Calcium silicate promotes bone stem cell energy changes that support bone formation and healing

The ScienCell Company (San Diego, USA) was the source of primary hBMSCs. The cells were cultivated and expanded in alpha minimum essential medium (α-MEM; Invitrogen, USA) enriched with 15% fetal bovine serum (FBS; Gibco, USA), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, USA) at 37°C in a humidified 5% COincubator. hBMSCs at passages 3-8 were utilized for subsequent analyses. 2

The mesenchymal phenotype of the hBMSCs was confirmed as per the criteria recommended by the International Society of Cellular Therapy (ISCT). Suspended hBMSCs were collected and stained with antibodies specific for CD73/90/105/34/45 (PE-conjugated antihuman antibodies; eBioscience, CA, USA) for 30 min in darkness at 4°C. Expression profiles were assessed using a flow cytometer system (BD Bioscience, San Jose, CA, USA).

For the preparation of CS extracts, CS powder (1 g) (Kunshan, China) was eluted in 50 ml of α-MEM at 37°C for 24 h. The extracts were subsequently collected and filtered via a sterile 0.22-μm microfiltration membrane. Various dilutions (final dilutions: 50%, 25%, 12.5% and 6.25%) of the medium extracts were then prepared in α-MEM.

Si, Ca, and phosphorus (P) ion levels in the CS extracts were assessed via inductively coupled plasma optical emission spectroscopy (ICP-OES) (iCAP 6500 Duo; Thermo Fisher Scientific, Florence, KY, USA).

For cell growth assessment, hBMSCs were cultivated in 96-well plates (3 × 10cells/well, = 5) in the presence or absence of various CS extract dilutions. After 1, 2, and 3 days in culture, Cell Counting Kit (CCK)-8 reagent was applied to a 96-well plate (10 μl/well) (Dojindo Molecular Technology; Japan), which was subsequently incubated at 37°C for 2 h. The optical density (OD) at 450 nm was subsequently assessed via an enzyme-linked immunosorbent assay (ELISA) plate reader. 3 n

hBMSCs were cultured in a 6-well plate with or without 25% CS extract for 24 h. To induce osteogenic differentiation, the medium was substituted with osteogenic medium (OM) containing 100 µM L-ascorbic acid 2-phosphate, 2 mM β-glycerophosphate, and 10 nM dexamethasone (all from Sigma-Aldrich). After 7 days, hBMSCs were rinsed with PBST (PBS +0.05% Tween-20) and fixed with 4% paraformaldehyde (PFA) for 2 min at room temperature (RT). Then, cell incubation was conducted with BCIP/NBT staining solution (Stemgent, USA) for 15–30 min at RT. Purple staining indicated alkaline phosphatase (ALP)-positive cells. After 14 days, matrix mineralization was evaluated via 1% alizarin red staining (ARS; Sigma-Aldrich) for 30 min at RT, and mineralized nodules were imaged.

Total RNA isolation from cultivated hBMSCs was conducted via the RNA-Quick Purification Kit (Yishan, China) per the manufacturer's protocol. cDNA synthesis was conducted via the PrimeScript™ RT Reagent Kit (TaKaRa, Japan). Real-time polymerase chain reaction was conducted with Fast SYBR Green Master Mix (Thermo Fisher, USA) and gene-specific primers (). The normalization of mRNA levels to β-actin was conducted via the 2method. Supplementary Table S1 –ΔΔCT

hBMSCs were stimulated with 25% CS extract for 24 h. MitoTracker Green, MitoTracker Red CMXRos (to assess the mitochondrial membrane potential), and MitoSOX (to detect mtROS) staining were subsequently conducted as per the manufacturer's guidelines (Invitrogen, USA). The mitochondrial profile of hBMSCs was estimated via a flow cytometer system (BD Bioscience, CA, USA).

