How tissue regeneration programs are triggered by injury has received limited research attention. Here we investigate the existence of enhancer regulatory elements that are activated in regenerating tissue. Transcriptomic analyses reveal that leptin b (lepb) is highly induced in regenerating hearts and fins of zebrafish. Epigenetic profiling identified a short DNA sequence element upstream and distal to lepb that acquires open chromatin marks during regeneration and enables injury-dependent expression from minimal promoters. This element could activate expression in injured neonatal mouse tissues and was divisible into tissue-specific modules sufficient for expression in regenerating zebrafish fins or hearts. Simple enhancer-effector transgenes employing lepb-linked sequences upstream of pro- or anti-regenerative factors controlled the efficacy of regeneration in zebrafish. Our findings provide evidence for ‘tissue regeneration enhancer elements’ (TREEs) that trigger gene expression in injury sites and can be engineered to modulate the regenerative potential of vertebrate organs.
At a glance
The capacity for complex tissue regeneration is unevenly distributed among vertebrate tissues and species. Salamanders and zebrafish possess remarkable potential to regenerate tissues like amputated appendages, resected heart muscle, and transected spinal cords1, 2. Investigations of gene expression and function have generated molecular models for regeneration in multiple contexts, yet there is a gap to be filled in our understanding of the regulatory events that activate tissue regeneration programs1, 2, 3, 4, 5.
Recent genome-wide chromatin analyses suggest that gene regulatory elements comprise a substantial portion of genomic sequence. Of these elements, distal-acting regulatory sequences, or enhancers, represent the most abundant class6, 7. Enhancers can direct expression of their target genes and have been predominantly examined as a means for stage- and tissue-specific regulation during embryonic development8, 9. Studies have also implicated enhancers in disease and as targets during evolution10, 11, 12, 13, 14, 15. Such findings raise the possibility that enhancer elements may also exist that engage with transcription factors in response to tissue damage to regulate genetic programs for regeneration. The identification of such elements could potentially inspire new solutions for manipulating regenerative events.
leptin b induction during fin and heart regeneration
To identify genes that are induced during tissue regeneration, we collected RNA from uninjured and regenerating tissues of adult zebrafish and sequenced transcriptomes. Our analyses identified 2,408 genes with significantly higher expression in tail fins at 4 days post-amputation (dpa), and 859 genes with significantly higher expression in cardiac ventricles 7 days after induced genetic ablation of half of all cardiomyocytes (Extended Data Fig. 1a and Supplementary Tables 1 and 2). In total, 360 genes were induced twofold or greater in both tissues compared to uninjured tissues (Extended Data Fig. 1a). Among these genes, 69 were present at low levels in uninjured fins and highly induced during regeneration (Supplementary Information). The gene leptin b (lepb), one of two zebrafish paralogues related to mammalian leptin, a secreted regulator of energy homeostasis16, had the highest relative change during fin regeneration of genes in this group (130-fold; Fig. 1c, Extended Data Fig. 2, and Supplementary Information). lepb transcripts were rare or undetectable in uninjured fins by semi-quantitative or quantitative RT–PCR (qPCR) or in situ hybridization (ISH), but induced in the regeneration blastema by 1 dpa (Extended Data Fig. 1b–d). Upon local injury of the cardiac ventricle by partial resection, lepb expression was induced in the endocardium, the endothelial lining of inner myofibres that has been implicated in regenerative events (Extended Data Fig. 1b, c, e)17, 18.
To capture the regulatory elements responsible for lepb induction, we replaced the first exon of lepb with an eGFP reporter transgene within a 150 kb BAC containing 105 kb of DNA sequence upstream of the start codon (Fig. 1d). Transgenic lepb:eGFP larvae had little or no detectable eGFP as viewed under a stereofluorescence microscope, and no fluorescence was detectable in fins or hearts throughout life (Fig. 1e and Extended Data Fig. 1h, i, l, m). Upon fin amputation, lepb:eGFP fluorescence was sharply induced in regenerating structures, where fluorescence localized to blastemal mesenchyme (Fig. 1e and Extended Data Fig. 1j, k). lepb:eGFP was also induced in wounds of resected ventricles, as well as in atrial tissue distant from the site of injury (Fig. 1g), a signature observed with other injury-induced markers17, 18. While sparse lepb:eGFP could be detected in epicardial tissue at 1 day post-resection (dpa; data not shown), cardiac lepb:eGFP fluorescence was predominantly endocardial by 3 dpa (Fig. 1g, h). Thus, sequences within a ~150 kb genomic region surrounding lepb direct regeneration-dependent expression in fin and cardiac tissues.
LEN-associated expression in injured mouse tissues
Analysis of regions upstream of leptin genes in murine and human genomes revealed limited primary sequence conservation of LEN (Extended Data Fig. 4e). This sequence divergence likely reflects rapid evolution of enhancers, reported in previous studies21, 22. To examine whether zebrafish LEN has activity in mammalian injury contexts, we fused it upstream of a construct containing a murine minimal hsp68 promoter and a lacZ reporter gene. We generated two stable lines, one of which displayed vascular endothelial X-gal staining in uninjured neonatal hearts and paws (Extended Data Fig. 6b). A second line had a small number of X-gal-positive cells in uninjured neonatal tissues and was selected for injury studies (LEN-hsp68::lacZ) (Fig. 3a, b). Neonatal digit tips amputated at P2 phalanges do not regenerate lost structures effectively23, whereas injured neonatal ventricles display a regenerative response24. Strikingly, amputated digit tips and damaged ventricles of all injured postnatal day 1 LEN-hsp68::lacZ neonates showed conspicuous X-gal staining in wounds 3 days after surgeries. A control transgenic line with an unrelated enhancer fragment also exhibited low basal expression in uninjured neonatal tissues, but unlike LEN-hsp68::lacZ animals, showed no detectable activation of the lacZ reporter upon injury to the digits or ventricle (Fig. 3a, b and Extended Data Fig. 6a). While future tests of LEN activity using a panel of promoters and transgene integration sites will be important, overall, these results suggest that zebrafish LEN sequences can interact with mammalian transcriptional machinery to enable injury-induced expression in mice.
