Biomimetic scaffolds with synergistic BMSC targeting and ROS scavenging for mitochondrial protection and effective bone-defect repair

Abstract

The reconstruction of large bone defects remains a significant clinical challenge, primarily owing to the insufficient mitochondrial protection and osteogenic activity of conventional implants. Exosomes (EXOs) derived from mesenchymal stem cells have emerged as promising tools for bone repair. This study reports a mitochondria-targeted therapeutic strategy utilizing EXOs derived from bone marrow mesenchymal stem cells (BMSCs). On MitoQ incorporation, these EXOs (EXO-MitoQ, EM) exhibit the targeted scavenging of mitochondrial reactive oxygen species; moreover, on surface decoration with the nucleic acid aptamer Apt 19 S (EM-Apt), they show the enhanced recruitment and precise delivery of BMSCs. The engineered EXOs show robust BMSC-targeting specificity and mitochondrial protective efficacy. To optimize their regenerative microenvironment and biomechanical properties further, these functionalized EXOs are integrated onto a 3D-printed β-tricalcium phosphate scaffold coated with a small intestinal submucosa (SIS) hydrogel, forming a composite system (TCP/SIS@EM-Apt). In a rat calvarial defect model, this TCP/SIS@EM-Apt scaffold increased the BV/TV by 1.9-fold compared to TCP/SIS, due to the combination of multiple multifunctional therapeutic effects (anti-inflammatory, angiogenic, and osteogenic). The mitochondria-targeting strategy proposed in this study presents a promising solution for the reconstruction of large bone defects and offers a synergistic approach for addressing complex regenerative challenges.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-025-04004-7.

Introduction

Large segmental bone defects pose a significant clinical challenge owing to their complexity and the limitations of current treatment options [1]. Although traditional methods, such as autografts and allografts, are commonly used to repair bone defects, they are associated with certain limitations, such as donor site morbidity and the risk of immune rejection. Advances in tissue engineering have led to the development of biomaterial scaffolds and cell-based therapies that offer new avenues for bone-defect repair [2]. However, an ideal bone-repair strategy should combine adequate mechanical support with the ability to promote vascularization and osteogenesis for optimal bone regeneration. Metabolic disorders of endothelial cells, immunocytes, and osteoblasts in the bone microenvironment compartment obstruct natural regeneration and repair after fractures [3, 4]. An increasing amount of clinical and experimental data indicates that the excessive accumulation of reactive oxygen species (ROS) in defective areas exacerbates mitochondrial dysfunction through feedback mechanisms, thereby affecting the osteogenic differentiation of osteoprogenitors and bone healing [5–8]. Therefore, interventions targeting mitochondrial function and the regulation of ROS have become a crucial strategy for the treatment of bone diseases [9–11].

Despite being considered a fundamental element of tissue engineering along with biomaterials and growth factors, stem cell therapy is associated with ethical limitations and quality control challenges [12]. Cell-free bioactive materials are promising candidates for diabetic bone-defect treatment. An effective treatment strategy comprises modifying scaffolds with bioactive components, such as growth factors, active drugs, and cellular components (such as extracellular matrix, exosomes (EXOs), and mitochondria) [13, 14]. Among these bioactive components, EXOs are particularly promising drugs for clinical treatment due to their excellent biocompatibility, stability, and universal modifiability [5, 15–17]. They efficiently transport regenerative molecules (proteins, lipids, and nucleic acids) with repair functions to damaged tissues [18–20]. Because of their natural lipid membranes with surface adhesion molecules, EXOs show certain unique characteristics, such as specific targeting, circulatory stability, and the ability to penetrate natural barriers. Besides exhibiting tissue repair capabilities, EXOs derived from mesenchymal stem cells can regulate immunity and metabolism, thereby exhibiting high potential in clinical therapeutics [19, 21]. However, due to the harsh inflammatory microenvironment at the injury site and the action of a large number of phagocytic cells, EXOs are rapidly decomposed and cleared from the injury site, limiting their accumulation and sustained local action. Moreover, the targeting ability of natural EXOs is insufficient, and their effects on mitochondrial dysfunction are limited [22, 23]. Bone marrow mesenchymal stem cells (BMSCs) exhibit immune and matrix regulatory functions with a capacity for efficient osteogenic differentiation, making them the primary endogenous stem cells in adult bone-defect repair [17]. Considering the dependence of the osteogenic repair process on high-energy metabolism, using BMSC-derived EXOs (BMSC-EXOs) to synergistically target mitochondrial protection is hypothesized to be a feasible strategy for promoting the repair of large bone defects.

MitoQ is a mitochondria-targeting antioxidant consisting of an antioxidant ubiquinone fragment covalently linked to a lipophilic triphenylphosphonium ion through a fatty carbon chain [24]. Ubiquinone, a natural electron carrier between mitochondrial membrane complexes, can reduce ROS production significantly. During mitochondrial respiration, electron transfer between the electron transport chains on the mitochondrial membrane increases the mitochondrial membrane potential (MMP), allowing the antioxidant modified with lipophilic cations to easily penetrate the phospholipid membrane and accumulate repetitively within the mitochondria [24]. Moreover, MitoQ can target and aggregate on the mitochondrial membrane, maintaining stable mitochondrial dynamics and protecting mitochondrial function by reducing ROS production and promoting mitochondrial autophagy and biogenesis [25, 26]. Animal experiments and clinical studies indicate that MitoQ can regulate the AMPK pathway to mitigate neurogenic damage in ischemic stroke [27], upregulate PGC-1α expression to enhance exercise-induced muscle peak power [26], and improve cigarette-mediated lung epithelial aging and fibrosis [28].