Total protein isolation from hBMSCs was conducted via RIPA lysis buffer with protease and phosphatase inhibitors, and protein levels were assessed via a BCA assay kit (all from Thermo Fisher Scientific, USA). For WB, separation of 20 μg of protein was conducted by SDS-PAGE (Invitrogen, USA) and transferred to 0.2 μm PVDF membranes (Millipore, USA). 5% bovine serum albumin (BSA) and 0.1% Tween-20 (Sigma, USA) were utilized to block the membranes for 1 h at RT, then an overnight incubation was conducted at 4°C with primary antibodies against human targets: LC3I/II (Cell Signaling Technology, Beverly, USA), p62 (Santa Cruz Biotechnology, Dallas, USA), and β-actin (Sigma-Aldrich, USA), and the latter was used as a loading control. A 1-h incubation of the membranes was conducted at RT with species-specific horseradish peroxidase (HRP)-conjugated secondary antibodies (1:10 000; Cell Signaling Technology). Immunoreactive bands were visualized via enhanced chemiluminescence (ECL) reagents (Millipore, USA).

Transfection of hBMSCs was conducted with a lentiviral plasmid encoding mCherry-GFP-LC3B (pLenti-CMV-mCherry-GFP-LC3B-IRES-Puro, OBiO Technology) to evaluate autophagic flux. Autolysosomes (mCherry-positive, red) and autophagosomes (mCherry/GFP double-positive, yellow) were detected via a laser-scanning confocal microscope (Zeiss LSM 980, Germany).

For immunostaining of translocase of the outer membrane 20 (Tom20) and/or lysosomal-associated membrane protein 1 (LAMP1), hBMSCs (2 × 10cells/well) were plated on chamber slides (Nunc, NY, USA). Subsequently, 4% PFA was utilized to fix the cells for 15 min, 0.1% Triton X-100 to permeabilize them, and 5% BSA to block them. The incubation of slides was conducted with primary antibodies overnight at 4°C and followed by incubation with Alexa Fluor 488- or 568-conjugated secondary antibodies (1:200, Invitrogen) for 1 h at RT. Lastly, the mounting of slides was conducted via Vectashield mounting medium with 4',6-diamidino-2-phenylindole (DAPI; Abcam, USA). 3

For immunostaining of osteopontin (OPN) and ALP, hBMSCs (1.5 × 10cells/well) were seeded on chamber slides and then induced to undergo osteogenic differentiation for 4 days. After fixation, permeabilization, and blocking of nonspecific sites, the incubation of cells was conducted with anti-OPN/ALP primary antibodies (1:300; Affinity, Jiangsu, China) and then with Alexa Fluor 649-conjugated secondary antibodies for 1 h at RT. The cytoskeleton was stained with ActinGreen 488 ReadyProbes (Thermo Fisher Scientific, USA). The slides were mounted using Vectashield mounting medium with DAPI, and observed under a confocal microscope (Zeiss LSM 980, Zeiss, Germany). The mean fluorescence intensity (MFI) was quantified using Image J software. 3

hBMSCs were treated with or without 25% CS extract for 24 h and seeded into Seahorse XF96 microplates at 8 × 10cells/well. Before the assay, cells were rinsed and kept at 37°C in a non-COincubator with XF base medium (Seahorse; 102353; Agilent) enriched with 2 mM L-glutamine and 2 mM sodium pyruvate (Gibco) for 1 h. Real-time OCR measurements were performed using an XF-96 Analyzer (Seahorse Bioscience) following sequential injection of Oligo (1 mM), FCCP (0.5 mM), and rotenone (Rot; 100 nM) plus AA (1 mM) (all from Sigma–Aldrich). Basal respiration equaled the OCR recorded prior to the first injection minus the lowest OCR observed after the final injection (nonmitochondrial oxygen consumption). ATP production was derived from basal OCR minus post- Oligo OCR. Maximal respiratory capacity (MRC) was calculated as peak OCR following FCCP administration minus nonmitochondrial OCR, while spare respiratory capacity was calculated as maximal respiration minus basal respiration. 4 2

After 24 h of stimulation with 25% CS extract, hBMSCs were collected by scraping and centrifugation. Cell pellets (1 × 10hBMSCs) were sent to the Beijing Genomics Institute (BGI) for metabolite profiling via LC–MS/MS in positive and negative ion modes. Data from both modes were merged, and duplicate metabolites were eliminated. The resulting dataset was analyzed using MetaboAnalyst 5.0 (McGill University, Canada). Principal component analysis (PCA) and heatmaps were generated using the "Statistical Analysis" tool, and metabolite set enrichment analysis (MSEA) was conducted via the "Enrichment Analysis" tool. 7

