XB-ART-55891Front Endocrinol (Lausanne) January 1, 2019; 10 194.
Opposite T3 Response of ACTG1-FOS Subnetwork Differentiate Tailfin Fate in Xenopus Tadpole and Post-hatching Axolotl.
Amphibian post-embryonic development and Thyroid Hormones (TH) signaling are deeply and intimately connected. In anuran amphibians, TH induce the spectacular and complex process known as metamorphosis. In paedomorphic salamanders, at similar development time, raising levels of TH fail to induce proper metamorphosis, as many "larval" tissues (e.g., gills, tailfin) are maintained. Why does the same evolutionary conserved signaling pathway leads to alternative phenotypes? We used a combination of developmental endocrinology, functional genomics and network biology to compare the transcriptional response of tailfin to TH, in the post-hatching paedormorphic Axolotl salamander and Xenopus tadpoles. We also provide a technological framework that efficiently reduces large lists of regulated genes down to a few genes of interest, which is well-suited to dissect endocrine regulations. We first show that Axolotl tailfin undergoes a strong and robust TH-dependent transcriptional response at post embryonic transition, despite the lack of visible anatomical changes. We next show that Fos and Actg1, which structure a single and dense subnetwork of cellular sensors and regulators, display opposite regulation between the two species. We finally show that TH treatments and natural variations of TH levels follow similar transcriptional dynamics. We suggest that, at the molecular level, tailfin fate correlates with the alternative transcriptional states of an fos-actg1 sub-network, which also includes transcription factors and regulators of cell fate. We propose that this subnetwork is one of the molecular switches governing the initiation of distinct TH responses, with transcriptional programs conducting alternative tailfin fate (maintenance vs. resorption) 2 weeks post-hatching.
PubMed ID: 31001200
PMC ID: PMC6454024
Article link: Front Endocrinol (Lausanne)
Species referenced: Xenopus
Genes referenced: actc1 actg1 akt1 ankrd1 aqr bcl6 brca1 cdh2 chst6 crispld2 dnajb5 egfr epb41l3 erbb2 fau fos fosl2 hhipl2 hspb1 irs2 itga11 junb klf13 klf17 klf9 lama3 lamc2 lsm2 map2k1 mcm5 mcoln1 mmp11 myh7 nr4a2 pla2g7 polr2h pprc1 prdm1 psme3 ptk2b rhof smad3 smarcd3 sox4 thra thrb tmprss4 tnn tpm2 ulk4 wnt10a znf395
Article Images: [+] show captions
|Figure 1. Biological context. The High Thyroid Hormone Period (HTP), that marks the end of tadpole stage (blue rectangle) in X. tropicalis, ignites metamorphosis and the resorption of larval tissues (such as tailfin). In Axolotl, the HTP does not induces tailfin resorption. “pre-HTP” animals will refer to class 3 Axolotl [as defined by Rosenkilde et al. (23)], when limb buds start growing (around 2 weeks post-hatching). At this stage, endogenous TH level is low and thyroid gland starts releasing TH. “mid-HTP” animals refer to class 8 Axolotl, with four toes on hind limbs (between 32 and 48 days post-hatching). This stage corresponds to the highest level of endogenous T4. “post-HTP” animals refer to class 12 Axolotl (around 3 months old), where T4 endogenous level dropped significantly. X. tropicalis tadpoles were staged according to the normal table of Xenopus laevis (Daudin) of Nieuwkoop and Faber (24). The TH levels are schematized from data of Leloup and Buscaglia (25) for Xenopus and of Rosenkilde et al. (23) for Axolotl. Digital paintings were carried out with BLENDER v2.8b.|
|Figure 2. Experimental design. (A) Experimental setup. Axolotl and X. tropicalis tadpoles were treated with T3 for 24 h and gene expression was measured by RNA-Seq, followed by system biology analysis of gene networks and developmental profiling of gene expression. Animal paintings not to scale. Tailfin is highlighted in green. “pre-HTP” animals will refer to 2 weeks post-hatching Axolotl when endogenous TH level is low and 2 weeks before the highest level of endogenous TH (23). X. tropicalis tadpoles were staged according to Nieuwkoop and Faber (24). Digital paintings were carried out with BLENDER v2.8b. (B) Data processing workflow.|
