Regulation of Zebrafish Fin Regeneration by Vitamin D Signaling

Abstract

Background

Vitamin D is an essential nutrient that has long been known to regulate skeletal growth and integrity. In models of major appendage regeneration, treatment with vitamin D analogs has been reported to improve aspects of zebrafish fin regeneration in specific disease or gene misexpression contexts, but also to disrupt pattern in regenerating salamander limbs. Recently, we reported strong mitogenic roles for vitamin D signaling in several zebrafish tissues throughout life stages, including epidermal cells and osteoblasts of adult fins. To our knowledge, molecular genetic approaches to dissect vitamin D function in appendage regeneration have not been described.

Results

Using a knock-in GFP reporter for the expression of the vitamin D target gene and negative regulator cyp24a1, we identified active vitamin D signaling in adult zebrafish fins during tissue homeostasis and regeneration. Transgenic expression of cyp24a1 or a dominant-negative vitamin D receptor (VDR) inhibited regeneration of amputated fins, whereas global vitamin D treatment accelerated regeneration. Using tissue regeneration enhancer elements, we found that local enhancement of VDR expression could improve regeneration with low doses of a vitamin D analog.

Conclusions

Vitamin D signaling enhances the efficacy of fin regeneration in zebrafish.

Introduction

Adult mammals including humans are restricted in their ability to regenerate complex tissues such as limbs. Injuries are often followed by limited tissue repair and permanent scar formation, causing functional impairment. Extensive research has been performed on animal models with a generally elevated capacity for regeneration, such as salamanders and teleost fish, to identify signaling pathways important for regeneration^1–3. In each case, rapid epidermal healing is followed by formation of a primordium called a blastema, a mass of proliferative cells that ultimately gives rise to new skeletal structures of the regenerating appendage. Over recent decades, many classical developmental signaling pathways, such as FGF, BMP, Notch and RA signaling, have been implicated in regulation of limb regeneration^4–8. While strides have been made in understanding the biological basis for limb regeneration, it is likely that therapeutic human limb regeneration is decades from realization.

To identify new regulators of heart regeneration, we previously performed an unbiased screen of a set of FDA-approved drugs to identify chemicals that enhanced live indicators of cardiomyocyte proliferation in zebrafish larvae^9,10. Analogs of vitamin D, a well-studied essential nutrient, had profound mitogenic effects on larval and adult cardiomyocytes. In cells, active vitamin D molecules bind to vitamin D receptor (VDR), which forms a heterodimer with retinoid X receptor (RXR) and directly regulates gene expression^11,12. Alternative molecular mechanisms by which VDR transduces signals have also been described¹³. CYP24A1 is a direct VDR target gene and acts as a negative feedback regulator by hydroxylating active vitamin D and targeting it for excretion¹⁴. In zebrafish, vitamin D not only enhances cardiomyocyte proliferation during larval development, adult homeostasis and injury-induced regeneration, but it can also broadly increase proliferation of many other cell types, including osteoblasts, epithelial cells, hematopoietic stem cells, and hepatocytes^9,15. While the role of vitamin D in bone growth and maintenance is widely accepted, its potential functions in limb regeneration are underexplored^16–18. Studies in zebrafish reported improvement of pectoral fin regeneration in a context of overexpression of the patterning factor hand2¹⁹, and in bone mineralization and ray lengths of regenerating caudal fins in a transgenic model of diabetes²⁰; however, neither study detected effects on regeneration in wild-type animals. Somewhat contrasting these results, studies in axolotls have found that exogenous vitamin D treatment causes skeletal abnormalities during limb regeneration^21,22. None of these previous reports applied loss-of-function or molecular genetic tools.

Here, we used a variety of molecular tools to explore the role of vitamin D following amputation of zebrafish caudal fins. A live indicator of vitamin D signaling revealed VDR activity in fin rays during homeostasis and injury-induced regeneration. Furthermore, loss-of-function studies indicated that vitamin D signaling is required pathway for fin regeneration, and that vitamin D analogs have the potential to accelerate fin regeneration. Finally, to test the possibility of enhancing regeneration locally while reducing toxic effects of high concentrations of vitamin D, we designed a tunable system employing regeneration-responsive regulatory sequences. Our study implicates vitamin D as a key pro-regenerative factor for appendage regeneration and presents a new, spatiotemporal targeting approach to improve regeneration in combination with a systemic, circulating drug or nutrient like vitamin D.