Transmission electron microscopy (TEM) was performed to assess mitochondrial morphology and the formation of autophagosomes and autolysosomes. Briefly, hBMSCs were treated with or without 25% CS extract for 24 h in complete αMEM. The collected cells were fixed with 2.5% glutaraldehyde and sliced into ultrathin sections. A Tecnai G2 spirit Twin transmission electron microscope (Czech Republic) was utilized to assess the sections at 100 kV.

Male BALB/c nude mice (18–21 g, 6–8 weeks old) were acquired from the Laboratory Animal Center of Sun Yat-Sen University (Guangzhou, China), with each animal serving as a single experimental unit. Mice were housed in specific pathogen-free circumstances at 23°C under a 12-h light/dark cycle with unrestricted access to food and water. The Institutional Animal Care and Use Committee of Sun Yat-Sen University gave its approval for each procedure (No. SYSU-IACUC-2023-000185).

Porous HA/β-TCP scaffolds (HA: TCP = 6:4, 4 mm diameter and 2 mm depth; National Engineering Research Center for Biomaterials, Chengdu, China) with pore sizes of 100–500 μm and 75% porosity were used as cell carriers. Control hBMSCs (5 × 10) or hBMSCs subjected to different treatments (5 × 10) were dropped onto each β-TCP scaffold under sterile conditions and culturedfor 24 h. After using 1% sodium pentobarbital (50 mg/kg) to anesthetize the nude mice, 20 mice were allocated to the following 4 groups ( = 5) in a random manner using a random number generator to ensure unbiased group assignment: (1) the control group, in which scaffolds seeded with untreated hBMSCs were applied; (2) the CS group, in which scaffolds seeded with CS-treated hBMSCs were applied; (3) the CS+Oligo group, in which scaffolds seeded with hBMSCs treated with the OXPHOS inhibitor Oligo in the presence of CS were applied; and (4) the CS+BafA1 group, in which scaffolds seeded with hBMSCs treated with the lysosomal inhibitor bafilomycin A1 (BafA1) in the presence of CS were applied. A 5–10 mm longitudinal incision was created in the skin using sterile surgical scissors. Then, the subcutaneous tissue was bluntly dissected with forceps to form a small pocket. The cell-loaded scaffolds were subsequently transplanted subcutaneously into the BALB/c nude mice. After eight weeks, the mice were euthanized, and the scaffolds were collected and fixed in 4% PFA. No animals were excluded from this experiment. 5 5 in vitro n

Control hBMSCs or hBMSCs subjected to different treatments (CS-treated hBMSCs, CS+Oligo-treated hBMSCs, or CS+BafA1-treated hBMSCs) were seeded onto PLGA scaffolds (Shandong Academy of Pharmaceutical Sciences, Shandong, China) for 24 h. 30 mice were allocated in a random manner to the following 5 groups ( = 6) using a random count generator to ensure unbiased group assignment: (i) the negative control (NC) group, in which scaffolds without cells were applied; (ii) the control group, in which scaffolds seeded with untreated hBMSCs were applied; (iii) the CS group, in which scaffolds seeded with CS-treated hBMSCs were applied; (iv) the CS+Oligo group, in which scaffolds seeded with hBMSCs treated with the OXPHOS inhibitor Oligo in the presence of CS were applied; and (v) the CS+BafA1 group, in which scaffolds seeded with hBMSCs treated with the lysosomal inhibitor BafA1 in the presence of CS were applied. After anesthetization with 1% sodium pentobarbital (50 mg/kg), an incision was made in the midline of the mouse cranium. Full-thickness defects (a 4-mm diameter) were formed using a dental drill on the right side of the mouse skull via saline irrigation, and cell-loaded scaffolds or scaffolds without cells (for the NC group) were implanted into the defects of BALB/c nude mice. The mice were sacrificed 8 weeks post-surgery, and their craniums were collected and fixed in 4% PFA. Subsequently, if a nude mouse died, if the incision or implanted scaffold caused skin ulceration and erosion, or if the incision failed to heal, the sample from that particular nude mouse was eliminated from the study. Herein, one nude mouse in the CS+BafA1 group exhibited an unhealed incision, and the scaffold was ejected; this mouse was excluded from further study and was humanely euthanized with an overdose of sodium pentobarbital. in vitro n