In this study, engineered and modified BMSC-EXOs were used for the targeted delivery of MitoQ, which synergistically enhanced the efficiency of osteogenesis by clearing ROS and promoting mitochondrial replication at the injury site. The membrane surface of the EXOs was modified with BMSC-specific nucleic acid aptamers (Apt 19 S) synthesized using a previously published protocol [29] to form a biomimetic nanomotor that could target and bind to the BMSC cell membrane to increase the EXO utilization efficiency. In addition, to increase the tissue retention time, a natural decellularized small intestinal submucosa (SIS) hydrogel [30, 31] was used to encapsulate the EXOs and coat 3D-printed β-tricalcium phosphate (β-TCP) scaffolds. The biomimetic dual-target scaffolds synthesized in this study meet the biological activity, microstructure, and mechanical property requirements for diabetic bone-defect repair.

The effects of the newly developed mitochondria-targeting material on the mitochondrial membrane protection in BMSCs was assessed in this study. Compared with the control group, the mitochondria-targeting material alleviated mitochondrial ROS damage and protected mitochondrial health. RNA sequencing (RNA-seq), used to investigate the molecular mechanisms underlying the bone regeneration process, indicated an increase in the activation of the Wnt pathway during bone repair. Furthermore, the regulatory effects of the biomimetic scaffold on the inflammatory response, microvascular system, and bone regeneration-repair in a rat model of calvarial defects was investigated (schematics are shown in Fig. 1).

Fig. 1.

Synthesis and therapeutic effects of TCP/SIS@EM-Apt. (A) BMSC-EXOs are synthesized by gradient ultracentrifugation and incorporated with mitochondria-targeted MitoQ (EM). Subsequently, the surface is modified with Apt 19 S to enable the nanovesicles to target BMSCs (EM-Apt). (B) Synthesis of the biomimetic mitochondrial-targeted engineering scaffold TCP/SIS@EM-Apt. (C) The therapeutic effects of the TCP/SIS@EM-Apt scaffold include mitochondrial protection, anti-inflammation activity, angiogenesis promotion, and osteogenesis

Materials and methods

BMSC culture

All experiments were approved by the Institutional Animal Care and Use Committee of Huazhong University of Science and Technology (HUST). BMSCs were isolated from Sprague-Dawley rats (4 weeks old) and maintained in DMEM/F-12 cell medium (Gibco, USA) supplemented with fetal bovine serum (10%) (Vivacell, Germany). An osteogenic induction medium (Cyagen, USA) was used to evaluate the osteogenic differentiation of the BMSCs. Cells from passages 2 to 4 were used.

Isolation and identification of exos

BMSC-EXOs were extracted using differential centrifugation. Briefly, the supernatant collected after centrifuging the culture supernatant of BMSCs at 500 g for 10 min at 4 ℃ was centrifuged at 2,000 g for 30 min at 4 ℃. The supernatant was collected and centrifuged at 10,000 g for 45 min at 4 ℃, and the supernatant was collected. The extract was filtered with a 0.22-µm sterile membrane filter (Millipore, Germany) and centrifuged at 100,000 g at 4 ℃ for 70 min; the supernatant was removed and the precipitate was resuspended in PBS and stored at − 80 ℃ for subsequent experiments. The morphology of the BMSC-EXOs was observed using TEM, exosome-specific labels (CD9, CD63, TSG101, and Calnexin) were detected using Western blotting, and the size distribution was detected by NTA.

Synthesis of dual-targeted drug-carrying vesicles

The extracted BMSC-EXOs were mixed with carboxypolyethylene glycol (PEG-COOH) (4%) at 25 °C for 24 h to obtain EXO@PEG-COOH. Hunan Precision Biotechnology Co. Ltd. synthesized the aptamer 19 S with BMSC-targeting specificity. The carboxyl groups on EXOs@PEG-COOH were activated for aptamer conjugation. Briefly, 100 µg of EXOs@PEG-COOH were mixed with 500 µL of 0.1 M MES buffer (pH ~ 6.0), followed by the addition of 5 mg 1-Ethyl − (3-dimethylaminopropyl) carbodiimide and 60 µg N-Hydroxysuccinimide. The reaction was gently mixed for 15 min and the excess solution was removed via ultracentrifugation. Then the activated EXOs@PEG-COOH was resuspended in 400 µL of fresh 0.1 M MES buffer. Then, 50 nmol of amine-modified Apt19S was introduced, and the conjugation was carried out at 25 °C for 2 h in a reciprocating oscillator.25 °C. MitoQ was introduced into the EXOs by electroporation followed by incubation at 37 ℃ for 15 min to restore the electroporated EXO membrane, resulting in an engineered nanovesicle EXO-MitoQ-Apt (EM-Apt) that can dually target the mitochondria of BMSCs. TEM, NTA particle-size analysis, and zeta potential detection were used for characterization.

The engineered nanovesicle, EXO-MitoQ-Apt (EM-Apt), was constructed by loading MitoQ into exosomes to achieve dual mitochondrial targeting in BMSCs. Briefly, MitoQ was loaded into exosomes via electroporation at a 1:1 mass ratio to form EXO-MitoQ. This was followed by a 15-minute incubation at 37 °C to facilitate membrane recovery. To determine the drug concentration, EXO-mitoQ was subjected to 30 min of sonication followed by centrifugation at 4 ℃ and 15,000 g for 20 min. The drug content in the EXOs was quantified using HPLC. The encapsulation efficiency was calculated according to the following equation (Supplementary Table S1):

where represents the total mass of the drug initially added to the system, and denotes the mass of the drug encapsulated within the exosomes.

Additionally, the drug loading efficiency was determined using the formula provided in Supplementary Table S2:

where refers to the mass of the loaded drug, and represents the total mass of the drug-loaded exosomes.”