Results

Visualization of vitamin D signaling in zebrafish fins

To monitor the activity of vitamin D pathway in live zebrafish, we previously knocked in an mGFP reporter into the translational start region of vitamin D target gene cyp24a1 (referred to as cyp24a1^mGFP)⁹. Fluorescent microscopy of intact fins revealed that cyp24a1^mGFP expression displayed distinct patterns along the proximodistal (PD) and anteroposterior (AP) axes in caudal, dorsal, pectoral and pelvic fins (Figure 1A–D). Along the PD axis, mGFP was localized predominantly in the distal regions of fin rays (Figure 1A–D), although the expression in lateral, longer fin rays was weaker than that of the middle, shorter rays of the caudal fin (Figure 1D). Consistent with a previous report¹⁹, the reporter displayed higher expression levels in the posterior rays of pectoral fins (Figure 1A). We also observed enhanced expression in the more posterior rays of dorsal and pelvic fins (Figure 1B, C). These results suggest that vitamin D signaling may be involved in maintenance of posterior and distal fin pattern or structural integrity.

Figure 1. Visualization of vitamin D signaling in uninjured and regenerating zebrafish fins.

(A-D) Expression of cyp24a1^mGFP in uninjured pectoral (A), pelvic (B), caudal (C), and dorsal (D) fins. A, anterior; P, posterior. (E) Immunostaining of transverse sections of caudal fins. Green represents cyp24a1-directed GFP fluorescence, and red represents Zns-5 immunofluorescence, marking osteoblasts. Arrow points to colocalization of signals. DAPI (blue) marks nuclei. (F) High magnification views of boxes in (E), indicating expression in rays that mainly colocalizes with Zns-5 (top), but also weak interray expression in non-osteoblast mesenchyme. (G-J) cyp24a1-directed fluorescence in the distal tips of regenerating caudal fins at different days post amputation (dpa). Dashed lines indicate approximate amputation planes.

To determine the cell types in which vitamin D signaling is active, we sectioned caudal fins of cyp24a1^mGFP reporter fish and co-stained the sections with a marker for osteoblasts (Zns-5). The majority of positive GFP signals overlapped with or surrounded Zns-5 positive cells, with some weaker expression evident in ray and interray mesenchymal cells (Figure 1E,F), indicating strong vitamin D signaling activity in osteoblasts. This expression is consistent with the known roles of vitamin D in bone homeostasis as well as its mitogenic effects on zebrafish fin osteoblasts⁹.

To monitor changes in vitamin D activity during fin regeneration, we amputated caudal fins of cyp24a1^mGFP reporter animals. We found cyp24a1-directed fluorescence in fin ray stumps by 2 dpa, and in the growing and bifurcating fin rays by 4 dpa, where it was maintained for the duration of regeneration and in regenerated structures. These findings suggested an active role for vitamin D signaling in regulation of zebrafish fin regeneration.

Inhibition of vitamin D signaling disrupts fin regeneration

To determine whether VDR activity is essential for zebrafish fin regeneration, we performed several approaches. First, we used a published transgenic line (hsp70:dn-vdra) that enables heat shock-inducible expression of a dominant-negative zebrafish vdra gene cassette⁹, adapted from a previously reported human dnVDR that can efficiently bind to RXR with limited ability to recruit coactivators and transactivate downstream genes²³. Considering the fact that RXR is also a binding partner for several other nuclear receptors including the retinoic acid (RA) receptor¹², we injected one-cell embryos with dn-vdra mRNA and performed in situ hybridization to examine expression of cyp24a1 and cyp26a1, a downstream target gene of RA signaling. While cyp24a1 expression was significantly reduced in dn-vdra injected embryos, no visible changes in cyp26a1 expression level were observed (Figure 2A), suggesting reliable selectivity of the inhibitory effects of dn-vdra. Caudal fins of hsp70:dn-vdra and wild-type clutchmates were amputated after an initial day of heat shock, and fin outgrowth was measured at 4 and 7 dpa during a regimen of daily heat shocks. Transgenic animals displayed ~41% reductions in lengths of regenerated fin rays at each time point compared to their wild type siblings, despite grossly normal morphology (Figure 2B, C). These results indicated that global disruption of VDR activity impairs fin regeneration.