LEN is separable into tissue-specific modules
To identify minimal sequences responsible for the activity of LEN, we tested the ability of various fragments to direct regeneration-activated expression. We found that more distal LEN fragments composed of approximate nucleotides 1–850, 450–1000, 450–850 or 660–850 could each drive eGFP expression from the lepb 2 kb promoter during fin regeneration (Fig. 4a, b and Extended Data Fig. 7). LEN fragments generated from the distal 1 kb portion also directed eGFP expression during fin regeneration when paired with the cmlc2 promoter (Extended Data Figs 5 and 9a, b). LEN fragments 1–850 and 450–1000 did not direct detectable eGFP expression during fin regeneration from the α-cry promoter in our experiments (Extended Data Fig. 5 and 9d–f), suggesting a repressive motif in α-cry upstream sequences. Intriguingly, none of these fragments directed endocardial expression after cardiac injury, although eGFP fluorescence was occasionally observed sparsely in epicardial cells or cardiomyocytes (Extended Data Fig. 8). Conversely, more proximal LEN fragments comprising approximate nucleotides 830–1350 or 1000–1350 directed endocardial expression during heart regeneration, but did not activate eGFP fluorescence in regenerating fins (Fig. 4a, b and Extended Data Figs 7 and 8). These proximal LEN fragments also could direct regeneration-associated expression in endocardial cells from cmlc2 and α-cry promoters (Extended Data Fig. 9c, h). Thus, our analyses suggested the presence of two separate, tissue-specific enhancer modules (Fig. 4c).
We analysed sequences of the minimal 190 nucleotide (nt) (fin) and 316 nt (heart) elements, and identified distinct sets of predicted transcription factor binding motifs. LEN(663–854) contains predicted AP-1, Sox, forkhead, and ETS binding sites, and we confirmed by transgenic reporter assays that a predicted AP-1 binding site at LEN(776–782) is necessary to direct expression in regenerating fins (Extended Data Fig. 9i, j). LEN(1034–1350) contains predicted NFAT, GATA, forkhead, and ETS binding sites, motifs associated with expression in endothelial cells25, 26 (Extended Data Fig. 9i). In total, our findings indicate a composite arrangement of regulatory elements with distinct tissue preferences within the LEN regeneration enhancer.
LEN element constructs control regenerative capacity
Recent studies have described new enhancer-target gene pairings caused by chromosomal rearrangements that underlie genetic diseases like cancer and neurological disorders10, 12, 15. To examine a parallel idea for experimentally guiding tissue regeneration, we designed transgenic constructs positioning LEN and the minimal lepb promoter upstream of pro- or anti-regenerative factors. A possible outcome is that LEN would limit embryonic expression of potent developmental influences to permit maturation from the one-cell stage to adulthood, but also trigger and sustain expression of these influences upon tissue damage.
To create enhancer-effector transgenes, we took advantage of the dependency of fin regeneration on signalling by fibroblast growth factors (Fgfs)4, 27. We first positioned LEN upstream of a cDNA encoding a dominant-negative form of fgfr1 (dnfgfr1)—a potent inhibitor of embryonic development27, 28—and injected this construct into wild-type embryos. We established stable lines of zebrafish harbouring either P2:dnfgfr1 or LENP2:dnfgfr1, demonstrating that dnfgfr1 expression was limited to developmentally insignificant levels. Adult P2:dnfgfr1 fins displayed no detectable dnfgfr1 induction after amputation and regenerated normally. By contrast, injury to LENP2:dnfgfr1 animals induced strong expression of dnfgfr1 (detectable by dnfgfr1–eGFP fusion protein fluorescence) that was restricted to the amputation plane. Moreover, these animals displayed conspicuous defects or outright failures in fin regeneration (Fig. 5a, b). In some cases, fin rays failed to regenerate even by 30 dpa and maintained dnfgfr1 expression in ray stumps, indicating persistent activation of LEN in the setting of regenerative failure (Fig. 5c and Extended Data Fig. 10b).
We complemented these experiments with a gain-of-function approach, based on the discovery that mutations in the fgf20a ligand gene, devoid of blastema (dob), arrest fin regeneration4. We positioned LEN and the minimal lepb promoter upstream of a fgf20a cDNA and injected this construct into one-cell dob embryos. We generated stable lines of control dob; P2:fgf20a and dob; LENP2:fgf20a animals, indicating that these constructs restricted ectopic fgf20a expression during embryonic development. Upon amputation of adult tail fins, dob; P2:fgf20a animals induced no additional detectable fgf20a and displayed regenerative blocks comparable to dob animals (Fig. 5d, f, g). By contrast, LENP2 sequences directed broad expression of fgf20a in mesenchymal cells upon fin amputation (Fig. 5d, f, g). Remarkably, blastemal cell proliferation was stimulated in amputated dob; LENP2:fgf20a fins, and these animals regenerated patterned structures that were often of normal length (Fig. 5e–g). In some cases, the lobed pattern of the tail fin was restored, and in no cases were there uncontrolled growth phenotypes (Fig. 5g).
Targeted cardiomyocyte proliferation by LEN
Heart regeneration occurs through injury-induced stimulation of proliferation by pre-existing cardiomyocytes29. Recent evidence indicates that the secreted factor neuregulin1 (Nrg1) is a cardiomyocyte mitogen during cardiac growth or repair in lower and higher vertebrates30, 31, 32. In zebrafish, nrg1 is present at very low levels in the heart, and it is induced upon injury at levels that remain undetectable by standard ISH methodology31. Strong transgenic overexpression of nrg1 in adult zebrafish cardiomyocytes activates overt cardiomyocyte proliferation and enlarges the ventricular wall31. To test whether LEN can influence heart regeneration, we created stable transgenic zebrafish lines with P2:nrg1 or LENP2:nrg1 constructs. Resection of the ventricular apex sharply increased nrg1 transcripts in injured portions of LENP2:nrg1, but not control P2:nrg1, ventricles (Fig. 6a, b). LEN-induced nrg1 expression was strongest in 7 dpa injury sites, slightly less prominent at 14 dpa, and scarcely detectable by 30 dpa, typically when a contiguous muscle wall has regenerated (Fig. 6a). To examine effects of targeted nrg1 enhancement, we quantified cardiomyocyte proliferation indices in LENP2:nrg1 and P2:nrg1 ventricles at 14 dpa. LENP2:nrg1 injury sites had a 52% increase in cardiomyocyte proliferation compared to P2:nrg1 wounds, indicative of improved muscle regeneration (Fig. 6c, d). By 30 dpa, when nrg1 levels approached baseline, regenerated ventricular walls appeared grossly normal (Fig. 6a). Thus, LEN can be designed to deliver mitogenic factors preferentially to areas of cardiac damage, boosting injury-induced cardiomyocyte proliferation.