Uptake of dual-targeted drug-carrying vesicles

Pre-labeling with the red fluorescent dye DiR (Sigma-Aldrich, USA) was used to examine MitoQ loading into EXOs. The BMSC-EXOs were co-cultured with DiR-labeled exosomes (100 µg/mL) and fixed with paraformaldehyde (4%) for 30 min at 25 °C. The MitoQ was first aminated with sodium triacetoxyborohydride, followed by FITC labeling, and their uptake by BMSCs was observed after 2 h of incubation (10 µg/mL). The nuclei were stained with Hoechst 33,342 for 20 min at 25 °C. Finally, the BMSCs were observed under a confocal microscope (Nikon A1, Japan).

Mitochondrial ROS level detection

To induce oxidative stress, BMSCs were treated with 200 µM H₂O₂ for 4 h in serum-free medium, followed by replacement with fresh complete medium for subsequent assays. Intracellular ROS were detected by flow cytometry and fluorescence microscopy using the MitoSOX kit (Invitrogen, USA). BMSCs were pretreated with H₂O₂ to simulate an oxidative stress microenvironment; subsequently, MitoQ and EM-Apt were added, and changes in the ROS levels were observed under a fluorescence microscope. After the same treatment, flow cytometry was used to quantitatively analyze the levels of intracellular ROS using the fluorescence intensity of MitoSOX.

MMP level detection

An MMP detection kit (JC-1) was used to observe the depolarization of the mitochondrial membrane. After pretreatment with H₂O₂, the BMSCs were incubated with MitoQ or EM-Apt. Subsequently, they were stained with the JC-1 working solution, and the fluorescence changes in JC-1 cells were observed using a fluorescence microscope. Flow cytometry was used to quantitatively analyze the MMP levels using the red and green fluorescence intensities of JC-1.

Mitochondrial function analysis

BMSCs were pretreated with H₂O₂ and incubated with MitoQ or EM-Apt. A MitoTracker deep red fluorescent probe was used to stain mitochondria to observe their fluorescence intensity and analyze their size and number. In addition, mitochondrial ATP levels and DNA quality were tested to comprehensively assess mitochondrial function after treatment.

Scaffolds characterization and extract preparation

The scaffolds were freeze-dried and coated with Au sputtering. The structural morphology of the scaffolds was observed using a scanning electron microscope (HITACHI S4800, Japan) and element mapping images. The scaffolds were manufactured according to previously reported methods [32]. The three types of sterilized scaffolds were soaked in the DMEM/F-12 medium (Gibco, USA) at a mass/volume ratio of 100 mg/mL and incubated at 37 ℃ for 24 h. The supernatants were collected, centrifuged, filtered with a 0.22-µm membrane filter (Millipore, Germany) to obtain the scaffold extracts.

Cell viability

The BMSCs were seeded into 24-well plates, and the scaffold extracts of each group were mixed with DMEM/F-12 at a ratio of 1:2 and used as a complete medium for cell culture. Cell proliferation and viability were examined at the indicated time points using a CCK-8 assay and Calcein-AM staining. A fluorescence microscope (Olympus IX73; Tokyo, Japan) was used to observed the stained images.

Cell morphology

BMSCs were seeded in 24-well plates, and a complete culture medium from each group was used for cell culture. After 5 days of culture, the BMSCs were stained with Hoechst and phalloidin (Beyotime, China) to visualize the cytoskeleton. The stained sections were observed under a fluorescence microscope.

Cell migration

A 24-well transwell system (Corning, NY, USA) was used to evaluate the effect of the scaffolds on BMSC migration. The cells were inoculated into the upper chamber of a transwell system, and the complete culture medium of each group was placed in the lower chamber. After 24 h of incubation, the cells were stained with crystal violet and imaged by an optical microscope.

ALP and ARS staining

BMSCs were seeded into 24-well plates, and the scaffold extracts of each group were mixed with osteogenic induction medium at a ratio of 1:2 to evaluate osteogenic differentiation. Cells were cultured in osteogenic induction medium for 7 days, stained using the BCIP/NBT ALP Kit (Beyotime, China), and imaged using an Olympus IX73 microscope. For ARS staining, the BMSCs were stained with ARS solution (0.1%) (Solarbio, China) after 14 days of culture, and the cell images were observed under a microscope.

Immunofluorescent staining

The seeded BMSCs were cultured in a mixed osteogenic induction medium for 14 days, fixed with paraformaldehyde, permeabilized with Triton X-100 PBS solution, and blocked with BSA (0.5%). Subsequently, the BMSCs were incubated overnight at 4 ℃ with primary antibody solutions of OCN and Runx2 (Beyotime, China), and the cytoskeleton and nucleus were stained with phalloidin and DAPI, respectively. Samples were photographed using a confocal laser scanning microscope (FV3000, Olympus).

Western blot analysis

The BMSCs were cultured in a mixed osteogenic induction medium for 14 days, and the total proteins of the cells/scaffolds were extracted. The supernatant was collected after centrifugation at 14,000 × g, mixed with loading buffer, and stored at − 20 ℃. After electrophoresis at a constant voltage, the PVDF membrane was transferred. The primary antibody was incubated overnight at 4 ℃, and the secondary antibody was incubated for 1 h at 25 °C and detected using a highly sensitive ECL luminometer (Gene Gnome XRQ, UK).

