Fig 2: RAW-exos suppress the expression of Arachidonate 15-lipoxygenase in osteoblasts by carrying miR-3102-5p. (A–E) mRNA expression level (**P<0.01, ***P<0.001, ****P<0.0001, N=3). mRNA expression level of Osteocalcin (Ocn) in osteoblasts (A). mRNA expression level of Runt-related transcription factor 2 (Runx2) in osteoblasts (B). mRNA expression level of Receptor activator of nuclear factor kappa-Β ligand (Rankl) in osteoblasts (C). mRNA expression level of Tumor necrosis factor-alpha (TNF-α) in osteoblasts (D). mRNA expression level of Interleukin-6 (IL-6) in osteoblasts (E). (F) Expression of ALOX15, SOD, GPX4, and OCN proteins in osteoblast. (G) Bioinformatics prediction of miRNAs binding to ALOX15 by TargetScan data(http://www.targetscan.org/) and miRWalk data (http://mirwalk.umm.uni-heidelberg.de/). (H–J) miRNA expression level of RAW-Exos (**P<0.01, N=3). (K) Binding site of miR-3102-5p, miR-7638-3p and Arachidonate 15-lipoxygenase (alox15). (L) Firefly luciferase reporter gene assay to detect the binding activity between miR-3102-5p and alox15 (*P<0.05, N=3).

Fig 3: Nuclear deformation promotes osteogenic differentiation of hMSCs.a Staining of nucleus (green) and F-actin (red) of hMSCs on flat and micropillar mPOC/HA surfaces. Insert: high magnification of cell nucleus. Dashed lines indicate micropillars. b Analysis of nuclear shape index of hMSCs. n = 117 (flat) and 132 (pillar) collected from 3 biological replicates, ****p < 0.0001. c Orthogonal view of cell nucleus on flat and micropillar surfaces. d Nuclear volume analysis based on 3D construction of the confocal images of cell nuclei. n = 35 cells collected from 3 biological replicates, ****p < 0.0001. e Initial cell adhesions on flat and micropillar surfaces. n = 5 biological replicates, N.S., no significant difference. f SEM images show the cell adhesions on flat and micropillar mPOC/HA surfaces. g Live/dead staining of hMSCs on flat and micropillar surfaces at 72 h in osteogenic medium. h Cell metabolic activity of cells on flat and micropillar surfaces tested by a MTT assay. n = 5 biological replicates, ****p < 0.0001. i Cell proliferation tested via DNA content after 72 h induction. n = 5 biological replicates, N.S., no significant difference. j ALP staining of hMSCs on flat and micropillar surfaces after 7 d induction. k ALP activity test of cells after 7 d osteogenic induction. n = 3 biological replicates. l Blot images of osteogenic marker OCN and RUNX2 in cells cultured on flat and micropillar implants. GAPDH is shown as a control. Quantification (m) OCN and (n). RUNX2 according to Western blot tests. n = 3 biological replicates, ****p < 0.0001. Data are presented as mean ± SD. Values from two groups were compared using a non-paired Student’s t-test (two-sided). Source data are provided as a Source Data file.

Fig 4: mPOC/HA micropillar implant promotes bone regeneration in vivo.a Image shows implantation of hMSC seeded flat and micropillar mPOC/HA scaffolds. b Staining images of nuclei (green) and F-actin (red) of cells on the implants. c Representative μCT images of a typical animal implanted with hMSC-seeded flat (left) and micropillar (right) scaffolds at 12-weeks post-surgery. d Regenerated bone volume in the defect region (n = 5 animals). e Trichrome staining of the defect tissue treated with flat and micropillar implants. f Average thickness of regenerated tissues with implantation of flat and micropillar scaffolds (n = 5 animals). IHC staining of osteogenic marker, g OPN and h. OCN, in regenerated tissues with flat and micropillar implants. Data are presented as mean ± SD. Values from two groups were compared using non-paired Student’s t-test (two-sided). Source data is provided as a Source Data file.

Fig 5: Comparative effects of BMP9 and BMP2 on osteogenic differentiation and osteoclastogenesis in vitro. (A) Real‐time PCR analysis of key osteogenic genes (Col1, Runx2, ALP, and OCN) in MC3T3‐E1 cells treated with 8 nM of BMP2 or BMP9 for 3, 5, and 7 days. All gene‐expression levels were normalized to GAPDH. (B) Western blot analysis of osteogenic marker proteins in cell lysates harvested after 7 days of treatment with BMP2 or BMP9. GAPDH was used as the loading control. Densitometric quantification of band intensities (integrated density) normalized to GAPDH is shown below the blots and presented as relative protein expression. (C) Western blot showing dose‐dependent p‐Smad1/5/9 in MC3T3‐E1 cells exposed to varying concentrations of BMP2 or BMP9. Phosphorylation was quantified by densitometry and expressed as fold change vs. control after normalization using [(p‐Smad1/5/9)/(total Smad1/5/9)] and further normalized to GAPDH, as shown in the graph below the blots. Asterisks indicate statistical significance for pairwise comparisons between BMP2 and BMP9 at the same concentration (****, p < 0.0001), unless otherwise indicated. (D) ALP activity and representative images of ALP staining in MC3T3‐E1 cultures after 7 days of induction with BMP2 or BMP9. (E) Alizarin Red S staining illustrating mineralized nodule formation after extended culture with BMP2 or BMP9. (F) Representative TRAP‐stained images of RAW 264.7‐derived osteoclasts following treatment with RANKL (3 nM), BMP2 (8 nM), or BMP9 (8 nM) for 5 days. TRAP‐positive multinucleated osteoclasts are indicated by arrows. Scale bar, 20 μm. (G) Quantification of TRAP‐positive multinucleated cells per well. Data are presented as the mean ± SD (n = 3 independent experiments), and p‐values were calculated using one‐way analysis of variance (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). BMP, bone morphogenetic protein; PCR, polymerase chain reaction; ALP, alkaline phosphatase; Col1, collagen type I; Runx2, runt‐related transcription factor 2; OCN, osteocalcin; GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase.