Sequential and Opposing Activities of Wnt and BMP Coordinate Zebrafish Bone Regeneration

SUMMARY Zebrafish fully regenerate lost bone, including after fin amputation, through a process mediated by dedifferentiated, lineage-restricted osteoblasts. Mechanisms controlling the osteoblast regenerative program from its initiation through reossification are poorly understood. We show that fin amputation induces a Wnt/β-catenin-dependent epithelial to mesenchymal transformation (EMT) of osteoblasts in order to generate proliferative Runx2+ preosteoblasts. Localized Wnt/β-catenin signaling maintains this progenitor population toward the distal tip of the regenerative blastema. As they become proximally displaced, preosteoblasts upregulate sp7 and subsequently mature into re-epithelialized Runx2−/sp7+ osteoblasts that extend preexisting bone. Auto-crine bone morphogenetic protein (BMP) signaling promotes osteoblast differentiation by activating sp7 expression and counters Wnt by inducing Dickkopf-related Wnt antagonists. As such, opposing activities of Wnt and BMP coordinate the simultaneous demand for growth and differentiation during bone regeneration. This hierarchical signaling network model provides a conceptual framework for understanding innate bone repair and regeneration mechanisms and rationally designing regenerative therapeutics.

Regeneration of caudal fins from Tg(sp7:EGFP) fish exposed to DMSO (upper panels) or 10 µM IWP-2 (lower panels) from 0-8 dpa. Each animal is shown before amputation and at 2, 4, and 8 dpa by Rotterman contrast and epifluorescence to visualize sp7:EGFP expression (green) in osteoblasts at 25x and at 120x magnification. Shown is one of three fish from each treatment group, within which phenotypes were indistinguishable. The experiment was repeated with three independent fish cohorts.    Regeneration of caudal fins from Tg(sp7:EGFP) animals exposed to DMSO (upper panels) or BMPRi (lower panels, 5 µM at 0-8 dpa). Images show the progression of regeneration in an individual fish before amputation and at 2, 4, and 8 dpa using Rotterman contrast and epifluorescence to visualize sp7:EGFP expression (green) in osteoblasts at 25x and 120x magnification. One of three fish, which all behaved similarly, from the two treatment groups are shown. The experiment was repeated using three independent sets of animals.

Immunostaining of sectioned tissue
To prepare paraffin sections, fins were fixed overnight in 4% paraformaldehyde (PFA) in PBS and then washed extensively in PBS. Fins were decalcified for 4 days in 0.5 M EDTA pH 8, which was replaced daily with fresh solution. Following decalcification, fins were rinsed extensively in PBS then dehydrated through an ethanol series and left overnight in 100% ethanol.
Ethanol was replaced with xylenes followed by paraffin embedding and sectioning at 7 µm thickness.
For immunostaining, paraffin sections were rehydrated and antigen retrieval was performed in antigen retrieval buffer (1 mM EDTA pH 8, 0.1% Tween-20) for 10' in a pressure cooker. The slides were blocked with PBS/0.1% Tween-20 (PBST) and 10% non-fat dry milk and then incubated overnight at 4°C with primary antibody diluted in blocking buffer. In the case of pSmad 1/5/8 staining, PBST contained 650 mM NaCl during binding of the primary antibody.
Slides were washed 3x5' in PBST and 1x30' in PBST containing 650 mM NaCl. Alexaconjugated secondary antibodies (Invitrogen) were used at 1:1000 diluted in blocking buffer and applied to sections for 1 hour at room temperature, followed by 3 x 5' washes. Nuclei were stained with Hoechst (Invitrogen) in PBST for 10' at room temperature followed by 2 x 5' washes. Slides were mounted using Fluoro-gel (Electron Microscopy Services) and visualized with an Olympus confocal microscope. For confocal imaging, antibody-stained sections were typically analyzed at 20x and 60x magnification. Optical sections were collected and processed using ImageJ software (NIH) with maximum intensity projections generated from z-stacks.
Where indicated, single optical sections were used to visualize nuclear-localized β-catenin.
For antibody staining of frozen sections, amputated fins were fixed overnight in 4% PFA/PBS, equilibrated in PBS, cryo-preserved in 30% sucrose/PBS, and frozen in agarose. Cryo-sections (16 μ m) were prepared and stored at -20°C until use. Sections were hydrated in PBST and blocked in 10% non-fat dry milk in PBST + 0.1% Triton X-100 for 1-4 hours at room temperature. Subsequent staining and imaging steps followed as for paraffin sections.
Antibodies were sourced and diluted as follows: anti-Runx2 (Santa Cruz Biotechnology, 27-K) 100 ng/ml, anti-sp7 (Santa Cruz Biotechnology, A-13) 20 ng/ml, anti-β catenin (Cell Signaling, To quantify osteoblast sub-types, 72 hpa paraffin sections, representing > 6 rays from multiple animals, were stained with Runx2 and sp7 antibodies, and imaged by confocal microscopy. Maximum intensity z-projections were generated using ImageJ (NIH) and processed in Adobe Photoshop. For scatter plot analysis, ImageJ was used to determine relative expression levels of Runx2 and sp7 in individual cells by first normalizing Runx2 and sp7 levels to nuclear staining intensity. Nuclei completely lacking expression of both Runx2 and sp7 were filtered out. Next, the mean normalized expression level was calculated for Runx2 and sp7. The deviation from the mean was determined for each nucleus by dividing the normalized expression level by the mean expression level for both sp7 and Runx2. Results were analyzed and plotted using GraphPad Prism.

