Changes in regeneration-responsive enhancers shape regenerative capacities in vertebrates

Abstract

Vertebrates vary in their ability to regenerate, and the genetic mechanisms underlying such disparity remain elusive. Comparative epigenomic profiling and single-cell sequencing of two related teleost fish uncovered species-specific and evolutionarily conserved genomic responses to regeneration. The conserved response revealed several regeneration-responsive enhancers (RREs), including an element upstream to inhibin beta A (inhba), a known effector of vertebrate regeneration. This element activated expression in regenerating transgenic fish, and its genomic deletion perturbed caudal fin regeneration and abrogated cardiac regeneration altogether. The enhancer is present in mammals, shares functionally essential activator protein 1 (AP-1)-binding motifs, and responds to injury, but it cannot rescue regeneration in fish. This work suggests that changes in AP-1-enriched RREs are likely a crucial source of loss of regenerative capacities in vertebrates.

Fig. 1.. Evolutionary changes and maintenance of the cis-regulome of regeneration in teleosts.

(A) The regenerating caudal fins at 1dpa and a phylogenetic tree showing the evolutionary relationships between the African killifish and zebrafish. Scale bar, 200 μm. (B) A tail blastema, indicated by the white arrow, forms by 1 dpa in both killifish and zebrafish. E-cadherin (ECAD) labels epithelial cells (red). Scale bar, 50 μm. (C) Heatmaps of regeneration-responsive H3K27ac peaks (non-promoter regions) in killifish and zebrafish. (D) Venn diagram showing 310 overlapping genes regulated by regeneration-responsive enhancers (RREs) between killifish and zebrafish. All the genes have H3K4me3-marked active promoters, which do not overlap with H3K27ac-defined RREs. P=6.2e-98 (hypergeometric test). (E) Large variations in the total number of regeneration-responsive genes (> 1.5-fold or < −1.5-fold, FDR < 0.01) between killifish and zebrafish. (F) Heatmaps of H3K4me3 peaks linked to the 528 shared genes at 1dpa. (G) A conserved regeneration response program (RRP) is composed of 49 RRE-regulated genes (red) with H3K4me3-marked active promoters and elevated gene expression. Three genes with known functions in zebrafish regeneration are highlighted. (H) Heatmap showing the dynamic expression of 49 RRP genes during killifish fin regeneration. (I) Heatmap showing the dynamic changes of GO terms enriched by 528 shared up-regulated genes during killifish regeneration. The top GO terms are highlighted in red.

Fig. 2.. The regeneration response program deployed by regeneration-specific cells is dysregulated in regeneration-incompetent animals.

(A) t-SNE plot showing 13 different cell clusters identified in early killifish regeneration (KR). 7208 cells were included in the analyses. (B) t-SNE plot showing 16 different cell clusters identified in early zebrafish regeneration (ZR). 8605 cells were included in the analyses. (C) Annotation of killifish and zebrafish cell clusters. The expression of fstl1 in the early killifish blastema cells was confirmed by in-situ hybridization. White dashed line indicates the amputation site. (D) The integrated single-cell analysis between killifish and zebrafish. Left, annotation of major cell types. Right, the percentage of cells contributed by killifish and zebrafish. (E) The expression of shared and species-specific blastema marker genes identified in the integrated analysis. (F) The expression of 528 shared genes in different cell types identified by scRNA-seq. 80 genes were specifically detected in the blastema cells. 232 genes were detected in two or more cell types. (G) Examples for the expression of RRP genes in t-SNE clustered killifish cells. Only the enriched clusters were displayed for each gene. (H) Differential regulation of 12 teleost-defined RRP genes between the regenerating ear pinna in the African spiny mice A. cahirinus (blue) and the non-regenerating ear pinna in the house mice M. musculus (green). Four representative are highlighted in red.

Fig. 3.. Regeneration-activated inhba expression is mediated through a regeneration-responsive enhancer K-IEN

(A) The dynamic expression of inhba(2of2) in killifish caudal fin regeneration. Right, the expression of inhba(2of2) in blastema cells at 2dpa. (B) A regeneration-responsive enhancer marked by H3K27ac peaks (red box) at inhba locus in killifish and zebrafish. (C) Transgenic constructs examined for regeneration-dependent expression in killifish caudal fin. Top, the design of a Tol2 transgenic vector. Constructs marked with green (K1159, K-S3 and K-S4) display enhancer activity in fin tissue. K-IEN (K-S4) is the minimal enhancer. (D) Images from the transgenic reporter line K-IEN:GFP. Left, the expression of GFP at 0 and 2 dpa. Right, co-staining of GFP (green) and E-cadherin (red) on 2 dpa cryo-sections. Scale bar, 50 μm. (E) The expression of K-IEN:GFP in different types of injury. Tissues were removed a by 1-mm diameter biopsy punch. Top, the damaged regions at 0 dpa are outlined (red). Bottom, GFP expression in the damaged regions (star) at 1 dpa. (F) The expression of K-IEN:GFP in response to proximal and distal amputation. The orientation of all caudal fin images is proximal to the bottom and distal to the top. Dashed line indicates the amputation site.

