Zebrafish transgenic lines of Tg(trap:GFP-CAAX) ^ou2031Tg and Tg(osterix:Lifeact-mCherry) ^ou2032Tg (Kobayashi-Sun et al., 2020a) were raised in a circulating aquarium system (AQUA) at 28.5°C in a 14/10 h light/dark cycle and maintained according to standard protocols (Westerfield, 1995). Goldfish are maintained in a circulating aquarium at 25°C. All experiments were performed in accordance with a protocol approved by the Committee on Animal Experimentation of Kanazawa University. For fracture stimulation, adult zebrafish were anesthetized in system water containing 0.02% tricaine (Sigma), and the epidermis portion of the scales was cut approximately 400 μm in length with fine scissors under a stereo microscope (Axiozoom V16, Zeiss).

The device for generating ELF-EMFs was newly constructed for the present study. Sinusoidal EMFs were generated by excitation of a unidirectional coil in orthogonal direction at a constant frequency of 60 Hz (Miyakawa et al., 2001). There is a space at the center of the device where a ring-shaped tank can be placed, allowing zebrafish or extracted scales to be uniformly exposed to EMFs.

For exposure of goldfish scales to ELF-EMFs, goldfish were anesthetized with 0.01% tricaine, and goldfish scales were extracted and placed in a 2 mL tube containing 500 μL of Eagle’s minimum essential medium (EMEM, ICN Biomedicals, Inc.) supplemented with 20 mM HEPES and 1% penicillin-streptomycin. The tubes containing goldfish scales were placed in a ring-shaped reservoir and exposed to ELF-EMFs for 24 h. Goldfish scales were then fixed with 10% formalin in 0.05 M cacodylate buffer (pH 7.4) and stored in 0.05 M cacodylate buffer at 4°C until use.

For exposure of zebrafish to ELF-EMFs, three zebrafish were transferred to a ring-shaped tank immediately after scale cutting, and the system water was circulated through the tank via a peristaltic pump to maintain the water temperature at 28.5°C. Zebrafish were exposed to ELF-EMFs for 4 h, then transferred to the circulating aquarium, and kept for 20 h. To induce higher eddy currents in system water by ELF-EMFs without stressing the zebrafish, NaCl was added to the system water at a concentration of 0.5%. After anesthetizing with 0.02% tricaine, fractured scales were extracted from the zebrafish and fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) overnight at 4°C. Fixed zebrafish scales were washed with a gradient of methanol in 0.1% Tween-20 (Sigma) in PBS (PBST) and kept in methanol at −20°C until use.

For ELF-EMF exposure to zebrafish fractured scales in the presence of IWR-1-endo, zebrafish with fractured stimuli were anesthetized with 0.02% tricaine, and fractured scales were extracted and placed in a 500 μL tube containing 500 μL conditioned medium (40% Leibniz’s L-15 medium (Wako), 32% Dulbecco’s modified Eagle’s medium (Wako), 12% Ham’s F12 medium (Wako), 8% FBS, 2 mM L-glutamine (Wako), 15 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma), 100 U/mL penicillin (Wako), and 100 μg/mL streptomycin (Wako) supplemented with 20 μM of IWR-1-endo or 0.2% DMSO for control. The tubes were then placed in a ring-shaped reservoir and exposed to ELF-EMFs for 4 h. The tubes were transferred to an incubator and incubated at 28.5°C for 20 h.

Fixed goldfish scales were transferred to each well of a 96-well plate and alkaline phosphatase (ALP) and tartrate-resistant acid phosphatase (TRAP) activity was measured as previously described (Suzuki and Hattori, 2002). For ALP activity, an aliquot of 200 μL of alkaline buffer (100 mM Tris-HCl, pH 9.5; 100 mM NaCl; 50 mM MgCl₂) was added to each well of a 96-well plate, followed by incubation at 20°C for 30 min with gentle agitation. After incubation, the reaction was stopped by adding 50 μL of 2 N NaOH. For TRAP activity, an aliquot of 200 μL of 10 mM para-nitrophenyl-phosphate and 20 mM tartrate in 0.1 M sodium acetate buffer (pH 5.3) was added to each well of a 96-well plate, followed by incubation at 20°C for 30 min with gentle agitation. After incubation, the reaction was stopped by adding 50 μL of 2 N NaOH. The 150 μL of colored solution was transferred to a new plate and absorbance was measured at 405 nm. The absorbance was converted to the amount of para-nitrophenol (pNP) produced using a standard curve for pNP. After measuring ALP and TRAP activities, the size of goldfish scales was measured with ImageJ. Thereafter, ALP and TRAP activities were normalized to the surface area (mm²) of each scale. The ALP and TRAP activity values were then shown relative to the mean value of the unexposed scale as 1.

