Red ginseng-derived nanovesicles to modulate osteoblast and osteoclastogenesis for osteoporosis therapy

Abstract

Osteoporosis is a skeletal disorder characterized by an imbalance between bone formation and resorption, which leads to progressive bone loss and increased fracture risk. While current treatments either inhibit bone resorption or stimulate bone formation, their long-term use is associated with adverse effects, necessitating alternative therapeutic approaches. In this study, we explore the use of red ginseng-derived nanovesicles (RGNVs) as a biocompatible nanotherapeutic strategy for treating osteoporosis. The RGNVs were successfully isolated and characterized, revealing a lipid bilayer structure enriched in bioactive ginsenosides and functional proteins. In vitro, RGNVs enhanced osteoblast proliferation, differentiation, and mineralization while suppressing osteoclast differentiation and bone resorption by modulating the BMP-2/Smad and MAPK signaling pathways. In an ovariectomy-induced osteoporosis mouse model, oral administration of RGNVs significantly restored bone volume and mineral density, and biodistribution studies confirmed their preferential accumulation in the bone tissue. Systemic toxicity evaluation indicated no adverse effects, supporting the safety of RGNVs for therapeutic use. These findings suggest that RGNVs regulate bone remodeling through a dual mechanism, to stimulate bone formation and inhibit bone resorption, thereby offering a promising and well-tolerated approach for osteoporosis management.

Graphical abstract

1. Introduction

Osteoporosis is a skeletal disease characterized by an imbalance between osteoclast-mediated bone resorption and osteoblast-driven bone formation, resulting in decreased bone density, compromised bone architecture, and increased fracture risk [[1], [2], [3]]. The prevalence of osteoporosis is increasing globally owing to the aging population [4]. Current pharmacological interventions are typically categorized as anti-resorptive agents which inhibit osteoclast activity; anabolic agents which stimulate osteoblast activity, or activated vitamin D analogs [5,6]. However, long-term use of these treatments can lead to adverse effects, such as atypical femoral fractures, osteonecrosis of the jaw and hypercalcemia [7]. This has promoted research into alternative therapeutic strategies, including those derived from natural products and herbal medicines [8,9].

Plant-derived exosome-like nanovesicles (PENs) are nanoscale lipid bilayer vesicles (30–300 nm in diameter), which carry proteins, lipids, mRNA, miRNAs and metabolites. PENs have been isolated from leaves, stems, roots, seeds and nuts of various plants [[10], [11], [12]], and has been shown to mediate intracellular communication and modulate biological responses across the biological kingdom [10,11]. Given their biocompatibility and stability, PENs have garnered interest for therapeutic applications in conditions such as inflammatory bowel disease, liver disorders, cancers and osteoporosis [[13], [14], [15], [16], [17]]. Notably, recent studies have shown that yam-derived exosome-like nanovesicles prevent osteoporosis in an ovariectomized (OVX)-induced osteoporosis mouse model by stimulating osteoblast activity, and that PENs derived from apples and plums enhance osteogenic differentiation in vitro [14,[18], [19], [20]]. These findings suggest that PENs promote bone formation. Moreover, PENs derived from various fruits and vegetables are stable across various pH and temperature conditions, and can resist gastrointestinal (GI) digestion [[21], [22], [23], [24]], suggesting their oral bioavailability and safety as therapeutic agents [25].

Korean ginseng (Panax ginseng Meyer) is a traditional medicinal herb [[26], [27], [28], [29], [30], [31]] widely used in East Asia owing to its broad range of health benefits [32,33]. Red ginseng, produced by steaming and drying harvested ginseng roots (typically 4–6 years old) [34], exhibits enhanced pharmacological efficacy compared with white ginseng [35]. Red ginseng has documented immunomodulatory, anti-fatigue, circulatory-enhancing, memory-improving, antioxidant, menopausal health-promoting, anticancer and anti-osteoporosis effects [26,[36], [37], [38], [39], [40], [41], [42]]. Ginsenosides are the primary bioactive components of red ginseng. In particular, ginsenosides Rg1 and Rb1 have been identified as potential therapeutic agents against osteoporosis because of their bone-protective activities [43,44]. However, the therapeutic potential of free ginsenosides is often severely limited by low oral bioavailability, susceptibility to degradation in the harsh GI environment, and rapid systemic clearance. Consequently, traditional extracts often fail to achieve the necessary therapeutic concentrations in bone tissue. These properties make red ginseng an attractive source for developing novel osteoporosis treatments only if these pharmacokinetic hurdles can be overcome. Recently, plant-derived nanovesicles have demonstrated exceptional stability in the GI tract and the ability to efficiently deliver encapsulated cargo to specific tissues via active endocytic pathways [[21], [22], [23]].