Fig. 4.. The regeneration-responsive enhancer K-IEN is required for regeneration

(A-C) Acid Fucshin Orange G (AFOG) staining on cryo-sections of uninjured (A), 7 dpi (B) and 18 dpi (C) killifish hearts to detect fibrin (red) and collagen (blue). A cartoon shows the resection of killifish heart ventricle (A). Magnified views of collagen deposition (white arrowhead) in the injured region are outlined with dashed box. n=10. (D) PCNA (red) and DAPI (blue) staining on cryo-sections of uninjured (top) and 5 dpi (bottom) killifish hearts. n=10. (E) The expression of K-IEN:GFP in 7 dpi killifish hearts. Top, merge of GFP, TPM1 and DAPI. The uninjured region is marked by tropomyosin (TPM1). A magnified view of GFP is outlined with dashed box. n=5. (F) Generation of homozygous K-IEN^−/− mutants. Top, schematic diagram showing the disruption of K-IEN through CRISPR/Cas9. Bottom, PCR genotyping of a homozygous K-IEN^−/− mutant. (G) Fin regeneration is significantly delayed in K-IEN^−/− mutants. Right, quantification of the regenerated tissue at 3dpa. n=10. **P < 0.01. (H) AFOG staining on cryo-sections of K-IEN^−/− mutant hearts at 18 dpi. n=10. (I) Injury-triggered cardiomyocyte proliferation was not altered in K-IEN^−/− mutant at 5 dpi. The percentages of myocardial nuclei undergoing DNA replication (PCNA staining) at the injury site were quantified. n=10. ***P < 0.001. n.s, not significant (P > 0.05). Student’s t-test was performed in (G) and (I). Dashed line indicates the injury site.

Fig. 5.. Evolutionary changes of K-IEN activities in vertebrates.

(A) VISTA alignment of inhba loci among killifish, zebrafish and human. Red peaks represent high levels of sequence conservation, and the absence of peaks indicates no significant conservation. The killifish RRE is marked in green. Bottom, schematic diagram showing the overlap between the zebrafish H3K27ac peak and the predicated enhancer (blue). (B) The GFP expression driven by the zebrafish enhancer Z-IEN at 2dpa in killifish caudal fin. (C) The expression of Z-IEN:GFP under different types of injury in killifish caudal fin. (D) The GFP expression driven by the human enhancer H-IEN was initially detected at 3dpa (middle) in killifish caudal fin. GFP is detected in the basal epidermal cells (arrow). (E) Regeneration-dependent expression of Z-IEN:GFP at 7 dpi in killifish hearts. Dashed box, a magnified view. (F) The expression of H-IEN:GFP is present during homeostasis and is not regeneration-dependent. Dashed line indicates the injury or amputation site. Scale bar, 50 μm.

Fig. 6.. Occupancy of AP-1 binding motifs is essential for RRE activities.

(A-B) Motifs enriched in the conserved (A) and species-specific (B) RREs identified in killifish (top) and zebrafish (bottom) caudal fin regeneration. AP-1 motifs are highlighted in red. Each dot in the graph represents a single binding motif. The sequence of AP-1 motifs are shown in fig. S19. (C) Identification of AP-1 motifs in the RREs K-IEN and Z-IEN. (D-E) The expression of GFP driven by K-IEN^M12 and Z-IEN^M12 is abolished at 2dpa in transgenic reporter lines. Right, quantification of the fluorescence intensity between wild-type and mutant enhancers. P < 0.001 (Student’s t-test). n=10. Dashed lines indicate the amputation site. (F) An RRE-based model for the loss of regenerative capacities during evolution. We propose a regenerative response to injury as the ancestral function of AP-1 motif enriched enhancers. In the course of evolution and speciation, regeneration and injury responses became dissociated from each other in some but not all enhancers. In extant species, regeneration-competent animals maintain the ancestral enhancer activities to activate both injury response and regeneration, while repurposing of ancestral enhancers in regeneration-incompetent animals led to loss of regenerative capacities.