Transcriptome sequencing

After pretreatment with H₂O₂, the BMSCs were co-cultured with extracts from the TCP/SIS@EM-Apt scaffolds. The total RNA was extracted using TRIzol (Invitrogen, USA). In the RNA sequencing experimental design, three biological replicates were established for each group. Samples were sent to Novogene Co. Ltd. for RNA extraction, library construction, and sequencing. RNA sample qualification was followed by reverse transcription and PCR amplification. The PCR products were purified using AMPure XP beads to generate the final sequencing library. A DNBSEQ-T7 high-throughput sequencer was used for mRNA sequencing. Raw reads were processed using the fastp software to obtain high-quality clean reads. The featureCounts software (v2.0.6) was used to calculate the FPKM for each gene. During the sequencing analysis phase, differentially expressed genes between the two groups were analyzed using the DESeq2 R package (v1.42.0) with significant differential expression thresholds set at a p-value ≤ 0.05 and |log2(fold change)| ≥ 1. The ClusterProfiler software was used to analyze the statistical enrichment of differentially expressed genes in the KEGG pathway, and the cnetplot function was used to visualize the KEGG pathway gene network.

RT-qPCR

The total mRNA was extracted using TRIzol (Takara, Japan) according to the manufacturer’s instructions. To obtain cDNA, the RevertAid kit (Takara, Shiga, Japan) was used for reverse transcription. An RT-PCR kit (Beyotime) was used for PCRs following the manufacturer’s protocol. The endogenous gene β-actin was used as a reference gene for normalization. The primers used in this study are listed in Table 1.

Table 1.

The primers used for RT-qPCR in this study

Gene Symbol	Forward primer	Reverse primer
IL-1β	CACCTCTCAAGCAGAGCACAG	GGGTTCCATGGTGAAGTCAAC
IL-6	TCTTGGGACTGATGTTGTTGC	AAGGCAACTGGATCGGATGC
IL-10	GCTCTTACTGACTGGCATGAG	CGCAGCTCTAGGAGCATGTG
Arg-1	CTGGCTTGCCGAGGCTGTTC	GGATCACCTCGCAGCTGGTC
Tgfbr1	CAGCCATCAGCATCT TCA CC	GCTGCAGTAGGCATCAAGGA
Smad3	GGAGGCTGAAGTCCAGAACG	CACTGCCCATCTTCTTCCTC
Wnt8b	TGCTGGAGGACTTTGTGATG	CAGGGTGTTGGTGTTGAGGT
Pdgfr1	TGCTCACCTTCACCACCTTT	CCAGGATGATGCCAAAGAGG
Osm	ATGGCTGCTGCTGTCTCTAC	TGAGGATGGCAGTGAGGAAG

Animal experiments

Calvarial defects of critical size were created in 30 Sprague-Dawley rats (male, 180–200 g) purchased from Beijing SPF Biotechnology Co. Ltd. (China). The animals were randomly divided into three groups: TCP/SIS, TCP/SIS@EM, and TCP/SIS@EM-Apt. The scaffolds used for in vivo implantation were fabricated as cylinders with a diameter of 5 mm and a thickness of 1 mm. Each TCP/SIS@EM and TCP/SIS@EM-Apt scaffold was loaded with a total of 100 µg of BMSC-EXOs protein [33]. The SD rats were anesthetized with pentobarbital, the skull was exposed by an incision on the scalp, bone defects with a diameter of 5 mm were created using trephine, and scaffolds were implanted in each group. The animals were sacrificed at 4 and 8 weeks after implantation, and their skulls were collected for analysis.

Micro-CT analysis and histological assessment

Micro-CT (SkyScan 1176, Bruker) was used to assess scaffold-induced bone regeneration in the collected skulls. Following 3D reconstruction after scanning, the bone volume percentage (bone volume/tissue volume, BV/TV) and BMD were analyzed using the CT analyzer software (Bruker, Billerica, MA, USA). The samples were then decalcified and embedded in paraffin. After sectioning, H&E and Masson staining as well as immunofluorescence were used analysis.

Statistical analysis

All data are expressed as mean ± standard deviation. GraphPad Prism (version 10.1.2) was used for statistical analysis. The statistical significance was determined using one-way ANOVA and p < 0.05 was considered statistically significant. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Results and discussion

Synthesis and characterization of EM-Apt

Here, BMSCs were used to mass-produce BMSC-EXOs. Briefly, BMSCs were ultracentrifuged, and EXOs were extracted from the supernatant. Transmission electron microscope (TEM) images indicate that the extracted EXOs exhibit a typical exosomal structure comprising spherical and membranous vesicles (Fig. 2A). Nanoparticle tracking analysis (NTA) indicates that the size distribution of these EXOs ranges from 50 to 300 nm, consistent with the literature (Fig. 2B). Western blotting shows that the extracted EXOs express exosomal surface markers, such as CD9, CD63, TSG101, and Calnexin (Fig. 2C). Collectively, these results confirm that the nanoparticles extracted from the supernatant of BMSC ultracentrifugation are EXOs. Furthermore, zeta potential analysis was used to confirm the loading of MitoQ into BMSC-EXOs and Apt 19 S modification on the vesicle surface. On incorporating MitoQ and Apt 19 S, the potential levels of EM and EM-Apt decrease gradually (Fig. 2D). Next, the ability of BMSCs to uptake DiR-labeled EXOs in vitro was assessed by observing red fluorescence within the cells. Fluorescence colocalization images obtained by laser confocal microscopy confirm MitoQ loading in EM-Apt (Fig. 2E). The encapsulation efficiency of MitoQ in exosomes was determined to be 52.24 ± 3.40% (Tab.S1). The encapsulation efficiency of MitoQ in EXOs was determined to be 34.29 ± 1.47% (Tab S2). The fluorescence colocalization of MitoTracker Red Deep and MitoQ demonstrates that EM-Apt has excellent mitochondria-targeting ability in BMSCs, comparable to other mitochondrial-targeting drugs [6]. (Fig. 2F).

Fig. 2.