Therefore, we investigated the potential of red ginseng-derived nanovesicles (RGNVs) as a therapeutic agent for osteoporosis that combines the pharmacological potency of ginsenosides with the superior stability and bone-targeting capability of a nanovesicular carrier. We isolated RGNVs and evaluated their effects on bone remodeling both in vitro and in vivo. As Fig. 1 shows, RGNVs are isolated from red ginseng via multistep ultracentrifugation and administered orally to the target bone tissue. We hypothesized that RGNVs accumulate in the bone, promote osteoblast differentiation via bone morphogenetic protein 2 (BMP-2) signaling, and concurrently inhibit osteoclast activation through receptor activator of nuclear factor-κB ligand (RANKL)-related signaling, thereby restoring the balance between bone formation and resorption (Fig. 1). We assessed the capability of RGNVs to enhance osteoblast function and inhibit osteoclastogenesis in cell culture and further validated their therapeutic efficacy in an OVX-induced osteoporosis mouse model. The biodistribution and safety of RGNVs were also examined. Our findings demonstrate that RGNVs exert a dual action on bone remodeling, highlighting their potential as a biocompatible treatment for osteoporosis.

Fig. 1.

A schematic to depict osteoporosis prevention effects of RGNVs. Schematic representation of RGNVs isolation from red ginseng and their systemic absorption following oral administration. The extraction of RGNVs from red ginseng occurs through a multi-step purification process and they are administered orally to investigate their effects on bone homeostasis. RGNVs accumulation in bone tissue leads to osteoblast differentiation via BMP-2 signaling and concurrent inhibition of osteoclast activation through RANK signaling, thereby restoring bone formation-resorption balance. These effects promote osteogenesis and inhibit bone resorption in an OVX-induced osteoporosis mouse model.

2. Materials and methods

2.1. Materials

Minimum essential medium alpha (MEM-α), fetal bovine serum (FBS), and penicillin-streptomycin were purchased from Gibco (Grand Island, NY, USA). Macrophage colony-stimulating factor (M-CSF), and receptor activator of nuclear factor-κB ligand (RANKL) were obtained from PeproTech (East Windsor, NJ, USA). Recombinant mouse epidermal growth factor (EGF) protein and recombinant mouse tumor necrosis factor-alpha (TNF-α) protein were supplied by ABclonal (Woburn, MA, USA). All other chemicals and reagents used in this study were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise specified.

2.2. Isolation and characterization of RGNVs

Red ginseng (Panax ginseng) roots were processed using a standardized method (ISO 19610). The roots were soaked in phosphate buffered saline (PBS) at 4 °C overnight and then ground into a homogenate. The homogenate was sequentially centrifuged twice to remove debris (500 × g for 10 min, 2000 × g for 20 min, and 10,000 × g for 30 min). The supernatant was obtained after filtration through 0.45 µm and 0.22 µm por-size filters to eliminate the remaining particulates, then ultracentrifuged at 100,000 × g for 1 h. The pelleted nanovesicles were washed, resuspended in PBS, and stored at −80 °C until use. The RGNVs’ size distribution and concentration were determined using nanoparticle tracking analysis (NTA; Nanosight NS300, Malvern, UK). The morphology of the negatively stained (0.75% uranyl formate, w/v) RGNVs was examined by transmission electron microscopy (TEM).

