The Glycogen Synthase Kinase-3β Inhibitor LSN 2105786 Promotes Zebrafish Fin Regeneration

The Wnt pathway has been shown to regulate bone homeostasis and to influence some bone disease states. We utilized a zebrafish model system to study the effects of a synthetic, orally bioavailable glycogen synthase kinase-3β (GSK3β) inhibitor LSN 2105786, which activates Wnt signaling during bone healing and embryogenesis. GSK3β inhibitor treatment was used to phenocopy GSK3β morpholino oligonucleotide (MO) knockdown in zebrafish embryos. Human and zebrafish synthetic mRNA injection were similarly effective at rescue of GSK3β MO knockdown. During caudal fin regeneration, bony rays are the first structure to differentiate in zebrafish fins, providing a useful model to study bone healing. Caudal fin regeneration experiments were conducted using various concentrations of a GSK3β inhibitor, examining duration and concentration dependence on regenerative outgrowth. Experiments revealed continuous low concentration (4–5 nM) treatment to be more effective at increasing regeneration than intermittent dosing. Higher concentrations inhibited fin growth, perhaps by excessive stimulation of differentiation programs. Increased Wnt responsive gene expression and differentiation were observed in response to GSK3b inhibitor treatment. Activating Wnt signaling also increased cell proliferation and osteoblast differentiation in fin regenerates. Together, these data indicate that bone healing in zebrafish fin regeneration was improved by activating Wnt signaling using GSK3b inhibitor treatment. In addition, caudal fin regeneration is useful to evaluate dose-dependent pharmacological efficacy in bone healing, various dosing regimens and possible toxicological effects of compounds.


Introduction
The Wnt signaling pathway controls bone formation, homeostasis, and repair [1]. Wnt ligands are cysteine-rich, secreted glycoproteins that regulate embryonic development by controlling cell growth, differentiation, and apoptosis [2]. In bone growth, canonical Wnt, β-catenin-dependent signaling is activated when the Wnt ligand binds to a transmembrane Frizzled (Fz) receptor with the co-receptors low density lipoprotein receptor-related proteins 5/6 (LRP 5/6). Receptor engagement activates Disheveled (Dsh) family proteins, which in turn, disrupts the β-catenin destruction complex, preventing cytoplasmic glycogen synthase kinase-3β (GSK3β) from phosphorylating β-catenin. In the rate. Fish with a mutation in one copy of the Wnt/β-catenin pathway inhibitor, axin1, display increased regenerative outgrowth at 7 dpa [25].
Interestingly, β-catenin independent, non-canonical Wnt signaling regulates fin regeneration. Blocking the non-canonical Wnt signaling pathway increased fin regenerative outgrowth, and over activation of non-canonical Wnt signaling completely blocks fin regeneration [25], showing that non-canonical Wnt signaling negatively feeds back on canonical/β-catenin signaling regulating normal regeneration.
Synthetic orally bioavailable GSK3β inhibitors, which also activate the canonical Wnt signaling pathway have been shown to enhance bone growth in rats [26]. In this study, we evaluated the ability of LSN 2105786 to enhance bone healing. The zebrafish caudal fin regeneration model was useful to evaluate dose-dependent pharmacological efficacy in bone healing, dosing regimens and possible toxicological effects. Experiments showed that a GSK3β inhibitor promoted caudal fin regeneration, and this model was useful to dissect the cellular and molecular mechanisms of GSK3β-inhibitor induced bone growth.

Zebrafish Husbandry
Zebrafish (Danio rerio) AB strain and Tg(TOP:GFP) [27] transgenic line were raised and housed under standard laboratory conditions [28]. Indiana University-Purdue University Indianapolis School of Science Institutional Animal Care and Use Committee (IACUC) approved animal care and use protocol for this work (SC212R, 5/20/2015). Zebrafish, 6-12 months of age, were obtained from EKKWill Waterlife Resources (Ruskin, FL, USA) for caudal fin regeneration experiments and housed under standard laboratory conditions [28].