Fixed specimens were subjected to Micro-computed tomography (micro-CT; μCT50, Scanco Medical, Switzerland) scanning and evaluation. All the samples were scanned at 70 kV and 114 mA, with a spatial resolution of 10 μm. Thereafter, volumetric three-dimensional (3D) reconstructions and histomorphometric analysis were conducted with built-in software. The bone volume fraction (BV/TV) was calculated within the regions of interest. The analysis was conducted by an investigator blinded to the group allocation.

After decalcification in 10% EDTA for 2 weeks at RT, samples were paraffin-embedded and sectioned at 5 μm. Hematoxylin and eosin (HE) staining was conducted to assess bone structure, and Masson's trichrome staining evaluated collagen maturity. Sections were scanned using a slice scanning microscope (Aperio AT2, Leica, Germany). Immunofluorescence staining employed goat anti-rabbit ALP (1:150, Affinity, Jiangsu, China) and anti-rabbit OPN (1:150, Affinity) antibodies. Nuclei were counterstained with DAPI, and images were observed via confocal microscopy (Zeiss LSM 980, Germany). Analyses were conducted by an investigator blinded to group distribution.

Statistical analyses were conducted via GraphPad Prism 9 (USA). Data are represented as mean ± standard deviation (SD). Parametric comparisons were performed using Student'stest or one-way analysis of variance (ANOVA). When the overall ANOVA was significant, pairwise comparisons between groups were performed using Bonferronitests. The Kruskal–Wallis rank sum test was applied for data with unequal variations.< 0.05 was deemed significant. t post hoc P

Flow cytometry demonstrated that up to 94% of the cells were positive for MSC surface markers, including CD73/90/105, but negative for the hematopoietic markers CD34/45, with 0.55% of the cells expressing these markers (). Thus, the cultured cells possessed a mesenchymal phenotype. To explore the impacts of CS extracts on the growth of hBMSCs, we cultivated cells in 50%, 25%, 12.5% or 6.25% CS extracts for 24, 48 or 72 h. The CCK-8 revealed no significant variation in cell growth after 24 h, although cell growth slightly increased after culture with various dilutions of CS extract. Interestingly, the groups treated with 25% CS extract exhibited greater proliferation than did the control and other groups after culture for 48 or 72 h (). Therefore, 25% was selected as the optimal level for the subsequent experiments. After 7 days of culture under osteogenic conditions, hBMSCs treated with 25% CS extract had significantly greater ALP activity, accompanied by upregulated mRNA expression of osteogenic markers, including,,and(), than did the control hBMSCs (). Furthermore, compared to the control+OM group, the CS+OM group presented a higher ARS intensity, indicating enhanced matrix mineralization post-osteogenic induction for 14 days (). Immunofluorescence staining further confirmed the elevated expression of OPN and ALP in CS+OM-treated cells (and). Figure 1A Figure 1B Figure 1C and D Figure 1C Figure 1E Supplementary Figure S1A Runx2 Osx ALP Collagen 1 Col-1

Moreover, CS can release large numbers of Ca and Si ions to trigger the osteogenic differentiation of BMSCs. Herein, the concentrations of Si (105.7 mg/L) and Ca (171.3 mg/L) ions were greater in the CS extract than in the α-MEM (less than 1 mg/L and 70.3 mg/L, respectively) (). However, the addition of CaCl(69.4 mg/L) to α-MEM, resulting in a Ca ion concentration identical to that in the 25% CS extract, did not significantly enhance the osteogenic differentiation of hBMSCs, as indicated by ALP and ARS staining (). Moreover, the addition of CaClupregulated only theexpression without altering the levels of,or(). These data indicate that, in our study, CS induced the osteogenic differentiation of hBMSCs, possibly mainlySi ion release. Table 1 Supplementary Figure S1B Supplementary Figure S1C 2 2 Col-1 Runx2 Osx ALP via