Cells were collected from zebrafish scales and analyzed by flow cytometry as previously described (Kobayashi-Sun et al., 2020a). Extracted scales were treated with Liberase TM (Roche) in PBS for 1 h at 37°C. The cells were then filtered through a 40 μm mesh and washed with 2% fetal bovine serum (FBS) in Hanks’ balanced salt solution (Wako) by centrifugation (300X g). Data were acquired on a FACS Aria III (BD Biosciences). Data analysis was performed using the Kaluza software (ver. 1.3, Beckman Coulter). The volume of fluid acquired per sec in FACS Area III was calculated by acquiring Accudrop beads (BD Bioscieces) whose concentrations were measured using a hemocytometer (Funakoshi). The absolute cell number was calculated based on the sample volume, acquisition events, acquisition times, and the percentage of each cell fraction.

Whole-mount immunohistochemistry of zebrafish scales was performed as previously described (Kobayashi-Sun et al., 2020b). Briefly, fixed zebrafish scales were permeabilized with acetone, blocked with 2% blocking reagent (Roche), and incubated overnight at 4°C with chicken anti-GFP antibody (Aves, GFP-1020) at 1:1,000 dilution and rabbit anti-RFP antibody (for mCherry staining) (Abcam) at 1:1,000 or rabbit anti-β-catenin antibody (Cell Signaling Technology) at 1:100. After washing with PBST, zebrafish scales were incubated overnight at 4°C with goat anti-chicken IgY Alexa Fluor 488-conjugated (Abcam) at 1:1,000 and donkey anti-rabbit IgG Alexa Fluor 647-conjugated (Abcam) at 1:1,000 dilution. Zebrafish scales were then stained with Hoechst 33342 at 5 μg/mL in PBS for 30 min and mounted with ProLong Gold Antifade Mountant (Thermo Fisher Scientific).

Immunofluorescence of isolated cells was performed as previously described (Kimura et al., 2022). Briefly, cells were collected from zebrafish scales as described above, smeared with Cyto-tek 2500 (Sakura), and fixed with 4% PFA in PBS for 20 min at room temperature. After blocking with 0.1% gelatin-PBS, cells were stained with rabbit anti-β-catenin antibody at 1:100 dilution for 40 min at room temperature. After washing with 0.1% gelatin-PBS, cells were stained with donkey anti-rabbit IgG Alexa Fluor 647-conjugated (Abcam) at 1:1000 dilution for 40 min at room temperature. After washing with 0.1% gelatin-PBS, cells were stained with Hoechst 33342 and mounted with ProLong Gold Antifade Mountant as described above.

Zebrafish scales and cells were imaged using an FV10i confocal microscope and Fluoview FV10i-SW software (ver. 2.1.1) (Olympus). To compare osterix:mCherry and trap:GFP expression, the mean fluorescent intensity of mCherry per square micrometer and the percentage of coverage area of GFP were calculated in each fractured scale.

Total RNA was extracted from zebrafish scales using RNeasy Mini Kit (QIAGEN) and cDNAs were synthesized with ReverTra Ace qPCR RT Master Mix (Toyobo). Quantitative real-time polymerase chain reaction (qPCR) assays were performed using TB Green Premix Ex Taq II (TaKaRa) on a ViiA 7 Real-Time PCR System according to the manufacturer’s instructions (Thermo Fisher Scientific). Expression of ef1a (ef1a1l1) was used to normalize transcript levels using the ΔΔCt method. Primer sequences used for qPCR are listed in Supplementary Table S1.

