Temperature-dependent vitamin D signaling regulates developmental trajectory associated with diapause in an annual killifish

Abstract

The mechanisms that integrate environmental signals into developmental programs remain largely uncharacterized. Nuclear receptors (NRs) are ligand-regulated transcription factors that orchestrate the expression of complex phenotypes. The vitamin D receptor (VDR) is an NR activated by 1α,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃], a hormone derived from 7-dehydrocholesterol (7-DHC). VDR signaling is best known for regulating calcium homeostasis in mammals, but recent evidence suggests a diversity of uncharacterized roles. In response to incubation temperature, embryos of the annual killifish Austrofundulus limnaeus can develop along two alternative trajectories: active development and diapause. These trajectories diverge early in development, from a biochemical, morphological, and physiological perspective. We manipulated incubation temperature to induce the two trajectories and profiled changes in gene expression using RNA sequencing and weighted gene coexpression network analysis. We report that transcripts involved in 1,25(OH)₂D₃ synthesis and signaling are expressed in a trajectory-specific manner. Furthermore, exposure of embryos to vitamin D₃ analogs and Δ4-dafachronic acid directs continuous development under diapause-inducing conditions. Conversely, blocking synthesis of 1,25(OH)₂D₃ induces diapause in A. limnaeus and a diapause-like state in zebrafish, suggesting vitamin D signaling is critical for normal vertebrate development. These data support vitamin D signaling as a molecular pathway that can regulate developmental trajectory and metabolic dormancy in a vertebrate. Interestingly, the VDR is homologous to the daf-12 and ecdysone NRs that regulate dormancy in Caenorhabditis elegans and Drosophila We suggest that 7-DHC-derived hormones and their associated NRs represent a conserved pathway for the integration of environmental information into developmental programs associated with life history transitions in animals.

Keywords: dormancy; life history; nuclear receptors; phenotypic plasticity.

Fig. 1.

WGCNA of RNAseq data in embryos of A. limnaeus developing along alternative developmental trajectories. To identify trajectory-specific gene expression patterns, embryos were reared 20 °C or 30 °C to induce the diapause or escape trajectories, respectively. Embryos (n = 3 groups of 20 embryos) were sampled at six morphological stages that bracket the temperature-sensitive window where developmental trajectory is determined: DC, dispersed cell stage; NK, neural keel stage; 6S, 6-somite embryo; 10S, 10-somite embryo; 16S, 16-somite embryo; 20S, 20-somite embryo; 24S, 24-somite embryo. (A) Heat map of module eigengene (based on weighted average of module gene expression) correlations to each stage of development along the diapause trajectory (positive is red, and negative is white). (Upper) Modules are significantly associated (r > 0.50; P value < 10⁻²) with a developmental stage. (Lower) Modules are not significantly associated. Modules with asterisks are not preserved between diapause and escape networks (Z_summary statistic < 10). (B) Heat maps of expression level reported as fragments per kilobase of transcript per million mapped reads (FPKM) for the light yellow module that contains a daf-36−like cholesterol desaturase (LOC106533739) and IL-1β as hub genes. (C) Cytoscape (version 3.5.1) visualization of the light yellow module, with yellow circles representing the genes and the connecting lines representing similarity strengths (thickness) (44), based on topological overlap determined by WGCNA.

Fig. 2.

Vitamin D₃ signaling gene transcripts in A. limnaeus. (A) Heat maps of expression levels (FPKM) for vitamin D₃ signaling genes in each developmental phenotype. Mammalian homologs are identified by gene symbol on the left. Text color for each gene symbol represents WGCNA module origin. Clustering was accomplished using uncentered Pearson correlation with complete linkage in Gene Cluster 3.0 (45, 46). Heat maps were generated with Java Treeview 1.1.6r4 using median-centered FPKM values (47). (B) Heat map showing significant FDR adjusted P values (FDR = 10%, P < 0.05, calculated in DESeq2) for transcripts differentially expressed in the two phenotypes; DC, dispersed cell stage; NK, neural keel stage; 6S to 24S indicates the number of somite pairs in each subsequent stage.

Fig. 3.

Vitamin D₃ analogs and their effect on developmental phenotype in embryos of A. limnaeus. (Left) Line graphs depict the transcript expression level of each enzyme critical for synthesis of active vitamin D₃ (FPKM values ± SEM). Gray shading indicates the critical window of development where temperature affects developmental trajectory. Asterisks identify differentially expressed genes in the two phenotypes (significant P values < 0.05 with FDR: *P, 10%; ***P, 1%). (Lower Left) The bar chart depicts levels of 1,25(OH)₂D₃ in 20- to 24-somite embryos developing at 20 °C or 30 °C (n = 3 groups of 20 embryos). The gray bars are from embryos incubated at 25 °C in the dispersed cell stage (DC), or in diapause (DII). (Middle) The canonical biochemical pathway for synthesis of 1,25(OH)₂D₃ is depicted. (Right) The bar graphs depict the effect of exogenous treatment of embryos with precursors or analogs of vitamin D₃ on the developmental phenotype of embryos incubated under conditions that should favor entrance into diapause. Embryos were incubated in solution with 1% DMSO or 1% EtOH, depending on the analog used. Control embryos were incubated in embryo medium, or with 1% DMSO, or 1% EtOH. Embryos were scored as diapause or escape by 38 dpf using morphological and physiological criteria as previously described (7); 12 to 24 embryos were used for each treatment and were incubated in individual wells of a 96-well plate with 150 μL of media that was changed daily. Each treatment has been validated several times with different batches of embryos.

Fig. 4.

Δ4-dafachronic acid induces the escape trajectory in A. limnaeus embryos. (A) Δ4-dafachronic acid causes embryos of A. limnaeus to follow the escape trajectory under conditions (25 °C) that should favor entrance into diapause. (B) Developmental rate and pattern of escape embryos induced by dafachronic acids are indistinguishable from normal escape embryos. However, embryos that follow the diapause trajectory despite treatment with dafachronic acids have a delayed development. Symbols are means ± SEM (n = 12 to 24). Dashed line at Wourms’ stage 33 is diapause II. dpf, days postfertilization.

Fig. 5.

Dafadine A induces diapause in A. limnaeus embryos. (A) Dafadine A causes embryos of A. limnaeus to enter diapause under conditions (30 °C) that should favor active development. (B) The effect of dafadine A on developmental trajectory can be reversed by addition of 100 pM 1,25(OH)₂D₃, although developmental rate is affected. Symbols are means ± SEM (n = 12 to 24). Dashed line at Wourms’ stage 33 is diapause II. dpf, days postfertilization.

Fig. 6.

Dafadine A induces a diapause-like state in zebrafish embryos. (A) Dafadine A arrests zebrafish development near the prim-5 stage of development and causes a phenotype that is strikingly similar to that of an A. limnaeus diapause embryo. Transfer of embryos treated for 48 h to dafadine-free media leads to a resumption of normal development. Symbols are means ± SEM (n = 3 groups of 20 embryos for normal development and 9 to 12 for recovery). (B) Images of zebrafish embryos exposed to control (1% DMSO) or 20 μM dafadine A. Zebrafish embryos were maintained at 28.5 °C and were initially exposed to dafadine A at the 1,000-cell stage. dpf, days postfertilization; hpf, hours postfertilization; Rec, recovery from dafadine A exposure. (Magnification, 40×.)