Engineered bone-targeting Lycium barbarum L.-derived extracellular vesicle-like nanoparticles deliver miR167a-5p to regulate mitochondrial dynamics and metabolic reprogramming in alleviating osteoporosis

Abstract

With the exacerbation of global population aging, the development of novel therapeutic strategies for osteoporosis (OP) has emerged as a pressing worldwide challenge. As an emerging natural nanomedicine, medicinal plant-derived extracellular vesicle-like nanoparticles (EVLPs) have shown considerable potential in OP treatment due to their advantages, such as good bioavailability, high biosafety characteristics, and natural targeting. Herein, we isolated Lycium barbarum L.-derived EVLPs (LB-EVLPs) from fresh LB via ultracentrifugation combined with sucrose gradient centrifugation and further functionalized LB-EVLPs with the bone-targeting peptide SDSSD (BT-LB-EVLPs) to treat OP. This engineered LB-EVLPs exhibited selective targeting of bone tissue and were effectively internalized by bone marrow mesenchymal stem cells (BMSCs), with a distinct propensity for mitochondrial localization. In ovariectomized (OVX)-induced osteoporotic mice, BT-LB-EVLPs alleviated bone loss, improved bone microstructure, and enhanced bone strength. Proteomic profiling indicated that BT-LB-EVLPs reprogram mitochondrial metabolism by enhancing oxidative phosphorylation while suppressing excessive glycolytic flux, thereby exerting anti-osteoporotic effects. In vitro experiments demonstrated that BT-LB-EVLPs attenuated oxidative stress, promoted mitochondrial fusion, inhibited mitochondrial fission, and facilitated metabolic reprogramming in BMSCs, ultimately restoring mitochondrial function and enhancing osteogenic differentiation. Through lentiviral-mediated overexpression of SLC25A26 combined with miR167a-5p mimic/inhibitor interventions, we verified that miR167a-5p derived from BT-LB-EVLPs directly targets the mitochondrial transporter gene SLC25A26, thereby regulating mitochondrial dynamics, sustaining energy metabolism balance, and promoting osteoblastogenesis. Additionally, in vivo knockdown of miR167a-5p exacerbated bone loss and bone microstructural damage, and abolished the anti-osteoporosis effect of BT-LB-EVLPs. Collectively, these findings emphasized this engineered LB-EVLPs as a promising targeted nanotherapeutic approach for OP treatment.

Graphical abstract

Engineered bone-targeting Lycium barbarum L.-derived extracellular vesicle-like nanoparticles (BT-LB-EVLPs) deliver miR167a-5p to target SLC25A26, thereby regulating mitochondrial dynamics and metabolic reprogramming, ultimately alleviating bone loss, improving bone microstructure, and enhancing bone strength in ovariectomized (OVX)-induced osteoporotic mice.

1. Introduction

Osteoporosis (OP) is a prevalent chronic bone metabolic disorder characterized by diminished bone mass, heightened bone fragility, and a consequently elevated risk of fracture [1]. According to the latest data, well over 200 million individuals worldwide suffer from OP, with an estimated prevalence of 18.3% [2]. Osteoporotic fracture (OF) is one of the most severe and debilitating consequences of OP, which poses a major public health burden due to its high associated morbidity, mortality, and socioeconomic cost [3,4]. Currently, the first-line clinical therapeutic agents for OP, such as bisphosphonates, monoclonal antibody RANKL, parathyroid hormone-related protein analogues, have achieved certain curative effects [5]. However, these therapeutic interventions are inevitably accompanied by adverse effects, including increased risk of cardiovascular diseases, atypical fractures, and osteonecrosis [6,7]. Therefore, there is an urgent need to identify natural bioactive products to serve as an alternative to clinical drugs used for the management and treatment of OP.

With increasing age and the levels of estrogen changing, systemic antioxidant defense capacity diminishes [8], leading to unrestrained reactive oxygen species (ROS) accumulation that disrupts the homeostatic balance between bone formation and bone resorption [9]. The elevated ROS levels inhibits the osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs), impairs the mineralization capacity of osteoblasts, and ultimately compromises the synthesis and mineralization of bone matrix [10]. Notably, osteogenic differentiation of BMSCs is an energy-demanding process that requires substantial metabolic activity [11]. Mitochondria act as the energy supply and metabolic center of cells, regulating various cellular functions, including ATP production, cellular redox homeostasis, and apoptosis [12]. In an oxidative stress microenvironment, the mitochondria continuously undergo dynamic fusion or fission to modulate the connectivity of the mitochondrial network and adapt to cellular energy demands [13]. Mitochondrial fusion promotes the integration of electron transport chain components, facilitating oxidative phosphorylation (OXPHOS) to meet energy requirements and thereby enhance BMSCs osteogenic differentiation [14]. Consequently, the development of natural bioactive compounds that target the osteogenic capacity of BMSCs and restore mitochondrial function represents a promising therapy for OP.

With the rapid development of nanotechnology in biomedicine, extracellular vesicles (EVs) and particles with nano-sized membranous structures have attracted considerable interest due to their considerable therapeutic potential in disease management [15,16]. EVs are capable of carrying and delivering a variety of bioactive molecules, including nucleic acids, proteins, and lipids, which play crucial roles in intercellular communication, mitochondrial energy metabolism, and the maintenance of organismal homeostasis [17,18]. Historically, investigations into EVs have predominantly concentrated on those derived from mammals, which have been extensively employed in targeted drug delivery, tissue engineering construction, and therapeutic interventions for various diseases [19,20]. However, the clinical translation of mammalian-derived exosomes (MDEs) faces challenges such as immunogenic toxicity, limited yields, and high production costs owing to large-scale cell culture [21]. These limitations motivated the search for alternative and sustainable sources of EVs, particularly those from plants. In contrast to their mammalian counterparts, plant-derived extracellular vesicle-like nanoparticles (EVLPs) exhibit low immunogenicity, wide availability, absence of human or zoonotic pathogens, and distinct therapeutic properties [22,23]. Recent research has demonstrated that EVLPs derived from plants such as Rhizoma Drynariae, Pueraria lobata, and Morinda Officinalis are proficient in delivering bioactive compounds, facilitating cross-species communication, and showing significant value in the prevention and treatment of OP [[24], [25], [26]]. Therefore, plant-derived EVLPs might be a promising next-generation therapeutic tools in the fields of osteoporosis biotherapy and targeted drug delivery.

Lycium barbarum L. (LB) is a well-known traditional medicine with the dual property of “medicine and food homology”, is widely utilized in China and other Asian countries, and has garnered increasing global attention for its health-promoting properties [27,28]. It was recorded to strengthen bones and reinforce kidney essence in the “Ben Cao Gang Mu (Compendium of Materia Medica)”. Modern pharmacological researches have demonstrated that LB exhibits multiple bioactivities, including anti-inflammatory, antioxidant, osteoprotective, and immunoprotective [[29], [30], [31]]. LB is enriched in diverse phytochemical constituents, including polysaccharides, flavonoids, phenolic acids, and alkaloids [32]. Among its constituents, lycium barbarum polysaccharides (LBPs) have been identified as the primary bioactive components, which can effectively mitigate bone loss and improve bone microstructure in OVX mice and naturally aging mice [33,34]. Notably, unlike single components, LB-derived EVLPs (LB-EVLPs) not only preserve the inherent bioactivity of herbs but also leverage the advantages of a nanoscale delivery system to achieve multifunctional therapeutic effects, including transmembrane transport and targeted regulation.

In the present work, LB-EVLPs were isolated from fresh LB via ultracentrifugation combined with sucrose gradient centrifugation. We further functionalized LB-EVLPs with the bone-targeting peptide SDSSD (BT-LB-EVLPs), which endowed the BT-LB-EVLPs with enhanced bone tissue specificity for the precise treatment of OP. The bioactive compositions of LB-EVLPs were systematically analyzed using lipidomics and proteomics sequencing. Through integrated pharmacological and biological approaches, we demonstrated that BT-LB-EVLPs could target bone tissues to mitigate bone loss, improve bone microstructure, and enhance bone strength in OVX-induced osteoporotic mice. Mechanistically, BT-LB-EVLPs promoted mitochondrial fusion and inhibited fission, balanced OXPHOS and glycolysis, thereby improved mitochondrial function, and ultimately enhanced BMSC osteogenic differentiation. Further analysis identified miR167a-5p as a critical functional component delivered by BT-LB-EVLPs, which targets SLC25A26 to regulate mitochondrial dynamics and energy metabolism, thereby decelerating the OP process. Overall, our findings elucidated the active constituents and mechanism of BT-LB-EVLPs in OP treatment, providing a solid foundation for its clinical application. We believed that the combination of plant-derived EVLPs with engineered bone-targeting technology represents a promising and pioneering strategy for developing effective therapeutics against bone metabolic disorders.

2. Materials and methods

2.1. Preparation and characterization of LB-EVLPs and BT-LB-EVLPs

Freshly harvested LB fruits were procured from Yinchuan, Ningxia Province, China, and thoroughly rinsed with deionized water. The LB was ground using a high-speed stirrer to prepare LB juice. The resulting juice was filtered to remove large debris and subsequently centrifuged at 1000 × g for 10 min, 3000 × g for 20 min, and 10000 × g for 60 min to eliminate smaller particulates. The supernatant underwent ultracentrifugation at 150,000 × g for 120 min, after which the pellet was resuspended in PBS to make the crude LB-derived vesicle extract. To purify LB-EVLPs, the precipitate was resuspended and subjected to sucrose density gradient centrifugation (8%, 30%, 45%, and 60% sucrose solutions, centrifuged at 150000 × g for 120 min). Finally, the LB-EVLPs suspensions were partially filtered using a 0.22 μm microporous membrane and stored at −80 °C until further use.

To prepare the bone-targeting LB-EVLPs (BT-LB-EVLPs), we first obtained the DSPE-PEG2000-SDSSD conjugate via covalent conjugation of the bone-targeting SDSSD peptide to the distal terminal of 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000] (DSPE-PEG2000) (Xi'an Ruixi Biological Technology Co., Ltd., China). Bone-targeted LB-EVLPs (BT-LB-EVLPs) were fabricated via a well-established hydrophobic insertion strategy, which enables stable peptide modification while preserving the intrinsic structure and bioactivity of LB-EVLPs. In brief, purified LB-EVLPs were incubated with DSPE-PEG2000-SDSSD in sterile PBS, with gentle end-over-end rotation for 12 h in the dark. To completely remove unbound free DSPE-PEG2000-SDSSD conjugates, the mixture was subjected to ultracentrifugation at 100,000 ×g for 70 min, and the pellet was washed 3 times with sterile PBS. The final BT-LB-EVLPs was resuspended in sterile PBS and stored at −80 °C for subsequent experiments.

The morphology of LB-EVLPs and BT-LB-EVLPs were examined using transmission electron microscopy (TEM) (Thermo Fisher Scientific, USA). The particle counts and size distributions were analyzed using nanoparticle tracking analysis (NTA) on a NanoSight NS300 system (Malvern Instruments, UK). The zeta potential of LB-EVLPs and BT-LB-EVLPs were assessed using a Malvern Zetasizer 3000 (Malvern, UK). Protein concentrations of LB-EVLPs and BT-LB-EVLPs were quantified employing a BCA protein assay kit. For visualization of proteins, the proteins of LB-EVLPs were separated on 10% SDS-PAGE and stained with Coomassie Brilliant Blue for 2 h, and then rinsed with bleaching solution 2-3 times. The stained gels were visualized using a multifunctional gel imaging system (Hercules, CA, USA).