Here, we used a profiling approach to identify small regulatory elements that direct gene expression in regenerating tissue, which we have termed tissue regeneration enhancer elements (TREEs). Recently, a ~18 kb region of the murine Bmp5 locus was reported to activate expression from minimal promoters in injury contexts33, suggesting it may harbour a TREE analogous to the LEN element we describe here. We suspect that diverse classes of TREEs exist, including elements activated during development and re-activated by injury34 or during regeneration, elements that activate expression preferentially during regeneration in multiple tissues, and regeneration-specific elements that are more tissue-restricted. The investigation of individual binding motifs within TREEs should identify upstream transcriptional regulators of regeneration, whereas genomic TREE locations can pinpoint novel downstream target genes.
Current methodologies to interrogate regenerative biology often have experimental disadvantages like multiple transgenes, ubiquitous promoters, irreversible expression, and/or stressful stimuli like oestrogen analogues, tetracycline analogues or heat shock35. By contrast, TREEs are single-transgene systems that can naturally induce and maintain target genes upon injury, and then naturally temper expression as regeneration concludes. Whereas LEN elements induce expression in fin mesenchyme and/or endocardium, we expect that future investigations will uncover a panel of regeneration-responsive TREEs representing additional distinct tissues. Thus, when combined with effectors or genome-editing enzymes, TREEs should facilitate targeted genetic manipulations that have been elusive to this point.
Multiple features of TREEs are appealing with respect to the design of potential regenerative therapies. Previous studies have implicated the manipulation of enhancer activity as a means to treat human genetic disease12, 36. In this study, we report that pro- or anti-regenerative factors directed by TREEs are capable of blocking regenerative growth, promoting cell proliferation, or even rescuing genetic defects in regeneration. With a TREE-based system, factor delivery is spatiotemporally defined and could permit therapeutic cycles as injury recurs. Notably, although Nrg1 impacts heart regeneration, systemic neuregulin delivery has the potential for neurological or oncogenic effects37, 38. Thus, enhancer-based targeting of Nrg1 to injury sites, as we model here in zebrafish, may represent a more effective regenerative medicine platform. We suggest that TREEs identified from natural regenerative contexts across vertebrate species can inform new strategies for precise factor delivery to injured human tissues.
Zebrafish maintenance and procedures
Wild-type or transgenic male and female zebrafish of the outbred Ekkwill (EK) strain were used for all experiments, with adults ranging in age from 3 to 12 months. Water temperature was maintained at 26°C for animals unless otherwise indicated. Fins were amputated to 50% of their original length using razor blades. As penetrance of the dob mutation was higher at 33°C than at 26°C, dob fish were maintained at 33°C after caudal fin amputation. To measure lengths of regenerates, lengths from the amputation plane to the distal tips of the third and fourth fin rays of dorsal and ventral caudal fin lobes were determined using ZEN software. Because some dob animals regenerated portions of the first and second fin rays of ventral lobes, regenerating caudal fin areas for Extended Data Fig. 10c were measured from the dorsal third fin ray to the ventral third fin ray and calculated using ZEN software. Partial ventricular resection surgeries were performed as described previously39, in which ~20% of the cardiac ventricle was removed at the apex. To ablate cardiomyocytes, cmlc2:CreER; bactin2:loxp-mCherry-STOP-loxp-DTA (Z-CAT) fish were used40. Z-CAT zebrafish were incubated in vehicle (0.01% EtOH) or 10 μM tamoxifen for 12 h. Work with zebrafish was performed in accordance with Duke University guidelines.
To generate lepb:eGFP BAC transgenic animals (full names, Tg(lepb:eGFP)pd120 and Tg(lepb:eGFP)pd121), the iTol2 cassette41 was integrated into the BAC clone DKEY-21O22 using Red/ET recombineering technology (GeneBridges). Then, the first exon of the lepb gene in the BAC clone DKEY-21O22 was replaced with an eGFP cassette by Red/ET recombineering. 5′ and 3′ homology arms were amplified by PCR (Supplementary Information) and subcloned into the pCS2-eGFP plasmid. One nl of 50 ng μl−1 purified, recombined BAC was injected into one-cell stage zebrafish embryos along with one nl of 30 ng μl−1 synthetic Tol2 mRNA41. To sort F0 transgenic animals injected with lepb:eGFP constructs, fin folds were amputated at 3 or 4 dpf, and embryos displaying eGFP fluorescence near the injury site at 1 dpa were selected (Extended Data Fig. 1f). After raising F0 zebrafish to adulthood, caudal fins were amputated and zebrafish displaying induced eGFP were selected for breeding (Extended Data Fig. 1g). Between 30–60 dpf, caudal fins of progeny from transgene-positive F0 fish were amputated, and eGFP+ transgenic animals were isolated to identify stable lines. Two lines were identified that had indistinguishable expression features.
To define LEN activity, over 60 additional new transgenic lines were established in this study, listed in Supplementary Data 1. To generate transgenic animals, DNA sequences were amplified by PCR with indicated primers (Supplementary Data 3) and subcloned into a pCS2-eGFP-I-sceI vector, in which I-SceI restriction sites were flanked by a multiple cloning site. As promoters, 2 kb, 1.6 kb, and 0.7 kb upstream sequences of lepb, cmlc2 (ref. 42), and α-cry43 genes were used, respectively. These constructs were injected into one-cell-stage wild-type or dob embryos using standard meganuclease transgenesis techniques. 2 kb lepb upstream sequences could induce transgene expression after fin fold amputation at larval stages, but never after caudal fin amputation in adults. To isolate stable lines, larvae were examined for transgene expression near injury site in response to fin fold amputation (2 kb lepb), in cardiomyocytes (1.6 kb cmlc2), and in lens (0.7 kb α-cry).