Adult Fin Amputation Assay
Zebrafish, 6-12 months of age, were used for caudal fin regeneration experiments. Fish were anesthetized in tricaine (Ethyl 3-aminobenzoate methanesulfonate), and approximately 50% of the caudal fin was amputated using a razor blade. Fish were placed in 2 L of water with various concentrations of GSK3β inhibitor compound (LSN 2105786) or dimethyl sulfoxide (DMSO; at least 1:1000 dilution) vehicle control and kept at 31 • C to promote rapid regeneration. Tank water was replaced daily, including fresh compound. Tanks and fish were rinsed between treatments to remove any residual compound.
At 4 and 7 dpa, fish were anesthetized and images of regenerating fins were collected using a Leica MZ12 microscope equipped with Leica DFC290 camera (Leica Microsystems Inc., Buffalo Grove, IL, USA). The length of the regenerate (from the amputation plane to the distal tip of the fin) at the third, fourth and fifth dorsal and ventral fin rays were measured using ImageJ software (NIH, http://rsb.info.nih.gov/ij/), and the average length of the regenerate was calculated for each fin.

Embryo Treatment with GSK3β Inhibitor
Embryos were incubated with LSN 2105786 (or vehicle controls) in embryo medium. Various concentrations of the drug were used to treat embryos from 6 to 28 h post fertilization (hpf), in Petri dishes wrapped with parafilm and maintained at 28.5 • C. Embryos were imaged using a Leica MZ12 microscope equipped with Leica DFC290 camera.

Whole Mount in situ Hybridization
Fins collected at various time points after amputation were fixed overnight at 4 • C in freshly made 4% paraformaldehyde in phosphate-buffered saline (PBS). Fins were then washed two times in PBS, dehydrated in methanol and stored at 20 • C overnight. Fins were rehydrated stepwise in methanol in PBS containing 0.1% Tween 20 (PBST). Next, fins were treated for 30 min in proteinase K (10 µg/mL) in PBST. Then, fins were washed two times in PBST, and post-fixed in 4% paraformaldehyde in PBS for 20 min. Fins were washed five times in PBST, and then prehybridized for 2 h at 70 • C in hybridization buffer (50% formamide, 5× SSC, 0.1% Tween 20, 50 µg/mL heparin, and 500 µg/mL yeast RNA). Following prehybridization, fins were hybridized overnight in hybridization buffer containing 0.5 µg/mL digoxigenin-labeled RNA probe at 70 • C. Then, fins were washed at 70 • C for 10 min each in 75% hybridization buffer/25% 2× SSC, 50% hybridization buffer/50% 2× SSC, 25% hybridization buffer/75% 2× SSC, and 2× SSC. Next, fins were washed two times in 0.05× SSC for 30 min each at 70 • C, followed by washes at room temperature for 5 min each in 75% 0.05× SSC/25% PBT, 50% 0.05× SSC/50% PBT, 25% 0.05× SSC/75% PBST, and PBST. Tissues were blocked for 2 h at room temperature in block buffer (2 mg/mL bovine serum albumin in PBST). Fins were then incubated overnight in a 1:5000 dilution of anti-digoxigenin antibody (Roche Diagnostics, Indianapolis, IN, USA) coupled to alkaline phosphatase in block buffer. The following day, fins were washed with PBST five times for 15 min each. Fins were then washed with alkaline phosphatase buffer (NTMT) three times for 15 min each, and then, fins were incubated with NBT/BCIP substrate in NTMT to allow color development. Development of the staining reaction was monitored carefully, and compound-treated and control fins were stopped at the same time to allow accurate comparison. Following color development, fins were washed five times in PBST and stored in 80% glycerol in PBST. Digoxigenin-labeled riboprobes for shh, lef1, bmp2b, bmp4 and sox9b were synthesized using DIG RNA Labeling Kit (Roche, Indianapolis, IN, USA) according to manufacturer's instructions. Images were collected using a Leica MZ12 microscope equipped with Leica DFC290 camera.