3′RNA-seq was performed according to the Lasy-Seq Ver 1.1 method (Kamitani et al., 2019; Kashima et al., 2022). Briefly, total RNA was extracted from zebrafish scales as described above. Reverse transcription (RT) was performed using Super Script IV (Thermo Fisher Scientific) and an RT primer containing oligo-dT, the index sequence, and 9-bases unique molecular identifiers (UMIs). RNA/cDNA hybrids were purified using AMpure XP beads (Beckman Coulter), and second-strand DNA was synthesized using DNA polymerase I (Enzymatics). dsDNAs were fragmented using WGS Fragmentation Mix (Enzymatics) and ligated into customized adapters with WGS Ligase (Enzymatics). Adapter-ligated dsDNAs were amplified with KAPA HiFi ReadyMix (Nippon Genetics) and SE PCR primers. Next-generation sequencing of cDNA libraries was performed by Novogene using an Illumina NovaSeq X (illumina). Primer and adaptor sequences used for RNA-seq were listed in Supplementary Table S2.

Sequence reads were trimmed with fastp (ver. 0.21.0). The trimmed reads were then mapped to the zebrafish reference sequences generated from Danio_rerio.GRCz11 genome and version 109 gff with gffread (ver. 0.12.7), using BWA mem (version 0.7.17-r1188). Duplicated UMIs per gene in bam files were deduplicated with UMItools dedup with a parameter --per-contig. UMI counts of each transcript in each sample were calculated using salmon (ver. 1.4.0). The data have been deposited in Gene Expression Omnibus (GEO) (National Center for Biotechnology Information) and are accessible through the GEO database (series accession number, GSE247669). Gene set enrichment analysis (GSEA) was performed using the GSEA software (Broad Institute, ver. 4.1.0). Hierarchical clustering of each subset was performed in R (ver. 4.1.2) with the Bioconductor gplots package.

Statistical significance between groups was determined by unpaired two-tailed Student’s t-test, one-way ANOVA with Dunnett’s test, or Pearson’s chi square test. A value of p < 0.05 was considered statistically significant.

We first investigated the effects of ELF-EMFs at 60 Hz on ALP activity of osteoblasts and tartrate-resistant acid phosphatase (TRAP) activity of osteoclasts. Goldfish scales are larger than zebrafish ones, making it easier to test ALP and TRAP activity in the scale. Eight to ten scales per goldfish were extracted, transferred to tubes containing EMEM, and unexposed (u.e.) or exposed to ELF-EMFs at 60 Hz with varying intensities of 3, 5, 8, 10, 20, and 30 mT for 24 h (Figure 1A). Figures 1B, C show the relative values of ALP and TRAP activity at each intensity, respectively. A significant increase in both ALP and TRAP activity was observed in the goldfish scale exposed to 10, 20, and 30 mT compared to the u.e. scale, suggesting that exposure to 10 mT of ELF-EMFs may be sufficient to enhance both ALP and TRAP activity. In contrast, a significant decrease in TRAP activity was observed in the goldfish scale exposed to 3 mT compared to the u.e. scale, whereas ALP activity was unchanged.