To test additional TREEs, we subcloned putative enhancer regions of il11a, plek, vcana, and cd248b upstream of 800 bp of lepb upstream sequence (P0.8). To define TREE activity, these constructs were injected into one-cell-stage wild-type embryos, Fin folds were amputated at 4 dpf, and eGFP fluorescence near the amputation plane was examined at 5 dpf (1 dpa).
Generation and analysis of transgenic mice
Transgenic mice (CD-1 strain) were generated by oocyte microinjection as described previously44. LEN-hsp68::lacZ transgenic mice were generated by subcloning the zebrafish LEN enhancer sequence into the transgenic reporter plasmid hsp68-lacZ45. Ctrl-hsp68::lacZ transgenic mice harbour a transgene, Prkaa2[mMEF2(1+2)]-hsp68-lacZ, which contains a modified version of a 931-bp enhancer sequence from the mouse Prkaa2 gene cloned into hsp68-lacZ (J. Hu and B. L. Black, unpublished observations). Partial apical resection injury in male and female neonatal mice at postnatal day 1 was performed similarly to previously described methods46. Hearts and paws were collected at postnatal day 4. All experiments with mice complied with federal and institutional guidelines and were reviewed and approved by the UCSF IACUC.
RNA isolation and quantitative PCR
RNA was isolated from dissected caudal fins and partially resected ventricles using Tri-Reagent (Sigma). cDNA was synthesized from 1 μg of total RNA using the Roche First Strand Synthesis Kit. Quantitative PCR was performed using the Roche LightCycler 480 and the Roche LightCycler 480 Probes Master. All samples were analysed in biological triplicates and technical duplicates. Primer sequences are described in Supplementary Information, and probe numbers for actb2, lepb, and nrg1 were 104, 156 and 76, respectively. lepb and nrg1 transcript levels were normalized to actb2 levels for all experiments.
Total RNA was prepared from two biological replicate pools of ablated Z-CAT ventricles and uninjured ventricles at 7 days post-ablation as per Gemberling et al.31, or regenerating and uninjured caudal fins. Generation of mRNA libraries and sequencing were performed at the Duke Genome Sequencing Shared Resource using an Illumina HiSeq2000. Sequences were aligned to the zebrafish genome (Zv9) using TopHat47. Differentially regulated transcripts were identified using EdgeR and an FDR cut-off of 0.1 (ref. 48). Accession numbers for transcriptome data sets are GSE75894 and GSE76564.
To identify candidate enhancer elements activated during heart regeneration, chromatin extracts were prepared from two biological replicate pools of 10 ablated Z-CAT ventricles and 10 uninjured ventricles. Chromatin was sonicated and immunoprecipitated with an antibody against H3K27ac (ActiveMotif) using the MAGnify ChIP system (Invitrogen). Sequencing libraries were prepared as per Bowman, et al.49. Sequencing was performed using an Illumina HiSeq2000, and 10–25 million 50 bp single end reads were obtained for each library. Sequences were aligned to the zebrafish genome (Zv9) using Bowtie2 (ref. 50). Differential peaks were identified using Model-based Analysis for ChiP-seq (MACS)51.
Histology and imaging
In situ hybridization on cryosections of 4% paraformaldehyde-fixed fins was performed as described previously52. To generate digoxigenin-labelled probes for lepb and fgf20a, we generated a fragment of lepb cDNA and a full length of fgf20a cDNA by PCR using primer sequences described in Supplementary Information. The nrg1 probe was prepared as described previously31. Immunohistochemistry was performed as described previously40. Primary and secondary antibodies used in this study were: anti-myosin heavy chain (mouse, F59, Developmental Studies Hybridoma Bank), anti-MEF2 (rabbit, sc-313, Santa Cruz Biotechnology), anti-PCNA (mouse, P8825, Sigma), anti-eGFP (rabbit, A11122, Life Technologies), anti-eGFP (chicken, GFP-1020, Aves Labs), anti-Raldh2 (rabbit, Abmart), anti-Ds-Red (rabbit, 632496, Clontech), anti-p63 (mouse, 4A4, Santa Cruz Biotechnology), Alexa Fluor 488 (mouse and rabbit; Life Technologies), Alexa Fluor 594 (mouse and rabbit; Life Technologies). For EdU incorporation experiments, zebrafish were injected intraperitoneally with 10 mM EdU (A10055, sigma), and caudal fins were collected at 1 h post-treatment. EdU staining was performed as previously described53. The secondary antibody used for EdU staining was Alexa 488 azide (10–20 μM, Sigma). Whole-mount images were acquired using an M205FA stereofluorescence microscope (Leica) or Axio Zoom (Zeiss). Images of tissue sections (10 μm for hearts and 14 μm for fins) were acquired using an LSM 700 confocal microscope (Zeiss). X-gal staining to detect β-galactosidase activity and counterstaining with nuclear fast red were performed with murine tissue as described previously44.
Data collection and statistics
Clutchmates were randomized into different treatment groups for each experiment. No animal or sample was excluded from the analysis unless the animal died during the procedure. Sample sizes were chosen on the basis of previous publications and experiment types, and are indicated in each figure legend or methods. No statistical methods were used to predetermine sample size. For expression patterns, at least five fish from each transgenic line were examined. At least 9 hearts of each group were pooled for RNA purification and subsequent RT–qPCR. Quantification of cardiomyocyte proliferation and calculation of statistical outcomes were assessed by a person blinded to the treatments. Other experiments were not blinded during experiments and outcome assessment. Sample sizes, statistical tests, and P values are indicated in the figures or the legends. One-way ANOVA tests were applied when normality and equal variance tests were passed. The Mann–Whitney rank sum test was applied in assays of cardiomyocyte proliferation.
- Advances in understanding tissue regenerative capacity and mechanisms in animals. Nature Rev. Genet. 11, 710–722 (2010)
- Limb regeneration: a new development? Annu. Rev. Cell Dev. Biol. 27, 409–440 (2011) &
- Molecular basis for the nerve dependence of limb regeneration in an adult vertebrate. Science 318, 772–777 (2007) , , , &
- fgf20 is essential for initiating zebrafish fin regeneration. Science 310, 1957–1960 (2005) , , &
- Signaling networks organizing regenerative growth of the zebrafish fin. Trends Genet. 31, 336–343 (2015) &
- The ENCODE Project Consortium, E. P. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012)
- Roadmap Epigenomics Consortium Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330 (2015)
- Mechanisms of transcriptional precision in animal development. Trends Genet. 28, 409–416 (2012) , &
- Genomic views of distant-acting enhancers. Nature 461, 199–205 (2009) , &
- A large genomic deletion leads to enhancer adoption by the lamin B1 gene: a second path to autosomal dominant adult-onset demyelinating leukodystrophy (ADLD). Hum. Mol. Genet. 24, 3143–3154 (2015) et al.
