FGF signalling plays similar roles in development and regeneration of the skeleton in the brittle star Amphiura filiformis

Abstract

Regeneration as an adult developmental process is in many aspects similar to embryonic development. Although many studies point out similarities and differences, no large-scale, direct and functional comparative analyses between development and regeneration of a specific cell type or structure in one animal exist. Here, we use the brittle star Amphiura filiformis to characterise the role of the FGF signalling pathway during skeletal development in embryos and arm regeneration. In both processes, we find ligands expressed in ectodermal cells that flank underlying skeletal mesenchymal cells, which express the receptors. Perturbation of FGF signalling showed inhibited skeleton formation in both embryogenesis and regeneration, without affecting other key developmental processes. Differential transcriptome analysis finds mostly differentiation genes rather than transcription factors to be downregulated in both contexts. Moreover, comparative gene analysis allowed us to discover brittle star-specific differentiation genes. In conclusion, our results show that the FGF pathway is crucial for skeletogenesis in the brittle star, as in other deuterostomes, and provide evidence for the re-deployment of a developmental gene regulatory module during regeneration.

Keywords: Biomineralization; Echinoderm; Regulatory networks; Signalling; Vegf.

Fig. 1.

Expression of FGF signalling components in embryos and early regenerating arm stages of A. filiformis. (A) Top: WMISH on embryos at blastula, mesenchyme blastula and gastrula stages of development showing expression of Afi-fgf9/16/20, Afi-fgfr1 and Afi-fgfr2. Bottom: schematic of major relevant cellular domains in corresponding stage embryos. (B) Top: WMISH on regenerates at stages 3, 4 and 5 showing the expression of Afi-fgf9/16/20, Afi-fgfr1 and Afi-fgfr2. Insets show detail of expression patterns. Bottom: Schematic of major relevant cellular domains at corresponding stages. Ce, coelomic epithelium; E, epidermis; Mu, metameric units; Rwc, radial water canal; S, skeletogenic cells in dermal layer. Images from the aboral view. Scale bars: 50 μm (A); 100 μm (B).

Fig. 2.

FGF signalling perturbation using the SU5402 inhibitor in brittle star embryos. (A) Experimental procedure for SU5402 treatment. (B) Phenotypic analysis of SU5402-treated A. filiformis embryos and controls at 58 hpf and 4 days post fertilization shows that perturbation of FGF signalling results in embryos with no skeletal spicules forming. Numbers at the bottom show counts for embryos observed with the represented phenotype/total embryos counted. Scale bars: 50 μm.

Fig. 3.

FGF signalling perturbation using SU5402 in brittle star regenerating arm explants. (A) Experimental procedure for SU5402 treatment. (B) Phenotypic analysis of SU5402-treated A. filiformis regenerates and controls at 24 h post treatment (stage 3) shows that perturbation of FGF signalling inhibits spicule formation. Insets show detail of spicules. Numbers at the bottom show counts for explants observed with the represented phenotype/total explants counted. Red line, amputation plane. Dashed lines, outline of regenerating bud. Scale bars: 50 μm.

Fig. 4.

FGF signalling perturbation interferes with arm regeneration in A. filiformis but not by reducing cell proliferation. (A) Phenotypic analysis of regenerating arm explants in control and SU5402 conditions at 24 h post treatment (hpt) and 48 hpt shows that skeletogenic spicules do not form and the arm ceases to regenerate further. Newly formed skeletal spicules are labelled by calcein in green. Arrows indicate spicules. (B) Confocal images of an EdU cell proliferation assay on control and treated regenerates shows no changes in the proportion of EdU-labelled nuclei in SU5402-treated explants both at 24 hpt and 48 hpt. (C) Quantification of the results in B showing no significant decrease in the proportion of EdU-labelled nuclei relative to all nuclei counted. Error bars show s.e.m. Scale bars: 100 μm.

Fig. 5.

Differential transcriptomic analysis and WMISH for SU5402-treated and control embryos. (A) MA-plot showing upregulated (top) and downregulated (bottom) genes in response to SU5402 treatment. (B) WMISH on embryos treated with SU5402 that were fixed at gastrula stage. Afi-αcoll was used as negative control and no change in expression was observed. Afi-egr, Afi-slc4a10, Afi-tetraspanin 19 (Afi-ttrspn19), Afi-msp130L and Afi-tr9107 are downregulated and Afi-alx/arx is upregulated in SU5402-treated samples. Embryos are all oriented with apical pole at the top and vegetal pole at the bottom. (C) Box plot summarizing differential gene expression in SU5402-treated embryos relative to DMSO showing consistency between transcriptome, qPCR and NanoString quantification strategies represented as log2(SU5402/DMSO). Scale bars: 50 μm.

Fig. 6.

Comparison of genes affected by FGF signalling perturbation in embryos and regenerating arms of A. filiformis. Boxplot of selected genes showing the median and data distribution (box, interquartile range; whiskers, maximum and minimum expression value; horizontal line, median) of gene quantification obtained in SU5402-treated embryos (grey) and regenerates (yellow) relative to DMSO controls, from at least three biological replicas. The relative abundance is expressed in log2(SU5402/DMSO) and threshold is set at ±1 corresponding to 2-folds of difference (grey horizontal line). Genes have been divided in functional categories: CC, cell cycle; S, signalling; SDU, skeletogenic downstream and unknown; TF, transcription factors. Stars under a gene indicate very low level of expression in control embryos (yellow; see Table S2) or in regenerating arms (grey; see Table S3).

Fig. 7.

Role of FGF signalling in skeletal cells in embryos and adult regenerating arms of A. filiformis. (A) Top: skeletal cellular arrangement in a gastrula embryo and a stage 3/4 regenerating arm when biomineralized skeleton is deposited. In both cases mesenchymal cells (red) adjacent to ectodermal/epidermal cells (blue) secrete the biomineralized skeleton (green) in the extracellular space. Bottom: representation of the signalling occurring from ectodermal/epidermal to mesenchymal cells. (B) Left: hypothetical gene regulatory network for skeletal cells built with data coming from this work and previous publications (Dylus et al., 2016, 2018; Czarkwiani et al., 2013, 2016). Genes are colour-coded and are represented by their cis-regulatory control system: green are orthologues of genes known to be essential in the biomineralization process in sea urchin; genes of unknown function but known expression domain are in black. Genes are connected by functional linkages, which are either inferred (dashed lines) or confirmed (solid lines) in this study. Arrows indicate positive inputs (activation) and barred lines indicate negative inputs (repression). Open circles represent post-transcriptional/biochemical interactions occurring in the cytoplasm (phosphorylation of the FGF and VEGF receptors upon binding to the ligand and the complex intracellular cascade of signalling events). Right: representation of the same network in the presence of an FGF signalling inhibitor (SU5402). Downregulated genes are shown in shaded colours. E indicates linkages present only in developing embryos and A indicates linkages present only in regenerating arms.