Components and characteristics of the double-targeted nanovesicles EM-Apt. (A) Characteristic transmission electron microscopy (TEM) scan images of BMSC-EXOs, EM, and EM-Apt (bar = 100 nm). (B) Particle diameters of BMSC-EXOs, EM, and EM-Apt. (C) Western blot confirms the typical protein components in BMSC-EXOs. (D) Potential levels of BMSC-EXOs, EM, and EM-Apt (n = 5). (E) Laser confocal fluorescence colocalization confirms MitoQ loading into EM-Apt (bar = 40 and 10 μm). (F) Laser confocal fluorescence colocalization confirms that EM-Apt targets BMSC mitochondria

EM-Apt scavenge ROS and protect mitochondrial function in BMSCs

The in vitro treatment of BMSCs with hydrogen peroxide (H₂O₂) leads to an intracellular ROS burst, which is observed as intense MitoSOX green fluorescence (Fig. 3A and B). Nevertheless, the fluorescence intensity in the MitoQ and EM-Apt treatment groups decreases significantly, indicating that EM-Apt can scavenge excessive ROS in mitochondria, safeguarding mitochondrial function. Flow cytometry analysis indicates that the EM-Apt treatment group effectively loads MitoQ, transporting it to the mitochondria, exerting an effect comparable to that of MitoQ treatment, thereby confirming the excellent performance of EM-Apt (Fig. 3C). Mitochondrial damage is a crucial marker of ROS-induced apoptosis. JC-1, a cationic mitochondria-selective dye, emits red and green fluorescence at high and low MMP levels, respectively. Thus, a reduction in red fluorescence and increase in green fluorescence indicates mitochondrial dysfunction. Mitochondrial damage following H₂O₂ treatment is characterized by high-intensity green fluorescence (Fig. 3E and G). The delivery of MitoQ in the EM and EM-Apt groups ameliorates mitochondrial function, as evidenced by high-intensity red fluorescence on approaching the normal state in the control group (BMSCs group). The results of flow cytometric analysis shown in Fig. 3F corroborate this finding.

Fig. 3.

Effects of the dual-targeted nanovesicles EM-Apt on mitochondrial ROS, the membrane potential, ATP, and the mitochondrial mass of BMSCs in vitro. The mitochondrial ROS levels are evaluated by MitoSOX fluorescence staining (A) and quantitative analysis (bar = 500 μm) (B). Green fluorescence indicates MitoSOX. The fluorescent probe DCFH-DA is used to label the ROS levels in different groups, and flow cytometry (C) is used for evaluation and quantitative analysis (D). (E) The MMP of the BMSCs is detected using JC-1 fluorescence staining images; mitochondrial aggregates and monomers are shown in red and green, respectively (bar = 500 μm). The qualitative (F) and quantitative (G) levels of JC-1 fluorescence are analyzed by flow cytometry. (H) Results of MitoTracker red fluorescence staining for the laser confocal imaging of mitochondria (bar = 10 μm). Quantitative analysis of mitochondrial length (I) and fluorescence intensity (J). Quantitative analysis of the ATP (K) and mitochondrial DNA (L) in the BMSCs. n = 5. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Typically, longer mitochondria show better energy-production capabilities and higher antioxidant capacities [34]. Conversely, shorter mitochondria are more likely to result in reduced energy production and increased oxidative stress [34]. Hence, in this study, the mitochondrial morphology was observed via MitoTracker staining. Similar to the MitoQ group, the EM-Apt group exhibits intense mitochondrial red fluorescence staining and long mitochondria, suggesting an increase in the number and length of mitochondria after damage recovery (Fig. 3H and J). Additionally, the ATP level and mtDNA concentration in the BMSCs after treatment were measured. Consistent with the results of previously discussed analyses, the EM-Apt group shows a good therapeutic effect. These results confirm the ability of the EM-Apt dual-targeting system to enhance mitochondrial function.

TCP/SIS@EM-Apt scaffolds promote BMSC recruitment and anti-inflammatory effects in vitro

We fabricated four types of scaffolds with exosome loading doses of 25 µg, 50 µg, 100 µg, and 200 µg, respectively. First, the extract method was used to conduct biocompatibility tests, and the results showed that all scaffolds with different loading doses exhibited good biocompatibility (Fig. S3). Subsequently, an antioxidant stress assay was conducted using a fluorescence microplate reader. The results showed that the antioxidant capacity of each experimental group increased with the rise in exosome loading concentration, yet no significant difference was observed between the 100 µg group and the 200 µg group (Fig. S2). Based on the above findings, scaffolds loaded with 100 µg of exosomes were used in all subsequent experiments.

Scanning electron microscopy (SEM) images were used to analyze the morphologies of the TCP/SIS, TCP/SIS@EXOs, and TCP/SIS@EM-Apt scaffolds (Fig. 4A). These scaffolds exhibit a highly porous structure with an average pore size of ~ 400 μm, which facilitates nutrient and oxygen transport. Moreover, the hydrogel coating on the surface of the TCP/SIS@EXOs and TCP/SIS@EM-Apt scaffolds effectively promotes cell growth, adhesion, and migration. Notably, elemental mapping via SEM confirms uniformity in scaffold synthesis (Fig. 4B). The accumulation curve in Fig. 4C indicates that MitoQ is slowly released from the scaffolds, and the MitoQ-loaded vesicles in the SIS@EM-Apt group exhibit a better controlled release rate and longer release time than those in the other groups.

Fig. 4.

