a, b Representative images of an intact a or fractured scale b in trap:GFP; osterix:mCherry double-transgenic zebrafish. Dotted lines in a show a boundary of the dermis (der) and epidermis (epd) area. The right panel in b shows a high magnification view of the white boxed area in the left panel. c Representative time-course changes of a fractured scale. The inset in the left panel shows a high magnification view of the blue boxed area. Images are orientated with the dorsal side to the top and anterior side to the left. Arrows indicate trap:GFP⁺ cells observed in the edge area of the fractured scale. Dotted lines in b and c show the fracture site. DIC differential interference contrast. Hoe Hoechst 33342; dpf days post-fracture; bars, 200 μm a; 10 μm (right panel in b); 100 μm (left panel in b, c). Experiments were performed twice with three biological replicates in each group a–c.

a Representative results of flow cytometric analysis of cells from intact (left panel) or fractured scales at 1 day post-fracture (dpf) (right panel). Red, orange, and green gate show trap:GFP^–osterix:mCherry⁺ (“mCh⁺”), trap:GFP^lowosterix:mCherry⁺ (“GFP^low”), and trap:GFP^high (“GFP^high”) cells, respectively. b Absolute number of mCh⁺, GFP^low, and GFP^high cells in an intact or fractured scale at 1 dpf. Error bars, s.e.m. (n = 9 for each group); n.s., no significance; *p < 0.05; **p < 0.001 by Student’s t-test. c–e Representative fluorescent images of mCh⁺c, GFP^lowd, and GFP^high cells e. Arrows indicate an osterix:mCherry⁺ particle observed in the cytoplasm. Numbers in bottom left of panels indicate the number of cells showing the displayed morphology over the total number of analyzed cells. DIC differential interference contrast. f Representative electron microscopic images of a GFP^high cell. Arrowheads show vesicles, which include secondary lysosomes, early endosomes, and multi-vesicular bodies. n nucleus; m mitochondrion; g Golgi apparatus; bars, 5 μm c–e; 1 μm f. Experiments were performed twice with nine biological replicates a, b and two biological replicates c–f in each group.

a Schematic diagram of transplantation assays. Hematopoietic cells from trap:GFP; osterix:mCherry double-transgenic zebrafish kidney were transplanted into wild type recipients irradiated with sublethal dose of X-ray (Transplantation-1, n = 3). Hematopoietic cells from trap:GFP single-transgenic zebrafish kidney were transplanted into osterix:mCherry single-transgenic recipients (Transplantation-2, n = 4). After 20 or 8 weeks post-transplantation, cells in scales at 1 day post-fracture (dpf) were analyzed by flow cytometry (FCM) and/or confocal microscopy. b Representative results of FCM analysis in fractured scales from a recipient in Transplantation-1 (left) and recipient in Transplantation-2 (midle). Red, orange, and green gate show trap:GFP⁻osterix:mCherry⁺ (“mCh⁺”), trap:GFP^lowosterix:mCherry⁺ (“GFP^low”), and trap:GFP^high (“GFP^high”) cells, respectively. GFP^high cells in a recipient of Transplantation-2 are displayed in an osterix:mCherry vs. side scatter (SSC) dot plot (right panel). c Representative images of a fractured scale from a trap:GFP; osterix:mCherry double-transgenic animal (left) and recipient in Transplantation-1 (right). Dotted lines indicate the fracture site. Both images showed merged channels of GFP and mCherry. Bars, 100 μm. Experiments were performed twice with three or four biological replicates in each group.

a Representative time-lapse imaging of the fractured scale in a trap:GFP; osterix:mCherry double-transgenic animal. Six sequences from Supplementary Movie 6 are presented, documenting the stepwise engulfment of OB-derived EVs (arrowheads) by a trap:GFP⁺ cell (arrows). b Single plane of a z-stack of a trap:GFP⁺ cell at the time point of 39 min. Images show a green (trap:GFP), red (osterix:mCherry), and merged channel. Bars, 10 μm. Experiments were performed six times with similar results.

