In vitro study on the promotion of osteogenic differentiation by mitochondrial-derived vesicles through activation of inflammation and reprogramming of metabolic pathways

Abstract

The secretion of mitochondrial-derived vesicles (MDVs) has been found to increase during osteogenic differentiation, but their role in intercellular communication and osteogenic promotion remains unclear. In this study, we extracted translocase of outer mitochondrial membrane 20 (Tomm20) + MDVs from bone marrow stromal cells (BMSCs) at different osteogenic culture days using differential centrifugation and immunoprecipitation, then co-cultured them with BMSCs to assess osteogenic differentiation, immune response and metabolic levels. The results showed that osteogenic differentiation enhances MDVs’ secretion and their mitochondrial DNA (mtDNA) content. In promoting osteogenic differentiation ability, osteogenic-induced MDVs (MDV-OMs, especially MDV-OM14 and MDV-OM21) significantly enhance mineralization with OD values 1.37-fold and 1.32-fold higher than those of MDV-OM7 (p < 0.05) after 21 days, respectively. However, these MDVs containing mtDNA activate immune responses by upregulating cGas, Sting, Caspase-9, Il-6, and Tnf-a mRNA levels, inducing cell apoptosis and oxidative stress. In addition, MDVs containing mitochondrial components also have metabolic regulatory functions. Metabolic level detection revealed that MDV-OMs downregulate lactate, promote tricarboxylic acid cycle (TCA) enzyme expression, and increase mitochondrial membrane potential. Among these MDVs, MDV-OM7, induced for 7 days, shows osteogenic function without strong immune response, possibly related to metabolic reprogramming. This study highlights the potential of osteogenic-induced MDVs for bone regeneration, cGAS-STING activation, and metabolic enhancement, and are expected to be used for the treatment of diseases such as tissue damage.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-025-05749-5.

Highlights

MDVs significantly promote the osteogenic differentiation of BMSCs.

MDVs exacerbate BMSC inflammatory response through cGAS-STING pathway.

MDVs co-culture enhance glycometabolism and oxidative phosphorylation in BMSCs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-025-05749-5.

Introduction

Osteogenesis is a critical step in the skeletal development of vertebrates, where osteoblasts synthesize and deposit organic matrix and subsequently facilitate its mineralization to build the skeletal framework [1–3]. In recent years, the role of mitochondria in this process has been increasingly elucidated. Beyond their well-known function in providing energy to support biosynthesis, mitochondria have been found to facilitate the formation of the mineralized matrix during osteogenic differentiation through multiple pathways, such as the formation of phosphate granules within mitochondria and the promotion of mineral component secretion via mitophagy [1, 4]. During the dynamic changes in the mitochondrial network, the upregulation of mitochondrial fusion and alterations in mitochondrial distribution fulfill the substantial energy demands associated with osteoblast differentiation. Research has revealed that the formation of mitochondrial donut structures is a unique morphological feature in physiological osteogenesis, which can trigger the generation of MDVs [5]. Moreover, the regulation of mitochondrial morphology towards fission and donut-shaped structures has been shown to support the maturation of osteoblasts [6]. The mechanism by which extracellular vesicles promote osteogenesis can be linked to their role in intercellular communication and the delivery of bioactive molecules [7–9]. This suggests that MDVs may serve as carriers of osteogenic signals, facilitating the coordination of bone formation by transferring critical regulatory factors between cells [5, 10]. Nevertheless, current studies have not yet explored the use of MDVs to enhance osteogenic differentiation. This unaddressed area has motivated us to elucidate the potential involvement of MDVs in intercellular communication and their role in promoting bone regeneration.

Neuspiel et al. were the first to discover that MDVs are single- or double-membrane vesicles generated from the mitochondrial outer membrane (OMM), inner membrane (IMM), and partial mitochondrial matrix under basal, stress, and hypoxic conditions [11]. Concurrently, Extensive research has shown that MDVs exhibit multifaceted roles in cellular physiology and pathology, exhibiting both protective and potentially deleterious effects [12]. MDVs are an essential component of mitochondrial quality control, selectively clearing damaged mitochondrial components and preventing cells from completely degrading mitochondria through mitophagy, thereby maintaining mitochondrial health and function [13–16]. They play a key role in inter-organelle communication, for example, by participating in peroxisome biogenesis, regulating immune responses, and protecting cells under hypoxic stress [11, 17–19]. Additionally, MDVs have demonstrated protective effects in various diseases, such as restoring cardiomyocyte energy metabolism during myocardial infarction [20] or clearing damaged mitochondria in alcohol-induced liver injury [21]. Under certain disease states, the abnormal release of MDVs may amplify inflammatory signaling pathways, further worsening the pathological process. For example, in hormone-resistant breast cancer, MDVs transfer mtDNA to promote the activation of cancer stem cells, leading to treatment resistance [22]. In some extreme cases, vesicles may transfer damaged mitochondrial components to other cells through horizontal transfer mechanisms, thereby spreading damage signals between cells and further disrupting tissue homeostasis [23–26]. Moreover, changes in the generation and release of MDVs are closely related to aging, autoimmune diseases, cancer, and infections [27]. Overall, the role of MDVs cannot be reduced to a binary classification but depends on their generation context within the cell, the cargo they carry, and the physiological or pathological environment in which they exist [12]. A deeper understanding of the complexity of MDVs is crucial for elucidating their roles in health and disease.

Under physiological and pathological conditions, cells encapsulate mitochondrial components in lipid membranes and release them into the extracellular environment as MDVs [28]. Various studies have identified multiple mitochondrial components in MDVs, including mtDNA, cardiolipin, mitochondrial transcription factor A (TFAM), translocase of outer mitochondrial membrane 20 (TOMM20), cytochrome C (CytC), pyruvate dehydrogenase (PDH), and voltage-dependent anion channel 1 (VDAC1) [29]. As cargo of MDVs, play pivotal roles in intercellular communication, thereby regulating numerous biological functions such as mitochondrial recovery, metabolic rescue, mitochondrial degradation, and inflammatory responses [30]. Collins et al. were the first to provide evidence that mtDNA acts as a damage-associated molecular pattern (DAMP) [31], and subsequent studies have demonstrated that mtDNA can activate various pattern recognition receptors (PRRs) [32]. Abhishek Jauhari et al. revealed that mtDNA release induces inflammatory responses by activating the cGAS/STING/IRF3 signaling axis [33]. When CytC is released into the cytosol, it drives the formation of the apoptosome, which subsequently initiates the intrinsic apoptotic program. Furthermore, A study have shown that extracellular vesicles derived from mesenchymal stem cells can stabilize mtDNA through the TFAM pathway, alleviating mitochondrial damage and inflammation, with therapeutic effects validated in acute kidney injury models [34]. On the other hand, inflammasome activation can enhance MDVs secretion as a self-protective mechanism to mitigate inflammation and prevent cellular damage [29]. These findings underscore the critical role of MDVs in immune regulation and disease, suggesting their involvement in inflammatory responses, immune cell activation, and the progression of pathological processes [18, 27]. This suggests that while studying the impact of MDVs on osteogenic differentiation, we also need to pay attention to the level of cellular inflammatory response.