Mitochondrial OXPHOS is vital for the osteogenic differentiation of MSCs []. We thus explored the bioenergetic profiles of hBMSCs treated with or without CS extract by measuring the OCR. Our results revealed greater OXPHOS activity, such as basal respiration, ATP generation, and MRC, in the CS-treated hBMSCs than in the control hBMSCs. However, the CS-treated hBMSCs did not exhibit significantly greater spare respiration than the control hBMSCs did (). To assess global metabolic changes, a comprehensive LC–MS/MS-based metabolomic analysis was performed on CS-treated and control hBMSCs (the original data are included in the). We verified that the metabolic profile of the CS-treated hBMSCs, which presented increases in the concentrations of the citric acid cycle metabolites, citric acid and L-malic acid, was distinct from that of the control hBMSCs (). MSEA demonstrated the enrichment of metabolites related to the OXPHOS pathway, such as pyruvate metabolism, transfer of acetyl groups into mitochondria and the citric acid cycle, in the CS-exposed hBMSCs (). Consistent with OCR analysis, MSEA further suggested metabolic commitment to OXPHOS in the CS-stimulated hBMSCs. ²⁰ Figure 2A and B Supplementary Materials Figure 2C, E and F Figure 2D

To further confirm whether the OXPHOS induction by CS contributes to the promoted BMSC osteogenic differentiation, we used the OXPHOS inhibitor Oligo. Oligo caused a significant decrease in the OCR in the CS-stimulated hBMSCs (). Furthermore, the capability of CS to trigger the osteogenic differentiation of hBMSC decreased in the presence of Oligo, as indicated by decreased ALP and ARS staining intensities (), as well as decreased expression of,,and(). Similarly, compared to that in the CS group, the protein expression of OPN and ALP in the CS+Oligo group was downregulated, as shown by decreased immunofluorescence staining intensity (). Collectively, these outcomes illustrate that CS promotes the osteogenesis of hBMSCs by inducing their metabolic preference for OXPHOS. Figure 2A Supplementary Figure S2A Supplementary Figure S2B Supplementary Figure S2C and D Runx2 Osx ALP Col-1

The structure of mitochondria is highly plastic and dynamic, and mitochondria undergo both fusion and fission procedures. Mitochondrial dynamics affect the efficiency of OXPHOS. Mitochondrial fragmentation can impair OXPHOS, whereas mitochondrial fusion enhances OXPHOS levels [,]. Thus, we investigated whether the morphology of mitochondria in hBMSCs is modulated by CS. We observed that compared with control hBMSCs, the CS-treated hBMSCs had elongated mitochondria, as illustrated by immunofluorescence staining of Tom20 and TEM (). Moreover, after CS treatment, the mRNA levels of the mitochondrial fission-associated factors mitochondrial fission protein 1 (Fis1) and mitochondrial protein 18 (MTP18) were decreased, whereas those of the mitochondrial fusion-related factor mitofusin 2 (Mfn2) were increased in hBMSCs. Nevertheless, no significant alterations were shown in the levels of dynamin-related protein 1 (Drp1) or Mitofusin 1 (Mfn1) (). These results demonstrate that CS can induce mitochondrial fusion in hBMSCs. ^{21 22} Figure 3A and B Figure 3C

Next, to further analyze the mitochondrial profile of hBMSCs, we used the mitochondrial membrane potential (ΔΨm)-independent probe MitoTracker Green in combination with the ΔΨm-sensitive probe MitoTracker Red []. According to flow cytometry, CS treatment reduced the percentage of dysfunctional mitochondria in hBMSCs (MitoTracker Green, MitoTracker Red) (). Loss of ΔΨm in polarized mitochondria causes a high mitochondrial reactive oxygen species (mtROS) generation [,]. Using the mitochondria-specific ROS probe MitoSOX, we detected a lower proportion of mtROS-producing hBMSCs (MitoTracker Green, MitoSOX) in the CS group than in the control group (). Moreover, mtROS levels were lower in the CS-treated hBMSCs than in the control hBMSCs, as shown by flow cytometry (). We also found that the CS-treated hBMSCs had fewer dysfunctional mitochondria and decreased mtROS levels in the presence of OM (). Our results indicate that CS can eliminate dysfunctional mitochondria and mtROS production in hBMSCs. ²³ Figure 3D ^{24 25} Figure 3E Figure 3F Supplementary Figure S3A and B +high +low +high +high