- Stepwise modification of a modular enhancer underlies adaptation in a Drosophila population. Science 326, 1663–1667 (2009) , , , &
- Long-range gene regulation and novel therapeutic applications. Blood 125, 1521–1525 (2015) , , , &
- Evolving new skeletal traits by cis-regulatory changes in bone morphogenetic proteins. Cell 164, 45–56 (2016) et al.
- Convergent evolution of complex regulatory landscapes and pleiotropy at Hox loci. Science 346, 1004–1006 (2014) , , , &
- A NOTCH1-driven MYC enhancer promotes T cell development, transformation and acute lymphoblastic leukemia. Nature Med. 20, 1130–1137 (2014) et al.
- Positional cloning of the mouse obese gene and its human homologue. Nature 372, 425–432 (1994) et al.
- Translational profiling of cardiomyocytes identifies an early Jak1/Stat3 injury response required for zebrafish heart regeneration. Proc. Natl Acad. Sci. USA 110, 13416–13421 (2013) et al.
- Retinoic acid production by endocardium and epicardium is an injury response essential for zebrafish heart regeneration. Dev. Cell 20, 397–404 (2011) et al.
- Histone modifications at human enhancers reflect global cell-type-specific gene expression. Nature 459, 108–112 (2009) et al.
- Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc. Natl Acad. Sci. USA 107, 21931–21936 (2010) et al.
- Enhancer evolution across 20 mammalian species. Cell 160, 554–566 (2015) et al.
- ChIP-seq identification of weakly conserved heart enhancers. Nature Genet. 42, 806–810 (2010) et al.
- The mouse digit tip: from wound healing to regeneration. Methods Mol. Biol. 1037, 419–435 (2013) , , , &
- Transient regenerative potential of the neonatal mouse heart. Science 331, 1078–1080 (2011) et al.
- Transcriptional regulation of endothelial cell and vascular development. Circ. Res. 112, 1380–1400 (2013) , &
- Combinatorial regulation of endothelial gene expression by ets and forkhead transcription factors. Cell 135, 1053–1064 (2008) et al.
- Fgf signaling instructs position-dependent growth rate during zebrafish fin regeneration. Development 132, 5173–5183 (2005) , , , &
- Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66, 257–270 (1991) , &
- Primary contribution to zebrafish heart regeneration by gata4+ cardiomyocytes. Nature 464, 601–605 (2010) et al.
- Neuregulin stimulation of cardiomyocyte regeneration in mice and human myocardium reveals a therapeutic window. Sci. Translat. Med. 7, 281ra245 (2015) et al.
- Nrg1 is an injury-induced cardiomyocyte mitogen for the endogenous heart regeneration program in zebrafish. eLife 4, e05871 (2015) , , &
- Neuregulin1/ErbB4 signaling induces cardiomyocyte proliferation and repair of heart injury. Cell 138, 257–270 (2009) , , &
- A distinct regulatory region of the Bmp5 locus activates gene expression following adult bone fracture or soft tissue injury. Bone 77, 31–41 (2015) et al.
- C/EBP transcription factors mediate epicardial activation during heart development and injury. Science 338, 1599–1603 (2012) et al.
- The zebrafish as a model for complex tissue regeneration. Trends Genet. 29, 611–620 (2013) , , &
- Reactivation of developmentally silenced globin genes by forced chromatin looping. Cell 158, 849–860 (2014) et al.
- Neuropathologic implication of peripheral neuregulin-1 and EGF signals in dopaminergic dysfunction and behavioral deficits relevant to schizophrenia: their target cells and time window. Biomed. Res. Int. 2014, 697935 (2014) , , , &
- Neuregulins and cancer. Clin. Cancer Res. 14, 3237–3241 (2008) et al.
- Heart regeneration in zebrafish. Science 298, 2188–2190 (2002) , &
- The regenerative capacity of zebrafish reverses cardiac failure caused by genetic cardiomyocyte depletion. Development 138, 3421–3430 (2011) et al.
- Transposon-mediated BAC transgenesis in zebrafish. Nature Protocols 6, 1998–2021 (2011) , , &
- High-throughput assay for small molecules that modulate zebrafish embryonic heart rate. Nature Chem. Biol. 1, 263–264 (2005) et al.
- Suppression of lens growth by αA-crystallin promoter-driven expression of diphtheria toxin results in disruption of retinal cell organization in zebrafish. Dev. Biol. 255, 113–127 (2003) et al.
- mef2c is activated directly by myogenic basic helix-loop-helix proteins during skeletal muscle development in vivo. Mech. Dev. 120, 1021–1032 (2003) , &
- Inducible expression of an hsp68–lacZ hybrid gene in transgenic mice. Development 105, 707–714 (1989) et al.
- Surgical models for cardiac regeneration in neonatal mice. Nature Protocols 9, 305–311 (2014) , , , &
- TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111 (2009) , &
- edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010) , &
- Multiplexed Illumina sequencing libraries from picogram quantities of DNA. BMC Genomics 14, 466 (2013) et al.
- Fast gapped-read alignment with Bowtie 2. Nature Methods 9, 357–359 (2012) &
- Model-based analysis of ChIP-seq (MACS). Genome Biol. 9, R137 (2008) et al.
- Maintenance of blastemal proliferation by functionally diverse epidermis in regenerating zebrafish fins. Dev. Biol. 331, 270–280 (2009) et al.