2.2. Stability evaluation of LB-EVLPs and BT-LB-EVLPs in vitro

The batch-to-batch consistency of LB-EVLPs was evaluated by quantifying the protein content of multiple independent preparations through a BCA assay. In addition, to assess storage stability, BT-LB-EVLPs were stored at temperatures of 4 °C, −20 °C, and −80 °C, and changes in zeta potential and particle size distribution were monitored by NTA on days 0, 7, and 14.

2.3. Lipidomic analysis of LB-EVLPs

Lipids were extracted from LB-EVLPs using the chloroform-methanol technique. To summarize, internal lipid standards were added to LB-EVLPs samples, which were then mixed with methanol and water, followed by liquid-liquid extraction with methyl tert-butyl ether (MTBE). After sonication and centrifugation, the upper organic phase was recovered, evaporated under nitrogen, and redissolved in a 90% isopropanol/acetone solution for subsequent lipidomic analysis. Lipid separation was performed using a UHPLC Nexera LC-30A system, followed by mass spectrometric detection on a Q-Exactive Plus instrument in both positive and negative ionization modes. Finally, LipidSearch was employed to identify peak identification, peak extraction, and lipid identification on the lipid molecules.

2.4. Proteomic analysis of LB-EVLPs

The protein samples extracted from LB-EVLPs were processed through the filter-aided sample preparation (FASP) technique. Following digestion, peptides were gathered via centrifugation, then desalted using C18 cartridges, and subsequently lyophilized. The resulting peptides were separated by a Vanquish Neo UHPLC system (Thermo Scientific). All mass spectrometry data were obtained in data-independent acquisition (DIA) mode and processed with DIA-NN software for protein identification and label-free quantification.

2.5. Animals experiments

8-week-old female C57BL/6J mice were obtained from the Qinglongshan Animal Breeding Farm in Nanjing, Jiangsu Province. All animal procedures were approved by the Animal Ethics Committee of Nanjing University of Traditional Chinese Medicine (Approval no.202408A050, Approval no. 202504A084). After one week of acclimatization, an osteoporotic model was established via bilateral ovariectomy. Briefly, mice were firstly anesthetized with isoflurane, and then bilateral abdominal incisions were made to expose the ovaries. After ligation of the ovarian vessels, the ovaries were excised. For sham-operated mice, only fat surrounding the ovaries was removed without ovarian excision. Each group of mice was administered antibiotics after the operation to prevent infection. Subsequently, the mice were randomly divided into the sham group (sham-operated + PBS), the OVX group (OVX + PBS), the LB-EVLPs group (OVX + LB-EVLPs, 1 mg/kg), the BT-LB-EVLPs group (OVX + BT-LB-EVLPs, 1 mg/kg), and the positive control group (OVX + zoledronate, 0.1 mg/kg). All treatments were administered via tail vein injection every 3 days for 8 weeks. To assess the therapeutic efficacy of miR167a-5p in mitigating bone loss, mice received either antagomir-NC or antagomiR-167a-5p (2 μL of antagomir diluted in 100 μL PBS) (Genepharma Shanghai, China) via the tail vein biweekly for a duration of 8 weeks.

2.6. Micro-CT analysis

The harvested femoral specimens were scanned using a micro-CT imaging system (SkyScan, Knotich, Belgium), and 3D reconstructions were performed. Bone microstructural parameters, including bone volume fraction (BV/TV), bone mineral density (BMD), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp), were calculated based on the 3D reconstructions.

2.7. Biomechanical testing

Three-point bending testing was employed to evaluate the bone strength of the mice femurs by a mechanical testing machine (CellScale, Univert S, Canada). Each femur was placed with its posterior surface facing downward and loaded at a constant displacement rate of 5 mm/min until fracture occurred. All mechanical testing was performed under ambient temperature conditions.

2.8. Histological analysis

The femoral tissues were isolated from mice and completely fixed in 4% paraformaldehyde for 24 h. Subsequently, all femoral specimens underwent standard procedures of decalcification, dehydration, paraffin embedding, and sectioning. To observe the morphology of the bone trabecular tissue, the sections were stained using H&E and Masson's trichrome. For immunohistochemical (IHC), the sections were blocked with 5% normal goat serum in PBS for 1 h and subsequently incubated overnight with primary antibodies: ALP (18507-1-AP, Proteintech, 1:100) and OCN (23418-1-AP, Proteintech, 1:200). After washing, the sections were incubated with secondary antibody for 1 h and the sections were counterstained with hematoxylin. Finally, stained sections were visualized and imaged under a light microscope.

2.9. Bio-distribution of LB-EVLPs and BT-LB-EVLPs in vivo

To evaluate the in vivo biodistribution of LB-EVLPs and BT-LB-EVLPs, both EVLP formulations were mixed with DiR dye (Yeasen, China) at 37 °C for 1 h, after which DiR-labelled LB-EVLPs (DiR-LB-EVLPs) and DiR-labelled BT-LB-EVLPs (DiR-BT-LB-EVLPs) were isolated by ultracentrifugation at 120,000 × g for 90 min at 4 °C and then resuspended in PBS. OVX mice were injected intravenously with PBS, DiR-LB-EVLPs, and DiR-BT-LB-EVLPs. At 6, 12, and 24 h after injection, the mice were sacrificed, and relevant organs were collected. The fluorescence intensity was detected using the IVIS in vivo imaging system (Perkin Elmer, USA).

2.10. In vivo safety evaluation

To evaluate the in vivo biosafety of the BT-LB-EVLPs, major organs (heart, spleen, lung, kidney, and liver) were harvested from mice at the final time point. Tissues were fixed in 4% paraformaldehyde solution, embedded in paraffin, sectioned, and subjected to H&E staining for histopathological analysis. In addition, serum was collected to quantify the levels of ALT, TP, BUN, and CREA.

2.11. Cell uptake and mitochondrial-targeting capacity assay of LB-EVLPs and BT-LB-EVLPs in vitro

To investigate the internalization of LB-EVLPs and BT-LB-EVLPs by BMSCs, LB-EVLPs and BT-LB-EVLPs were fluorescently labelled with PKH67. Briefly, the stock solution of PKH67 was sequentially diluted: it was first 10-fold in Diluent C, followed by a 25-fold dilution in PBS to prepare the working solution. The LB-EVLPs and BT-LB-EVLPs were incubated in this solution for 30 min, and free dye was eliminated by ultracentrifugation. Afterwards, the PKH67-labelled LB-EVLPs and BT-LB-EVLPs were incubated with BMSCs in darkness for 2 h, 4 h, and 8 h. After incubation, cells were analyzed using a confocal laser scanning microscope (CLSM) and a flow cytometer (Beckman Coulter, USA).

To further assess the mitochondrial-targeting capacity of BT-LB-EVLPs, the PKH67-labelled BT-LB-EVLPs were incubated with BMSCs in darkness for 2 h, 4 h, and 8 h. Next, cells were stained with Mito-tracker Red (Beyotime, C1049B) for 30 min. Fluorescence images were immediately photographed using a CLSM.

2.12. Cell culture and treatment

The BMSCs were cultured in α-MEM basal medium supplemented with 10% FBS and 1% penicillin-streptomycin, under conditions of 37 °C and 5% CO₂. To establish an oxidative stress model, BMSCs were seeded in 96-well plates and intervened with several concentrations of hydrogen peroxide (H₂O₂) for 12 h. After treatment, CCK-8 solution was added and incubated in the dark for 1 h. The OD value was measured at 450 nm using a microplate reader. Based on these viability assay results, an optimal H₂O₂ concentration was selected for subsequent experimental procedures. To evaluate the protective effects of LB-EVLPs and BT-LB-EVLPs, the H₂O₂-injured BMSCs were co-cultured with LB-EVLPs and BT-LB-EVLPs at different concentrations for 24 h. Cell viability was then evaluated under the same experimental conditions using the CCK-8 assay.

2.13. Cell viability, proliferation, and osteogenic differentiation of BMSCs

To evaluate the cell viability of BMSCs, a live/dead cell staining assay was utilized. For this assessment, BMSCs were incubated for 24 h. The samples were stained with calcein-AM solutions for 15 min and EthD-1 solutions for 3 min and then observed under a fluorescence microscope.

The EDU kit (Beyotime, C0075S) was utilized to evaluate the number of newly proliferating cells. The cells were co-cultured with LB-EVLPs and BT-LB-EVLPs for 24 h, after which they were fixed, permeabilized, and fluorescently labelled. DAPI was employed to stain the cell nuclei, and fluorescent images were captured using a fluorescence microscope.

BMSCs were placed in a 6-well plate and cultured in osteogenic induction medium for 7-21 days, with the medium changed every 48 h. For the alkaline phosphatase (ALP) staining, at day 7 of induction, the cells were washed with PBS, fixed, and subsequently incubated with BCIP/NBT working solution (Beyotime, C3206) until color developed. For alizarin red S (ARS) staining, at day 21 of induction, the cells were fixed with 4% paraformaldehyde for 15 min, followed by washing with distilled water, and then stained with ARS solution (Beyotime, C0148) for 30 min. Following the removal of excess dye, calcium nodules were observed and imaged under an optical microscope.

2.14. Assessment of intracellular ROS

Intracellular ROS were assessed by a fluorescent probe-based ROS Assay Kit (Abcam, abl13851). The BMSCs were subjected to different interventions and then incubated with 10 μM DCFH-DA for 30 min. And fluorescent images were captured using a fluorescence microscope.

2.15. Assessment of mitochondrial function

The mitochondrial ROS were measured using the MitoSOX™ Red (Invitrogen, M36009). The cells were incubated with 5 μM MitoSOX reagent for 30 min under light-protected conditions, and fluorescence images were captured by a fluorescence microscope.

The mitochondrial membrane potential of BMSCs was evaluated with a JC-1 assay kit (Beyotime, C2006). Following washing with PBS, the cells were treated with 5 μM JC-1 dye for 30 min, and evaluations were conducted through flow cytometry and fluorescence microscopy.

To observe the mitochondrial morphology, cells were incubated with MitoTracker Red (Beyotime, C1049B) and subsequently imaged using a confocal microscope. For TEM, BMSCs were fixed with 2.5% glutaraldehyde. Samples were then dehydrated through a graded ethanol series, embedded in epoxy resin, and sectioned into ultrathin slices. Sections were stained with uranyl acetate and lead citrate, and observed under a TEM microscope (Hitachi, HT7800).

Intracellular ATP content was quantified by an ATP assay kit (Abcam, ab83355). After treatment, BMSCs were lysed with ice-cold extraction buffer and centrifuged. The resulting supernatant was transferred to a 96-well plate, mixed with the ATP detection working solution, and luminescence was measured with a microplate reader.

Mitochondrial calcium levels were detected by a Rhod-2 AM probe (Beyotime, S1062S). The cells were treated with 5 μM Rhod-2 Red in serum-free DMEM for 20 min under light-protected conditions, and fluorescence intensity was assessed by a fluorescence microscope.