For proteomic analysis, peptides were separated using a Vanquish Neo UHPLC system coupled to an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific). Peptides were loaded onto an Acclaim PepMap 100 trap column (75 µm × 20 mm) and separated on an EASY-Spray PepMap Neo analytical column (75 µm × 500 mm). The mobile phases consisted of solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile). Peptides were eluted at a flow rate of 300 nl/min using the following gradient: 5% solvent B from 0 to 5 min, increased to 30% solvent B over 100 min, and then to 45% solvent B over the next 20 min. Raw data were processed using Proteome Discoverer (version 3.0.1.27) and searched against the Panax ginseng UniProt database combined with an in silico in-house database. Protein identification required at least one unique peptide per protein group, with a false discovery rate (FDR) set to < 0.01 at both peptide and protein levels. The mass spectrometry proteomics data are being deposited to the ProteomeXchange Consortium via the PRIDE partner repository. Data are available via ProteomeXchange with identifier PXD075980.

2.3. Stochastic optical reconstruction microscopy analysis of RGNVs

To confirm the presence of extracellular vesicle’s marker proteins on the lipid membrane of RGNVs, stochastic optical reconstruction microscopy (STORM) was performed using the EV Profiler Kit (#EV-MAN-1.0, ONI) in accordance with the manufacturer’s protocol. All imaging data were processed and analyzed using NimOS software (Version 1.19, ONI).

2.4. Cell culture and osteogenic/osteoclastic differentiation

MC3T3-E1 cells [mouse calvarial pre-osteoblast cell line, source: American Type Culture Collection (ATCC, USA), supplier: Penielbio (Daegu, Republic of Korea), date obtained: 09-09-2020, RRID: CVCL_5440, official name: MC3T3-E1 subclone4] were cultured in MEM-α supplemented with 10% FBS and 1% penicillin-streptomycin at 37 °C in a humidified 5% CO₂ incubator. Osteogenic differentiation was initiated at confluence by adding osteoblast differentiation medium (MEM-α with 10% FBS, 100 µg/ml ascorbic acid, and 10 mM β-glycerophosphate). Primary mouse hematopoietic stem cells (mHSCs) were isolated by flushing the femurs and tibias of 6-week-old male C57BL/6 mice. The cells were cultured in the medium described above. Osteoclast differentiation was induced by culturing the cells in medium supplemented with M-CSF 30 ng/ml and RANKL 100 ng/ml.

For osteoblast differentiation, cells were cultured in osteogenic induction medium containing ascorbic acid (50 µg/ml) and β-glycerophosphate (10 mM). The medium was replaced every 2–3 d during the differentiation period. A BMP inhibitor and activators of the ERK, JNK/p38 MAPK and NF-κB pathways were used, including LDN-193189 hydrochloride, recombinant mouse EGF, anisomycin and recombinant mouse TNF-α, respectively (Table S1).

2.5. Cell proliferation MTT assay

The MC3T3-E1 cells were seeded in 96-well plates at 1 × 10⁴ cells/well and allowed to attach for 24 h. Cells were then treated with RGNVs at 0, 1, 5 and 10 µg/ml and cultured for 3 or 7 d. At each time point, cell proliferation was assessed using the [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT) assay. Briefly, MTT reagent was added to cells in each well and incubated for 4 h at 37 °C. The absorbance of the resulting purple formazan crystals dissolved in dimethyl sulfoxide (DMSO) was measured at 570 nm using a microplate reader (TECAN, Männedorf, Switzerland) to quantify viable cell metabolic activity.

2.6. Animal ethics and housing

All animal procedures were approved by the Andong National University Animal Ethics Committee (approval number: 2024-3-0930-04) and were carried out in accordance with the institutional guidelines on laboratory animal care. Four-week-old female C57BL/6 N mice (Koatech, Pyeongtaek, Korea) were housed under standard conditions (12 h light/dark cycle, with food and water provided ad libitum).

2.7. Osteogenic differentiation, mineralization assays and related staining ALP activity

MC3T3-E1 cells and mHSCs were treated with or without RGNVs during differentiation. Alkaline phosphatase (ALP) activity, an early osteoblast differentiation marker, was measured on Day 3 and 7 using a colorimetric assay. The cell lysates (for intracellular ALP) and culture supernatants (for secreted ALP) were collected. ALP activity was determined by the enzymatic conversion of p-nitrophenyl phosphate (pNPP) to p-nitrophenol (pNP) in alkaline buffer, as described previously [37]. The reaction product, pNP, was measured at 405 nm and normalized to the total protein content measured by using bicinchoninic acid (BCA) protein assay. ALP activity was expressed as nmol pNP produced per minute per milligram of protein (intracellular) or per milliliter of medium (secreted).