Whole Mount Immunostaining and Image Analysis
Fins were collected at various time points post-amputation and fixed overnight at 4 • C in 4% paraformaldehyde in PBS. Following fixation, fins were briefly washed several times with PBS. Fins were incubated overnight at room temperature in block solution, and then, fins were incubated with primary antibody in block solution for 1 to 3 days. Primary antibodies used were: mouse anti-Zns5 (Zebrafish International Resource Center, Eugene, OR, USA) at 1:200; and rabbit anti-phospho-histone H3 (Chemicon International, Inc.) at 1:500. After incubation in primary antibody, fins were washed with PBST six times for 10 min each, and then incubated overnight with secondary antibody in block solution. Goat anti-mouse Texas Red (Invitrogen), and goat anti-rabbit Alexa 488 (Invitrogen) were used. Nuclei were visualized by incubating fins in 1 µM To-PRO3 iodide for 1 h at room temperature. Confocal images were acquired using a Zeiss Observer Z1 LSM 700 confocal microscope (40× 1.1 NA W or 20× 0.8 NA objectives; Carl Zeiss Microscopy, Thornwood, NY USA). To measure cell proliferation phospho-histone 3 (PH3)-positive cells were counted from at least three fins per group. The number of PH3-positive cells per unit area was calculated for each fin using Volocity software (version 6.0.1, Perkin Elmer, Waltham, MA, USA).

DNA Sequence Alignment and 3D Protein Structure Prediction
DNA sequence alignment was done by using Clustal Omega program, The European Bioinformatics Institute (EMBL-EBI), Hinxton, Cambridgeshire, UK. 3D-protein structure prediction was done by Phyre2 web portal for protein modeling, prediction and analysis (Structural Bioinformatics Group, Imperial College, London, UK) [29].

Knockdown of gsk3β by Morpholino and Injection of Zebrafish gsk3β or Human GSK3β mRNA
Morpholino oligonucleotide (MO) was synthesized by Gene Tools, LLC (Philomath, OR, USA) to block the translation of gsk3β (this morpholino targets both gsk3βa and gsk3βb). The previously published sequence of the MO is GTTCTGGGCCGACCGGACATTTTTC [30].
MO concentration was determined spectrophotometrically. Different concentrations (2-6 ng/nL) of the MO were injected into 2-4 cell stage embryos to determine optimal effective concentrations. Injection of 1 nL of 3.5 ng/nL MO consistently produced defective embryos.
Zebrafish gsk3β resistant to MO and human GSK3β genes were synthesized in vitro at Lilly Research Laboratories and cloned in pAN104b and pcDNA3.1 vectors, respectively. Messenger RNAs were synthesized using T7 polymerase. Different concentrations of mRNA ranging from 100 to 500 ng/nL were injected along with the MO into 2-4 cell stage embryos. Co-injection of 200 ng/nL of either zebrafish or human synthetic GSK3β mRNA reduced the severity with MO (3.5 ng/nL) induced defects.

Statistical Analysis
Analyses on fin regenerate length and cell proliferation were performed using unpaired two-tailed Student's t-test (GraphPad Software, La Jolla, CA, USA).

Human GSK3β is Structurally and Functionally Equivalent to Zebrafish Gsk3β
The human GSK3B protein consisting of 420 amino acids is 95% identical and 98% similar to the zebrafish GSK3β protein consisting of 421 amino acids (Gsk3βa; Gsk3βb is 419 amino acids; Figure 1A). 3D-protein structure analyses predicted highly similar protein structures for zebrafish Gsk3β and human GSK3B proteins ( Figure 1B,C). To test whether human GSK3β is functionally equivalent to zebrafish Gsk3β, endogenous Gsk3β was knocked down in zebrafish embryos by injecting Gsk3β translation blocking morpholino (MO; 3.5 ng/nL) [30], and synthetic mRNA (200 ng/nL) encoding either zebrafish Gsk3β or human GSK3β protein was injected into the embryos at 2-4 cell stage. Embryos were examined at 1 and 2 days postfertilization (dpf). MO injected embryos were categorized into two phenotypic groups, severe and moderate, depending on the severity of the defects that include short bent body, small eye, pericardial edema (PE), brain ventricle edema and curly blistered tail ( Figure 1E,F,K). Co-injection of the same amount of zebrafish or human synthetic GSK3β mRNA (200 ng/nL) with MO (3.5 ng/nL) reduced the severity of the defects, produced higher percentage of embryos with mild PE and with only slightly bent tails ( Figure 1I,J). Zebrafish or human mRNA (200 ng/nL) injection alone produced only a few embryos with mild PE ( Figure 1G,H). This experiment showed that injection of human synthetic GSK3β mRNA into zebrafish embryos rescued the zebrafish Gsk3β knockdown phenotype.