2.16. Seahorse

Oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) were assessed using an XF‐96 Extracellular Flux Analyser (Agilent Tecnologies, USA). BMSCs (5 × 10⁴ cells/well) were seeded in Seahorse XF96 Cell Culture Microplates. For the measurement of OCR, 1.5 μM oligomycin, 1.5 μM FCCP, and 1 μM rotenone/antimycin A were utilized. To examine ECAR, 10 mM glucose, 1 μM oligomycin, and 50 mM 2-deoxy-glucose (2-DG) were added to the solution.

2.17. Assessment of NAD⁺/NADH ratio and lactic acid content

The NAD⁺/NADH ratio and lactic acid content in BMSCs were assessed using a NAD⁺/NADH quantification kit (Abcam, ab65348) and a lactate dehydrogenase assay kit (Beyotime, P0393S), respectively, in accordance with the recommended protocols.

2.18. MiRNA profiling of BT-LB-EVLPs

The extraction and purification of total RNA from BT-LB-EVLPs were conducted according to the manufacturer's procedure. The RNA amount and purity were assessed on a NanoDrop ND-1000 spectrophotometer (NanoDrop, USA). The RNA fragment integrity was assessed with a Bioanalyzer 2100 system (Agilent, USA). Small RNA fractions were isolated by PAGE electrophoresis for library preparation and sequencing. Sequencing libraries were constructed and sequenced on an Illumina NovaSeq 6000 platform. Raw reads were processed with ACGT101-miR (v4.2) to eliminate adapters, low-quality sequences, and non-coding RNAs and repeats. Afterwards, unique sequences ranging from 18 to 25 nucleotides in length were then mapped to specific precursor sequences in miRBase 22.1, allowing for the identification of both known and novel 3p- and 5p-derived miRNAs.

2.19. MiR167a-5p target gene prediction and dual-luciferase reporter assay

To identify the direct targets of miR167a-5p, we initially performed in silico prediction using the miRanda, PITA, and RNAhybrid databases. Among the candidate targets, SLC25A26 was selected for further experimental validation. Subsequently, a luciferase reporter plasmid containing the 3′-UTR sequence of SLC25A26 was constructed (Biosune, Shanghai, China). To evaluate the binding between miR167a-5p and SLC25A26, a dual luciferase reporter assay was performed. BMSCs were seeded in a 96-well plate and co-transfected with reporter constructs along with either a miR167a-5p mimic or a negative control. After a 48 h, luciferase activity was measured by the Dual-Luciferase reporter system (Promega, Madison, USA).

2.20. Proteomics analysis

To explore the mechanism by which BT-LB-EVLPs alleviate OP, femoral tissues were collected from the sham group, OVX group, and OVX + BT-LB-EVLPs group. The frozen samples were maintained on ice and homogenized in protein lysis buffer containing protease inhibitors using a high-throughput tissue homogenizer. We then centrifuged the homogenates and gathered the supernatants for protein quantification with a BCA assay kit. The quantified proteins were then reduced and digested with dithiothreitol, iodoacetamide, and trypsin, respectively. After digestion, the peptides were vacuum-dried, reconstituted in 0.1% TFA, and desalted via HLB solid-phase extraction cartridges. Following peptide quantification, analyses were performed on the peptides using a VanquishNeo system coupled with an Orbitrap Astral mass spectrometer (Thermo, USA).

2.21. Western blot (WB)

The proteins of bone tissues and cells were lysed in RIPA buffer supplemented with 1% (v/v) protease and phosphatase inhibitors, and protein concentration was quantified using a BCA Protein Assay Kit (Beyotime Biotechnology, China). Aliquots containing 60 μg protein per lane were resolved by 8-15% SDS-PAGE and electrophoretically transferred to PVDF membranes. Membranes were blocked with 5% non-fat milk in TBST for 1 h and subsequently incubated overnight at 4 °C with primary antibodies: RUNX2 (#12556, CST, 1:1000), ALP (#DF6225, Affinity, 1:1000), OPN (#88742, CST, 1:1000), ATP5A1 (14676-1-AP, Proteintech, 1:5000), COX5A (#38563, CST, 1:1000), NDUFS1 (#70264, CST, 1:1000), SDHA (#5839, CST, 1:1000), UQCRB (10756-1-AP, Proteintech, 1:1000), SLC25A26 (PA5-55452, Invitrogen, 1:1000), OPA1 (27733-1-AP, Proteintech, 1:5000), MFN1 (13798-1-AP, Proteintech, 1:5000), MFN2 (12186-1-AP, Proteintech, 1:5000), DRP1 (#8570, CST, 1:1000), FIS1 (10956-1-AP, Proteintech, 1:5000), PGC-1α (66369-1-Ig, Proteintech, 1:5000), TFAM (22586-1-AP, Proteintech, 1:5000), NRF1 (12482-1-AP, Proteintech, 1:3000), and β-actin (20536-1-AP, Proteintech, 1:4000). Afterwards, the membranes were incubated with secondary antibodies. Protein bands were visualized using enhanced chemiluminescence (ECL; Bio-Rad, USA).

2.22. Real-time quantitative PCR (RT-qPCR)

Total RNA was isolated from bone tissues and intervened cells using a Total RNA Extraction Kit. Subsequently, RNA was reverse-transcribed into cDNA using a PrimeScript RT reagent kit. The resulting cDNA was amplified using a SYBR Premix Ex Taq kit. RT-qPCR was conducted using the SYBR® Premix Ex Taq™ (Takara, Japan). The primer sequences (forward and reverse) used are listed in Table S1.

2.23. Immunofluorescence analysis

Tissue sections and cell smears were fixed with 4% PFA, permeabilized with 1% Triton X-100, and blocked with 3% BSA. Following these steps, the sections were incubated overnight with primary antibodies: ATP5A1 (14676-1-AP, Proteintech, 1:800), RUNX2 (#12556, CST, 1:800), ALP (#DF6225, Affinity, 1:200), OPN (#88742, CST, 1:800), and OPA1 (27733-1-AP, Proteintech, 1:5000) overnight. Afterwards, the sections were incubated with species-appropriate secondary antibodies. DAPI was used to stain the cell nuclei, and the sections were observed and photographed under a fluorescence microscope.

2.24. Overexpression plasmids and miRNA transfection

The SLC25A26 overexpression plasmid (OE-SLC25A26) and its corresponding negative control plasmid (OE-NC) were synthesized by Heyuan Biology Co. Ltd. (Shanghai, China). For plasmid transfection, BMSCs were seeded in 6-well plates and transfected at approximately 60% confluence using Lipofectamine 3000 (Invitrogen, USA) in accordance with the manufacturers’ protocol. The sequence of the SLC25A26 overexpression plasmid was as follows: Table S2.

The miR167a-5p mimic, mimic negative control (NC-mimic), miR167a-5p inhibitor, and inhibitor negative control (NC-inhibitor) were synthesized by GenePharma (Shanghai, China). For miRNA transfection, BMSCs were seeded in 6-well plates until they reached 80% confluence, after which transfections were conducted using a commercial transfection reagent lipofectamine 2000 (Invitrogen, USA), following the manufactures’ instructions.

2.25. Statistical analysis

The statistical analysis was performed using GraphPad Prism 10.4 software. Student's t-test and one-way ANOVA, followed by Tukey's post hoc test, were employed to establish significance levels.

3. Results

3.1. Preparation and characterization of LB-EVLPs and BT-LB-EVLPs

LB-EVLPs were extracted from fresh LB using the sequential ultracentrifugation and sucrose gradient centrifugation method, and LB-EVLPs are mainly distributed and collected in a 30-45% sucrose density gradient. Furthermore, the bone-targeting functionalization of the LB-EVLPs was achieved through covalent surface conjugation of the bone-targeting peptide (SDSSD) (Fig. 1A). TEM revealed that LB-EVLPs exhibited a characteristic cup-shaped morphology with an intact membrane structure (Fig. 1B). NTA indicated that the LB-EVLPs had an average diameter of 144.8 nm and a concentration of 1.03 × 10¹² particles/mL (Fig. 1C and D). And different batches of LB-EVLPs showed consistent protein concentrations (Fig. S1A). SDS-PAGE analysis indicated that the proteins in LB-EVLPs ranged in molecular weight from 10 to 180 kDa (Fig. 1E). Following this, the comprehensive composition of the purified LB-EVLPs was examined. A nontargeted lipidomic analysis was conducted to profile the lipid composition of LB-EVLPs (Table S3). The results indicated that the predominant lipid constituents in LB-EVLPs included triglyceride (TG, 29.69%), diacylglycerol (DG, 9.27%), hexosylceramide (Hex1Cer, 9.27%), monogalactosyldiacylglycerol (MGDG, 9.06%), and ceramide (Cer, 8.33%). Proteomic analysis identified 5172 proteins in LB-EVLPs, with 97.25% falling within the 10-180 kDa range (Fig. 1F–Table S4). Subcellular localization analysis demonstrated that the top five protein categories were derived from the membrane (48.85%), cytoplasm (29.73%), ribosome (11.08%), endoplasmic reticulum (3.42%), and golgi apparatus (2.77%) (Fig. 1G). GO enrichment revealed that LB-EVLP proteins were primarily involved in biological processes, including small molecule metabolic, amino acid metabolic, and organic acid metabolic. For cellular components, the enriched terms focused on the cytoplasm, membrane coat, and coated membrane, while molecular functions were dominated by catalytic activity, purine ribonucleoside triphosphate binding, and pyrophosphatase activity (Fig. 1H). KEGG enrichment analysis further highlighted significant enrichment in carbon metabolism, biosynthesis of amino acids, metabolic pathways, and glycolysis/gluconeogenesis (Fig. 1I).

Fig. 1.

Preparation and characterization of LB-EVLPs and BT-LB-EVLPs. (A) An preparation flowchart of BT-LB-EVLPs. (B) TEM image of LB-EVLPs (Scale bars, 100 nm). (C–D) Particle size and distribution of LB-EVLPs. (E) LB-EVLPs proteins were extracted and separated using SDS-PAGE and then stained with Coomassie brilliant blue dye. (F) The lipid compositions of LB-EVLPs were detected by lipidomic analysis. (G) The protein compositions of LB-EVLPs were detected by proteomic profiling. (H–I) GO enrichment and KEGG analysis results based on LB-EVLPs proteomic. (J) TEM image of BT-LB-EVLPs (Scale bars, 100 nm). (K) Zeta potential of LB-EVLPs and BT-LB-EVLPs. (L) Size distribution of LB-EVLPs and BT-LB-EVLPs. Data are presented as mean ± SD (n = 3).