Studies have revealed that MDVs encapsulate mitochondrial enzymes, including PDH, mitochondrial matrix enzymes involved in the tricarboxylic acid cycle (TCA) and fatty acid β-oxidation, as well as the antioxidant enzyme superoxide dismutase 2 (SOD2) [35]. Hazan and D’Acunzo detected subunits of complex V in yeast MDVs and confirmed their ATP-generating capability [36]. Osteogenic differentiation is a biologically demanding process with significant energy requirements that necessitates substantial ATP consumption to facilitate cell proliferation, matrix synthesis, and mineralization. And mitochondria generate energy through oxidative phosphorylation (OXPHOS) and glycolysis to meet the energy demands of osteoblasts at different stages of differentiation [9]. Furthermore, metabolic byproducts such as lactate, ATP, and NAD⁺ act as signaling molecules to regulate osteoblast differentiation and function, while moderate levels of oxidative stress are maintained through balanced mitochondrial function [5]. Therefore, our rough speculation about the impact of MDVs on osteogenic differentiation is based on the influence of metabolic pathways.

The present research explores the effects of MDVs secreted by BMSCs under varying culture conditions on osteogenic differentiation, inflammatory responses, and mitochondrial metabolic function through in vitro experiments. It underscores the therapeutic potential of MDVs in promoting bone remodeling and lays the groundwork for further investigations into their regulatory roles in the process of osteogenic differentiation.

Materials and methods

Primary cell extraction and culture of bone marrow stromal cells

BMSCs have been extracted in our previous research [37]. In accordance with the ethical care of experimental animals, 2-week-old male SD rats were anesthetized by 3% sodium pentobarbital intraperitoneal overdose and sterilized by immersion in 75% ethanol for 5 min. After aseptically removing the tibia and femur of SD rats, the bone marrow tissue was flushed to obtain a cell suspension. Erythrocyte lysate was added and incubated for 5 min followed by centrifugation at 800 rpm. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, USA) containing 10% exosome-free FBS (SBI, USA) and 1% penicillin-streptomycin (10000 U/mL, Gibco, USA) for 24 h and the medium was changed to obtain pure BMSCs. Subsequently, the medium being changed every 2 days. After passaging the cells to the third generation, the BMSCs were harvested for further experiments.

Osteogenic differentiation of BMSCs

For the BMSCs induced by osteogenic medium (BMSCs-OM) group, the cells were cultured under an osteogenic induction medium containing β-glycerophosphate (10 mM, Sigma, USA), 2-Phospho-L-ascorbic acid trisodium salt (50 µM, Sigma, USA), and dexamethasone (100 nM, Sigma, USA). For the BMSCs induced by proliferation medium (DMEM containing 10% exosome-free FBS and 1% penicillin-streptomycin) (BMSCs-PM) group, the cells were not treated.

Alizarin red S staining

BMSCs-PM and BMSCs-OM, cultured for 7, 14, and 21 days, were stained with alizarin red. The cells in each group were seeded onto 24-well plates. The cells in the wells were washed with phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde for 15 min, and subsequently stained with 1 mL alizarin red staining solution. After static staining at room temperature for 30 min, the cells were washed with distilled water and imaged using a microscope. The optical density (OD) was measured using a microplate reader for quantification.

Separation of MDVs by ultracentrifugation and immunoaffinity

The MDVs isolation method was established in our preliminary research [38]. The cell culture medium from BMSCs-PM and BMSCs-OM at 7, 14, 21 days were collected, the medium in each plates was 50 mL, and EVs were purified by ultracentrifugation. The medium was centrifuged at 300× g for 10 min at 4℃, then centrifuged at 2,000× g for 10 min at 4℃, to transfer the supernatant and then centrifuged at 10,000× g for 10 min at 4℃ to remove dead cells and cell debris. The supernatant was carefully collected and transferred to a 50 mL open-ended ultracentrifugal tube for 4℃ and 100,000× g for 70 min, and the supernatant was discarded. The precipitate was suspended with 2 mL PBS and then transferred to a 12.5 mL tube for centrifugation at 4℃ for 70 min at 100,000× g. Each centrifuge tube was suspended with 2 mL PBS.

After EV separation, we further isolate MDVs with immunoaffinity separation. The magnetic beads (abs9905, absin, China) were prewashed with PBS to remove the storage buffer and reduce nonspecific binding. A magnetic stand was used to separate the beads from the PBS. The supernatant was discarded, and the washing step was repeated twice. The prewashed magnetic beads were incubated with an anti-Tomm20 antibody (sc-17764, Santa Cruz, Germany) using 5 µg per mg of beads with gentle rotation for 1 h at room temperature to allow antibodies to bind to the beads. After the antibody binding, the beads were washed three times with IP buffer to remove any unbound antibodies. Previously isolated EVs (EVs from BMSCs-OM or BMSCs-PM cell culture medium) were added to the antibody-conjugated beads. The mixture was incubated with gentle rotation overnight at 4℃ to allow the Tomm20 antibodies to bind to MDVs. After incubation, the tube was placed on a magnetic stand to collect the beads bound to MDVs, and the supernatant containing unbound EVs was removed. Bead-bound MDVs were washed three times with an ice-cold IP buffer to remove unbound EVs. After the final wash, PBS was added to the beads and gently resuspended to collect bead-bound MDVs for downstream analyses. If intended for cell culture, serum-free DMEM was used for resuspension. The isolated MDVs were stored at -80 ℃ for long-term storage. The separated MDVs are named according to the culture conditions of BMSCs, such as MDV-PM 7, 14, 21 and MDV-OM7, 14, 21.

Nanoparticle tracking analysis (NTA)

NTA was conducted using ZetaView PMX 110 (Particle Metrix, Germany) and the corresponding software (ZetaView 8.02.28). The MDVs were appropriately diluted in PBS and loaded in a sample cell chamber, followed by size distribution and concentration measurement under a wavelength of 405 nm. Finally, the size of the particles was quantified and recorded.

BCA protein concentration assay

An amount of 20 µL MDVs solution from each group were taken for reaction with the BCA (Beyotime, China) working solution and incubated at 37℃ for 30 min. The OD values of the sample group and the standard group at 562 nm were read by the microplate reader, the standard curve was plotted, and the protein concentration of the sample was calculated. The samples were adjusted to the same protein concentration for subsequent detection.

Transmission Electron Microscope (TEM)

TEM was performed for morphological analyses per standard operating procedures. MDVs suspensions were placed onto TEM grids, and the samples were negatively stained for 30 s with 2% uranyl acetate. The resulting images were observed using a TEM instrument (HT7700, Hitachi, Japan).