LSN 2105786 Treatment Upregulates Wnt Signaling in Zebrafish and Produces Eyeless Embryos
It was previously shown that over-activation of Wnt signaling after the mid-blastula transition produces small eye or eyeless phenotype [31]. To test whether LSN 2105786 acts upon the target enzyme GSK3β in zebrafish and upregulates Wnt signaling, embryos were treated with various concentrations (300 nM-5 µM) of the compound from 6 to 48 h post-fertilization (hpf). Treated embryos were examined at 1 and 2 dpf. LSN 2105786 treatment produced eye and other defects in zebrafish embryos, in a dose dependent manner, similar to the defects caused by over-activation of Wnt signaling. The defect was first observed at 1 dpf, which was more obvious at 2 dpf. Treatment with higher concentrations (750 nM-5 µM) of LSN 2105786 resulted in 100% eyeless embryos (n = 60 (three replicates); n, no. of embryos). The lower concentrations used (350-500 nM) resulted in variable phenotypes, ranging from reduced eye size to eyeless phenotypes. Treatment with 300 nM LSN 2105786 produced nearly normal appearance to eyes (92% near normal, 8% small eye; n = 66 (three replicates) (Figure 2A-T). To directly measure LSN 2105786 effects on a Wnt responsive reporter gene, TOP:GFP zebrafish transgenic strain embryos, which carries β-catenin-Tcf/Lef promoter binding sites that drive GFP gene expression [27], were treated with 350-500 nM compound from 6 to 28 hpf. Midbrain-hindbrain boundary region that expresses GFP was imaged at 28 hpf, and GFP intensity was analyzed using confocal microscopy. LSN 2105786 treated embryos showed increased GFP expression, and highest intensity was exhibited in the 450 nM compound treated embryos ( Figure 2U-Y). GFP intensity decreased in 500 nM LSN 2105786 treated embryos ( Figure 2Y), perhaps due to attenuation of Wnt signaling due to over-activation. phenotypes, ranging from reduced eye size to eyeless phenotypes. Treatment with 300 nM LSN 232 2105786 produced nearly normal appearance to eyes (92% near normal, 8% small eye; n = 66 (three 233 replicates) (Figure 2A−T). To directly measure LSN 2105786 effects on a Wnt responsive reporter 234 gene, TOP:GFP zebrafish transgenic strain embryos, which carries β-catenin-Tcf/Lef promoter 235 binding sites that drive GFP gene expression [27], were treated with 350−500 nM compound from 6 236 to 28 hpf. Midbrain-hindbrain boundary region that expresses GFP was imaged at 28 hpf, and GFP

LSN 2105786 Continuous Treatment after Fin Amputation Accelerates Fin Regeneration
To test the effects of LSN 2105786 on the rate of zebrafish fin regeneration, half of the caudal fin was amputated from adult zebrafish (6 to 12 months old) and the fish were treated either with compound or with vehicle DMSO immediately following amputation. To ascertain the optimal treatment regimen for maximal regenerative outgrowth, fish were either treated intermittently (1, 2, 4, 6 and 8 hr/day) with higher concentrations (50-300 nM) or continuously with lower concentrations (1-50 nM) of LSN 2105786. Continuous exposure with high concentration of LSN 2105786 killed the fish. At 4 and 7 dpa, fins were imaged, and regenerative outgrowth was measured. Continuous exposures with low concentrations of LSN 2105786 were more effective at stimulating regeneration than intermittent exposure, as detailed below.
The optimal dosing regimen for maximal regenerative outgrowth was ascertained when caudal fin was amputated in fish that were treated continuously with various concentrations of LSN 2105786

255
To test the effects of LSN 2105786 on the rate of zebrafish fin regeneration, half of the caudal fin 256 was amputated from adult zebrafish (6 to 12 months old) and the fish were treated either with 257 compound or with vehicle DMSO immediately following amputation. To ascertain the optimal 258 treatment regimen for maximal regenerative outgrowth, fish were either treated intermittently (1, 2,