To evaluate the effects of ELF-EMFs on fracture healing by osteoblasts and osteoclasts in vivo, we established an experimental system in which zebrafish were placed in a ring-shaped tank and exposed to ELF-EMFs while the system water was circulating at constant temperature. The epidermis portion of the scales in trap:GFP; osterix:mCherry double-transgenic zebrafish was cut with fine scissors to induce fracture, and these animals were unexposed or exposed to ELF-EMFs at 60 Hz with varying intensities of 3, 10, or 30 mT for 4 h. Zebrafish scales were then extracted at 24 h post-fracture (h.p.fr.) and analyzed by confocal microscopy (Figure 2A). Figure 2B shows representative expression patterns of osterix:mCherry or trap:GFP in the extracted scale of zebrafish unexposed or exposed to ELF-EMFs. We observed increased expression of osterix:mCherry throughout the fractured scale from zebrafish exposed to 10 mT (hereafter denoted as the “10 mT fractured scale”), but not in the 30 mT fractured scale. The mean fluorescent intensity of osterix:mCherry was approximately 1.73-fold higher in the 10 mT fractured scale compared to the u.e. fractured scale, whereas that in the 3 mT or 30 mT fractured scale was unchanged (Figure 2C), suggesting that exposure to 10 mT ELF-EMFs increases osteoblast activity. Similar results were also obtained in trap:GFP ⁺ osteoclasts. A large number of trap:GFP ⁺ osteoclasts accumulated at the fracture site in the 10 mT fractured scale, but not in the 3 or 30 mT fractured scale (Figure 2B). To compare osteoclast activity in each condition, we measured the percentage of coverage area of trap:GFP ⁺ within the fractured scale. The coverage area of trap:GFP ⁺ was approximately 2.06-fold higher in the 10 mT fractured scale than the u.e. fractured scale, but tended to be reduced in the 3 mT fractured scale, which was consistent with TRAP activity in the goldfish scale. The coverage area of trap:GFP ⁺ was unchanged in the 30 mT fractured scale (Figure 2D).

We followed the changes in osterix:mCherry and trap:GFP expression upon exposure to 10 mT ELF-EMFs until 3 days post-fracture (d.p.fr.). In the u.e. fractured scale, the coverage area of trap:GFP ⁺ increased gradually until 3 d.p.fr. In contrast, after a high increase by 2 d.p.fr., the coverage area of trap:GFP ⁺ began to decrease on 3 d.p.fr. in the 10 mT fractured scale, suggesting that bone resorption at the fracture site peaked out by 2 d.p.fr. in the 10 mT fractured scale. The mean fluorescent intensity of osterix:mCherry was unchanged by exposure to 10 mT ELF-EMFs at 2 and 3 d.p.fr. (Supplementary Figures S1A–C). These data suggest that exposure to 10 mT ELF-EMFs may have at least the effect of shortening the bone resorption period.

Since ELF-EMF exposure generates eddy currents strongly in saline water but weakly in freshwater, it may be possible that the effects of ELF-EMFs on the goldfish scale in EMEM and those on the zebrafish scale in freshwater are distinct. Therefore, we investigated the expression of osterix:mCherry and trap:GFP under conditions in which zebrafish are unexposed or exposed to 10 or 30 mT ELF-EMFs in system water containing 0.5% NaCl. As in freshwater, higher expression of osterix:mCherry and a greater number of trap:GFP ⁺ osteoclasts was observed in the 10 mT fractured scale in 0.5% NaCl compared to the unexposed or 30 mT fractured scale in 0.5% NaCl (Figures 3A, B). The mean fluorescent intensity of osterix:mCherry and the coverage area of trap:GFP ⁺ was approximately 1.82-fold and 2.09-fold higher in the 10 mT fractured scale compared to the 30 mT fractured scale in 0.5% NaCl, respectively, which were similar to those in freshwater (1.51-fold and 1.54-fold, respectively). We observed no significant differences in either the mean fluorescent intensity of osterix:mCherry or the coverage area of trap:GFP ⁺ between freshwater and 0.5% NaCl at 10 or 30 mT as well as unexposed controls (Figures 3C, D). These results suggest that eddy currents generated within the tissues/cells of the zebrafish scale, but not outside the body, promote the activity of both osteoblasts and osteoclasts.