Western blot

Protein samples (20 µg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes (Millipore, Germany). After blocking with 5% skimmed milk for 1 h at room temperature, they were incubated with the corresponding primary antibody overnight at 4℃. Next, the blots were incubated with secondary antibodies and developed for 2 min using the ECL Plus Substrate (Thermo Scientific, USA). The signal was detected using an enhanced chemiluminescence detection system (Tanon5200, Tanon, China) per the instructions of the manufacturer. Western blot was used to detect the positive MDVs markers TOMM20 (1:10000, 11802-1-AP, Proteintech, USA), TIM44(1:3000, 13859-1-AP, Proteintech, USA), VDAC1 (1:2000, 10866-1-AP, Proteintech, USA); EVs markers CD9 (1:500, ab307085, Abcam, UK), CD63 (1:5000, 67605-1-Ig, Proteintech, USA), CD81 (1:2000, ab109201, Abcam, UK), HSP70 (1:10000, 10995-1-AP, Proteintech, USA) and ALIX (1:10000, 12422-1-AP, Proteintech, USA) were used as control.

ROS fluorescence staining

After co-culture MDVs with BMSCs for 3 days, a DCFH-HA Kit (Beyotime, China) was used to measure ROS production. Briefly, the supernatant was removed by centrifugation and 1 × 10⁶ cells/mL were collected. The 2’,7’ dichlorofluorescin diacetate probe (10 µM) was added and incubated for 20 min at 37 °C. Confocal microscopy was used to detect cell fluorescence.

Live/Dead cell viability assay

After 3 days of co-culture, the cell viability of BMSCs was detected using a Live/Dead Cell Viability/Cytotoxicity Assay Kit (Beyotime, China), as per the manufacturer’s instructions. Briefly, BMSCs were washed 3 times with PBS and stained with Live/Dead working solution in serum-free DMEM. After incubating the cells for 30 min at 37℃, the stained cells were observed and photographed using an confocal microscope.

qRT-PCR

RNA from each group was extracted using the Trizol reagent solution (Invitrogen, USA). Complementary DNA (cDNA) was synthesized using a transcriptase first strand cDNA Synthesis Kit (Marligen Biosciences, USA). Next, qRT-PCR was performed using the AceQ qPCR SYBR Green Master Mix (Vazyme, China). The cycling conditions are as follows: initiation at 95 °C for 10 min, cycling 40 times at 95℃ for 15 s, and 60℃ for 1 min. Relative expression of multiple genes was determined compared with the internal control of the β-actin gene. The primer sequences used for qRT-PCR are listed in Table S1(Supporting Information).

Detection of ATP generation

The ATP concentration was measured using a colorimetric ATP assay kit (Abcam, UK) per the kit protocol. Each group of cells was lysed with 150 µL of ATP buffer. The samples were then centrifuged at 13,000 ×g for 5 min at 4℃ to remove insoluble material and loaded onto 96-well plates. The plates were incubated at room temperature for 30 min in the dark. The absorbance was measured at 570 nm using a microplate reader.

Detection of lactate generation

Intracellular L-lactate levels were quantified using the L-Lactate Assay Kit (Abcam, UK) following the manufacturer’s protocol. Cells were harvested and washed with cold PBS. The cell pellets were resuspended in 4× volumes of L-Lactate Assay Buffer (provided in the kit) and subjected to repeated pipetting to ensure complete homogenization. The lysates were then centrifuged at maximum speed for 2–5 min at 4℃ to remove insoluble material, and the supernatants were subsequently collected for further analysis. To eliminate interference from endogenous lactate dehydrogenase (LDH) and other proteins, the supernatants were deproteinized using the Deproteinizing Sample Preparation Kit (ab204708, Abcam). The deproteinized samples were stored on ice until further analysis. A standard curve was prepared using the 1 mM L-lactate standard provided in the kit, with final concentrations ranging from 0 to 10 nmol per well. The assay was performed in a 96-well plate, with 50 µL of each standard and sample added in duplicate. The reaction mix was prepared by combining 46 µL of L-Lactate Assay Buffer, 2 µL of Developer Solution (L-Lactate Substrate Mix), and 2 µL of Enzyme Mix (L-Lactate Enzyme Mix). A total of 50 µL of the reaction mix was added to each standard and sample well, while a background reaction mix (without Enzyme Mix) was added to background control wells. The plate was incubated at room temperature for 30 min, and the absorbance was measured at 450 nm using a microplate reader. The concentration of intracellular L-lactate was calculated based on the standard curve.

Hexokinase activity assay

The activity of hexokinase (HK) in cells was determined using the Hexokinase Activity Assay Kit (Solarbio, China), following the manufacturer’s protocol. The assay is based on the principle that HK catalyzes the conversion of glucose to glucose-6-phosphate, which is subsequently oxidized by glucose-6-phosphate dehydrogenase to produce NADPH, generating a characteristic absorbance peak at 340 nm, detectable by a microplate reader. Cells were harvested and centrifuged at 8000×g for 10 min at 4℃, and the supernatant was discarded. The cell pellet was resuspended in extraction buffer at a ratio of 200 µL buffer per 1 × 10⁶ cells. The cells were lysed using an ultrasonic processor (200 W, 3 s on/10 seconds off, for 30 cycles) on ice. The lysate was then centrifuged again at 8000×g for 10 min at 4℃, and the supernatant containing soluble proteins was collected for the measurement of HK activity. The absorbance was measured at 340 nm using a microplate reader.

Mitochondrial membrane potential determination

The JC-1 mitochondrial membrane potential kit (Beyotime, China) was used to detect MDVs and BMSCs co-cultured for 2 days. The cells in each group were seeded in a confocal microscopy dish. JC-1 staining was conducted according to the kit’s instructions and added to the plate. The cells were incubated at 37℃ for 15 min. The negative control group was set as BMSCs-CCCP with 10 µM CCCP. An inverted laser confocal microscope was used to stimulate and photograph at 488 and 594 nm. It has been established that JC-1 exists in the mitochondrial matrix as a polymer when the mitochondrial membrane potential is high, and the polymer fluorescence can be measured at 594 nm. However, JC-1 exists as a monomer when the mitochondrial membrane potential is low and cannot be aggregated in the mitochondrial matrix. JC-1 is mostly dissociated in the cytoplasm, and the monomer fluorescence can be measured at 488 nm. Fluorescence images were quantified by the ratio of two fluorescence channels (594/488 nm).

Statistics and reproducibility

All values are presented as mean ± standard error of the mean (SEM). One-way ANOVA with Tukey’s post hoc test was performed for statistical analysis of more than 2 groups, while two-way ANOVA with Tukey’s HSD test was performed for statistical analysis of time grouping. Statistical analyses were performed using graphpad prism 9.3. A p-value < 0.05 was considered statistically significant. Samples sizes are given for each experimental condition in the figure legends. For representative data and images given in the figures, at least three or more independent experiments were conducted, showing similar results.