|Figure 3. T3 regulates different gene sets in Axolotl and X. tropicalis at the post-embryonic transition. (A) Overlap between differentially expressed genes in both species. (B) Heatmap of differentially expressed genes in Axolotl (A.m) and X. tropicalis (X.t). (C) Log2 ratio of TH-induced gene expression changes. (D) Gene ontology analysis. Top: Number of genes for each GO term (not shown), on both species. Blue: Number (Nbr) of terms found in Axolotl. Red: Number of terms found in X. tropicalis. Bottom: ratio of the number of terms found in both species for each GO term (in the same order as the top panel). Positive and negative values correspond to terms mostly found in Axolotl or X. tropicalis gene set, respectively.|
|Figure 4. T3 affect a similar network of pathways in both species, despite regulating different sets of genes. Networks of KEGG pathways affected in Axolotl (A) and X. tropicalis (B). The reconstructed network for Axolotl is composed of 3,305 nodes and 12,776 edges, with a densely connected component (2,274 nodes and 12,554 edges), some weakly connected genes (273 genes and 222 edges) and a set of singletons (758 nodes). The X. tropicalis network is composed of 3,561 nodes and 16,237 edges, with a highly interconnected component (2,443 nodes and 15,950 edges), some weakly connected genes (318 genes and 287 edges) and a set of singletons (800 nodes). Nodes correspond to gene products, linked together by the functional interactions described in the pathways (edges). Individual node size is proportional to the number of nodes connected to it. Large nodes thus correspond to hubs between KEGG pathways. Red: nodes in common to both networks. Blue: nodes only found in one (or the other) network. Layout computed with the prefuse force directed algorithm. (C) Overlap between the node (gene product) content of the two networks. (D) Cumulative distribution of node connectivity (degree). In both species, T3 do not target (or avoid) specific network components.|
|Figure 5. Differential gene expression at the Actg1-Fos subnetwork, in Axolotl and X. tropicalis. (A) Axolotl subnetwork. (B) X. tropicalis subnetwork. The subnetworks are composed of the first (laid out in circle) and second neighbors of Actg1 and Fos nodes. Hubs (nodes with degree >20) are shown with rounded squares. Node size is proportional to their degree (connectivity). Colors indicate differentially expressed genes (red: induced, blue: repressed). Homologous nodes are located at the same place in both networks. (C) RT-qPCR analysis of DE genes in the Axolotl subnetwork. (D) RT-qPCR analysis of DE genes in the X. tropicalis subnetwork. Statistical significance (Mann-Whitney test) with *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.|
|Figure 6. Developmental time course of gene expression. Normalized gene expression changes (log2 Fold Change) before, during or after the endogenous peak of TH, corresponding to pre-, mid-, and post-HTP animals. Statistical significance (Mann-Whitney test) with *p ≤ 0.05.|
|Figure 7. Axolotl tailfin transcriptional responses to T3 at pre-HTP and paedomorph stages. (A) Overlap between T3-responsive gene sets at pre-HTP and 6 months old paedomorph stages, measured by RNA-Seq. (B) Heatmap of differentially expressed genes at pre-HTP and paedomorph stages. (C) Expression fold changes at pre-HTP vs. old paedomorphs. (D) Gene ontology analysis. Top: Number of genes for each GO term (not shown), at both stages. Red: Number (Nbr) of terms found at pre-HTP. Blue: Number of terms found in paedomorph. Bottom: ratio of the number of terms found at each stage for each GO term (in the same order as the top panel). Positive and negative values correspond to terms mostly found at pre-HTP or paedomorph gene set, respectively. (E) RT-qPCR normalized gene expression changes (log2 Fold Change) after T3 treatment. Tailfin transcriptional response to T3 differs between class 3 larvae and 6 months old paedomorphs. Statistical significance (Mann-Whitney test) with *p ≤ 0.05.|
References [+] :
Anders, Differential expression analysis for sequence count data. 2011, Pubmed