To enhance the bone-targeting ability of LB-EVLPs, we functionalized the membrane surface of LB-EVLPs with the SDSSD via hydrophobic insertion. TEM imaging revealed that the BT-LB-EVLPs maintained a well-defined, uniform spherical nanostructure with intact vesicular membranes (Fig. 1J). And SDS-PAGE analysis indicated that the protein distributions of LB-EVLPs and BT-LB-EVLPs were highly comparable, indicating that the key protein cargo components were largely retained after functionalization engineering [35,36] (Fig. S2). Noticeably, the surface zeta potential shifted from −27.17 ± 0.33 mV to −30.71 ± 1.68 mV (Fig. 1K), a change attributable to the conjugation of negatively charged SDSSD peptides, which also suggests enhanced colloidal stability and dispersibility. Concurrently, DLS analysis revealed that SDSSD functionalization resulted in a moderate increase in the average hydrodynamic diameter, from 124.50 nm for unmodified LB-EVLPs to 163.37 nm for BT-LB-EVLPs (Fig. 1L). Furthermore, BT-LB-EVLPs stored at varying temperatures (4 °C, −20 °C, and −80 °C) for 2 weeks retained stable surface charge, protein content, and nanostructural integrity (Fig. S1B–D). Collectively, these results confirmed the successful engineering of BT-LB-EVLPs that exhibit stable physicochemical properties and contain multiple active ingredients for potential biological application.

3.2. Biodistribution of LB-EVLPs and BT-LB-EVLPs in vivo and cellular uptake

To track the in vivo distribution of LB-EVLPs and BT-LB-EVLPs, both EVLP formulations were labelled with the fluorescent dye DiR. OVX mice were intravenously administered with PBS, DiR-labelled LB-EVLPs, or DiR-labelled BT-LB-EVLPs, respectively. At specific intervals (6 h, 12 h, and 24 h) post-injection, major organs (heart, liver, spleen, lung, kidneys, and femurs) were harvested, and the fluorescence signals of the collected tissues were visualized and analyzed using the IVIS in vivo imaging systems (Fig. 2A). As shown in Fig. 2B, the fluorescent signal mainly accumulated in the liver, spleen, and lung in the DiR- LB-EVLPs group, with only a faint fluorescence signal detectable in the femurs at 12 h post-administration, which was completely absent and undetectable by 24 h. Notably, we observed that DiR-BT-LB-EVLPs exhibited a prominent bone-targeting capacity, with preferential accumulation in the femurs. The fluorescence intensity in the femurs reached its maximum 12 h post-administration and gradually declined thereafter, and a substantial quantity of DiR-BT-LB-EVLPs remained detectable in the femurs at 24 h. These results demonstrated the inherent and sustained bone-targeting capacity of BT-LB-EVLPs.

Fig. 2.

In vitro cellular uptake behavior and in vivo biodistribution. (A) A schematic diagram of an in vivo imaging design. (B) Living fluorescence imaging of various tissues of OVX mice after intravenous injection of PBS, DiR-LB-EVLPs, and DiR-BT-LB-EVLPs. (C) A schematic diagram of the BT-LB-EVLP uptake by BMSCs. (D) Representative fluorescence images for cellular uptake of LB-EVLPs (Scale bars, 50 μm). (E) Representative fluorescence images for cellular uptake of BT-LB-EVLPs (Scale bars, 50 μm). (F–G) Representative fluorescence images and flow cytometry images of BT-LB-EVLPs internalization after pretreatment with various inhibitors (Scale bars, 5 μm). (H–K) Representative images of the colocalization of PHK67-labelled BT-LB-EVLPs (green fluorescence) and mitochondria (red fluorescence) at different times (2 h, 4 h, 8 h) in BMSCs (Scale bars, 50 μm). (L) Schematic diagram of BT-LB-EVLPs targeting mitochondria. Data are presented as mean ± SD (n = 3).

To investigate the osteogenic potential of LB-EVLPs and BT-LB-EVLPs, we initially examined the uptake of LB-EVLPs and BT-LB-EVLPs by BMSCs (Fig. 2C). Following co-culture with PHK67-labelled LB-EVLPs and PHK67-labelled BT-LB-EVLPs, it can be clearly observed that LB-EVLPs and BT-LB-EVLPs were internalized by BMSCs in a time-dependent manner, with fluorescence primarily localized into the cytoplasm (Fig. 2D and E). Subsequently, to elucidate the mechanism of internalization of BT-LB-EVLPs, we employed various inhibitors to block the endocytosis pathways in BMSCs. Fluorescence and flow cytometry results indicated a marked decrease in the internalization efficiency of BT-LB-EVLPs upon exposure to chlorpromazine (clathrin endocytosis inhibitor). In contrast, the presence of amiloride (giant pinocytosis specific inhibitor), cytochalasin D (actin-dependent endocytosis inhibitor), and β-cyclodextrin (caveolin endocytosis inhibitor) did not affect the internalization of BT-LB-EVLPs [9,26]. Hence, the cellular uptake of BT-LB-EVLPs primarily depended on the clathrin endocytosis pathway (Fig. 2F and G). In addition, we further assessed the mitochondrial-targeting potential of BT-LB-EVLPs. Colocalization analysis revealed significant colocalization between the green fluorescence of BT-LB-EVLPs and the red fluorescence of Mito-Tracker, as evidenced by Pearson's correlation coefficients (PCC), which increased over time from 0.41 at 2 h to 0.64 at 4 h and 0.78 at 8 h (Fig. 2H–K), indicating that BT-LB-EVLPs possess inherent mitochondrial-targeting activity (Fig. 2L).

3.3. BT-LB-EVLPs ameliorate bone loss and improve bone strength in OVX-induced osteoporotic mice

To investigate the efficacy of BT-LB-EVLPs in ameliorating OP induced by estrogen deficiency, an OVX mouse model was employed. Following 1 week post-OVX, BT-LB-EVLPs were administered intravenously every 3 days for 8 weeks. Zoledronate, a widely used clinical medication for osteoporosis, was used as a positive control (Fig. 3A). Micro-CT and 3D reconstruction analyses indicated significant trabecular bone loss and fractured trabeculae in the OVX group, confirming the effective establishment of the OVX-induced osteoporosis model. Quantitative analysis further demonstrated that the LB-EVLPs, BT-LB-EVLPs, and zoledronate treatment groups all exhibited significantly increased BMD, BV/TV, Tb.Th and Tb.N, while reducing Tb.Sp compared with the OVX group (Fig. 3B). These results demonstrated that BT-LB-EVLPs mitigate osteoporotic bone loss and alleviate bone microstructural deterioration in OVX mice (Fig. 3C–G). Bone biomechanical testing is a critical aspect of bone parameters for evaluating OP. The results of the three-point bending test demonstrated that the maximum load and stiffness values of femurs in both the LB-EVLPs and BT-LB-EVLPs groups were significantly elevated, whereas the maximum deflection was markedly reduced compared with those in the OVX group (Fig. 3H), indicating that BT-LB-EVLPs improved bone mechanical properties and resistance to fracture in osteoporotic mice. Furthermore, femur samples were stained with H&E and Masson to visualize newly formed trabecular bone and fibrous tissue. As shown in Fig. 3I, the OVX group exhibited the pronounced adipocytic vacuolization within bone marrow cavities and sparse trabecular distribution. In contrast, both the LB-EVLPs and BT-LB-EVLPs groups demonstrated thicker and more continuous trabeculae along with reduced bone marrow cavities. Furthermore, Masson staining indicated a prominent increase in osteoid matrix deposition in both the LB-EVLPs and BT-LB-EVLPs groups, suggesting improved bone quality and metabolic activity. These biomechanical and histological results demonstrated that BT-LB-EVLPs potently reinforce bone strength and quality, thereby conferring robust protection against OVX-induced skeletal fragility.

Fig. 3.

BT-LB-EVLPs ameliorate bone loss and improve bone strength in osteoporotic mice. (A) The experimental design and corresponding flowchart. (B) Representative micro-CT images of bone scan (n = 6). (C–G) Quantitative analysis of key bone microstructure parameters of femurs (n = 6). (H) Biomechanical properties analysis of femurs (n = 3). (I) H&E and Masson staining images of femurs (Scale bars, 50 μm and 200 μm) (n = 3). (J–L) Immunohistochemistry and semi-quantitative analysis of ALP and OCN expressions of femurs (n = 3). (M–N) WB images and quantitative analysis of RUNX2, ALP, and OPN (n = 3). (O) RT-qPCR of RUNX2, ALP, and OPN mRNA levels. Data are presented as mean ± SD. Compared with the Sham group, ^#P < 0.05, ^##P < 0.01. Compared with the OVX group, ∗P < 0.05, ∗∗P < 0.01.

Enhancing osteogenic activity and restoring impaired bone formation capacity are pivotal strategies for ameliorating OP and improving bone mechanical integrity [37]. To further investigate osteogenic regulatory effects of BT-LB-EVLPs, we conducted immunohistochemical staining, WB, and RT-qPCR to detect the expressions of key osteogenic markers. As presented in Fig. 3J–L, immunohistochemical staining revealed markedly reduced ALP and OCN expression in the OVX group, whereas their expression was significantly upregulated in BT-LB-EVLP-treated groups. Furthermore, WB analysis demonstrated that the levels of osteogenic-related proteins, including RUNX2, ALP, and OPN, were considerably lower in the OVX group compared to the sham group. However, BT-LB-EVLPs intervention substantially increased the expression of these pro-osteogenic proteins (Fig. 3M and N). Consistent with these findings, RT-qPCR revealed that the mRNA levels of RUNX2, ALP, and OPN were diminished in the OVX group, an effect that was significantly reversed following BT-LB-EVLPs intervention (Fig. 3O). In addition, TRAP staining revealed a significantly increased osteoclast surface/bone surface (Oc.S/BS) in the OVX group compared with the sham group, while BT-LB-EVLPs treatment markedly reversed this pathological elevation, indicating effective suppression of bone resorptive activity (Fig. S3A–B). Collectively, these data suggested that BT-LB-EVLPs treatment ameliorates bone loss, improves bone structural integrity, enhances bone mechanical strength and regulates bone homeostasis in osteoporotic mice.

Subsequently, we evaluated the biosafety of BT-LB-EVLPs administration. H&E staining indicated no obvious pathological abnormalities in major organs (heart, liver, spleen, lung, and kidney) across all experimental groups after 8 weeks of BT-LB-EVLPs intervention (Fig. S4A). Besides, serum biochemical assays revealed that the levels of hepatic and renal function indicators in different groups, including BUN, TP, ALT, and CREA, were all within the normal reference range. (Fig. S4B–E). These findings suggested that BT-LB-EVLPs exhibited excellent biocompatibility and biosafety profiles, highlighting their promising biomedical applications in OP treatment.