Result

Isolation and characterization of MDVs from BMSCs

To investigate the secretion levels and roles of MDVs at different stages of osteogenic differentiation, we first performed mineralization staining on BMSCs, the source cells of MDVs. The Alizarin Red S staining results in Fig. 1 A showed that BMSCs in the osteogenic induction group secreted calcium salt deposits under the microscope from day 7. However, BMSCs in the normal culture group only showed a few calcium nodules at day 21.

Fig. 1.

Isolation and characterization of MDVs from BMSCs (A) Alizarin red S staining of BMSCs under different culture conditions. (B) Schematic image of MDVs isolation using differential centrifugation and immunoaffinity. (C) NTA detection of MDV-PM7, MDV-PM14 and MDV-PM21. (D) NTA detection of MDV-OM7, MDV-OM14 and MDV-OM21. (E) BCA protein concentration detection of MDVs samples(n = 3 each). (F) Morphology of MDVs observed by TEM, scale bar = 30 nm. (G) WB of the MDVs and EV markers. (H) Detection of mtDNA in MDVs (n = 3 each). ns: p > 0.05

After constructing BMSCs cell models under different osteogenic differentiation states, we used them as source cells and isolated MDVs from the cell culture supernatant. Specifically, referring to our previous research, extracellular vesicles were initially isolated after different centrifugation, and then immunoaffinity separation was performed on the universal marker Tomm20 of MDVs (Fig. 1B). Different induction conditions (proliferation medium, PM) (osteogenic induction medium, OM) and induction days were used to name MDVs.

We use various methods to characterize MDVs. NTA results shown that the particle sizes of the MDVs were mainly concentrated in the range of 50 to 100 nm (Fig. 1 C, D). In addition, we observed that the number of MDVs produced by BMSCs cultured in osteogenic induction medium (MDV-OM21 and MDV-OM14) was significantly higher than that of BMSCs cultured in proliferation medium (MDV-PM14 and MDV-PM21). Therefore, BCA protein assay was used to detect the amount of MDVs 100mL cell culture supernatant (Fig. 1E). The results showed that BMSCs wound secrete more MDVs in the medium containing osteogenic inducers compared to normal medium for the same number of days of culture. At day 21, the MDVs concentration was 47.8 ± 2.97 µg/mL in the MDV-PM21 group and 83.97 ± 7.92 µg/mL in the MDV-OM21 group, which was 1.76-fold difference. The result suggests that osteogenic inducers may have a significant effect on MDV secretion level of BMSCs. The extracted MDVs all showed a cup-shaped morphology containing a double-layer membrane structure (Fig. 1 F), and expressed MDVs marker Tomm20, Tim44, Vdac1, Cd63, but did not express Cd81, Cd9, Hsp70 or Alix (Fig. 1G). The above results confirmed the successful isolation of MDVs from cell culture supernatant.

Furture, we detected the mtDNA content of MDVs in each group. At the early stage of culture (day 7), there was no significant difference in mtDNA content between the two groups of MDVs, whereas at the middle and late stages of induction (day 14 and 21), mtDNA was significantly higher in the MDV-OM groups than in the MDV-PM groups (Fig. 1 H). And the expression of mtDNA increased with the increase in the number of culture, which was 1.38-fold higher in MDV-OM21 compared to MDV-OM7.

The above results indicate that we extracted MDVs secreted by BMSCs with different osteogenic induction days (7, 14, 21 days) and found that the secretion level of MDVs increased significantly in the late stage of osteogenic differentiation, and the level of mtDNA contained also increased. However, there was no significant difference in the morphology and biomarkers of MDVs. As the process of osteogenic differentiation progresses, we will investigate whether the MDVs secreted at different stages mediate differentiated intercellular communication behavior.

MDVs exacerbate BMSC inflammatory response through cGAS-STING pathway

Following 3 days of co-culturing MDVs with BMSCs, we utilized the DCFH-DA fluorescent probe to evaluate ROS production in various treatment groups (Fig. 2 A). The ROS fluorescence intensity exhibited significant variance among treatment groups with increasing culture duration. In both MDV-PM7 and MDV-OM7 groups, ROS levels remained low. In contrast, the ROS fluorescence intensity increased significantly by 2.9- and 2.23-fold in the MDV-OM14 and MDV-OM21 groups, respectively, compared to the MDV-PM14 and MDV-PM21 groups (Fig. 2B). As seen in Fig. 2C, the percentage of ROS-positive cells was significantly low in both the MDV-PM7 and MDV-OM7 groups, with no significant intergroup differences noted. In the MDV-PM14 and MDV-OM14 groups, the percentage of ROS-positive cells increased significantly, with the MDV-OM14 group showing a particularly substantial increase, being 5.63 times higher than that in the MDV-PM14 group. In the MDV-PM21 and MDV-OM21 groups, the percentage of ROS-positive cells further escalated, with values of 26.28 ± 5.69% and 38.94 ± 3.68%, respectively, indicating a significantly higher percentage in the MDV-OM21 group as compared to the MDV-PM21 group. The co-culture of BMSCs with MDV-PM14, MDV-PM21, MDV-OM14, and MDV-OM21 significantly increased ROS production, with this effect becoming more pronounced after treatment with MDV-OM14 and MDV-OM21. Conversely, MDV-PM7 and MDV-OM7 treatments had no significant effect or a minimal effect on ROS levels.

Fig. 2.

MDVs exacerbate oxidative stress, reduce cellular activity and promote expression of inflammation-related genes in BMSCs. (A) ROS immunohistochemical staining of BMSCs with DCFH-DA (green, for ROS detection) and Hoechst (blue, for nuclei) on day3, scale bar = 200 μm. (B) Quantification of ROS fluorescence instensity. (C) Percentage of ROS-positive cells. (D) Fluorescence imaging of live (green) and dead (red) BMSCs on day3, scale bar = 100 μm. (E) Percentage of green fluorescent cells. (F) Cell density of cells. (G) cGas expression of co-cultured BMSCs at 7, 14, and 21d. (H) Sting expression of co-cultured BMSCs at 7, 14, and 21d. (I) Caspase-9 expression of co-cultured BMSCs at 7, 14, and 21d. (J) IL-6 expression of co-cultured BMSCs at 7, 14, and 21d. (K) Tnf-α expression of co-cultured BMSCs at 7, 14, and 21d. (L) Schematic diagram illustrating the inflammatory response and apoptosis induced by MDVs. (n = 3 each, ns: p > 0.05.)

Additionally, we employed Calcein-AM/PI staining to assess cell viability across different treatment groups (Fig. 2D). The cell viability in the MDV-PM7, MDV-PM14, MDV-PM21, MDV-OM7, and MDV-OM14 groups approached 100% (Fig. 2E), indicating no significant impact on cell viability. Nevertheless, the cell viability of the MDV-OM21 group was 94.12 ± 2.04%, significantly lower than other groups, suggesting a potential negative impact of MDV-OM21 treatment on cell viability. In summary, MDV-OM21 slightly reduces cell viability and may induce cell damage, while treatments with MDV-PM7, MDV-PM14, MDV-PM21, MDV-OM7, and MDV-OM14 had no significant effect on BMSCs’ cellular activity. BMSCs co-cultured with MDV-OM21 exhibited a significantly lower cell density, averaging 70.46 ± 2.85%, compared to other groups where cell densities exceeded 90%, representing a 20% difference. This indicates that the cells are washed away during the staining process due to cell death (Fig. 2 F).