To further investigate the effect of ELF-EMFs on osteoblasts and osteoclasts, we next quantified the number of osteoblasts and osteoclasts by flow cytometry. We have previously shown that mature osteoclasts with one to three nuclei are detected in the trap:GFP ^high fraction at 24 h.p.fr., whereas immature osteoclasts are in the trap:GFP ^low fraction. Due to uptake of osteoblast-derived extracellular vesicles, the majority of trap:GFP ⁺ osteoclasts are detected in the osterix:mCherry ⁺ fraction (Kobayashi-Sun et al., 2020a) (Supplementary Figure S2). The percentage of both osterix:mCherry ⁺ osteoblasts (“mCh⁺”) and trap:GFP ^high osteoclasts (“GFP^high”) was increased in the u.e. fractured scale compared to the intact scale, as shown in our previous study (Kobayashi-Sun et al., 2020a). These percentages were further raised in the 10 mT fractured scale under freshwater conditions (Figures 4A, B). We also quantified the absolute numbers of osterix:mCherry ⁺ osteoblasts and trap:GFP ^high osteoclasts per zebrafish scale in each condition. Despite an increase in the percentage of osterix:mCherry ⁺ osteoblasts with fracture stimulation, there was no significant change in the number of osteoblasts between the u.e. fractured scale and the intact scale. This is attributed to the decreased number of total cells in the fractured scale due to scale cutting. However, there was a significant increase in the number of osterix:mCherry ⁺ osteoblasts in the 10 mT fractured scale compared to the intact scale. There was also a significant increase in the number of trap:GFP ^high osteoclasts in the 10 mT fractured scale compared to the intact scale or the u.e. fractured scale (Figure 4C). These data suggest that exposure to 10 mT ELF-EMFs leads to increase in the number of both osteoblasts and osteoclasts in the fractured scale.

The percentage and number of osterix:mCherry ⁺ osteoblasts and trap:GFP ^high osteoclasts were also examined in the intact scale unexposed or exposed to 10 mT ELF-EMFs. An increase in the percentage and number of osteoclasts was observed in 10 mT intact scales compared to u.e. intact scales, while the increase in osteoblasts was not significant (Supplementary Figures S3A–C).

To investigate the molecular mechanisms underlying the effect of 10 mT ELF-EMFs on osteoblasts and osteoclasts, gene expression analysis was performed in the zebrafish scale. After fracture stimulation, zebrafish were unexposed or exposed to 10 mT ELF-EMFs for 4 h, followed by extraction of zebrafish scales at 24 h.p.fr. for qPCR analysis. As shown in Figure 5A, the expression of both osteoblast marker genes (alpl, col1a1a, and rankl) and osteoclast marker genes (trap, nfatc1, and rank) increased in the 10 mT fractured scale compared to the u.e. fractured scale or the intact scale. Since the number of osteoblasts and osteoclasts was increased by exposure to 10 mT ELF-EMFs, it is likely that proliferation signals for osteoblasts and/or osteoclasts may be increased in the 10 mT fractured scale. Therefore, fractured scales were extracted from zebrafish immediately after exposure to 10 mT ELF-EMFs and examined the mRNA expression of some critical signaling molecules at 4 h.p.fr. We found that the expression of canonical wnt genes, wnt10b, and Notch ligand genes, such as jag1a and jag1b, increased in the 10 mT fractured scale compared to the u.e. fractured scale. Particularly, wnt10b was expressed approximately 3.6-fold higher in the 10 mT fractured scale than in the u.e. fractured scale. In contrast, the expression of opg and il1b is upregulated by fracture stimulation but is reduced by exposure to 10 mT ELF-EMFs (Figure 5B). To examine whether Wnt and/or Notch signaling is enhanced by ELF-EMF exposure, RNA-seq analysis was performed on scales from zebrafish unexposed or exposed to 10 mT ELF-EMFs at 24 h.p.fr. Gene set enrichment analysis (GSEA) revealed that both Wnt target genes and Notch target genes significantly increased in the 10 mT fractured scale compared to the u.e. fractured scale (Figure 5C). Since the β-catenin-dependent canonical Wnt pathway has been shown to promote proliferation and differentiation of osteoblast precursors (Kato et al., 2002; Clément-Lacroixet al., 2005), the increased numbers of osteoblasts and osteoclasts by ELF-EMF exposure may be a consequence of enhanced canonical Wnt signaling. Indeed, some Wnt target genes involved in cell proliferation, such as mycn, jun, igf2a, edn1, and cd44a, were highly expressed in the 10 mT fractured scale compared to the u.e. fractured scale (Figure 5D).