3.4. BT-LB-EVLPs alleviate osteoporosis by modulating metabolic reprogramming

To clarify the mechanism of BT-LB-EVLPs in the treatment of OP, proteomic sequencing was performed on femur tissues of mice. Principal component analysis (PCA) demonstrated distinct clustering among the three experimental groups, sham, OVX, and BT-LB-EVLPs (OVX + BT-LB-EVLPs), based on their protein expression profiles, indicating a relatively high sample quality (Fig. S5A). Subsequently, Venn analysis identified 1098 differentially expressed proteins between the sham versus OVX and OVX versus BT-LB-EVLPs comparisons (|log₂ FC|>1.5 and adj p-value<0.05) (Fig. 4A). Sequencing analysis was performed to obtain differentially expressed proteins across the sham, OVX, and BT-LB-EVLPs groups (Fig. S5B). Specifically, the volcano plot illustrated that compared to the sham group, the OVX group had 907 upregulated and 548 downregulated proteins. In contrast, the BT-LB-EVLPs group exhibited 1235 upregulated and 1627 downregulated proteins (Fig. 4B and C). Furthermore, KEGG pathway analysis pinpointed that the proteins significantly altered following BT-LB-EVLPs treatment were involved in OXPHOS and reactive oxygen species (Fig. 4D). And GO enrichment analysis indicated that compared with the OVX group, the mitochondria, mitochondrial membrane, and mitochondrial inner membrane under BT-LB-EVLPs treatment were the most enriched cellular components (Fig. 4E). Subsequently, the heatmap illustrated that BT-LB-EVLPs markedly reduced the levels of specific proteins associated with the OXPHOS pathway and mitochondrial function (Fig. 4F). Additionally, GESA analysis confirmed that there was positive enrichment of OXPHOS and tricarboxylic acid cycle, while there was negative enrichment of reactive oxygen species and glycolysis/gluconeogenesis in the BT-LB-EVLPs group (Fig. 4G). In summary, our data indicated that BT-LB-EVLPs might reprogram mitochondrial metabolism by enhancing mitochondrial respiration while suppressing excessive glycolytic flux, thereby exerting anti-osteoporotic effects.

Fig. 4.

BT-LB-EVLPs alleviate osteoporosis by modulating metabolic reprogramming. (A) The Venn diagram illustrates the differential expression of 1098 proteins between the sham versus OVX and OVX versus BT-LB-EVLPs comparisons. (B–C) The volcano plot highlights the substantially upregulated (red) and downregulated (blue) proteins in the OVX group compared to the sham group and the BT-LB-EVLPs group compared to the OVX group. (D) KEGG enrichment of proteomic sequencing. (E) GO enrichment of proteomic sequencing. (F) A correlational heatmap depicts the expression of markers associated with OXPHOS and mitochondria. (G) The GSEA of the OXPHOS, tricarboxylic acid cycle, reactive oxygen species, and glycolysis/gluconeogenesis. (H–I) Immunofluorescence images and semiquantitative analysis of ATP5A1 in mice distal femurs (Scale bars, 50 μm) (n = 3). (J–K) WB images and quantitative analysis of NDUFS1, SDHA, UQCRB COX5A, and ATP5A1 (n = 3). Data are presented as mean ± SD. Compared with the Sham group, ^#P < 0.05, ^##P < 0.01. Compared with the OVX group, ∗P < 0.05, ∗∗P < 0.01.

As is well known, mitochondria are crucial for energy metabolism within the cell, which generates ATP for cellular energy through OXPHOS, and play central roles in energy metabolism homeostasis [38]. To validate these proteomic findings, we conducted immunofluorescence to assess the expression of ATP synthase f1 subunit alpha (ATP5A1), which is a key part of the mitochondrial respiratory chain complexes. The results confirmed that BT-LB-EVLPs significantly increases the expression of ATP5A1 in OVX mice (Fig. 4H and I). Further WB results indicated that BT-LB-EVLPs upregulated the protein expression of the five complexes (complexes I-V) in the OXPHOS system: NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1 compared to the OVX group (Fig. 4J and K). In summary, these results suggested that BT-LB-EVLPs exert anti-osteoporotic effects by modulating mitochondrial OXPHOS to restore cellular energy metabolism, thereby mitigating OVX-induced metabolic dysfunction in bone tissue.

3.5. BT-LB-EVLPs promote the proliferation and osteogenic differentiation of BMSCs in vitro

Next, we assess the osteogenic differentiation capacity of BT-LB-EVLPs within the osteoporotic microenvironment. Based on previous studies [39], we utilized H₂O₂ to trigger oxidative stress, thereby simulating the detrimental microenvironment of OP. Initially, BMSCs were exposed to a gradient of H₂O₂ concentrations. The CCK-8 assay indicated a concentration-dependent reduction in cell viability, with the 200 μM H₂O₂ treatment reducing viability to approximately 50%, this concentration was therefore selected for the subsequent experiments (Fig. S6A). Following this, we assessed the impact of LB-EVLPs and BT-LB-EVLPs on H₂O₂-impaired BMSCs viability (Fig. 5A). Both treatments restored cell viability in a dose-dependent manner, with the most prominent protective effect observed at 20 μg/mL (Fig. S6B–C). Therefore, we employed these concentration for further experiments. Furthermore, the live/dead cell staining indicated that the LB-EVLPs and BT-LB-EVLPs groups exhibited prominent green fluorescence (indicative of viable cells), with only negligible red fluorescence signals (corresponding to dead cells) detected (Fig. S7A–B), which suggested that LB-EVLPs and BT-LB-EVLPs possess good biocompatibility characteristics.

Fig. 5.

BT-LB-EVLPs promote the proliferation and osteogenic differentiation of BMSCs in vitro. (A) Schematic of BT-LB-EVLPs treatment in H₂O₂-stimulated BMSCs. (B–C) The proliferation of BMSCs was analyzed using EDU (Scale bars, 200 μm). (D) ALP and ARS staining of BMSCs in different groups after osteo-induction (Scale bars, 200 μm). (E) Quantitative analysis of ALP activity. (F) Semiquantitative statistics of positively stained areas in ARS staining. (G–J) WB images and quantitative analysis of RUNX2, ALP, and OPN in BMSCs. (K–N) Representative confocal images and semiquantitative analysis of RUNX2, ALP, and OPN in BMSCs (Scale bars, 50 μm). (O–Q) RT-qPCR of RUNX2, ALP, and OPN mRNA levels. Data are presented as mean ± SD (n = 3). Compared with the control group, ^#P < 0.05, ^##P < 0.01. Compared with the H₂O₂ group, ∗P < 0.05, ∗∗P < 0.01.

The functional modulatory effects of BT-LB-EVLPs on BMSCs were evaluated through proliferation, osteogenic differentiation, and level of osteoblast-related genes and proteins. EDU assays indicated that the oxidative stress microenvironment suppressed the proliferative viability of BMSCs, whereas BT-LB-EVLPs noticeably ameliorated this inhibitory effect (Fig. 5B and C). To assess the osteogenic potential of BT-LB-EVLPs, ALP and ARS staining were performed on days 7 and 21 of osteogenic induction, respectively. As shown in Fig. 5D–F, H₂O₂ + BT-LB-EVLPs group exhibited higher levels of ALP activity and calcium deposition area compared to the H₂O₂ group, indicating BT-LB-EVLPs possess increased mineral synthesis and generation. Furthermore, the protein and mRNA expressions of osteogenesis-related markers such as RUNX2, ALP, and OPN were thoroughly evaluated by WB, RT-qPCR, and immunofluorescence. Both WB and immunofluorescence consistently revealed that H₂O₂ induction leads to a reduction of osteogenic differentiation, specifically indicated by the downregulation of osteogenesis-related proteins. Notably, BT-LB-EVLPs treatment effectively counteracted the suppression induced by H₂O₂ (Fig. 5G–N). Subsequently, the mRNA levels of osteogenesis-related genes across various groups were comprehensively profiled via RT-qPCR, which was similar to the experimental results of WB and immunofluorescence (Fig. 5O–Q). In summary, these findings demonstrated that BT-LB-EVLPs exhibit strong osteogenic differentiation capabilities in BMSCs under the oxidative stress microenvironment.

3.6. BT-LB-EVLPs attenuate oxidative stress-induced mitochondrial dysfunction via regulating mitochondrial dynamics

Mitochondria serve not only as the energy centers of the cell but also as critical triggers that determine cell fate under oxidative stress conditions [40]. To determine whether BT-LB-EVLPs alleviate cellular oxidative stress, RT-qPCR analysis indicated that BT-LB-EVLPs substantially upregulate the mRNA expression of several antioxidant factors, including NRF2, CAT, SOD1, and HO1 in H₂O₂-induced BMSCs, indicating an effective mitigation of cellular oxidative stress (Figure S8 A-D). A prominent early feature of mitochondrial dysfunction is the induction of mitochondrial oxidative stress. We employed a fluorescent probe (MitoSOX™ Red) to detect the capacity of BT-LB-EVLPs to mitigate mt-ROS in BMSCs. The findings revealed that the levels of mt-ROS in BMSCs were substantially elevated in the H₂O₂ group, while BT-LB-EVLPs led to a marked reduction in red fluorescence (Fig. 6A and B). Furthermore, both immunofluorescence and flow cytometry revealed that the ratio of JC-1 aggregate/monomer fluorescence in the H₂O₂ group was considerably reversed compared with the control group. Importantly, BT-LB-EVLPs intervention effectively attenuated this decline (Fig. 6C–E), suggesting the essential role of BT-LB-EVLPs in preserving mitochondrial membrane integrity in BMSCs under oxidative stress. In addition, TEM was conducted to observe the mitochondrial ultrastructure. As shown in Fig. 6F, prominent mitochondrial structural impairments were observed in the H₂O₂ group, characterized by outer mitochondrial membrane rupture, inner membrane cristae fragmentation, apparent mitochondrial swelling, and increased mitochondrial fragmentation. However, BT-LB-EVLPs significantly alleviated these oxidative stress-induced mitochondrial ultrastructural damages, restoring mitochondrial membrane integrity, cristae arrangement, and normal morphological features.

Fig. 6.

BT-LB-EVLPs mitigate oxidative stress-induced mitochondrial dysfunction by promoting mitochondrial fusion and suppressing mitochondrial fission. (A–B) Representative immunofluorescence images and semiquantitative analysis of MitoROS (Scale bars, 200 μm). (C) Representative immunofluorescence images of mitochondrial membrane potential (Scale bars, 50 μm). (D–E) Flow cytometry of mitochondrial membrane potential. (F) Representative TEM images of BMSCs (red arrows indicates mitochondria) (Scale bars, 2 μm and 500 nm). (G) Representative fluorescence images of mitochondrial structure in BMSCs (Scale bars, 5 μm). (H–I) Representative immunofluorescence images and semiquantitative analysis of MitoTracker Red staining and immunofluorescence staining of OPA1 (Scale bars, 50 μm). (J–K) WB images and quantitative analysis of OPA1, MFN1, and MFN2 in BMSCs. (L–M) WB images and quantitative analysis of DRP1 and FIS1 in BMSCs. (N) RT-qPCR of OPA1, MFN1, MFN2, DRP1, and FIS1 mRNA levels in BMSCs. Data are presented as mean ± SD (n = 3). Compared with the control group, ^#P < 0.05, ^##P < 0.01. Compared with the H₂O₂ group, ∗P < 0.05, ∗∗P < 0.01.