LSN 2105786 Treatment Increased Expression of Wnt Targets and Downstream Genes in Regenerating Fin Tissue
To determine the spatial and temporal effects of GSK3β inhibition on Wnt responsive gene expression in fin regenerates, the Wnt/β-catenin signaling target gene, lef1, and downstream genes shh, bmp2b, and bmp4 were analyzed by using in situ hybridization. Previous studies showed that lef1 controls blastema formation [23,32], and it is expressed in the basal epidermal layer surrounding the blastema at 1 dpa. During the regenerative outgrowth phase (48 hpa and beyond), lef1 is expressed in the distal blastema in two sets of cells on each side of the fin ray [23]. LSN 2105786 treated fin showed strongly expanded and more intense lef1 expression at 1 and 2 dpa compared to vehicle treated control fins ( Figure 4A-D). At 1 and 2 dpa, lef1 was restricted to the area immediately surrounding the blastema in DMSO treated fins; whereas, LSN 2105786 treated fins showed an expansion of lef1 expression in to the wound epidermis and inter ray tissue. At 3 dpa, lef1 staining in two groups was similar ( Figure 4E,F). controls blastema formation [23,32], and it is expressed in the basal epidermal layer surrounding the blastema at 1 dpa. During the regenerative outgrowth phase (48 hpa and beyond), lef1 is expressed 286 in the distal blastema in two sets of cells on each side of the fin ray [23]. LSN 2105786 treated fin 287 showed strongly expanded and more intense lef1 expression at 1 and 2 dpa compared to vehicle

308
Previous fin regeneration studies showed that bmp2b and bmp4 are expressed in the blastema [33,34].

309
During regeneration, bmp2b is expressed in differentiating scleroblasts and in the basal layer of the 310 epidermis [33], and bmp4 is expressed in the distal blastema [34]. The expression of bmp2b in the LSN Previous fin regeneration studies showed that bmp2b and bmp4 are expressed in the blastema [33,34]. During regeneration, bmp2b is expressed in differentiating scleroblasts and in the basal layer of the epidermis [33], and bmp4 is expressed in the distal blastema [34]. The expression of bmp2b in the LSN 2105786 treated group was nearly indistinguishable from DMSO treated control group at 3 and 7 dpa ( Figure 4G-J). In contrast, bmp4 expression at the blastema region in 3 and 4 dpa fins showed weak bmp4 expression at the blastema in controls, but LSN 2105786 treated group displayed more intense expression of bmp4 ( Figure 4K,L). At 4 dpa, bmp4 expression was similarly expressed in the blastema, comparing in LSN 2105786 and DMSO treated control groups ( Figure 4M,N), but higher expression in distal most epidermis was detected. The developmental signaling molecule shh, which is involved in patterning of various structures during vertebrate development, was previously shown to be expressed in the basal layer of wound epidermis in the regenerative outgrowth [33,35]. Similar to lef1, shh expression is typically restricted to two sets of cells on each side of the distal tip of the fin ray during the regenerative outgrowth phase. In the LSN 2105786 treated fins, there was a slight expansion of the shh expression domain at 3 dpa, with shh transcripts being detected in the tissue proximal to the blastema and extending to near the amputation plane ( Figure 4O,P). At 4 dpa, the intensity of shh staining remained high in the LSN 2105786 treated group as compared to the DMSO treated control group. By 7 dpa, shh expression was similar in LSN 2105786 and DMSO treated control groups ( Figure 4Q,R).

LSN 2105786 Increases Cell Proliferation in Regenerating Fins
The Wnt signaling pathway controls cell proliferation throughout development [36]. Continuous LSN 2105786 treatment was tested in the regenerating fin tissue, and cell proliferation was examined. Phosphorylated histone-3, a marker for chromatin condensation during M phase, was used examine proliferation in the 1 and 2 dpa regenerating fin. Proliferating cells were counted, and the number of nuclei per unit area was calculated. Numbers of proliferating cells at 1 dpa was nearly identical in treated and control fins ( Figure 5A-F,M), but at 2 dpa, there was an approximately three-fold increase in numbers of proliferating cells in LSN 2105786 treated fins as compared with controls ( Figure 5G-M).