Figure 2. Inhibition of vitamin D receptor signaling impairs fin regeneration.

(A) In situ hybridization experiments for expression of cyp24a1, a VDR target gene, or cyp26a1, an RAR target gene. Embryos microinjected with 25 ng/μL dn-vdra mRNA at the single-cell stage displayed visibly reduced cyp24a1, but not cyp26a1, expression level at 24 hpf. n = 20 for each group. (B, C) Whole-animal induction of dominant-negative VDR in transgenic fish significantly delayed fin regeneration at 4 and 7 dpa. n = 8 for each group. (D-G) Transient, local inhibition of vitamin D signaling using regeneration-responsive lepb regulatory elements to induce cyp24a1 (D, F) or dn-vdra (E, G) impairs fin regeneration. In cyp24a1 overexpression studies, n = 13 for wild-type fish at 4 dpa; n = 12 for wild-types at 7 dpa; n = 11 for transgenic fish at both time points. In dn-vdra overexpression groups, n = 11 for wild-types; n = 13 for transgenic fish. (H) Fin regeneration was modestly delayed by 7 dpa in vdra/vdrb double knockout zebrafish. Values on right incorporate corrections for the tail fin sizes of heterozygous and (the smaller) mutant animals, as explained in the main text, which is why relative lengths are plotted. Values in (H) were normalized to the heterozygous 7 dpa average lengths (set at 1.0). n = 14 for each group. Dashed lines indicate approximate amputation planes.

To test whether a local reduction of vitamin D pathway activity can affect fin regeneration, we generated transgenic lines that express cyp24a1, the negative pathway regulator, or dn-vdra, under the control of the upstream 7 kb of regulatory sequences of the leptin b gene (lepb). This sequence contains a tissue regeneration enhancer element, or TREE, named LEN, that directs little or no expression in uninjured tissues, but can target expression specifically to regenerating mesenchymal tissue of fins upon amputation²⁴. While uninjured fish displayed normal organismal and fin development, transgenic fish with TREE-directed cyp24a1 or dn-vdra displayed significantly inhibited fin regeneration. The lengths of regenerated rays were decreased by 22% and 17% at 4 and 7 dpa, respectively, in lepb:cyp24a1 transgenic fish, as compared to wild-type siblings (Figures 2D, F). Similarly, 51% and 60% reductions in fin ray lengths were observed at 4 and 7 dpa, respectively, in lepb:dn-vdra transgenic fish (Figures 2E, G). These data indicate that local vitamin D signaling activity is important for fin regeneration.

Finally, to test the effects of genetic mutation in VDR genes, we assessed fin regeneration in vdra/vdrb double mutant fish previously generated in our lab, which grow to adulthood but more slowly than wild-type and double heterozygous zebrafish⁹. Interestingly, the effects of null mutations on fin regeneration were less severe than the transgenic manipulations described above. At 7 dpa, fin ray lengths were on average 11% shorter in vdra^−/−;vdrb^−/− fish than in vdra^+/−;vdrb^+/− fish (Figure 2H). Because of the effects of VDR mutations on overall growth, we also calculated the lengths of regenerates relative to the lengths of the middle fin ray stumps, relying on the close attention we paid to removing 50% of each appendage. Relative lengths of vdra^−/−;vdrb^−/− rays were 9.7% shorter than that of vdra^+/−;vdrb^+/− fish at 7 dpa, indicating that the phenotype of impaired regeneration is not simply a result of tissue size. Collectively, these data demonstrate that intact VDR signaling is essential for fin regeneration.