To evaluate the impact of co-culturing MDVs with BMSCs for 7, 14, and 21 days on inflammation-associated genes, we measured the mRNA levels of cGas, Sting, Caspase-9, Il-6, and Tnf-α. The PM group served as the negative control, while the OM group served as the positive control. As in Fig. 2G, cGas expression increased in the MDV-OM14, and MDV-OM21 groups on day 14, peaking on day 21. Furthermore, these two groups demonstrated the most significant increases, with comparable cGas expression levels. Sting expression remained stable across all groups at day 7, but significantly increased in the MDV-OM14, and MDV-OM21 groups by day 14, peaking at day 21, particularly in the MDV-OM21 groups (Fig. 2 H). At day 14, the expression level of Caspase-9 was more than twofold in both the MDV-OM14 and MDV-OM21 groups compared to day 7, and the increases in these two groups were significantly greater than those in the OM group (Fig. 2I). At day 21, the highest expression level was observed in the MDV-OM21 group, followed by the MDV-OM14 group. Il-6 expression was significantly altered at day 7 in the OM, MDV-OM14, and MDV-OM21 groups, peaking at this time point, but was lower in the MDV-OM groups than in the OM group and gradually decreased as the incubation period lengthened (Fig. 2 J). Tnf-α expression in the OM, MDV-OM14, and MDV-21 groups also peaked at day 7 and showed an increasing then decreasing trend (Fig. 2 K). At day 7, the OM group exhibited a higher elevation than the MDV-OM14 and MDV-21 groups, but at days 14 and 21, the expression levels in the OM, MDV-OM14, and MDV-21 groups significantly decreased and became more uniform.

The above results indicate that whether in proliferation medium or osteogenic induction medium. MDVs secreted by BMSCs will exacerbate oxidative stress and cell death of target cell after long term culture. This phenomenon was particularly significant in the MDVs (MDV-OM14, MDV-OM21) groups secreted after 14 days of osteogenic induction. Furthermore, the expression variations of cGas, Sting, Caspase-9, Il-6, and Tnf-α as inflammatory-related genes in the MDV-PM and MDV-OM7 groups were not significant across different culture periods. However, MDV-OM14, and MDV-OM21 groups were significantly expressed during osteogenic differentiation. Specifically, cGas, Sting, and Caspase-9 began to increase midway through osteogenic differentiation, i.e., at day 14, and peaked at day 21. In contrast, Il-6 and Tnf-α showed significant changes early in osteogenic differentiation, i.e., at day 7, followed by a gradual decline in expression in the middle and late stages. The cascade of inflammation and apoptosis induced by MDVs in BMSCs highlights the potential regulatory mechanisms of MDVs and emphasizes their significance in the regulation of bone remodeling (Fig. 2 L). The aforementioned results elucidate the dynamic changes in the expression of inflammation-related genes during the co-culture of MDVs and BMSCs, indicating that the MDV-OM14 and MDV-OM21 groups exhibited higher levels of mtDNA. It is plausible that the delivery of mtDNA by MDVs activate the innate immune response of cells, suggesting that inflammatory reactions may play distinct roles at various stages of osteogenic differentiation.

MDV-OM promote osteogenic differentiation of BMSCs

To clarify whether MDVs have a promoting effect on the osteogenic differentiation of BMSCs, we co cultured MDVs isolated under different conditions with BMSCs cell and detected them by Alizarin Red staining at 7, 14 and 21 day time points (Fig. 3A, B). The microscopic image after BMSCs staining is shown in Fig. 3A, B. Very little mineralization staining was observed after MDV-PMs induction, suggesting low levels of mineralization. After 14 days of MDV-OMs induction, a large number of deeply stained mineralized nodules appeared. It is worth nothing that then MDV-OM14 and MDV-OM21 were co cultured with BMSCs for 21 days, cell density decreased and open areas appeared due to cell death and shedding, as indicated by the yellow arrow.

Fig. 3.

The effects of MDVs on the osteogenic differentiation level of target cells (A) Co-culture of BMSCs with MDV-PM7, MDV-PM14, MDV-PM21 followed by Alizarin red staining at 7, 14, 21d. (B) Co-culture of BMSCs with MDV-OM7, MDV-OM14, MDV-OM21 followed by Alizarin Red Staining at 7, 14, 21d. The yellow arrow indicated cell death and shedding. (C, D) Quantitative analysis of OD value of Alizarin red staining. (E)Comparison diagram of co-culture between MDVs and BMSCs. (F) Ocn expression of co-cultured BMSCs at 0, 7, 14, and 21d. (G) Runx2 expression of co-cultured BMSCs at 0, 7, 14, and 21d. (H) Opn expression of co-cultured BMSCs at 0, 7, 14, and 21d. (I) Col-1 expression of co-cultured BMSCs at 0, 7, 14, and 21d. (n = 3 each, ns: p > 0.05.)

The quantitative analysis of Alizarin Red staining is calculated by reading the OD value of the staining (Fig. 3 C, D). By day 7 of co-culture, there was an increase in calcium salt deposition within the extracellular matrix of MDV-OM-treated BMSCs, indicative of the onset of cellular mineralization. Although calcium nodule formation was not markedly evident at this stage, early mineralization was indicated, with no significant variation among treatment groups. Following 14 days of induction, osteogenic differentiation of BMSCs became more pronounced. Notably, BMSCs treated with MDV-OM14 and MDV-OM21 displayed the most pronounced mineralization, with OD values 1.19-fold and 1.28-fold greater than those of the MDV-OM7 group, respectively. After 21 days of induction, the MDV-OM14 and MDV-OM21 groups continued to demonstrate the most significant osteogenic differentiation, with OD values 1.37-fold and 1.32-fold that of the MDV-OM7 group, respectively. The above results demonstrate the interactions and effects of MDVs and BMSCs from distinct sources during co-culture, revealing notable differences in apoptosis and osteogenic differentiation outcomes (Fig. 3E).