- A chemical method for fast and sensitive detection of DNA synthesis in vivo. Proc. Natl Acad. Sci. USA 105, 2415–2420 (2008) &
We thank J. Burris, N. Lee, and T. Thoren for zebrafish care; A. Knecht, and J. Savage for technical advice or assistance; and M. Bagnat, C. Chen, F. Conlon, D. Fox, and M. Mokalled for comments on the manuscript. J.H. was supported by an AHA postdoctoral fellowship (12POST11920060), R.K. by an NIH Clinical Investigator Award (K08 HL116485), V.A.T. by an NSF Graduate Research Fellowship (1106401), and J.A.G. by an NIH postdoctoral fellowship (F32 HL120494). This work was supported by NIH grants to B.L.B. (R01 HL089707 and R01 HL064658) and K.D.P. (R01 GM074057 and R01 HL081674), who acknowledges support from HHMI.
Extended data figures and tables
Extended Data Figures
- Extended Data Figure 1: lepb transcripts sharply increase during fin and heart regeneration. (724 KB)
a, Venn diagram displaying numbers of genes with significantly increased transcript levels during fin and heart regeneration. b, RT–PCR of samples from 2 days post-fertilization (dpf) and 4 dpf embryos, and uninjured and regenerating adult tissues. lepb was not detected during embryogenesis and in uninjured tissues, but induced during regeneration. β-act2 is used as loading control. Uninj, Uninjured. c, Left: relative expression of lepb in uninjured, 1, 2, and 4 dpa fin regenerates. lepb transcript levels are increased at 1 and 2 dpa. Right: relative expression of lepb in uninjured or 3 dpa cardiac ventricles, assessed by qPCR. d, e, Endogenous lepb expression assessed by in situ hybridization in sections of fins (d) and cardiac ventricle and atrium (e). Arrowhead, amputation plane. Arrows, endocardial lepb expression. Left: uninjured tissues, Right: regenerating tissues. dpa: days post-amputation. f, g, F0 animals, injected with the transgenic lepb:eGFP BAC reporter construct at the one-cell stage, induced eGFP after larval fin fold amputation (f) and during adult fin regeneration (g). Note that lepb:eGFP is mosaically expressed. Arrowheads, amputation planes. h, i, Expression pattern of lepb:eGFP stable transgenic animals. lepb:eGFP was not detected in fin and heart during embryogenesis (2 dpf (h); 4 dpf (i)). Below ‘i’ are enlargements of the boxed areas, which show heart (left) and fin fold (right). Dotted line, outline of fin fold. The yolk is autofluorescent. j, k, Section images of lepb:eGFP caudal fin regenerates at 2 dpa (j) and 4 dpa (k). The majority of lepb:eGFP-positive cells are mesenchymal cells, overlapping partially with cells that incorporate EdU (collected 60 min after injection; red). l, m, Lack of detectable expression of lepb:eGFP in hearts of uninjured (l) or sham-operated (m) lepb:eGFP animals. n = 8 and 5 for uninjured and sham-operated hearts, respectively. Arrowheads, amputation planes. Scale bars represent 10 μm (d, f, h–k); 50 μm (e, l, m); 500 μm (g).
- Extended Data Figure 2: Leptin signalling during fin and heart regeneration. (725 KB)
a–e, Expression pattern of lepr:lepr–mCherry BAC reporter line. a, Schematic of the lepr:lepr–mCherry BAC transgenic construct. mCherry is fused at the C terminus of Lepr. b, mCherry fluorescence in the lepr:lepr–mCherry BAC reporter strain is induced during fin regeneration. n = 5; all animals displayed a similar expression pattern. c, Section images of 4 dpa lepr:lepr–mCherry caudal fin regenerates. The majority of Lepr–mCherry+ cells are epidermal cells, overlapping partially with p63+ basal and suprabasal cells (left). In addition, some putative vascular cells in the intra-ray region have Lepr–mCherry signals (right). d, e, Confocal images of sections through uninjured (d) and regenerating (e) lepr:lepr–mCherry hearts. Lepr–mCherry fluorescence co-localizes with MHC+ cardiomyocytes in uninjured and 3 dpa hearts (arrows). Note that these expression patterns are similar to leptin receptor expression in mice (see Supplementary Information). n = 7 and 6 for uninjured and 3 dpa hearts, respectively. f–j, Analysis of fin and heart regeneration in lepbpd94 mutants. f, A schematic representation of Lepb, showing the effects of the pd94 mutation. Lepb is composed of 5 alpha-helix domains. lepbpd94 has a 19 bp insertion and a 3 bp deletion at the third α-helix (helix C). g, Sequencing of wild-type and lepbpd94 alleles revealed an indel (red highlight). h, A comparison of the amino acid sequences of leptin genes of human, mice, and zebrafish. The predicted amino acid sequence of the lepbpd94 gene product is shown at the bottom, with the predicted truncation sites indicated in red. The predicted lepbpd94 protein product lacks the majority of C-terminal amino acids. Asterisk indicates identical amino acid residue between three species. i, Quantification of regenerated fin lengths from lepbpd94 and wild type siblings at 4 dpa. n = 12 each of lepbpd94 and wild-type. j, Quantification of cardiomyocyte proliferation at 7 dpa. n = 7 (lepbpd94) and 8 (wild-type). Data are represented as mean ± s.e.m. NS, not significant.
- Extended Data Figure 3: Identification of LEN and tests of regulatory sequences near lepb. (688 KB)
a, Schematic depicting the genomic region surrounding lepb (corresponding to the lepb BAC used in this study) with the profiles of RNA-sequencing and H3K27ac marks from uninjured and regenerating heart tissues. b, Enlargement of the boxed area in a. lepb is the only upregulated gene in this genomic region during regeneration. H3K27ac-enriched peaks in regenerating samples are present in a ~1 kb region (red bar) that is ~7 kb upstream of the start codon. c, Schematic representation of transgenes to examine regulatory sequence activity. Fin and endocardial expression during regeneration and the number of stable lines are indicated. Asterisk indicates that one LENP2:eGFP line showed occasional, weak endocardial eGFP expression in uninjured hearts, whereas eGFP signal in this line was broad and strong during regeneration. EC, endocardial cells. d, Images of representative 0 dpa fins from lines indicated in c. eGFP fluorescence is not detectable in fins at 0 dpa or in uninjured fins, but is induced in regenerating ray blastemas in P7:eGFP and LENP2:eGFP lines. P6:eGFP regenerates displayed weak eGFP expression below the amputation plane during regeneration, with very weak or undetectable expression in regenerating portions (see Fig. 2c). e, LENP2:eGFP expression pattern during fin regeneration. eGFP is detectable as early as 12 hpa, but is undetectable at 30 dpa. n = 5; all animals displayed a similar expression pattern. Arrowheads, amputation planes. f, Section images of representative uninjured and regenerating hearts from P2:eGFP, P6:eGFP, P7:eGFP, and LENP2:eGFP animals. eGFP fluorescence is rarely detectable in uninjured P2:eGFP, P6:eGFP, P7:eGFP, or LENP2:eGFP hearts, except in one line of LENP2:eGFP (mentioned above). Upon injury, P2 drove weak, occasional expression in cardiomyocytes and epicardium but not in endocardium, whereas P7 and LEN drove endocardial eGFP expression in ventricle and atrium. i, ii, enlargements of boxes areas in regenerating ventricle and atrium, respectively. Scale bars: 500 μm (d, e); 50 μm (f).