As a specialized type of autophagy, mitophagy is vital for mitochondrial quality control through the targeted removal of dysfunctional or excessive mitochondria []. To verify whether CS promotes mitophagy to eliminate dysfunctional mitochondria in hBMSCs, we stained the mitochondria and lysosomes of hBMSCs with Tom20 (green) and LAMP1 (red), respectively. Colocalization of mitochondria and lysosomes displayed a significant elevation in CS-treated hBMSCs compared to untreated hBMSCs, indicating enhanced mitophagy (). To assess whether CS induces autophagic flux (including autophagosome formation, fusion with lysosomes, and the eventual autophagic substances degradation), hBMSCs were exposed to the lysosomal inhibitor BafA1 in the presence of CS. BafA1 specifically inhibits vacuolar-type H-ATPase (V-ATPase), a proton pump that maintains the acidic pH of intracellular organelles, including lysosomes, and the extracellular environment. Its ability to block lysosomal acidification and subsequent protein degradation makes it a widely used tool for evaluating autophagic flux [,]. Intriguingly, WB analysis illustrated that CS treatment suppressed LC3-II and p62 expression (). LC3-II suppression may indicate either suppressed autophagy or increased lysosomal degradation. The blockage of autophagic flux with BafA1 resulted in similarly elevated LC3-II and p62 levels in both control and CS-treated hBMSCs (), suggesting that CS enhances lysosomal clearance rather than autophagosome formation. To further assess autophagic flux, hBMSCs were transduced with a lentiviral mCherry-GFP-LC3B construct. Compared to controls, CS-treated hBMSCs showed a higher ratio of red puncta (autolysosomes, as GFP is quenched in acidic lysosomes) and fewer total LC3 puncta, indicating increased autolysosome formation and elevated autophagic flux. In accordance with WB data, BafA1 inhibited GFP degradation in both control and CS groups, as evidenced by similar LC3 puncta counts (). We further observed the cell ultrastructure using TEM. Both autophagosomes and autolysosomes appeared in the cytoplasm of control hBMSCs, indicating the occurrence of early and late stages of mitophagy. In contrast, a markedly increased number of autolysosomes was observed in the CS-treated group, whereas few autophagosomes were present (). These data demonstrate that CS may promote the fusion of autophagosomes with lysosomes, eventually resulting in accelerated lysosomal degradation and elevated autophagic flux in hBMSCs. ²⁶ Figure 4A ^{27 28} Figure 4B Figure 4B Figure 4C Supplementary Figure S4 +

Mitophagy can reprogram stem cell differentiation by modulating the metabolic profile []. Our above findings prompted us to further explore whether CS-induced mitophagy promotes OXPHOS to enhance the osteogenesis of hBMSCs. We found that repressing the formation of autolysosomes by BafA1 impaired CS-mediated mitophagy in hBMSCs, as indicated by increased dysfunctional mitochondria, with or without OM (and). Furthermore, when hBMSCs were exposed to BafA1, CS treatment failed to decrease mtROS production (and) or increase the OCR (). CS also showed a decreased ability to promote osteogenic differentiation in the presence of BafA1, as indicated by its impaired ability to increase the ALP and ARS staining intensity or the expression of,,and(). Moreover, the immunofluorescence staining intensities of OPN and ALP were decreased in the existence of BafA1 (and). These findings indicate that CS induces the osteogenic differentiation of hBMSCs by promoting the mitophagy-mediated metabolic transition toward OXPHOS. ²⁹ Figure 5A Supplementary Figure S3A Figure 5B Supplementary Figure S3B Figure 5D Figure 5C and E Figure 5F Supplementary Figure S3C Runx2 Osx ALP Col-1