To futher assess the effect of MDV in inducing osteogenic differentiation, we analyzed the changes in the mRNA expression levels of osteogenesis-related genes Ocn, Runx2, Opn, and Col-1 using quantitative RT-PCR. BMSCs were co-cultured with MDVs for 0, 7, 14, and 21 days and then compared with the proliferation medium (PM) used for negative control, and the osteogenic inducing medium (OM) were used for positive control. The results showed that Ocn expression began to increase on day 7 in the OM group, while it started to rise on day 14 in the MDV-OMs group (Fig. 3 F). By day 21, Ocn expression reached its peak in both the OM and MDV-OMs groups, with the relative expression in the OM group being 4.35, higher than that in the MDV-OMs group. Within the MDV group, MDV-OM14 exhibited the highest relative expression at 3.51, followed by MDV-OM21 at 3.16. Runx2 expression significantly increased and peaked on day 14 in the OM, MDV-OM14, and MDV-OM21 groups, then gradually declined. There was no statistically significant difference in Runx2 expression between the MDV-OM14 and MDV-OM21 groups on days 14 and 21 (Fig. 3G). Opn expression was significantly elevated in the OM and MDV-OM groups on day 7 and continued to increase slowly as incubation time extended, but the expression in the MDV-OMs groups remained consistently lower than that in the OM group, with MDV-OM14 and MDV-OM21 showing the highest relative expression (Fig. 3 H). Col-1 expression gradually increased in both the OM and MDV-OMs groups, peaking on day 21, but the expression level in the MDV-OMs groups was lower than that in the OM group (Fig. 3I).

The results indicated that, compared to day 0, there was no significant change in the expression of osteogenesis-related genes in the MDV-PMs group on days 7, 14, and 21. In contrast, the expression in the MDV-OMs groups significantly increased, further confirming that the sorted MDV-OMs possesses a certain degree of osteogenic potential. Notably, the contribution to bone differentiation was more significant in the MDV-OM14 and MDV-OM21 groups.

Taken together, the MDV-OM groups was more effective in enhancing the osteogenic differentiation of BMSCs compared to the MDV-PM groups, with the extent of differentiation increasing over extended co-culture periods. Particularly, the MDV-OM14 and MDV-OM21 groups exerted a more potent induction effect on BMSC osteogenic differentiation.

MDV-OMs co-culture enhance glycometabolism and oxidative phosphorylation in BMSCs

After 2 days of co-culture with MDVs, we examined the glycometabolism products, glycolytic enzyme activity, tricarboxylic acid cycle enzyme mRNA levels and mitochondrial membrane potential in BMSCs. The data obtained after two days of incubation of BMSCs in the proliferation medium constituted the PM group, serving as the negative control. Compared with the PM group, the MDV-OM14 and MDV-OM21 groups exhibited the highest ATP production, with the most significant increase, reaching 1.38 times that of the negative control. However, lactate production in the MDV-OM14 and MDV-OM21 groups was 0.70 and 0.67 times that of the PM group, respectively, whereas no significant differences were observed between the other groups and the PM group (Fig. 4A). Additionally, the MDV-OM14 and MDV-OM21 groups had the highest activity of hexokinase (HK) with values increased by 32% and 39% compared to the negative control group (Fig. 4B). In this study, the mRNA expression levels of isocitrate dehydrogenase (Idh), citrate synthase (Cs) and α-ketoglutarate dehydrogenase complex (α-Kgdh) in the tricarboxylic acid cycle were examined using qRT-PCR (Fig. 4 C). Compared with the negative control group, the mRNA expression levels of the aforementioned enzymes were significantly higher in both the MDV-OM14 and MDV-OM21 groups, being approximately twice as high as those in the negative control group. The aforementioned results indicated that aerobic glucose metabolism was significantly enhanced in BMSCs in the MDV-OM14 and MDV-OM21 groups. To assess the changes in mitochondrial membrane potential in BMSCs, the present study utilised fluorescent staining with the JC-1 probe for detection (Fig. 4D, E). In the mitochondrial membrane potential assay, when the mitochondrial membrane potential was high, the JC-1 probe aggregated in the mitochondrial matrix as polymers and emitted red fluorescence; whereas at low mitochondrial membrane potential, the JC-1 probe could not enter the mitochondrion and only showed green fluorescence as a monomer. Therefore, the ratio of aggregates (red) to monomers (green) can be used to reflect the mitochondrial membrane potential. BMSCs treated with the mitochondrial depolarisation inducer CCCP exhibited higher green fluorescence intensity and lower mitochondrial membrane potential, serving as a positive QC control group (Fig. 4 F). Furthermore, the OM group exhibited a significantly reduced mitochondrial membrane potential compared to the PM group. The mitochondrial membrane potential was maintained at a high level in the MDV-OM14 and MDV-OM21 groups (Fig. 4G). However, there was no significant difference in the mitochondrial membrane potential levels within the MDV-PM groups. In summary, the mitochondrial membrane potential of BMSCs was significantly rised after co-culture with MDV-OM14 and MDV-OM21, indicating that the cells might be undergoing oxidative phosphorylation, thereby driving the synthesis of ATP. ()

Fig. 4.

MDV-OMs co-culture enhance glycometabolism and oxidative phosphorylation in BMSCs (A) Relative ATP and LA production of co-cultured BMSCs at 2d. (B) Relative HK activity of co-cultured BMSCs at 2d. (C) Idh, Cs, α-Kgdh expression of co-cultured BMSCs at 2d. (D, E) Fluorescence image of mitochondrial membraane potential JC-1 staining, scale bar = 50 μm. (F, G) Fluorescence ratio quantitative analysis of mitochondrial membrane potential JC-1 staining. (n = 3 each, ns: p>0.05.)

Discussion

In this study, MDVs were isolated from BMSCs at various stages of osteogenic differentiation, and it was observed that the quantity of MDVs secreted by BMSCs increased significantly during the mid-to-late stages of osteogenic differentiation. The promotive effect of MDVs on osteogenic differentiation was further validated through co-culture experiments, alongside the potential pathological and physiological effects of their associated components on cells. These findings suggest that MDVs play a crucial role in osteogenesis, offering a scientific foundation for using MDVs as a novel approach to promote osteogenic differentiation. However, they may also exert adverse effects on cells, a matter that requires further investigation.