Characterization of scaffolds and in vitro cytotoxicity, cell recruitment, and anti-inflammatory effects. (A) SEM images of the TCP/SIS, TCP/SIS@EXOs, and TCP/SIS@EM-Apt scaffolds. (B) Elemental mapping images of the TCP/SIS@EM-Apt scaffold. (C) MitoQ is released from the TCP/SIS@EM-Apt scaffold (n = 3). (D) Calcein-AM staining of BMSCs cultured on TCP/SIS, TCP/SIS@EXOs, and TCP/SIS@EM-Apt scaffolds for 1 and 5 days. (E) Fluorescence images of BMSCs cultured on scaffolds. (F) Viability of BMSCs cultured on TCP/SIS, TCP/SIS@EXOs, and TCP/SIS@EM-Apt scaffolds for 1, 3, and 5 days (n = 3). (G) Representative images of cell migration in the scaffold groups. (H) qRT-PCR analysis of M1- and M2-related gene expression in macrophages (n = 5). **p < 0.01, ***p < 0.001, ****p < 0.0001

Calcein-AM staining images of BMSCs cultured on the scaffolds indicate that a large number of cells survive on all three types of scaffolds after 1 and 5 days of culture (Fig. 4D). The morphology of the BMSCs on the scaffolds was examined by labeling the actin cytoskeleton with phalloidin. Typical fluorescent images show that the BMSCs are well-attached to the scaffolds in a spreading morphology (Fig. 4E). Cell proliferation was quantitatively evaluated using the CCK-8 assay. The proliferation rate of the BMSCs remains consistent across all scaffolds (Fig. 4F). These results confirm the excellent cellular compatibility of the TCP/SIS, TCP/SIS@EXOs, and TCP/SIS@EM-Apt scaffolds. Stem-cell recruitment is a crucial factor in tissue regeneration, and promoting BMSC recruitment is critical in bone-defect repair [29]. In this study, the results of the transwell assay show that the TCP/SIS@EM-Apt group contains a significantly increased number of migrating cells compared with the TCP/SIS and TCP/SIS@EXO groups (Fig. 4G). Among the three groups, the TCP/SIS@EM-Apt group targets more MitoQ in the BMSCs, thereby exhibiting a better therapeutic effect. Finally, TCP/SIS@EM-Apt and TCP/SIS@EXOs promote M2-related gene expression and inhibit M1-related gene expression in macrophages after TNF-α stimulation (Fig. 4H). Interestingly, the engineered exosomes show more significant changes, indicating enhanced inflammatory regulatory effects owing to engineering upgrades.

TCP/SIS@EM-Apt promotes the osteogenic differentiation of BMSCs

To clarify the potential underlying mechanisms of bone repair, the expression levels of osteogenic genes and proteins, including Runx2, a specific marker of osteogenic differentiation, and osteocalcin (OCN), a marker of late-stage osteogenic differentiation, were investigated using immunofluorescence staining and Western blot analysis (Fig. 5). Immunofluorescence images (Fig. 5A and C) and the quantitative analysis of immunofluorescence staining (Fig. 5B and D) show that the BMSCs on the TCP/SIS@EM-Apt scaffold express higher levels of Runx2 and OCN proteins than those on the TCP/SIS and TCP/SIS @EXO scaffolds. The osteogenic differentiation of BMSCs cultured on the scaffolds was evaluated based on the expression of specific markers, such as alkaline phosphatase (ALP), an early-stage osteogenic marker. ALP staining and quantitative analyses of BMSCs were used to assess the osteogenic differentiation status on day 7 (Fig. 5E and G). Compared with the other two groups (TCP/SIS and TCP/SIS@EXO scaffolds), the TCP/SIS@EM-Apt scaffolds exhibit significantly enhanced ALP staining. In addition, Alizarin Red S (ARS) staining was used to detect extracellular-matrix mineralization, an indicator of the efficiency of late-stage osteogenesis. ARS staining microscopy images (Fig. 5F) and quantitative analysis (Fig. 5H) on day 14 show that the BMSCs produce a mineralized matrix on all three scaffolds. Among the three systems, the formation of mineralized nodules is most prominent on the TCP/SIS@EM-Apt scaffold, indicating excellent osteogenic potential. Finally, Western blotting confirms that the TCP/SIS@EM-Apt scaffold plays the most significant role in enhancing the expression of related proteins, namely Runx2, OCN, and ALP, in the BMSCs (Fig. 5I and J). These results confirm the outstanding osteogenic bioactivity of the TCP/SIS@EM-Apt scaffold.

Fig. 5.

The osteogenic differentiation of BMSCs cultured on the scaffolds. Immunofluorescence staining images of the osteogenic marker proteins (A) OCN and (C) Runx2 in BMSCs cultured on TCP/SIS, TCP/SIS@EXOs and TCP/SIS@EM-Apt scaffolds at day 14. The nuclei are stained in blue; OCN and Runx2 proteins are stained in red; and cellular microfilament proteins are stained in green. Quantitative analysis of the (B) OCN and (D) Runx2 immunofluorescence shown in (A, C). (E) ALP staining at day 7 and (F) Alizarin red S staining at day 14. (G) (J) Quantitative analysis of (E) and (F), respectively. (I) Western blot of the osteogenic protein expression (OCN, Runx2, and ALP) at day 14. (J) Quantitative analysis of the Western blot shown in (I). n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Exploration of the osteogenic mechanism of TCP/SIS@EM-Apt scaffolds in vitro

To elucidate the potential mechanism of TCP/SIS@EM-Apt intervention in BMSCs, H₂O₂-pretreated BMSCs were collected for bioinformatics analysis after TCP/SIS@EM-Apt treatment (Fig. 6). In the H₂O₂-treated group, 1,407 differentially expressed genes (DEGs) are identified, of which 729 are upregulated and 678 are downregulated (Fig. 6A). Notably, among the significantly differentially expressed genes, those associated with osteogenic differentiation, including Tgfbr1, Smad3, Osm, Jak2, Wnt8b, Pdgfra, Camk2g, Mapk10, and Ndufaf6, are significantly upregulated. Conversely, genes related to mRNA decay, glycolysis, and inflammation, including Ddx6, Eno3, and Il12rb2, are significantly downregulated (Fig. 6B).