We further confirmed whether CS-induced mitochondrial OXPHOS enhances the osteogenic differentiation of hBMSCs. hBMSCs, CS-treated hBMSCs or Oligo/BafA1+CS-treated hBMSCs were seeded onto HA/β-TCP scaffolds and then transplanted subcutaneously into nude mice. After 8 weeks, micro-CT scanning of the implanted scaffolds revealed more areas with a high density of dots in the scaffolds loaded with CS-induced hBMSCs than in the control scaffolds (), suggesting more mineralized tissue deposition. However, Oligo/BafA1+CS-treated hBMSCs failed to enhance mineralization deposition and displayed a density similar to that of control hBMSCs. In addition, more collagen fibers were found in the CS group (), as indicated by HE and Masson's trichrome staining. Stronger immunofluorescence staining for the osteogenic markers ALP and OPN was also observed in the CS group compared with the control group (, and). In contrast, collagen fiber formation was markedly inhibited in the group treated with CS+Oligo or BafA1 compared with the group treated with CS alone (). Similarly, immunofluorescence staining for OPN and ALP was impaired in the Oligo/BafA1+CS-treated hBMSC groups (, and), demonstrating that CS significantly enhances the osteogenesis of hBMSCsand that the mitophagy-mediated metabolic transition toward OXPHOS plays a critical role in these processes. in vivo in vivo Figure 6A Figure 6B and C Figure 6D and E Supplementary Figure S5A Figure 6B and C Figure 6D and E Supplementary Figure S5A

A critical-sized calvarial defect mouse model was constructed to evaluate the bone repair potential of CS-treated hBMSCs. Micro-CT 3D reconstructions showed markedly enhanced bone regeneration in the CS-treated hBMSC group compared to the NC group (PLGA scaffold alone), with bone formation extending beyond the calvarial margin. In contrast, untreated hBMSCs induced only modest new bone formation. Corresponding BV/TV values displayed a significant elevation in the CS group compared to controls (26.14 ± 1.37% vs. 16.66 ± 1.83%, < 0.001) (). New bone and collagen fiber generation in the bone tissue defect area were observed via HE and Masson's trichrome staining. In the NC group, fibrous tissue made up most of the defect area, and few collagen fibers formed. Compared to the NC group, the scaffolds in the untreated BMSC group presented a small amount of collagen fibers among abundant fibrous tissues. Notably, scaffolds loaded with CS-induced hBMSCs exhibited extensive new bone and collagen fiber formation in the defect area and alongside the undegraded scaffolds (). We assessed the levels of the osteogenic hallmarks OPN and ALP via immunofluorescence staining. OPN-positive areas were barely observed in the NC group and untreated hBMSC group, whereas the CS-treated hBMSCs exhibited remarkable OPN and ALP expression, indicating active and robust osteogenesis. We also observed OPN- and ALPpositive cells, especially around the newly formed bone, indicating that osteoblasts lined the bone margins (, and). These data confirm the enhanced osteogenic or bone repair capacity of CS-treated hBMSCs P - in vivo. Figure 7A Figure 7B and C Figure 7D and E Supplementary Figure S5B

Next, we applied the OXPHOS inhibitor Oligo or mitophagy inhibitor BafA1 to CS-treated hBMSCs to further confirm that CS enhances the bone repair capacity of hBMSCs through mitophagy-mediated metabolic transitionNew bone formation in calvarial defects was impaired in the groups implanted with hBMSCs exposed to CS and Oligo or BafA1 compared to the group implanted with hBMSCs treated with CS alone, as indicated by the micro-CT images in, together with the decreased BV/TV (26.14 ± 1.37% in the CS group vs. 16.83 ± 2.59% in the CS+Oligo group,< 0.001; 26.14 ± 1.37% in the CS group vs. 17.02 ± 1.31% in the CS+BafA1 group,< 0.001). Histological findings also revealed that the addition of Oligo/BafA1 largely diminished new bone and collagen fiber formation in CS-treated hBMSCs () and led to weak OPN and ALP expression (, and). These results further indicate that the enhanced osteogenic or bone repair capacity of CS-treated hBMSCsdepends on the mitophagy-mediated metabolic transition toward OXPHOS. in vivo. P P in vivo Figure 7A Figure 7B and C Figure 7D and E Supplementary Figure S5B