Activation of vitamin D signaling enhances fin regeneration

Treatment of adult zebrafish with vitamin D analogs induced osteoblast and epithelial cell proliferation in uninjured caudal fins⁹, concomitant with an elevation in vitamin D signaling activity we report here as indicated by increased cyp24a1^mGFP expression (Figure 3A). To test the extent to which excess vitamin D signaling can promote fin regeneration, we first administered daily injections of a high dose (0.8 μg) of the analog alfacalcidol following fin amputation. We found that this treatment slightly increased lengths of wild-type regenerated fins, by 5% at 4 dpa, compared to vehicle-treated animals (Figure 3B). We also assessed effects of alfacalcidol treatment during fin regeneration in osx:H2A-mCherry; osx:Venus-hGeminin transgenic animals²⁵, which revealed a 30% increase in osteoblasts in proliferative cell cycle phases in 4 dpa blastemas (Figure 3C). Notably, these experiments were limited by observations that continued injections with high doses of alfacalcidol appear to stress the animals. Possible explanations include documented toxicity of vitamin D caused by increased downstream gene transcription and/or hypercalcemia^26–28. Thus, we suspected that a system for local activation of vitamin D signaling was likely to be informative.

Figure 3. Systemic and targeted elevation of vitamin D signaling promotes fin regeneration.

(A) Three daily injections of 0.8 μg alfacalcidol increases cyp24a1-directed GFP fluorescence in uninjured caudal fins. (B) The lengths of 4 dpa regenerates increased by 5% with daily injections of 0.8 μg alfacalcidol injection, compared to vehicle (10% EtOH) injection. Values were normalized to the 4 dpa average lengths of vehicle-injected fish to correct for batch differences and plotted as relative lengths, as pooled data from three experimental groups were used. n = 41 for each group. (C) Daily injections of 0.8 μg alfacalcidol to osx:H2A-mCherry; osx:Venus-hGeminin transgenic fish following fin amputation increased osteoblast proliferation compared to 10% EtOH treated ones in regenerating fins at 4 dpa. Red, osteoblast nuclei. Green, osteoblast nuclei in S/G2/M. n = 9 for each group. (D) Microinjection of 25 ng/μL vdra RNA into one-cell stage embryos sensitized the response to 24 hours treatment with 0.1 μM alfacalcidol. Green represents cyp24a1-directed fluorescence, a readout for VDR activity, which is higher in the RNA-injected group. Vehicle for RNA injection is water, and vehicle for alfacalcidol treatment is 10% DMSO. Representative embryos shown, n = 20 for each group. (E, F) A low dose (0.4 μg) daily injection of alfacalcidol is sufficient to improve fin regeneration in lepb:vdra and lepb:vdrb transgenic lines, which induce wild-type VDR expression in regenerating fins. Wild-type animals show no difference between vehicle- (10% EtOH) and alfacalcidol-injected groups. n = 10 for all groups. Dashed lines indicate approximate amputation planes.

We postulated that conditional overexpression of a wild-type vitamin D receptor may sensitize zebrafish tissue to lower levels of vitamin D analogs. Moreover, targeting this overexpression specifically to injury sites could potentially enable enhancement of vitamin D signaling specifically in regenerating tissues. To first test the idea that increased VDR expression levels can sensitize tissue responses, we injected single cell-stage cyp24a1^mGFP reporter embryos with vdra mRNA and treated the embryos with alfacalcidol for 24 hours immediately after injection. Interestingly, while vehicle treatment led to a comparable level of cyp24a1-direct fluorescence in vdra-injected and water-injected embryos, embryos injected with vdra mRNA displayed clearly higher fluorescence in response to a low concentration (0.1 μM) of alfacalcidol (Figure 3D). This result suggested that increasing VDR presence can sensitize zebrafish tissue to vitamin D pathway activation.