LSN 2105786 Treatment Increases Osteoblast Differentiation
To test the effect of GSK3β inhibitor LSN 2105786 on osteoblasts, fins were immunostained with the monoclonal antibody Zns-5, which labels differentiated osteoblasts and nuclear stain To-PRO3 iodide. Confocal microscopy images of fins treated with 5 nM LSN 2105786 showed larger Zns-5 positive area in the blastema compared to DMSO treated fins at 2 dpa ( Figure 6A-F). At 7 dpa, regenerated bony rays of 5 nM LSN 2105786 treated fins were strongly stained with Zns-5 antibody. Zns-5 positive osteoblasts were distributed in the lateral regions of the hemirays of DMSO treated fish, but 5 nM LSN 2105786 treated fish exhibited Zns-5 positive osteoblasts in the lateral regions of the hemirays as well as in the joints between the segments ( Figure 6G-L). High magnification images showed more Zns-5 staining in the distal region of the bony rays of 5 nM LSN 2105786 treated fish compared to DMSO treated fish ( Figure 6M-R).

Discussion
Previously, Flemming et al. [37] and others showed that the small size, ease of care and relatively low cost of zebrafish are important advantages to using fish for bone research and drug discovery screening [38,39]. Zebrafish afford the important opportunity to ascertain quickly efficacious doses of compounds on target tissues, optimal dosing regimens and toxicity of novel synthetic molecules, which may or may not be off target. Additionally, caudal fin regeneration studies allow the generation of data on bone healing mechanisms, particularly on the earliest responses to injury.
Our studies are in agreement with Stoick-Cooper et al., showing that activating Wnt/β-catenin signaling improves zebrafish caudal fin regenerative outgrowth [22]. LSN 2105786 treatment increased regenerate length at 7 dpa by 13% as compared to controls. Additionally, previous studies showed that increased Wnt/β-catenin signaling in heterozygous axin1 +/fish had increased in regenerative outgrowth, as compared to wild type fish at the same time point [25]. Interestingly, intermittent LSN 2105786 treatment was less effective than continuous treatment, which is in agreement with previously published data showing that pulses of a wnt8 transgene expression was sufficient to increase cell proliferation, but did not affect the length of fin regenerates [25]. A possible limitation of this study is that zebrafish can regenerate fins normally, without intervention. So, the extent to which regenerative outgrowth can be augmented in healthy fish may be limited. Therefore, in future studies, it may be useful to test GSK3β inhibitor and other compounds that promote regeneration in defective fish with impaired regenerative processes, or with abnormal bone phenotypes, perhaps due to aging or chronic alcohol exposure.
The inhibitory effects on regeneration observed following continuous LSN 2105786 treatment at higher concentrations above 5 mM may indicate that over-activation of the Wnt/β-catenin pathway excessively stimulates differentiation programs, perhaps depleting the progenitor cell population and preventing outgrowth. Additional studies aimed at determining effects of higher concentrations of GSK3β inhibitor on Wnt/β-catenin signaling in fin regeneration are needed, and these studies may provide insight into modulation of the regeneration process.
GSK3β inhibitor effects recapitulated previous studies using ectopic Wnt signaling phenotypes in embryos, showing LSN 2105786 acts specifically. For example, upregulation of the direct Wnt target, lef1 provides strong evidence that LSN 2105786 activates canonical Wnt signaling during fin regeneration. Experiments using human and zebrafish GSK3β synthetic mRNA to rescue Gsk3β morpholino oligonucleotide knockdown phenotypes also strongly supports the conclusion that the human and zebrafish enzyme are interchangeable in the zebrafish embryo. Together, these results indicate that the zebrafish is a useful model to study GSK3β effects on Wnt/β-catenin signaling pathways.
LSN 2105786 treatment was sufficient to activate cell proliferation and expression of downstream genes in Wnt/β-catenin signaling. LSN 2105786 treatment, which induces blastema cell proliferation and outgrowth, may help induce and maintain cell proliferation and regenerative outgrowth. Expression of shh was expanded at 3 dpa, which induces proliferation and alignment of osteoblasts in the regenerating fin [35].
In summary, our analysis of the effects of GSK3β inhibition on zebrafish fin regeneration showed that enhanced Wnt/β-catenin signaling produced Wnt/β-catenin target gene expression; cell proliferation; and more rapid fin regeneration. Taken together, the results indicate that chemical biology of zebrafish fin regeneration will help us understand Wnt-mediated bone growth and repair. In particular, mechanistic features can be studied and applied to clinically important questions for potential bone healing therapies.