Fig. 6.

TCP/SIS@EM-apt treatment enhances the expression of osteogenic differentiation-related pathways in BMSCs. (A) volcano plot of the DEGs in BMSCs after TCP/SIS@EM-Apt treatment with osteogenic- and inflammatory-related genes marked in red and blue, respectively. (B) Heat map showing representative upregulated and downregulated DEGs following TCP/SIS@EM-Apt treatment (H represents H₂O₂, T represents TCP/SIS@EM-apt). (C) Representative KEGG pathways are significantly upregulated and downregulated on TCP/SIS@EM-apt treatment. (D) Cnetplot of the representative upregulated and downregulated KEGG pathways. (E) qPCR quantification of key osteogenic-related genes. n = 3. *p < 0.05, ** p < 0.01, *** p < 0.001

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis indicates that the DEGs in the TCP/SIS@EM-Apt-treated group exhibit increased activity in pathways related to cell proliferation (amino and nucleotide sugar metabolism), skeletogenesis (axon guidance, gap junctions), cellular metabolism (phospholipase D and apelin signaling pathways), and osteogenic differentiation (Wnt and ErbB signaling pathways) compared with the H₂O₂ group. In contrast, these DEGs show decreased activity in pathways associated with glycolysis (glucagon signaling pathway), RNA decay (RNA degradation), inflammatory responses (inflammatory bowel disease), and adipogenic differentiation (adipocytokine signaling pathway) (Fig. 6C). Furthermore, KEGG pathway interaction network analysis indicates that the osteogenic differentiation pathways are interconnected with the skeletogenesis and cellular metabolism pathways (Fig. 6D). These findings suggest that TCP/SIS@EM-Apt treatment promotes the osteogenic differentiation and suppresses the adipogenic differentiation of BMSCs under oxidative stress. Moreover, the high expression levels of representative osteogenesis-related genes, including Tgfbr1, Smad3, Wnt8b, Pdgfr1, and Osm (Fig. 6E), indicate that TCP/SIS@EM-Apt treatment promotes endogenous stem-cell osteogenic differentiation through multiple pathways, including the TGF-β/Smad3 and Wnt signaling pathways.

Beyond these canonical signaling cascades, mitochondrial protection may also directly modulate osteogenic differentiation. By preserving mitochondrial integrity, TCP/SIS@EM-Apt maintains intracellular redox balance and ATP generation, which are essential for osteogenic lineage commitment [35]. Proper mitochondrial function supports the activation of MAPK and JAK/STAT pathways, both of which enhance osteogenic gene expression (e.g., Osm, Jak2, Mapk10) and promote matrix mineralization [36, 37]. Furthermore, improved mitochondrial metabolism reduces glycolytic flux and inflammatory signaling (e.g., Eno3, Il12rb2 downregulation), thereby shifting cell fate from adipogenesis toward osteogenesis [38]. These findings suggest that mitochondrial protection exerts a multifaceted regulatory effect on osteogenic differentiation through energy metabolism and signal transduction networks in addition to the Wnt and TGF-β/Smad pathways.

TCP/SIS@EM-Apt scaffolds promote bone-defect repair in rats

The in vivo osteogenic capacity of the synthesized scaffolds was investigated using a 5-mm skull-defect rat model. Micro-CT was used to assess bone repair at 4 and 8 weeks. In the scaffold groups, new bone growth is observed to initiate from the defect periphery and progress toward the defect center. Among the tested systems, the TCP/SIS@EM-Apt scaffold shows the best micro-CT performance (Fig. 7A). The results of microarchitecture parameter analysis are shown in Fig. 7B. The growth of new bone tissue in all groups is significantly higher at 8 weeks than at 4 weeks. The bone tissue volume/total tissue volume (BV/TV) and bone mineral density (BMD) of the new bone in the TCP/SIS@EM-Apt scaffold group are significantly higher than those of the new bone in the TCP/SIS @EXO and TCP/SIS scaffold groups. These findings suggest that the TCP/SIS@EM-Apt scaffold induces bone regeneration more rapidly and effectively than the other systems.

Fig. 7.

Evaluation of in vivo bone regeneration. (A) Representative micro-CT and pseudo-color images of calvarial defects implanted with the TCP/SIS, TCP/SIS@EXOs and TCP/SIS@EM-Apt scaffolds at 4 and 8 weeks after implantation. (B) Quantification of the newly formed bone with BV/TV, and the BMD in the repaired cranial defect area. (C) Representative IVIS images of Cy3-EM-Apt-loaded and Cy3-labeled hydrogels in the mouse cranial defect model. (D) Quantitative analysis of IVIS imaging. (E) Representative H&E and Masson staining images of each group at week 4 and 8 post-implantation. (F) Quantitative analysis of the new bone area. n = 5. **p < 0.01, ***p < 0.001, ****p < 0.0001

Subsequently, the controlled-release behavior of the hydrogel on EM-Apt was investigated in vivo. EM-Apt and hydrogel were labeled with Cy3 to study their localization. Cy3-EM-Apt-loaded and Cy3-SIS hydrogels were implanted into mouse cranial defects, and Cy3-labeled EM-Apt was directly injected into untreated cranial defects as a control. After hydrogel implantation, an in vivo imaging system (IVIS) was used for regular observation. Free EM-Apt is rapidly cleared from the defect site, whereas hydrogel-wrapped EM-Apt and the hydrogel alone persist at the defect site for more than 3 weeks (Fig. 7C and D). These results suggest that the SIS@EM-Apt hydrogel functions as an effective carrier that extends the existence time of EM-Apt and maintains a high local concentration of the moiety in the surrounding tissues.