To produce contexts with elevated VDR presence specific to regenerating fin tissue, we generated transgenic lepb:vdra animals. Adult fish were injected daily with 10 μL of a reduced dose of alfacalcidol (0.4 μg) or 10% ethanol vehicle as a control, immediately following fin amputation. While no significant difference in the lengths of 4 dpa fin regenerates was observed in wild-type fish injected with alfacalcidol or vehicle, a 16% increase attributable to alfacalcidol was observed in lepb:vdra fish. We also generated lepb:vdrb transgenic animals, with regulated expression of the other zebrafish vdr paralog, seeing a 23% increase attributable to alfacalcidol (Figures 3E, F). Importantly, the acceleration of fin regeneration in transgenic lines treated with 0.4 μg alfacalcidol was greater than that in wild-type fish treated with twice the dose of alfacalcidol. Together, these experiments indicate that local augmentation of vitamin D signaling can enhance fin regeneration.

Discussion

Here, we provide evidence using expression studies, pharmacological reagents, and genetic tools that local activation of vitamin D signaling is both necessary for normal regeneration and sufficient to enhance its efficacy. Our data indicate that while some level of regeneration occurs in the absence of vitamin D signaling, normal or enhanced presence of this pathway helps optimize regeneration outcomes.

The relatively weak phenotypes we observed in vdra/vdrb double mutants were unexpected. However, that these fish were able to survive and breed as adults hints at the possibility that there are compensatory mechanisms that allow adaptation to a chronically low level of vitamin D signaling, potentially through transcriptional regulatory mechanisms²⁹. By contrast, the induced dn-vdr and cyp24a1 cassettes in transgenic lines are more acute loss-of-function events, likely circumventing potential compensatory mechanisms and causing more severe phenotypes after injury. We cannot rule out any nonspecific effects of ectopic gene overexpression could also be contributors to the discrepancy in phenotypical response. In contrast to the high selectivity and specificity of cyp24a1 enzyme for its substrate³⁰, a high level of dn-VDR expression could potentially disrupt the activity of other nuclear receptors that interact with RXR. Although our data indicate little or no effects on RA signaling, it is possible that other subfamily 1 nuclear receptors, including the constitutive androstane receptor, bile acid receptor, liver X receptor, peroxisome proliferator-activated receptors, pregnane X receptor and thyroid hormone receptor are impacted in some way^12,31–33. Future experiments that inhibit endogenous vitamin D signaling activity, such as conditional vdr gene knockouts, may further illuminate what we report here.

Whereas this study has focused on the description of endogenous vitamin D pathway activity during appendage regeneration and the effects of modulating pathway activity, there remains much to explore on the mechanistic side. The pro-regenerative effect of vitamin D on fin regeneration is at least in part a local rather than systemic effect, and possible mechanisms by which vitamin D may act are regulation of calcium signaling³⁴, energy metabolism and control of oxidative stress^35,36, or stimulation of FGF ligand secretion by secreted factors from chondrocytes³⁷. Although most research studies on vitamin D have assessed mammals, these mechanisms could individually or together contribute to mitogenic effects during fin regeneration in teleosts.

Vitamin D is an essential nutrient but toxic at high doses; thus, it is important to identify methods beyond systemic administration. Recent studies reported that systemic vitamin D treatment induced abnormalities during axolotl limb regeneration^21,22. A possible explanation for this result may be linked to the toxicity of vitamin D and its effects on RA signaling, another pathway that is critical for tissue regeneration^8,12. We found in this study that spatiotemporal targeting of vitamin D signaling activity by TREEs can enhance fin regeneration under conditions of reduced vitamin D exposure. We expect that use of TREEs can be augmented by optimization of enhancer sequences and ligand-receptor interactions. In this way, engineering of vectors to boost vitamin D signaling at injury sites, or more generally to enhance pro-regenerative signaling, can inspire new ideas to tackle challenges in regeneration like therapeutic human limb renewal.