a Representative flow cytometric analysis of cells in fractured scales at 1 day post-fracture (dpf) from an osterix:mCherry single-transgenic zebrafish. Gated regions in the left panel indicate the osterix:mCherry⁺ Hoecht 33342^high (mCh⁺ Hoe^high) cell fraction and mCh⁺ Hoe^low EV fraction. mCh⁺ Hoe^high cells and mCh⁺ Hoe^low EVs are displayed in a forward scatter (FSC) vs. side scatter (SSC) dot plot (middle and right panels, respectively). b Representative electron microscopic images of isolated large and small EVs. c Representative image of isolated EVs negatively stained. Arrows indicate an EV. d Percent size distribution of EVs (n = 223). e Absolute number of mCh⁺ Hoe^low EVs in an intact and fractured scale at 1 dpf. **p < 0.01 (n = 9 for each group). Bars, 1 μm b; 2 μm c. Experiments were performed twice with two biological replicates a–d and nine biological replicates e in each group.

a Representative flow cytometric analysis of cells in scales at 1 day post-fracture (dpf) from a trap:GFP; osterix:mCherry double-transgenic animal. Gated regions in the left panel indicate the Hoechst 33342^high (Hoe^high) cell fraction and osterix:mCherry⁺ Hoe^low fraction (“EV”). Cells in the Hoe^high fraction are displayed in the right panel to further divide into three populations, trap:GFP⁻osterix:mCherry⁺ (“mCh⁺”), trap:GFP^lowosterix:mCherry⁺ (“GFP^low”), and trap:GFP^high (“GFP^high”). b Principal component analysis (PCA) based on the read per million (RPM) of each sample. c, d Hierarchical clustering of selected OC-related c and OB-related genes d in the mCh⁺, GFP^low, GFP^high, and EV fraction in the fractured scale. e Gene ontology enrichment analysis of highly expressed genes in the mCh⁺, GFP^low, GFP^high, and EV fraction. Experiments were performed once with two biological replicates.

a Representative flow cytometric analysis of cells in scales at 1 day post-fracture (dpf) from a osterix:mCherry single-transgenic animal. Gated regions indicate the mCh⁺ Hoe^high cell fraction and mCh⁺ Hoe^low EV fraction. b mCh⁺ Hoe^high cells and mCh⁺ Hoe^low EVs were displayed in an Annexin-V-FITC vs. Sytox Red dot plot. mCh⁺ Hoe^high cells were subdivided into three fractions, “live”, “pre-apoptotic”, and “apoptotic”, whereas mCh⁺ Hoe^low EVs were divided into two fractions, “microvesicle” (MV) and “apoptotic body” (AB). c Schematic diagram of in vitro cell culture assays. Kidney marrow cells (KMCs) from trap:GFP single-transgenic zebrafish were co-cultured with OBs, MVs, or ABs from fractured scales of osterix:mCherry single-transgenic zebrafish in a fibronectin-coated plate. At 2 days of co-culture, the number of trap:GFP⁺ cells was counted in each well. Non-co-cultured KMCs were used as a control. d The average number of trap:GFP⁺ cells in each type of wells. Error bars, s.d. (n = 4 for each group). e Representative images of trap:GFP⁺ cells co-cultured with OBs (left panel), MVs (middle panel), or ABs (right panel). trap:GFP⁺ cells contained OB-derived EVs in the cytoplasm (arrows). Bars, 5 μm. f Schematic diagram of zebrafish rankl locus. gRNA target sites and primer recognition sites are shown in red bars and blue arrows, respectively. g qPCR analysis of rankl in the adult fin of wild type control and rankl gRNA-injected animals. Results from seven individual gRNA-injected animals are shown. Data are mean ± s.d. from three independent experiments. h The average number of trap:GFP⁺ cells in non-co-cultured KMCs (control), or KMCs co-cultured with OBs or EVs from wild type or rankl gRNA-injected animals. Error bars, s.d. (n = 4 for each group); n.s., no significance; *p < 0.05; **p < 0.01; ***p < 0.001 by one-way ANOVA followed by Dunnett’s test. Experiments were performed twice with four biological replicates in each group d, h.