Drug treatments and imaging of regenerating zebrafish
For long-term small molecule inhibitor studies, at t = 0, Tg(sp7:EGFP)b1212 fish from the same clutch were anesthetized in Tricaine and their caudal fins were imaged on a Leica M165FC stereo microscope with epifluorescent illumination. Caudal fins were then amputated with a razor and the animals returned to fish water containing 10 µM IWP-2, 100 nM Wnt-C59, 5 µM BMPRi, or DMSO vehicle. At each indicated time point, animals were anesthetized with Tricaine and re-imaged as described above. Water, containing fresh drug or DMSO, was changed every 24 hours for the duration of the study.

Heat shock studies
Heat shock experiments were performed by amputating fins of control and Tg(hsp70l:dkk1b-GFP)w32 animals and allowing regeneration to proceed for 48 h. At 48 hpa, fish were transferred to a water bath and heated to 38°C for 45 minutes. Fish were removed and placed at 28°C. This heat shock regimen was repeated once more at 64 hpa and fins were collected at 72 hpa. For heat-shock studies (n = 3), cohorts of 3 or 4 animals were used and each analyzed at the completion of the study, and images shown are representative examples of each cohort.

In vivo EdU labeling
To analyze cell proliferation, fish were injected intraperitoneally with 12.5 µl of a 1 mg/ml solution of EdU (Invitrogen) in sterile saline 6 hours prior to fin harvesting. EdU was detected on paraffin-sectioned fins using the Click-iT proliferation assay kit (Invitrogen).

Isolation of osteoblasts from zebrafish caudal fins
Osteoblasts were isolated by first amputating the distal end of approximately 50 adult caudal fins.
The cell suspension was then passed through a 40 µm cell strainer (BD Biosciences) and plated onto a 15 cm cell culture dish. Cells were allowed to attach for 20 minutes before the medium and unattached cells were removed. Adhered cells were gently washed 3 times with PBS and osteoblast-enriched cells were recovered with 3 ml 0.25% Trypsin-EDTA for 5 minutes, followed by adding 3 ml of OM. The cells were centrifuged as above, washed, resuspended in OM, and plated onto multi-well dishes containing coverslips coated with rat-tail collagen (Invitrogen). Cells were cultured at 30°C in 5% CO 2 and 5% O 2 . To determine cell purity at 4 days post isolation, cells were fixed, immunostained with Runx2 and sp7 antibodies, and scored.

Treatment and staining of zebrafish osteoblasts
To assay β-catenin localization, zebrafish osteoblasts were maintained in OM for 48 hours before addition of either 300 nM LDN193189 (BMPRi) and/or 40 ng/ml recombinant mouse Wnt3a (R&D Systems) for 24 hours. Wnt3a dosage was determined according to the manufacturer's recommendation and dose-response pilot studies. Cells were then washed twice in PBS and fixed at room temperature for 20 minutes in 4% PFA in PBS. Cells subsequently were permeabilized in 0.2% Triton X-100 in PBS for two minutes and blocked with 5% Normal Goat Serum (NGS, MP Biomedicals) in PBS for 30 minutes. Primary antibodies used were anti-EGFP 1:1000, anti-sp7 20 ng/ml, anti-Runx2 100 ng/ml, anti-β-catenin 1:250, and anti-phospho-SMAD 1/5/8 1:500 and were applied for 1 hour at room temperature in blocking buffer. To visualize nuclei, the slides were incubated with 2 µg/ml Hoechst solution in PBS for 10-15 minutes. Coverslips were mounted with Fluoro-gel mounting medium (Electron Microscopy Services).
To quantify nuclear β-catenin levels, cells first were imaged at 40x magnification. Then, the ratio of nuclear signal as a percentage of the total β-catenin signal for individual cells was determined using FIJI software. To normalize for background signal, we imaged cells treated with secondary antibody alone. The same acquisition exposure-time parameters were used to capture images for all conditions. For fluorescence intensity measurements the following formula was used: The mean percent nuclear β-catenin from at least 25 cells from each treatment was plotted with error bars representing the standard deviation.