Methods

Zebrafish maintenance and procedures

Wild-type or transgenic male and female zebrafish of the outbred Ekkwill (EK) strain were used for all experiments, with adult animals ranging from 4 to 12 months. For RNA injection experiments, vdra mRNA was transcribed from an in vitro transcription vector containing full-length vdra coding sequences and injected into single-cell stage embryos at 25 ng/μL⁹. Water temperature was maintained at 28°C. For heat shock experiments, fish were maintained at 28°C and heat shocked at 38°C for 30 min daily as described³⁸. For regeneration essays, fin tissues were amputated at 50% of the original length with razor blades. Alfacalcidol treatment was performed using previously published protocols⁹. All procedures with animals were approved by the animal care and use committee at Duke University. Transgenic and mutant lines used in this study were cyp24a1^{mGFP pd2729}; hsp70:dn-vdra^pd2759; vdra^pd3079; vdrb^pd3089; osx:H2A-mCherry^pd31025; osx:Venus-hGeminin^pd27125; Tg(lepb:cyp24a1–2A-TagBFP;ins:GFP) (referred to as lepb:cyp24a1^pd327); Tg(lepb:dn-vdra-2A-TagBFP;ins:GFP) (referred to as lepb:dn-vdra^pd329); Tg(lepb:vdra-2A-TagBFP;ins:GFP) (referred to as lepb:vdra^pd331); and Tg(lepb:vdrb-2A-TagBFP;ins:GFP) (referred to as lepb:vdrb^pd333).

Generation of lepb:dn-vdra, lepb:cyp24a1, lepb:vdra and lepb:vdrb zebrafish

MultiSite Gateway cloning was used for construction of all plasmids with lepb regulatory sequences. The 7 kb leptin b upstream regulatory sequences were amplified from zebrafish genomic DNA using the following primers: attB1-lepb-f (5’-GGGGACAAGTTTG TACAAAAAAGCAGGCTCTCGAGTACTCGCCAATTTGCTTCT-3’) and attB2-lepb-r (5’-GGGGACCACTTTGTACAAGAAAGCTGGGTACCGGTATTTCTGCAAAAGACCAAATGAAAT-3’), and recombined into the 5’ entry vector using BP recombination reaction. A 1.8 kb selection marker ins:GFP was amplified from the ZIP promoter JM6 plasmid³⁹ with primer SalI-Ins-f (5’-GTCGTCGACTTCGAAGCCCACAGTCTAGTTTAG-3’) and primer StuI-GFP-r (5’-CTCAGGCCTCCTTGTACAGCTCGTCCAT-3’) and inserted into the 3’ entry vector. The dn-vdra-2A-TagBFP, cyp24a1–2A-TagBFP, vdra-2A-TagBFP and vdrb-2A-TagBFP cassettes were digested with restriction enzymes from existing P2A-TagBFP vectors containing these fragments respectively⁹ and inserted into the middle entry vector. Then, a MultiSite Gateway LR recombination reaction was performed to simultaneously transfer all three DNA fragments into a destination vector with I-SceI sites. Purified plasmids were linearized with I-SceI meganuclease and injected into single-cell stage embryos. Stable transgenic lines were screened for positive ins:GFP signals and examined for BFP fluorescence in fin regenerates using a fluorescence microscope.

Histology and imaging

Whole-mount embryo in situ hybridization was performed as described previously⁴⁰. Whole-mount fin tissues and embryos were imaged with a Zeiss AxioZoom V16 microscope. For immunofluorescence experiments, fin tissues were fixed with 4% paraformaldehyde and cryosectioned at 14 μm. Immunostaining was performed as described previously⁴¹, using antibodies against Zns-5 (ZIRC, Zns-5, 1:200) and GFP (Life Technologies, A11122, 1:200). Confocal images were acquired with a Zeiss LSM 880 microscope.

Quantification and statistical analysis

For all experiments, clutchmates were randomized into control groups and experimental groups. At least two biological replicates were performed, with at least 7 animals in each group for each experiment. At least two independent lines for the same transgene were used for all studies. No data were eliminated from the results unless an animal died during the experiments. For fin regeneration essays, fin ray lengths are calculated by averaging the measured distances from the amputation plane to the distal tips of the third and fourth fin rays of dorsal and ventral caudal fin lobes, measured using ZEN software. For quantification of osteoblast proliferation, maximum intensity projections of confocal images with z-stacks were generated. Proliferation indices were calculated by quantification of the number of Venus-positive cells divided by mCherry-positive cells from each projection. All statistical results are indicated as Mean ± Standard Deviation. Quantified results were tested with equal variance tests, and two-tailed Student’s t-tests and/or Mann-Whitney tests were used for statistic evaluation.