To elucidate the role of MDVs in osteogenic differentiation, it is essential to isolate and purify a substantial quantity of these vesicles. To achieve this, the present study developed a method to isolate MDVs from BMSCs, followed by characterization and investigation of the isolated MDVs. Initially, these extracellular vesicles were isolated using density gradient centrifugation, a standard separation technique, followed by further purification through immunoaffinity assay to isolate MDVs. This method has been utilized in previous studies as well [39]. As previously reported in the literature, MDVs are mitochondria-derived vesicles that primarily originate from the outgrowth process of the OMM and certain components of the IMM. Tomm20 is a characteristic marker protein for the OMM, and most MDVs contain this membrane component, classified as TOMM20+, suggesting that MDVs commonly express Tomm20. This is why Tomm20 was selected as the immunoaffinity antibody in this study. Additionally, some vesicles bud off from the IMM, known as vesicles derived from the inner mitochondrial membrane (VDIMs). According to Akriti Prashar et al., VDIMs are a newly identified vesicular structure derived from the IMM, which can be selectively cleared by lysosomes and are specific products of IMM damage [40]. However, this protocol may not isolate VDIM-like vesicles originating from the inner mitochondrial membrane during MDV isolation. NTA revealed that the diameter of the nanoparticles ranged from 50 to 100 nm, which is consistent with the typical size of mitochondrial vesicles [5, 11]. TEM imaging further confirmed this size range and confirmed the presence of a bilayer membrane structure. Several studies have demonstrated that MDVs can carry mtDNA and other signaling molecules, and are subsequently released outside the cell via the multivesicular body pathway (MVB) to facilitate functions such as intracellular material transport, mitochondrial quality control, and cellular communication [10, 41, 42]. Consequently, the western blot results indicated that the extracted MDVs expressed the exosome marker CD63, along with mitochondrial proteins. Collectively, the sizes and marker proteins of the isolated vesicles were consistent with expectations. It was observed that the secretion level of isolated MDVs increased with prolonged culture time in both proliferation and osteogenic induction media, with the most significant increase in MDV production observed in osteogenic differentiation medium. Joonho Suh et al. reported that during bone formation, osteoblasts promote the secretion of mitochondria and MDVs via CD38/cADPR signaling, and these secreted substances can enhance the differentiation and maturation of osteogenic progenitor cells [6]. This finding is consistent with the present study, where MDV secretion was also enhanced during osteogenic differentiation.

MDVs were initially identified in 2008 as vesicles that selectively transport cargo, including mtDNA [11]. Existing literature suggests that mtDNA is transported via MDVs as DAMPs, triggering an inflammatory response that plays a crucial role in inflammation activation [29]. Collins et al. provided the first evidence in 2004 that mtDNA acts as DAMPs, demonstrating this by inducing inflammatory arthritis in 6- to 8-week-old mice via mtDNA injection into their knees [31]. Following this, multiple studies have confirmed that mtDNA can trigger activation of various PRRS. Initially, mtDNA is recognized by cGAS, a double-stranded DNA (dsDNA) sensor, which produces 2’3’-cGAMP. The latter directly binds to the stimulator of interferon genes (STING). Subsequently, the activation of STING induces the phosphorylation of interferon regulatory factor 3 (IRF3). IRF3 then translocates to the nucleus and triggers the expression of multiple IFN-stimulated genes. Additionally, STING also activates the NF-κB signaling pathway via phosphorylation, resulting in increased expression of IL-6 and TNF-α [43–45]. The internalization of mtDNA via MDVs by BMSCs may involve mechanisms such as endocytosis, phagocytosis, and/or transmembrane diffusion. However, further studies are required to elucidate the mechanisms by which MDVs carrying mtDNA cargo activate intracellular PRRs signaling factors during osteogenic differentiation. Importantly, although previous studies have reported the upregulation of inflammatory factors during osteogenic differentiation, the involvement of MDVs in this process has not been previously explored. Our results suggest that this phenomenon may be associated with the regulatory effects of MDVs. This suggests that future research should pay attention to the immune response caused by MDVs and conduct safety assessments.

This study investigated the role of MDVs in the osteogenic differentiation of BMSCs. MDV-OMs isolated from the osteogenic differentiation medium effectively promoted the osteogenic differentiation of target cells. Microscopic observation revealed voids in the mineralized nodules, which were presumed to be gaps left by apoptosis. Thus, MDVs may induce apoptosis while promoting osteogenic differentiation, potentially having detrimental effects on target cells. Biomineralization is initiated by the penetration of amorphous calcium phosphate (ACP) precursors, produced in mitochondria, into the extracellular collagen matrix, where they are converted into intrafibrillar carbonated apatite [1, 2, 4]. Mitochondria, as the primary source of mineral particles in osteoblasts, deliver ACP via mitochondrial autophagy and participate in the biomineralization process mediated by the BMP/Smad signaling pathway [6]. MDVs are a mechanism of mitochondrial quality control and also serve as a complementary form of mitophagy [10, 42, 46]. This may be associated with the observed increase in MDVs numbers during osteogenic differentiation. We hypothesize that during osteogenic differentiation, BMSCs initiate biomineralization through mitophagy, which in turn leads to an increase in MDVs secretion. These MDVs may further promote osteogenic differentiation in surrounding cells via intercellular communication, forming a cascade effect that accelerates osteogenic differentiation. However, the specific mechanisms involved still need to be determined through the detection of mitochondrial dynamics, mitophagy flux, and the presence of ACP in MDVs. Recent studies have shown that MDVs are involved in the regulation of bone formation following bone defects under pathological conditions [6]. MDVs derived from adipocytes are able to activate antioxidant stress defense mechanisms in cardiomyocytes [23]. In diabetic foot ulcers, MDVs have been shown to exacerbate oxidative stress [39]. Under specific induction conditions, MDVs can influence cardiomyocyte apoptosis via the Bcl-2 pathway [16]. Moreover, MDVs may release cytochrome C (CytC) into the cytoplasm, where it binds to apoptotic protease-activating factor-1 (Apaf-1) to form the apoptosis, thereby inducing apoptosis [47]. Integrating the results from previous studies, we found that MDV-OM7 is a relatively safe extracellular vesicle that can promote osteogenic differentiation. However, MDV-OM14 and MDV-OM21 strongly activate inflammatory signals, which can easily induce cell apoptosis and thus may not be suitable as vesicular therapeutic agents for osteogenic differentiation. Additionally, the presence of dexamethasone in the OM environment may result in the residual osteogenic inducers in the extracted MDVs, which could mislead the interpretation of their osteogenic-promoting functions. Therefore, it is necessary to exclude the interference of this confounding factor through detection methods such as high-performance liquid chromatography.

MDVs originate from mitochondria, a crucial source of cellular metabolism and energy production [27], and in this study, the regulation of the metabolic profile of BMSCs by MDVs was also observed. The results indicated that MDVs promoted the aerobic metabolism of glucose in BMSCs while decreasing lactate production from glycolysis during the early stage of co-culture. This result contrasts with previously reported pathways of cellular energy metabolism during bone remodeling [5, 48]. During osteogenic differentiation, cells primarily generate energy through the glycolytic pathway [49]. However, related literature has also reported that enhancing aerobic metabolism in mitochondria can promote osteogenic differentiation [5], a finding consistent with the results of the present study. The mitochondrial membrane potential, formed by the ion concentration gradient across the inner mitochondrial membrane, serves as a sensitive indicator of mitochondrial function and can provide an indirect reflection of the activity of cellular energy metabolism. In terms of mitochondrial membrane potential [50], the MDV-OM14 and MDV-OM21 groups exhibited significantly higher levels than the OM group, indicating enhanced aerobic metabolism, whereas the OM group exhibited lower levels. In conjunction with the results from the earlier stages of this study [48, 50], this finding further suggests that during osteogenic differentiation within the MDV-OM14 and MDV-OM21 groups, BMSCs primarily generate ATP through aerobic glycolysis. This indicates that MDVs have the capacity to modulate metabolic pathways, thereby indirectly influencing energy metabolism.