Canonical Wnt signaling stabilizes β-catenin by preventing β-catenin phosphorylation from the destruction complex, which results in translocation of β-catenin into the nucleus to transcribe Wnt target genes (MacDonald et al., 2009; Liu et al., 2022). To examine whether β-catenin expression is enhanced by exposure to 10 mT ELF-EMFs, zebrafish with fracture stimulation were unexposed or exposed to 10 mT ELF-EMFs for 4 h, and fractured scales were extracted and stained with anti-β-catenin antibody. The expression of β-catenin was strongly detected in the 10 mT fractured scale, whereas it was very weak in the u.e. fractured scale. Interestingly, β-catenin expression signals were predominantly observed along the fractured site within the zebrafish scale (Figure 6A), suggesting that Wnt signaling is induced at the fracture site by ELF-EMF exposure. To determine the percentage of β-catenin (+) cells and the localization of β-catenin within the cell, cells collected from the zebrafish scale were smeared and stained with anti-β-catenin antibody. The expression of β-catenin was frequently detected not only in the cytoplasm but also in the nucleus in cells from the 10 mT fractured scale (Figure 6B). The percentage of β-catenin (+) cells was approximately 20.1% in the 10 mT fractured scale, whereas it was approximately 9.3% in the u.e. fractured scale (Figure 6C). Moreover, the percentage of nuclear β-catenin (+) cells out of total β-catenin (+) cells was approximately 66.7% in the 10 mT fractured scale, which was significantly higher than the u.e. fractured scale (46.5%) (Figure 6D). We also examined β-catenin expression in osterix:mCherry ⁺ osteoblasts and trap:GFP ⁺ osteoclasts. The percentage of osterix:mCherry ⁺ osteoblasts within total β-catenin (+) cells was approximately 6.6% ± 2.2% and 7.2% ± 5.1% (±s.e.m.; n = 6) in the u.e. fractured and 10 mT fractured scale, respectively, whereas that of trap:GFP ⁺ osteoclasts was less than 0.1% in both types of the fractured scale. The percentage of nuclear β-catenin (+) osteoblasts out of total osteoblasts was approximately 48.5% in the 10 mT fractured scale, significantly higher than in the u.e. fractured scale (17.3%). In contrast, there was no significant difference in the percentage of nuclear β-catenin (+) in osteoclasts (Supplementary Figures S4A, B). Although more than 90% of β-catenin (+) cells were not osteoblasts or osteoclasts, elevated expression of β-catenin in osteoblasts was observed in the 10 mT fractured scale. Taken together, these data suggest that exposure to 10 mT ELF-EMFs enhances Wnt/β-catenin signaling particularly at the fracture site of the zebrafish scale.

To test whether Wnt/β-catenin signaling promotes the proliferation of osteoblasts and osteoclasts in the fractured scale, we inhibited this signaling by treatment with IWR-1-endo, which promotes β-catenin phosphorylation via stabilization of the Axin-scaffold destruction complex (Huang et al., 2009). After fracture stimulation, zebrafish scales were extracted and transferred to conditioned medium supplemented with dimethyl sulfoxide (DMSO) or IWR-1-endo and exposed to 10 mT ELF-EMFs for 4 h. Cells were then collected from fractured scales at 24 h.p.fr. and analyzed by flow cytometry (Figure 7A). The number of both osterix:mCherry ⁺ osteoblasts and trap:GFP ^high osteoclasts was significantly increased by exposure to 10 mT ELF-EMFs, but it was decreased by IWR-1-endo treatment. Similar results were also obtained in the percentage of each cell type (Figures 7B–E). We also performed gene expression analysis of an osteoblast marker gene (col1a1a), an osteoclast marker gene (nfatc1), and three Wnt target genes (fos1b, ccn1, and mmp9) in four different types of scales, intact DMSO-treated, unexposed fractured DMSO-treated, 10 mT fractured DMSO-treated, and 10 mT fractured IWR-1-endo-treated. All of these five genes showed similar expression patterns; they increased by exposure to 10 mT ELF-EMFs, but decreased by IWR-1-endo treatment (Figure 7F). Taken together, these data suggest that Wnt/β-catenin signaling enhanced by ELF-EMF exposure promotes proliferation of both osteoblasts and osteoclasts under fracture stress.