Histological analysis using hematoxylin and eosin (H&E) and Masson’s trichrome (Masson) staining collectively indicate the presence of newly formed bone within the defects (Fig. 7E). After implantation, significant inflammatory reactions or necrosis are not detected in any of the groups. The defects treated with the scaffolds are filled with newly formed tissue, and mineralized new bone tissues grow along the surface of the scaffolds. In each scaffold group, the quantity of new bone tissue increases with increasing implantation time. The amount of new bone in the TCP/SIS@EM-Apt scaffold group is greater than that in the TCP/SIS and TCP/SIS@EM scaffold groups (Fig. 7F).

TCP/SIS@EM-Apt scaffolds promote anti-inflammatory response, angiogenesis, and osteogenic regeneration

To evaluate the inflammatory response following scaffold implantation, bone tissue sections from rats were harvested seven days post-treatment. Both TCP/SIS@EM-Apt and TCP/SIS@EXOs significantly increase the M2 macrophage population (Fig. 8A and B), confirming the immunomodulatory effects of BMSC-EXOs. Among the tested systems, the EM-Apt group exhibits a more pronounced enhancement in M2 polarization, likely due to a MitoQ-mediated reduction in ROS levels. ROS are primarily generated in M1 macrophages and function as critical inflammatory mediators regulating the M1–M2 phenotypic balance [39, 40]. Furthermore, an increased proportion of M2-polarized macrophages facilitates tissue repair and osteogenesis [11][]. To confirm the in vivo osteogenic activity of the TCP/SIS@EM-Apt scaffold, immunofluorescence assays on all sections were used to determine the expression of the osteogenic marker OCN and angiogenic marker CD31. Owing to the strong correlation between osteogenesis and angiogenesis, vascularization is vital for bone regeneration. The vascular endothelial growth factor is known to facilitate blood vessel formation, and CD31 is a specific marker for vascular endothelial cells [41]. CD31 positive staining in the TCP/SIS@EM-Apt scaffold group is more intense than that in the other groups (Fig. 8C), indicating that the TCP/SIS@EM-Apt scaffold promotes angiogenesis in vivo. This outcome may be attributed to the targeted delivery of MitoQ to BMSCs, which improves the condition of the bone microenvironment, thereby promoting angiogenesis. Additionally, OCN is strongly expressed in the area of new bone formation in the TCP/SIS@EXO and TCP/SIS@EM-Apt scaffold groups, whereas positive staining is not observed in the TCP/SIS scaffold group. This indicates that the delivery of MitoQ significantly promotes bone regeneration in the defect area, consistent with the findings of in vitro experiments. Therefore, the results of in vivo analysis confirm the excellent ability of the TCP/SIS@EM-Apt scaffold to promote anti-inflammatory response, vascularization, and bone repair in a rat model of skull defects.

Fig. 8.

Bone repair after treatment. (A) Representative immunofluorescence staining of M2 macrophages identified by CD206 after different treatments. (B) Quantitative analysis of M2 macrophages. (C) Representative OCN and CD31 immunofluorescence staining images at week 8 post-implantation. Quantitative analysis of (D) CD31 and (E) OCN expression levels. n = 5. ***P < 0.001, ****p < 0.0001

Conclusion

In this study, the impact of a newly synthesized mitochondria-targeting material on the mitochondrial ROS, ATP content, and MMP of BMSCs was comprehensively evaluated. In addition, the osteogenic properties of the material were analyzed in vitro and in vivo. The TCP/SIS@EM-Apt scaffold outperformed both TCP/SIS and TCP/SIS@EXOs, significantly mitigating mitochondrial ROS damage, safeguarding mitochondrial health, and enhancing the bone-regeneration microenvironment through targeted delivery mechanisms. In vitro and in vivo experiments consistently confirmed the excellent bone-inductive ability of the TCP/SIS@EM-Apt scaffold. In vivo analysis indicated that this scaffold enhances angiogenesis within the defect area, and RNA-seq analysis indicated an increased activation of the Wnt pathway. In conclusion, mitochondria-targeting scaffolds represent an emerging biomaterial with potent biological activity to promote vascularized bone regeneration. Besides effectively improving the bone microenvironment for excellent bone regeneration, the mitochondria-targeting delivery platform constructed in this study could offer novel perspectives in the field of bone tissue engineering.

Author contributions

S.Y.: Writing-original draft, Methodology, Funding acquisition. L.Z.: Methodology, Data curation, Formal analysis. W.H.: Formal analysis. J.L.: Visualization. X.C.: Formal analysis. Z.X.: Formal analysis, Funding acquisition. Y.Z.: Methodology. S.H.: Data curation. B.Y.: Methodology, Data curation, Formal analysis. T.S.: Project administration, Resources, Funding acquisition, Supervision. X.G.: Writing-review &editing, Funding acquisition, Project administration, Resources.

Funding

This work was financially supported by the National Natural Science Foundation of China (82272460; 82472440), Natural Science Foundation of Hubei Province of China (2025 AFB290), and the Innovation and Entrepreneurship Training Program for College Students of Huazhong University of Science and Technology (S202410487741). We would like to thank the Medical Subcenter of the Huazhong University of Science and Technology Analytical & Testing Center. We thank the Laboratory Animal Center of Huazhong University of Science and Technology.

Data availability

The datasets generated during this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were approved and performed according to the regulations and guidelines of the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (IORG No.: IORG0003571) and conducted according to the Animal Research: Reporting of In Vivo Experiments (ARRIVE) and Institutional Animal Care and Use Committee (IACUC) guidelines.

Consent for publication

All authors have consented to the publication of this article.

Competing interests

The authors declare no competing interests.