Mitochondrial dynamics, including fission and fusion, closely regulate structural mitochondrial homeostasis [41]. In the OVX group, we observed a marked decrease in the expression of mitochondrial fusion-related proteins (OPA1 and MFN1), accompanied by a concomitant increase in mitochondrial fission-related proteins (DRP1 and FIS1). Notably, administration of BT-LB-EVLPs effectively abrogated these pathological alterations in mitochondrial dynamic protein expression (Figure S9 A-E). To investigate whether BT-LB-EVLPs treatment modulates these key mitochondrial dynamic processes in BMSCs under oxidative stress, we first evaluated the integrity of mitochondria using MitoTracker staining. In contrast to the control group, which displayed a highly interconnected mitochondrial network, H₂O₂ stimulation induced excessive mitochondrial fragmentation, shifting the morphology from tubular to punctate structures. Notably, BT-LB-EVLPs robustly reversed this damage, increasing both mitochondrial content and network connectivity and restoring tubular length (Fig. 6G). These phenomena indicated that BT-LB-EVLPs might regulate the balance of mitochondrial dynamics under oxidative stress. Immunofluorescence and colocalization analysis further confirmed that BT-LB-EVLPs treatment enhanced the increased colocalization of OPA1 (green fluorescence) and MitoTracker (red fluorescence), indicating a facilitation of mitochondrial fusion (Fig. 6H and I). Furthermore, WB analysis demonstrated that BT-LB-EVLPs significantly upregulated the protein levels of OPA1, MFN1, and MFN2 (Fig. 6J and K), while downregulating the protein expressions of DRP1 and FIS1 (Fig. 6L and M). Subsequent RT-qPCR analysis corroborated these findings at the transcriptional level, showing similar expression trends for mitochondrial dynamics-related genes across treatment groups (Fig. 6N). Collectively, these results indicated that BT-LB-EVLPs keep the balance between mitochondrial fusion and fission, ultimately restoring mitochondrial functional activities.

3.7. BT-LB-EVLPs promote oxidative phosphorylation and inhibit glycolysis in oxidative stress-induced BMSCs

The fusion and fission of mitochondria in a dynamic manner influence respiratory activity and modulate the metabolic patterns, thereby satisfying the ever-changing bioenergetic demands of the cell [42]. Prior KEGG enrichment and GSEA analysis of proteomic sequencing confirmed that the effects of BT-LB-EVLPs are associated with OXPHOS and glycolysis (Fig. 4D–G). Consequently, we further investigated whether BT-LB-EVLPs regulate the mitochondrial OXPHOS process to sustain cellular bioenergetics. RT-qPCR revealed that BT-LB-EVLPs increased the mRNA levels of the five complexes in the OXPHOS system: NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1 in H₂O₂-induced BMSCs (Fig. 7A). Additionally, the mRNA expressions of glycolysis genes, including HK2, PKM2, PFKL, GLUT1, ENO3, and LDHA, were suppressed by BT-LB-EVLPs in H₂O₂-induced BMSCs (Fig. 7B). Further WB results demonstrated that in comparison to the control group, BT-LB-EVLPs upregulate the protein expressions of OXPHOS proteins, including NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1 (Fig. 7C and D). Additionally, compared to the H₂O₂-induced group, BT-LB-EVLPs treatment increased cellular ATP content (Fig. 7E) and upregulated the OXPHOS biomarker NAD⁺/NADH ratio while downregulating the glycolysis indicator lactic acid content (Fig. 7F and G). Mitochondrial biogenesis is the cellular process responsible for increasing mitochondrial mass and function, and is essential for meeting energy demands and maintaining homeostasis [43]. We therefore assessed the protein expressions of three key metabolic regulators including, PGC-1α, TFAM, and NRF1. WB results indicated that BT-LB-EVLPs treatment significantly reversed the H₂O₂-induced downregulation of PGC-1α, TFAM, and NRF1, suggesting that BT-LB-EVLPs may promote functional mitochondrial biogenesis to drive metabolic reprogramming and sustain cellular energy metabolism (Fig. S10). To clarify the causal relationship between mitochondrial dynamics and metabolic reprogramming driven by BT-LB-EVLPs, we performed pharmacological intervention experiments using mitochondrial fusion inhibitor MYLS22. After blocking mitochondrial fusion with MYLS22, the beneficial upregulatory effects of BT-LB-EVLPs on key OXPHOS proteins, including NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1, were significantly abolished. These results suggested that BT-LB-EVLPs-driven remodeling of mitochondrial fusion-fission homeostasis acts as the essential upstream initiating event to drive metabolic reprogramming (Fig. S11).

Fig. 7.

BT-LB-EVLPs regulate metabolic reprogramming in oxidative stress-induced BMSCs. (A) OXPHOS-related genes NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1 mRNA levels in BMSCs were detected via RT-qPCR. (B) RT-qPCR of glycolysis-related genes HK2, PKM2, PFKL, GLUT1, ENO3, and LDHA mRNA levels in BMSCs. (C–D) WB images and quantitative analysis of NDUFS1, SDHA, UQCRB COX5A, and ATP5A1 in BMSCs. (E–G) The ATP content, NAD⁺/NADH ratio, and lactic acid content in BMSCs were examined. (H) The OCR was detected by seahorse. (I) Quantification of basal respiration, ATP production, and maximum respiration. (J) The ECAR was detected by seahorse. (K) Quantification of glycolysis, glycolytic capacity, and glycolytic reverse. (L–M) The concentrations of mitochondrial and cytosolic Ca²⁺ were determined by Rhod-2 AM assay. (N) Schematic of BT-LB-EVLPs regulating BMSCs metabolic reprogramming. Data are presented as mean ± SD (n = 3). Compared with the control group, ^#P < 0.05, ^##P < 0.01. Compared with the H₂O₂ group, ∗P < 0.05, ∗∗P < 0.01.

Furthermore, to investigate how BT-LB-EVLPs restore mitochondrial energy metabolism, we assessed OCR and ECAR using a Seahorse XF analyzer. Compared to the H₂O₂-induced group, BT-LB-EVLPs treatment restored intracellular OCR levels and suppressed ECAR levels (Fig. 7H). Specifically, BT-LB-EVLPs markedly enhanced baseline respiration, maximal respiration, and ATP generation (Fig. 7I). In contrast, BT-LB-EVLPs treatment significantly inhibited glycolysis, glycolytic capacity, and glycolytic reverse (Fig. 7J and K). Given that mitochondrial calcium overload can disrupt the OXPHOS process [44], we next evaluated mitochondrial Ca²⁺ level using the fluorescent probe Rhod-2 AM (red fluorescence). Our findings indicated a notable overload of mitochondrial Ca²⁺ in association with the increased stimulation from H₂O₂, indicating that oxidative stress triggered Ca²⁺ outflow into cytoplasm and mitochondria, promoting mitochondrial Ca²⁺ overload. Nonetheless, BT-LB-EVLPs effectively mitigated mitochondrial Ca²⁺ overload (Fig. 7L and M). Collectively, these data indicated that BT-LB-EVLPs elicit a metabolic shift in BMSCs from glycolysis toward OXPHOS under oxidative stress, thereby sustaining cellular energy homeostasis (Fig. 7N).

3.8. BT-LB-EVLPs-derived miR167a-5p restores mitochondrial dysfunction and promotes osteogenic differentiation by targeting SLC25A26

Plant-derived EVLPs are known to contain diverse miRNAs that mediate cross-kingdom communication and exert bioactive functions in recipient organisms [45]. The miRNA sequencing was conducted to identify 79 unique miRNAs in BT-LB-EVLPs that regulate mitochondrial function (Table S5). Fig. 8A showed the levels of the top ten most abundant miRNAs in BT-LB-EVLPs. GO enrichment and KEGG enrichment further indicated that the predicted target genes of these miRNAs were significantly linked to OXPHOS and ATP binding (Fig. S12A–B), consistent with our previous proteomic sequencing results. We next evaluated the osteogenic potential of the five most abundant miRNAs, including miR162a-3p, miR167a-5p, miR166a-3p, miR166a-5p, and miR827-5p. Among these, miR167a-5p significantly upregulated the mRNA levels of RUNX2 and ALP, and increased the ALP-positive staining area, suggesting a pronounced role in promoting osteogenic differentiation (Fig. 8B, Fig. S13A–B). Subsequently, to identify potential downstream targets of miR167a-5p, we employed the miRanda, PITA, and RNAhybrid databases, which collectively predicted 85 candidate target genes (Fig. 8C). Next, at the intersection of these candidates with the top 500 mitochondrial-related genes from the GeneCards database, SLC25A26 was the most likely candidate (Fig. 8D).

Fig. 8.

BT-LB-EVLPs-derived miR167a-5p restores mitochondrial dysfunction and promotes osteogenic differentiation by targeting SLC25A26. (A) The ten most abundant miRNAs enriched in BT-LB-EVLPs were exhibited by the histogram based on miRNA sequencing. (B) The mRNA levels of RUNX2 and ALP in mimics treated group (miR162a-3p, miR167a-5p, miR166a-3p, miR166a-5p, and miR827-5p) were detected via RT-qPCR. (C) Overlap of potential targets of miR167a-5p predicted by the miRanda, PITA, and RNAhybrid databases. (D) Venn diagram of putative target genes of miR167a-5p overlapping with the top 500 mitochondria-related genes from the GeneCards database. (E) Schematic of the predicted binding site between the 3′-UTR of SLC25A26 mRNA and miR167a-5p based on the TargetScan database. (F) The dual luciferase reporter gene assay validated SLC25A26 as the direct target of miR167a-5p. (G) WB images of SLC25A26 in BMSCs. (H–I) WB images and quantitative analysis of RUNX2, ALP, and OPN in BMSCs. (J–K) Representative images and semiquantitative ALP staining of BMSCs (Scale bars, 200 μm). (L–M) Representative images and semiquantitative ARS staining of BMSCs (Scale bars, 200 μm). (N–O) WB images and quantitative analysis of NDUFS1, SDHA, UQCRB COX5A, and ATP5A1 in BMSCs. Data are presented as mean ± SD (n = 3). Compared with the control group, ^#P < 0.05, ^##P < 0.01. Compared with the H₂O₂ + NC-mimic group, ∗P < 0.05, and ∗∗P < 0.01. Compared with the H₂O₂ + NC-inhibitor group, ^&P < 0.05, ^&&P < 0.01.

SLC25A26 is the gene encoding S-adenosylmethionine carrier (SAMC), which is the sole known mitochondrial transporter responsible for importing S-adenosylmethionine (SAM) from the cytosol into mitochondria [46]. To determine whether SLC25A26 is a direct target of miR167a-5p, we created a luciferase reporter plasmid containing the 3′-UTR of SLC25A26, along with a corresponding mutant construct (Fig. 8E). Dual-luciferase reporter assays confirmed that miR167a-5p mimic considerably diminished the luciferase activity of wild-type SLC25A26 3′-UTR, whereas it exerted no influence on the activity of mutant SLC25A26 3′-UTR (Fig. 8F). Consistent with this finding, WB analysis revealed that BT-LB-EVLPs treatment downregulated SLC25A26 protein expression, an effect that was abolished by co-treatment with a miR167a-5p inhibitor (Fig. 8G, Fig. S14), indicating that SLC25A26 is a direct functional target of miR167a-5p.