- Extended Data Figure 4: Additional putative regeneration enhancer elements. (409 KB)
a, Cartoon depicting the distal upstream regions of il11a, cd248b, vcana, and plek. RNA-sequencing profiles indicate that these genes are upregulated during heart regeneration. The red bar indicates putative enhancer regions that are enriched with H3K27ac marks in regenerating tissue. Two of these putative enhancers, near il11a and vcana, showed primary sequence conservation in other non-mammalian vertebrates but not in mammals. b, Scheme depicting assays in injected F0 transgenic animals. At 4 dpf, eGFP expression in the uninjured fin fold was examined, and then the fin fold was amputated. eGFP expression near the amputation plane was examined at 5 dpf. c, Table indicating injected constructs and the number of animals with eGFP+ cells near the amputation plane. d, Images of representative 4 dpf (uninjured) and 5 dpf (regenerating) fin folds from animals in c. e, Vista plot of genomic regions from mir129 to lepb based on LAGAN alignment with reference sequence zebrafish. Sequence comparison indicates that this region is not highly conserved between zebrafish and mammals. Arrowheads, amputation planes.
- Extended Data Figure 5: Transient transgenic assays examining lepb-linked regeneration enhancer fragments in combination with different promoters (fin regeneration). (212 KB)
a, Scheme depicting assays in injected F0 transgenic animals. Transgene-positive larvae were selected by detection of eGFP in response to fin fold amputation (lepb promoter), in cardiomyocytes (cmlc2 promoter), or in lenses (α-cry promoter). Caudal fins of F0 transgenic positive zebrafish were amputated at 60–90 days post-fertilization (dpf), and eGFP expression was examined at 2 dpa. b, Schematic representation of the transgenic constructs to examine fin regeneration enhancer activity. Expression during fin regeneration and the number of assessed F0 animals are indicated. Many embryos transgenic for LEN(1–850), LEN(450–1000), LEN(450–850), and LEN(660–850) coupled with the lepb or cmlc2 promoter showed activity during fin regeneration. One of eleven LENα-cry:eGFP animals displayed fin eGFP expression, but LEN(1–850)α-cry:eGFP and LEN(450–1000)α-cry:eGFP did not drive eGFP expression during fin regeneration, indicating that there may be repressive motifs in the α-cry promoter fragment that affect fin regeneration enhancer activity (See also Extended Data Fig. 9). ND, not determined.
- Extended Data Figure 6: X-gal staining in stable transgenic mouse lines. (431 KB)
a, Additional whole mount images of X-gal stained hearts from neonatal LEN-hsp68::lacZ (line 13, presented in Fig. 3) and control animals injured at postnatal day 1 and assessed at postnatal day 4. X-gal staining is undetectable in sham-operated hearts of LEN-hsp68::lacZ mice (n = 6; representative image shown) and injured hearts of control mice, but strong in partially resected hearts of LEN-hsp68::lacZ mice (arrows). Dashed red lines indicate injury area, positioned facing the front. Arrows, injury-dependent β-galactosidase expression. dpi, days post-injury. b, Whole-mount images of X-gal stained hearts and paws from LEN-hsp68::lacZ line 6, which exhibited vascular endothelial expression in uninjured hearts and paws. Scale bars represent 1 mm.
- Extended Data Figure 7: Transgenic assays examining lepb-linked regeneration enhancer fragments in combination with lepb P2 (fin regeneration). (345 KB)
a, Schematic representation of the transgenic constructs to examine LEN fragments that direct expression during fin regeneration. Expression during fin regeneration and the number of stable lines is indicated. b, Images of representative 0 dpa and 2 dpa fins from a. eGFP fluorescence is rarely detectable in uninjured fins. LEN(1–850), LEN(450–1000), LEN(450–850), and LEN(660–850) coupled with P2 directed eGFP expression during fin regeneration. *LEN(830–1350)P2:eGFP lines exhibited very weak eGFP expression in fin regenerates, detectable with long exposure times and at high magnification (data not shown), suggesting the possibility of minor fin regeneration enhancer elements in 850–1000. At least 5 fish from each transgenic line were examined, and all animals displayed a similar expression pattern. Arrowheads, amputation planes.
- Extended Data Figure 8: Images of heart sections from uninjured and regenerating transgenic lines that employ lepb-linked enhancer fragments. (884 KB)
a–h, eGFP fluorescence is rarely detectable in uninjured hearts in all transgenic lines. One exception is LEN(1000–1350)P2:eGFP, which showed occasional, weak endocardial eGFP expression in uninjured hearts. LEN(1–850)P2:eGFP (a), LEN(450–1000)P2:eGFP (b), LEN(450–850)P2:eGFP (d), and LEN(660–850)P2:eGFP (g) transgenic lines, which include distal LEN elements, directed eGFP expression from promoters in a subset of epicardial cells and/or cardiomyocytes, but not endocardial cells. LEN(450–660)P2:eGFP lines (e) showed regeneration-dependent enhancer activity in cardiomyocytes near the injury site, but not in endocardial cells. Our data indicated that the activities of LEN(1–850)P2:eGFP (a), LEN(450–1000)P2:eGFP (b), and LEN(450–850)P2:eGFP (d) lines were not as strong as LEN(450–660)P2:eGFP (e), suggesting that there might be repressive elements for cardiomyocyte expression outside of sequences 450–660. LEN(830–1350) (c) and LEN(1000–1350) (h), which did not activate expression from promoters during fin regeneration, could direct endocardial expression in both ventricle and atrium during regeneration, similar to the reference reporters lepb:eGFP and LENP2:eGFP. Arrows in c, h, endocardial eGFP. i, ii, Enlargements of the boxed areas in regenerating ventricle and atrium, respectively. At least 5 fish from each transgenic line were examined, and all animals displayed a similar expression pattern. Scale bars represent 50 μm.
