al, 2019; Oikonomou et al, 2023). Briefly, electroporation solutions were prepared by diluting pCAG-H2B-EGFP, pCAG-EGFP-CAAX (Addgene #86056), pCAG-myr-mRFP (Addgene #32604), or pCAG-DeAct-SpvB (a fusion protein of DeAct-SpvB with EGFP, subcloned into pCAG from Addgene #89446) plasmids to 3.5 µ g/ µ L in molecular grade ddH2O with 5% sucrose and 0.1% Fast Green FCF.…”

“…Briefly, electroporation solutions were prepared by diluting pCAG-H2B-EGFP, pCAG-EGFP-CAAX (Addgene #86056), pCAG-myr-mRFP (Addgene #32604), or pCAG-DeAct-SpvB (a fusion protein of DeAct-SpvB with EGFP, subcloned into pCAG from Addgene #89446) plasmids to 3.5 µ g/ µ L in molecular grade ddH2O with 5% sucrose and 0.1% Fast Green FCF. Following delivery of the DNA solution to the ventral surface by micropipette, embryos were electroporated using a Nepa 21 transfection system (Nepa Gene, Ichikawa City, Japan) with a sequence consisting of three 35V poring pulses of 0.2 msec duration separated by 50 msec with a decay rate of 10% between successive pulses, followed by five 4V transfer pulses of 5.0 msec duration separated by 50.0 msec with a 40% decay rate (Nerurkar et al, 2019; Oikonomou et al, 2023). Experiments were carried out after 6-8 hours incubation ex ovo, when embryos had reached HH stage 13.…”

“…Image analyses on time series movies obtained from the mechanical testing experiments were performed using the open-source image visualization Python library Napari (Sofroniew et al, 2022), preprocessing functions from the sci-kit image library (van der Walt et al, 2014), the segmentation algorithm Cellpose (Stringer et al, 2020) (starting from weights of the cyto2 model, trained with manual annotations for 1000 epochs on a workstation equipped with an NVIDIA T400 GPU), and btrack, a Python library for multi-object Bayesian tracking (Ulicna et al, 2021) (particle configuration). Where necessary, time lapse videos were stabilised against drift using custom code as in (Oikonomou et al, 2023), inspired by Fast4Dreg (Pylvänäinen et al, 2023). Cell morphometry features generated through this pipeline were aggregated across biological replicates and plotted in UMAP space, generally following the approach of cellPLATO (Shannon et al, 2023) and adapting code from ColabTracks (Jacquemet, 2024).…”