To clarify the role of miR167a-5p in modulating cell fate and differentiation of BMSCs, both a miR167a-5p inhibitor and a miR167a-5p mimic were introduced into H₂O₂-induced BMSCs. The transfection efficiency of miR167a-5p mimic and miR167a-5p inhibitor was examined using RT-qPCR (Fig. S15A–B). WB analysis revealed that the miR167a-5p mimic enhanced the protein levels of osteogenic-related proteins RUNX2, ALP, and OPN, whereas the miR167a-5p inhibitor decreased these osteogenic-related proteins’ expression compared to the negative control group (Fig. 8H and I). Correspondingly, ALP staining revealed a notable increase in the ALP-positive area in the miR167a-5p mimic group, while a marked decrease was observed in the miR167a-5p inhibitor group (Fig. 8J and K). The area of calcium deposition was reduced in the miR167a-5p inhibitor group, whereas miR167a-5p mimic markedly increased the formation of mineralized nodules (Fig. 8L and M). In addition, immunofluorescence analysis indicated that the miR167a-5p mimic markedly increased the ratio of JC-1 red/green fluorescence in the H₂O₂-induced BMSCs. Conversely, the miR167a-5p inhibitor notably decreased this ratio, which confirms the functional significance of miR167a-5p in mitigating oxidative stress-induced mitochondrial dysfunction (Fig. S16A–B). Further WB results indicated that the miR167a-5p mimic robustly enhanced the levels of OXPHOS proteins, including NDUFS1, SDHA, UQCRB, COX5A, and ATP5A1, whereas a pronounced reduction in these protein levels was observed in the miR167a-5p inhibitor group (Fig. 8N and O). Collectively, these findings demonstrated that miR167a-5p delivered from BT-LB-EVLPs directly targets SLC25A26 to restore mitochondrial function and enhance osteogenic differentiation in BMSCs.

3.9. miR167a-5p regulates oxidative phosphorylation to promote osteogenic differentiation of BMSCs by suppressing SLC25A26

To elucidate the precise mechanism by which SLC25A26 regulates mitochondrial function, SLC25A26 overexpression plasmids and miR167a-5p mimic were transfected into H₂O₂-induced BMSCs. First, we verified the successful overexpression of SLC25A26 (Fig. S17A–B). Compared with the H₂O₂ + OE-NC group, the miR167a-5p mimic (H₂O₂ + OE-NC + Mimic group) markedly attenuated both cellular and mitochondrial ROS levels and significantly increased the JC-1 red/green ratio, indicating that miR167a-5p improved the mitochondrial membrane potential of BMSCs under oxidative stress (Fig. 9A–D, Fig. S18A–B). Notably, SLC25A26 overexpression substantially abrogated the protective effect of miR167a-5p on mitochondrial integrity in oxidative stress-induced BMSCs. Based on this, we further assess the impact of miR167a-5p on mitochondrial dynamics and energy metabolism under oxidative stress. As depicted in Fig. 9E and F, miR167a-5p mimic significantly downregulated the expressions of DRP1 and FIS1, while upregulating the levels of OPA1, MNF1, and MFN2 in H₂O₂-induced BMSCs. However, these proteins were reversed when SLC25A26 was overexpressed concurrently. In addition, H₂O₂ stimulation inhibited the protein expression of five key OXPHOS complex subunits, including ATP5A1, COX5A, NDUFS1, SDHA, and UQCRB, while the miR167a-5p mimic increased the levels of these OXPHOS proteins, which was similarly attenuated by SLC25A26 overexpression (Fig. 9G and H).

Fig. 9.

miR167a-5p regulates oxidative phosphorylation to promote osteogenic differentiation of BMSCs by suppressing SLC25A26. (A–B) Representative immunofluorescence images and semiquantitative analysis of MitoROS (Scale bars, 200 μm). (C–D) Representative immunofluorescence images and semiquantitative analysis of mitochondrial membrane potential (Scale bars, 50 μm). (E–F) WB images and quantitative analysis of DRP1, FIS1, OPA1, MFN1, and MFN2 in BMSCs. (G–H) WB images and quantitative analysis of ATP5A1, COX5A, NDUFS1, SDHA, and UQCRB in BMSCs. (I–J) Representative images of ALP and ARS staining of BMSCs in different groups (Scale bars, 200 μm). (K–L) Semiquantitative statistics of all positively stained areas in ALP and ARS staining. (M–N) WB images and quantitative analysis of RUNX2, ALP, and OPN in BMSCs. Data are presented as mean ± SD (n = 3).Compared to the control + OE-NC group, ^#P < 0.05, ^##P < 0.01. Compared to the H₂O₂ + OE-NC group, ∗P < 0.05, and ∗∗P < 0.01. Compared to the H₂O₂ + OE-NC + Mimic, ^&P < 0.05, ^&&P < 0.01.

Given that miR167a-5p attenuates H₂O₂-induced mitochondrial dysfunction by suppressing SLC25A26, we next examined whether its regulation of BMSC osteogenic differentiation and cell fate depends on SLC25A26. Compared with the H₂O₂ + OE-NC group, the H₂O₂ + OE-NC + Mimic group exhibited a significantly increased positive area of ALP and ARS staining (Fig. 9I–L), along with elevated expressions of osteogenesis-related markers (Fig. 9M and N). However, this promoting effect of miR167a-5p on BMSCS osteogenic differentiation was reversed by overexpressing SLC25A26. These results demonstrated that miR167a-5p regulates mitochondrial dynamics to sustain OXPHOS by suppressing SLC25A26, thereby promoting osteogenic differentiation of BMSCs.

3.10. BT-LB-EVLPs deliver miR167a-5p to exert anti-osteoporosis effects in vivo

To investigate whether miR167a-5p plays a crucial role in BT-LB-EVLPs-mediated alleviation of bone loss in OVX‐induced osteoporosis, OVX mice were administered antagomir-NC or antagomir-167a-5p via tail vein injection every two weeks for 8 weeks (Fig. 10A). Micro-CT analysis indicated that compared to the OVX + antagomiR-NC group, antagomir-167a-5p significantly accelerated OVX-induced trabecular bone loss and structural deterioration, as evidenced by decreased BMD, BV/TV, Tb.N, and Tb.Th, along with increased Tb.Sp and Tb.Pf (Fig. 10B–H). Furthermore, treatment with antagomir-167a-5p significantly inhibited the therapeutic benefits of BT-LB-EVLPs in alleviating bone loss in OVX-induced osteoporosis when compared to the OVX + antagomiR-NC + BT-LB-EVLPs group (Fig. 10B–H). Furthermore, by using H&E and Masson staining to evaluate the alterations in bone microarchitecture, we further discovered that the alleviating effect of BT-LB-EVLPs on bone architecture changes could be counteracted by the knockdown of miR167a-5p (Fig. 10I). Additionally, immunohistochemical staining demonstrated that the antagomir-167a-5p treatment substantially decreased the expression of ALP and OCN in comparison to the OVX + antagomiR-NC group. However, the pro-osteogenic differentiation effect of BT-LB-EVLPs was obviously abolished with the injection of antagomir-167a-5p (Fig. 10J–L). In summary, all these findings confirmed that miR167a-5p is essential for mediating the anti-osteoporotic effects of BT-LB-EVLPs.

Fig. 10.

BT-LB-EVLPs deliver miR167a-5p to exert anti-osteoporosis effects in vivo. (A) The experimental design and corresponding flowchart. (B) Representative micro-CT images of bone scan (n = 5). (C–H) Quantitative analysis of key bone microstructure parameters of femurs (n = 5). (I) H&E and Masson staining images of femurs (Scale bars, 50 μm and 200 μm) (n = 3). (J–L) Immunohistochemistry and semi-quantitative analysis of ALP and OCN expressions (n = 3). Data are presented as mean ± SD. Compared with the Sham group, ^#P < 0.05, ^##P < 0.01. Compared with the OVX + Antagomir-NC group, ∗P < 0.05, and ∗∗P < 0.01. Compared with the OVX + Antagomir-NC + LB-EVLPs, ^&P < 0.05, ^&&P < 0.01.

4. Discussion

OP is a systemic skeletal disorder characterized by low bone mass, deterioration of bone microarchitecture, and consequent increased fracture risk [47]. Currently, most anti-osteoporotic drugs lack bone-targeting specificity, resulting in significant accumulation in organs, such as the liver and kidney, which often causes significant toxic side effects [48,49]. Therefore, there is an urgent need to develop bone-targeting therapies or novel strategies that minimize the adverse effects associated with conventional treatments. The recent advances in nanotechnology, particularly in ultra-high-speed centrifugation, nanoscale characterization and functional modification, are driving a paradigm shift in the application of traditional Chinese medicine [50,51]. Among the emerging natural nanoplatforms, plant-derived EVLPs have garnered substantial interest as natural nanomaterials. Their promise as therapeutic agents stems from advantageous properties such as scalable production, high bioavailability, and excellent biosafety characteristics [52]. In this study, we pioneered the isolation of EVLPs from the medicinal and edible LB and further functionalized LB-EVLPs with the bone-targeting peptide SDSSD (BT-LB-EVLPs), thereby endowing these nanoparticles with enhanced bone tissue specificity. We demonstrated that BT-LB-EVLPs effectively ameliorate bone loss, improve bone microarchitecture, and enhance bone strength in osteoporotic mice. Notably, the engineered LB-EVLPs exhibited robust and specific bone-targeting ability in OVX mice, which improved the effectiveness of the treatment without causing any systemic toxicity. Considering that BMSCs are the major source of osteoblasts and critically regulate bone formation, we further investigated the impact of BT-LB-EVLPs on BMSCs proliferation and osteogenic differentiation. We observed that BT-LB-EVLPs were rapidly internalized by BMSCs via the clathrin endocytosis pathway, and BT-LB-EVLPs exhibit a potent pro-osteogenic effect on BMSCs under the oxidative stress microenvironment. These findings not only demonstrated the promise of BT-LB-EVLPs as a green and efficacious approach for OP but also laid the foundation for the future clinical use of plant-derived EVLPs in OP treatment.

In this study, compositional analysis revealed that LB-EVLPs are enriched in miRNAs, proteins, and lipids, which are functionally associated with metabolic processes, including small molecule metabolic, amino acid metabolic, and organic acid metabolic. Furthermore, to elucidate the therapeutic mechanisms of BT-LB-EVLPs in OP, we performed proteomic sequencing of femoral tissues. GO and KEGG enrichment analyses indicated that significant alterations in proteins associated with “OXPHOS” and “reactive oxygen species” in the OVX + BT-LB-EVLPs group compared with the OVX group. Metabolic reprogramming within BMSCs has been increasingly recognized as a key driver of osteoporosis progression [53]. Within the bone microenvironment, the differentiation and mineralization of BMSCs are highly energy-demanding processes, requiring substantial ATP to support collagen synthesis and bone matrix production [54,55]. Under oxidative stress, BMSCs shift their metabolic shift from OXPHOS toward glycolysis, which results in the accumulation of LDHA and an increase in ROS generation [56]. Mechanistically, glycolysis converts glucose to pyruvate and NADH. Pyruvate can be directed by pyruvate dehydrogenase into acetyl-CoA for the TCA cycle or alternatively converted by LDHA to lactate [57]. Our data demonstrated that BT-LB-EVLPs increase the expression of the five OXPHOS complexes (complexes I-V), OCR levels, and ATP, while suppressing key glycolytic genes, ECAR levels, and lactic acid content in H₂O₂-induced BMSCs. These findings suggested that BT-LB-EVLPs exert anti-osteoporosis effects by regulating the balance between OXPHOS and glycolysis to restore energy metabolism of BMSCs.