- Extended Data Figure 9: Transgenic assays to examine lepb-linked enhancer fragment activity in combination with cmlc2 and α-cry promoters. (838 KB)
a, Schematic representation of the transgenic constructs to examine enhancer fragment activity in combination with the cmlc2 promoter. Expression during fin regeneration and the number of stable lines is indicated. b, Images of representative 0 dpa and 2 dpa fins from a. eGFP fluorescence was very weak or undetectable in 0 dpa or uninjured fins. (1–850), (450–1000), (450–850), and (660–850) LEN fragments coupled with the cmlc2 promoter activated blastemal eGFP fluorescence (arrows) during fin regeneration. One LEN(1–850)cmlc2:eGFP line did not show fin regeneration enhancer activity. Arrowheads, amputation planes. At least five fish from each transgenic line were examined, and all animals displayed a similar expression pattern except for the following: For two strains of LEN(450–850)cmlc2:eGFP, 4 of 5 animals induced eGFP fluorescence at 2 dpa; For LEN(660–850)cmlc2:eGFP, 4 of 7 animals induced eGFP fluorescence at 2 dpa. c, Left: schematic diagram of the LEN(1000–1350)cmlc2:eGFP transgenic construct. Right: images of sections from uninjured and regenerating LEN(1000–1350)cmlc2:eGFP hearts. eGFP is expressed mosaically in cardiomyocytes via the cmlc2 promoter. Uninjured hearts had no detectable endocardial eGFP fluorescence, whereas 3 dpa hearts displayed induced endocardial eGFP fluorescence (arrows). Arrowheads indicate cardiomyocyte eGFP fluorescence driven by cmlc2 promoter activity. d–h, Schematic representation of the transgenic constructs to examine enhancer fragment activity in combination with the α-cry promoter. Expression during fin regeneration and injury-activated endocardial expression, and the number of stable lines are indicated. At least 5 fish from each transgenic line were examined, and all animals displayed a similar expression pattern. EC, endocardial cells. One LENα-cry:eGFP line showed regeneration-dependent expression (arrows) in fins (e); yet, unlike when coupled with lepb and cmlc2 promoters, the LEN(450–1000) fragment did not direct expression during fin regeneration (d and data not shown). This suggests a possible repressive motif within α-cry sequences. Asterisk indicates that one LENα-cry:eGFP line showed weak endocardial eGFP expression in uninjured hearts, but the eGFP signal (arrows) was stronger and broader during regeneration (g). Two LEN(830–1350)α-cry:eGFP lines had no detectable eGFP fluorescence in regenerating fins (f) or uninjured hearts (h), but displayed induced endocardial eGFP fluorescence (arrows) during heart regeneration (h). i, ii, Enlargements of the boxed areas in regenerating ventricle and atrium, respectively. i, LEN sequences annotated with putative binding sites in fin (663–854) and cardiac (1034–1350) regeneration enhancer modules. j, A predicated AP-1 binding site is necessary for fin regeneration enhancer activity. Top, schematic representation of the LEN(450–850-AP-1mut)P2 transgenic construct, in which the predicted AP-1 binding site (TGACTCA) is mutated to AAAAAA. Two LEN(450–850-AP-1mut)P2 lines had no detectable eGFP fluorescence in regenerating fins. Scale bars represent 50 μm.
- Extended Data Figure 10: Pairing LEN with potent developmental influences can control regenerative capacity. (404 KB)
a, Images of representative F0 transgenic zebrafish injected with P2:dnfgfr1 (left) or LENP2:dnfgfr1 (right) constructs, shown at 3 dpa. The dn-fgfr1 cassette is fused in frame to eGFP. Whereas zero of 27 P2:dnfgfr1 F0 animals displayed defective regeneration, 7 of 67 LENP2:dnfgfr1 F0 zebrafish had impaired fin regeneration in some fin rays, corresponding to eGFP fluorescence (arrow). b, Additional examples of LENP2:dnfgfr1 fins at 30 dpa, from experiments with a stable line. Inset in b, high magnification view of the boxed area, showing eGFP fluorescence. c, Quantification of regenerated area from dob; LENP2:fgf20a F0 transgenic zebrafish (n = 45, 44 at 5, 10 dpa, respectively), dob mutants (n = 19, 19 at 5, 10 dpa, respectively), and dob; P2:fgf20a F0 transgenic zebrafish (n = 40, 40 at 5, 10 dpa, respectively) at 5 dpa and 10 dpa. Dotted line indicates 500,000 μm2. d, Images of representative dob; LENP2:fgf20a F0 transgenic zebrafish, dob mutants, and dob; P2:fgf20a F0 transgenic zebrafish at 5 dpa. e, Confocal images of tissue sections of 3 dpa fin regenerates. Mosaic regenerates indicate expression of the linked ef1α:nls–mCherry marker construct (red), and EdU incorporation (collected 60 min after injection; green). DAPI, blue. F0 mosaic dob; LENP2:fgf20a regenerates show evidence of distal growth and blastemal EdU incorporation. Arrow, blastema. Dotted lines, amputation planes. i, ii, Enlargements of the boxed areas. f, In situ hybridization in sections of 3 dpa fin regenerates from dob; P2:fgf20a (left) and F0 mosaic dob; LENP2:fgf20a (right) animals, indicating LEN-induced fgf20a expression in mesenchymal cells and regenerative growth (arrows). fgf20a is rarely detected in dob; P2:fgf20a regenerates. Arrowheads, amputation planes.
- Supplementary Table 1 (1.4 MB)
RNA-sequencing data of differentially expressed genes during fin regeneration.
- Supplementary Table 2 (1.3 MB)
RNA-sequencing data of differentially expressed genes during heart regeneration.
- Supplementary Data (518 KB)
Lists of reagents, additional experimental data, and references.