Vesicles are well-known to be transferred between cells, playing roles in various cellular activities, including cell-cell communication, pathogenicity, cargo transport, and genetic material transfer [11, 27]. However, the functions of MDVs remain largely unexplored. To evaluate the functional significance of MDVs in osteogenic differentiation, a series of investigations into isolated MDVs were conducted in this study. It was found that the mitochondrial cargo effect of MDVs may act as a double-edged sword, promoting osteogenic differentiation on one hand while possibly triggering cellular oxidative stress on the other, leading to early apoptosis and enhanced inflammatory responses. Consequently, MDV-OM7 has the potential to serve as a vesicle therapeutic agent for osteogenic differentiation. Tomm20 is currently recognized as a canonical marker of MDVs. However, MDVs originating from the mitochondrial inner membrane may lack Tomm20 [40]. Therefore, this subtype may be overlooked during the isolation and extraction of MDVs. However, further investigation is required to fully evaluate their potential therapeutic applications. In cases where MDVs strongly activate immune responses, they could be considered for research as anti-tumor therapeutic agents. Compared to existing osteogenic materials and mitochondrial therapies [51, 52], MDVs, which are naturally secreted by cells, offer notable advantages in bone tissue engineering. MDVs facilitate intercellular communication, modulating the function of recipient cells. They demonstrate excellent biocompatibility, enabling efficient uptake by recipient cells and minimizing immunogenic responses. This characteristic is especially beneficial for long-term implantation, reducing immune rejection and inflammatory reactions commonly associated with traditional materials. Furthermore, MDVs can be optimized through physical stimulation, genetic editing, or modification with biomaterials, enhancing their functionality and therapeutic efficacy. For example, electroporation technology can load growth factors or gene fragments into MDVs, further enhancing their capacity to promote bone repair [53, 54]. Taken together, these features underscore the potential of MDVs as a promising therapeutic strategy in bone tissue engineering.

Conclusion

This study, through co-culturing BMSCs with MDVs, provides novel insights into the role of MDVs in promoting osteogenic differentiation and emphasizes the dual effects of their mitochondrial cargo: on one hand, it promotes osteogenic differentiation, while on the other, it induces cellular oxidative stress, resulting in early apoptosis and heightened inflammatory responses. However, this lack of in vivo validation restricts the immediate clinical applicability of these results. This gap underscores the necessity for further in vivo studies to comprehensively assess the osteogenic efficacy and biosafety profile of MDVs. In conclusion, these findings not only offer a theoretical foundation for utilizing MDVs as a potential therapeutic strategy for osteogenesis but also provide a fresh perspective on understanding the role of MDVs in osteogenic differentiation.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ACP

Amorphous Calcium Phosphate

Alix

ALG-2-interacting protein X

ANOVA

Analysis of Variance

ATP

Adenosine Triphosphate

BCA

Bicinchoninic Acid Assay

BMSCs

Bone Marrow Stromal Cells

Caspase-9

Cysteine Aspartate Specific Protease-9

CCCP

Carbonyl Cyanide m-Chlorophenyl Hydrazine

CD81

Cluster of Differentiation 81

CD38

Cluster of Differentiation 38

CD63

Cluster of Differentiation 63

CD9

Cluster of Differentiation 9

Col-1

Collagen Type I Alpha 1 Chain

CS

Citrate Synthase

cADPR

Cyclic ADP-Ribose

cGAMP

Cyclic GMP-AMP

cGAS

Cyclic GMP–AMP Synthase

CytC

Cytochrome C

DAMP

Damage-Associated Molecular Pattern

DMEM

Dulbecco’ s Modified Eagle Medium

EVs

Extracellular Vesicles

FBS

Fetal Bovine Serum

Hsp70

Heat-Shock Protein 70

IDH

Isocitrate Dehydrogenase

IL-6

Interleukin-6

IMM

Inner Membrane

IRF3

Interferon Regulatory Factor 3

LDH

Lactate Dehydrogenase

MDVs

Mitochondria-Derived Vesicles

MVB

Multivesicular Body Pathway

mtDNA

Mitochondrial DNA

NAD⁺

Nicotinamide Adenine Dinucleotide

NTA

Nanoparticle Tracking Analysis

Ocn

Osteocalcin

OD

Optical Density

OM

Osteogenic Induction Medium

OMM

Outer Membrane

Opn

Osteopontin

OXPHOS

Oxidative Phosphorylation

PBS

Phosphate-Buffered Saline

PDH

Pyruvate Dehydrogenase

PM

Proliferation Medium

PRRs

Pattern Recognition Receptors

qRT-PCR

Quantitative Real-Time Polymerase Chain Reaction

ROS

Reactive Oxygen Species

Runx2

Runt-Related Transcription Factor 2

SD

Standard Deviation

SOD2

Superoxide Dismutase 2

STING

Stimulator of Interferon Genes

TCA

Tricarboxylic Acid Cycle

TEM

Transmission Electron Microscope

TFAM

Transcription Factor A

Tim44

Translocase of the Inner Mitochondrial Membrane 44

TNF-α

Tumor Necrosis Factor-Alpha

TOMM20

Translocase of the Outer Mitochondrial Membrane 20

VDAC1

Voltage-Dependent Anion Channel 1

VDIMs

Vesicles Derived from the Inner Mitochondrial Membrane

WB

Western Blot

Author contributions

Chun Pan: Writing-review & editing, Writing-original draft, Visualization, Investigation, Data curation. Cheng Cheng: Writing-review & editing, Writing-original draft, Visualization, Investigation. Shu Zhong: Writing-review & editing, Writing-original draft, Visualization, Investigation. Shiyu Li: Conceptualization, Writing-review & editing, Writing-original draft, Supervision, Project administration, Resources, Funding acquisition. Wei Tan: Writing-review & editing, Writing-original draft, Supervision, Project administration, Resources, Funding acquisition. Yachao Yao: Writing-review & editing, Writing-original draft, Supervision, Project administration, Resources, Project administration.

Funding

This work was supported by the National Natural Science Foundation of China (82300018 to S.L.), Research Project of China Disabled Persons’ Federation - on assistive technology (2023CDPFAT-04 to W.T.), the Scientific research project of Traditional Chinese Medicine Bureau of Guangdong Province (20232007 to S.L.), the Fundamental Research Funds for the Central Universities (21624220 to S.L.), the Guangxi Key Laboratory of Birth Defects Research and Prevention (GXWCHZDKF-2022-21 to W.T.), the Guangdong Medical Research Foundation (A2023127 to W.T.) and the Guangdong Provincial Basic and Applied Basic Research Fund-Provincial Enterprise Joint Fund Project (SL2024A03J01479 to W.T.).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval and consent to participate

All primary cell studies were approved by the Animal Committee of Guangzhou Women and Children’s Medical Center (Ethics No. RSDW-2024-02006). All efforts were made to minimize animal suffering.

Consent for publication

All authors agree to be published.

Competing interests

The authors declare no competing interests.