Mitochondria serve as crucial centers for the cellular energy supply and are pivotal for cellular functions, including OXPHOS, glycolysis, calcium homeostasis, and oxidative homeostasis [58]. Mitochondria continuously undergo dynamic fusion or fission to preserve the integrity of the mitochondrial network, facilitating adaptive adjustments to the cellular energy supply [13]. Fusion of mitochondria encourages the integration of electron transport chain complexes, thereby enhancing OXPHOS for energy generation. On the other hand, mitochondrial fission results in the expansion of mitochondrial cristae, which separates the electron transport chain complexes and shifts cellular metabolism towards glycolysis [14]. In our study, we found that BT-LB-EVLPs possess inherent mitochondrial-targeting activity and reduced oxidative stress-induced mitochondrial membrane hyperpolarization and the accumulation of mitochondrial superoxide to promote BMSCs fitness and osteogenic differentiation, which offers a promising therapeutic strategy for drug delivery in OP. Previous studies have reported that BMSCs display elongated and interconnected mitochondria associated with OPA1, MFN1, and MFN2 during the osteogenic differentiation process [59]. Our findings demonstrated that the fragmented mitochondria harbored highly disorganized and swollen cristae in H₂O₂-induced BMSCs. In contact, BT-LB-EVLPs treatment robustly reversed this damage and increased both mitochondrial content and network connectivity, restoring tubular length. Furthermore, BT-LB-EVLPs promoted fusion while inhibiting fission, resulting in restoring the disrupted mitochondrial homeostasis caused by oxidative stress and enhancing osteoblast differentiation. Mitochondrial biogenesis is essential for maintaining mitochondrial function and cellular homeostasis under diverse pathological conditions [60]. As a core co-activator of mitochondrial biogenesis, PGC-1α co-activates NRF1 to drive TFAM-mediated mitochondrial DNA transcription and OXPHOS complex assembly [61]. Our results showed that BT-LB-EVLPs treatment significantly upregulated the protein levels of PGC-1α, TFAM, and NRF1 compared to the H₂O₂ group. Based on the findings mentioned above, we speculated that BT-LB-EVLP-mediated mitochondrial dynamics and mitochondrial biogenesis may establish a synergistic regulatory axis that restores the osteogenic potential of BMSCs by improving mitochondrial energy metabolism. Future studies employing in vitro and in vivo genetic intervention approaches are warranted to rigorously validate this proposed mechanistic model and to fully delineate the regulatory network through which BT-LB-EVLPs exert their anti-osteoporotic effects.

The therapeutic effect of EVLPs is largely attributed to their biomolecular cargo, which encompasses functional miRNAs, lipids, and proteins [62,63]. Emerging evidence has highlighted plant-derived miRNAs as essential regulators of post-transcriptional gene expression that participate in diverse biological processes, including cell proliferation, cycle regulation, and inflammatory responses [64,65]. Preclinical studies have further demonstrated that plant EVLPs-derived miRNAs can effectively modulate human cellular processes and pathological signaling pathways via targeted gene regulation, highlighting their potential as innovative therapeutic interventions [66]. In our study, miRNA sequencing identified miR167a-5p as a highly abundant miRNA in BT-LB-EVLPs and exhibiting a strong ability to promote osteogenic differentiation compared to other miRNAs. miR167a is involved in multiple abiotic stress responses in plants [67,68]. miR167a-5p, a mature miRNA derived from the pre-miR-167 precursor, was highly conserved among different species [69]. To functionally characterize this miRNA, we transfected BMSCs with either a miR167a-5p mimic or inhibitor. The upregulation of miR167a-5p enhanced ALP activity and promoted osteogenic marker expression, including RUNX2, ALP, and OPN. Moreover, the miR167a-5p mimic effectively scavenged mitochondrial ROS and stabilized mitochondrial membrane potential in H₂O₂-induced BMSCs, indicative of improved mitochondrial redox homeostasis under oxidative stress. In addition, to further validate the functional contribution of miR167a-5p in vivo, OVX mice were co-administered with either antagomir-NC or antagomir-167a-5p. The results revealed that the antagomir-167a-5p obviously attenuates the protective effects of BT-LB-EVLPs against bone loss and microstructural deterioration. These findings indicated that miR167a-5p is a critical regulator responsible for BT-LB-EVLPs-mediated alleviation of bone loss, emphasizing its promise as a therapeutic target for OP treatment.

To investigate the potential downstream mechanisms of miR167a-5p, we conducted a bioinformatic analysis that identified SLC25A26 as a key target gene of miR167a-5p. Consistent with this finding, SLC25A26 expression was downregulated in BMSCs treated with BT-LB-EVLPs and upregulated upon transfection with a miR167a-5p inhibitor. SLC25A26 is an inner mitochondrial membrane transporter, facilitating the transport of SAM into the mitochondrial matrix [70]. Aberrant expression of SLC25A26 may disrupts mitochondrial SAM-dependent methylation processes, resulting in elevated oxidative stress, mitochondrial dysfunction, bioenergetic deficits, and metabolic dysregulation [71]. Our findings demonstrated that SLC25A26 overexpression could inhibit mitochondrial fusion and mitochondrial oxidative respiration while promoting mitochondrial division, thereby exacerbating the mitochondrial homeostasis induced by oxidative stress. Notably, SLC25A26 overexpression in BMSCs counteracted the positive effects of the miR167a-5p mimic on mitochondrial oxidative stress, metabolism homeostasis, and mitochondrial dynamics. Moreover, we found that the overexpression of SLC25A26 in BMSCs negated the pro-osteogenic effect of the miR167a-5p mimic. Together, these findings supported a mechanistic pathway in which miR167a-5p-mediated SLC25A26 suppression contributes to an imbalance in mitochondrial dynamics and energy metabolism disorders, ultimately enhancing osteogenic differentiation of BMSCs.

In the development of anabolic anti-osteoporosis drugs, precise delivery of therapeutic agents to bone tissue and efficient cellular internalization by recipient cells are critical for optimizing therapeutic efficacy. Previous attempts have focused on encapsulating bioactive molecules (e.g., bisphosphonates, parathyroid hormone fragments, or natural products) into synthetic nanocarriers such as liposomes or polymer nanoparticles to enhance solubility, prolong circulation, and improve bone targeting [72]. However, these synthetic platforms often face inherent limitations, including suboptimal biocompatibility, poor stability, high toxicity, and suboptimal targeted delivery capabilities. By contrast, plant-derived EVLPs represent a promising alternative due to their innate biocompatibility, structural stability, and intrinsic targeting capabilities. Specifically, LB-EVLPs integrate the natural bioactivity of herbal components with the structural and functional merits of nanovesicles, exhibiting distinct advantages for OP therapy. First of all, they feature a native bilayer membranous structure that enhances their stability and protects encapsulated bioactive cargo from degradation [73]. Besides, as nanoscale vesicles with inherent membrane permeability, LB-EVLPs can efficiently traverse biological barriers and be internalized by BMSCs [74]. And the bone-targeted engineering of LB-EVLPs enabled their preferential accumulation in bone tissues to improve local drug bioavailability, which minimizes off-target effects on non-bone organs and enhances therapeutic precision. Last but not least, LB-EVLPs are derived from a medicinal and edible plant (Lycium barbarum L.) with a long history of safe use, ensuring low immunogenicity and favorable biocompatibility, which is particularly critical for long-term OP management. In the future, advances in plant-derived EVLPs through surface modification and cargo loading techniques are expected to further enhance the targeting specificity and drug encapsulation efficiency, which strengthens their translational potential for OP therapy [75,76].

Although significant progress has been made in this study, several limitations should be acknowledged. First, the study did not establish a bone-targeting peptide control group (e.g., a control group treated with SDSSD peptide alone or DSPE-PEG2000 empty vector), making it difficult to rule out non-specific binding of the vector itself or peptide fragments to bone tissue or therapeutic effects. Second, the causal relationship between BT-LB-EVLPs-mediated mitochondrial dynamics and downstream metabolic reprogramming has not been fully and accurately verified at the in vivo organismal level. Future investigations employing tissue-specific genetic interventions, such as osteoblast-specific SLC25A26 overexpression or conditional OPA1 knockout mouse models to further explore the in vivo causal chain linking the miR167a-5p/SLC25A26 axis, mitochondrial dynamic remodeling, and the anti-osteoporosis effect of BT-LB-EVLPs. Third, all in vivo efficacy and mechanistic validations were performed only in an OVX mouse model, the therapeutic potential of BT-LB-EVLPs in other clinically prevalent osteoporosis subtypes, including type 2 diabetic osteoporosis (T2DOP) and glucocorticoid-induced osteoporosis (GIOP), has not been explored. Meanwhile, this strategy remained in the early preclinical stage, and systematic long-term safety assessment (beyond 3 months) and comprehensive immunogenicity profiling have not been conducted. In future investigations, we will systematically validate the efficacy and long-term safety of BT-LB-EVLPs in diverse osteoporosis models and in large animal models including, rabbits and beagle dogs, to steadily advance the preclinical development of this targeted nanotherapy.

5. Conclusion

We successfully isolated LB-EVLPs from LB and engineered bone-targeted BT-LB-EVLPs via surface functionalization with the bone-homing peptide SDSSD, with markedly enhanced bone tissue specificity. Systemic administration of BT-LB-EVLPs effectively mitigated bone loss, enhanced bone structural integrity, and increased bone strength in osteoporotic mice. Mechanistically, BT-LB-EVLPs restored mitochondrial dynamic balance and drove metabolic reprogramming, thereby enhancing the osteogenic differentiation of BMSCs. Furthermore, we identified miR167a-5p as a pivotal functional component within BT-LB-EVLPs, responsible for improving mitochondrial function and energy metabolism, ultimately alleviating bone loss. Additionally, BT-LB-EVLPs exhibited excellent biocompatibility and biosafety profiles, highlighting their promising biomedical applications in OP treatment. Overall, these findings not only provided a scientific foundation for developing BT-LB-EVLPs-based precise nanotherapeutic strategies in OP management but also highlighted the promising role of plant extracellular vesicle-derived miRNAs as pro-osteogenic agents for improving bone health.

Ethics approval and consent to participate

The procedure was conducted in accordance with the Committee on the Ethics of Animal Experiments of Nanjing University of Chinese Medicine (Approval no.202408A050, Approval no. 202504A084).

Funding

This work was funded by the National Famous Traditional Chinese Medicine Expert Inheritance Studio Construction Project (National Traditional Chinese Medicine Education Letter No. 75, 2022), the third batch of Jiangsu Province Traditional Chinese Medicine Leading Talent Training Program (Jiangsu Traditional Chinese Medicine Science and Education [2023] No. 17), the Key Project of Jiangsu Province's Traditional Chinese Medicine Science and Technology Development Plan (ZD202202), and the Postgraduate Research & Practice Innovation Program of Jiangsu Province (SJCX25_1021).

CRediT authorship contribution statement

Shuai Chen: Data curation, Investigation, Methodology, Writing – original draft, Writing – review & editing. Miaochao Qin: Investigation, Methodology. Qilong Zhao: Data curation, Methodology. Ningsheng Tian: Investigation, Methodology. Lingling Yu: Methodology. Liansheng Shao: Conceptualization. Su Yan: Methodology. Xinyu Ding: Methodology. Pengfei Sun: Methodology. Peng Ma: Data curation. Junwu Wang: Supervision, Writing – review & editing. Penghua Fang: Supervision, Writing – review & editing. Wen Min: Funding acquisition, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.