2.7.1. Von Kossa staining

Mineral deposition in osteogenic cultures was evaluated by Von Kossa staining to detect phosphate minerals. Cells treated with RGNVs (0, 1, 5, and 10 µg/ml) and boiled RGNVs at 100 °C for 30 min for 3 and 7 d were fixed in 70% ethanol, followed by incubation with 5% silver nitrate under ultraviolet (UV) light for 1 h. After washing, the presence of mineralized nodules was indicated by black/brown staining of the calcium phosphate deposits. A light microscope (Leica, Wetzlar, Germany) visualized the stained cultures.

2.7.2. Alizarin Red S staining

Calcium deposition was assessed using Alizarin Red S staining. RGNV-treated cells (0, 1, 5 and 10 µg/ml) and boiled RGNVs at 100 °C for 30 min were fixed after 3 or 7 d and stained with 40 mM Alizarin red solution (pH 4.2, Sigma-Aldrich, St. Louis, MO, USA) for 10 min. Excess dye was washed off, and calcium-rich deposits appeared as red nodules. The stained cultures were observed microscopically (Leica, Nussloch, Germany), and the bound dye was eluted with acetic acid and ammonium hydroxide and measured spectrophotometrically at 405 nm (TECAN, Männedorf, Switzerland).

2.8. Cellular uptake and endocytosis pathway of RGNVs

RGNVs were labeled with DID, a red fluorescent dye. For the cellular uptake study, 1 × 10⁴ cells were seeded into each well of eight-chamber slides and cultured for 24 h. The cells were then treated with RGNVs at 0, 1, 5 and 10 µg/ml for an additional 24 h. After incubation, non-internalized RGNVs were removed by washing with PBS, followed by fixation with 4% paraformaldehyde for 30 min, The cells were then stained with DAPI and visualized using a fluorescence microscope.

To investigate the endocytic pathway involved in RGNV uptake, cells were pretreated with 10 µM methyl-β-cyclodextrin (MβCD; Sigma Aldrich #C4555) or 22.5 µM of chlorpromazine (CPZ; Sigma Aldrich #C8138) for 15 min to selectively inhibit cholesterol/lipid raft- or clathrin-mediated endocytosis, respectively. The inhibitors were subsequently removed by PBS washing, and the cells were incubated with 10 µg/ml DID-labeled RGNVs for 4 h. Cellular uptake was then assessed using fluorescence microscopy.

2.9. qRT-PCR

Total RNA was extracted from the cultured cells using a RNeasy Mini Kit (Qiagen, Valencia, CA, USA). Nanodrop® ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) was employed to determine RNA concentration and purity. cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). Quantitative real-time polymerase chain-reaction (qRT-PCR) was performed on a QuantStudio 1 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) using SYBR Green PCR Master Mix. Gene expression levels of osteogenic markers (Runx2, ALP, osteopontin [OPN], and procollagen I [ProCOL1]) and signaling molecules (BMP-2/4, Smad1, Smad5, and Smad8) were quantified, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal control. The primer sequences for each target gene are listed in Table S2.

2.10. Western blot analysis

Cultured osteoblasts and osteoclasts were lysed in 1 × radio-immune precipitation assay (RIPA) buffer containing protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitors (Sigma-Aldrich). Equal amounts of protein from each sample were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Membranes were blocked with 5% non-fat milk in Tris-buffered saline (TBS-T; with 0.1% Tween-20) for 2 h, then incubated overnight at 4 °C with primary antibodies (Table S2), targeting osteoblast differentiation (Runx2, ALP, OPN, ProCOL1, BMP-2/4, phosphorylated Smad1/5/9, and total Smald1); osteoclastogenesis (nuclear factor of activated T cells 1 [NFATc1], c-Fos, Cathepsin K, phosphorylated ERK1/2, phosphorylated JNK, p38 MAPK, and phosphorylated IκBα); and loading control (β-actin, Table S3). After washing three times with TBS-T, the membranes were washed three times with TBS-T, followed by incubation with horseradish peroxidase-conjugated secondary antibodies (1:1000; sc-516102, Santa Cruz Biotechnology) for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) substrate (Thermo Fisher, Waltham, MA, USA) and imaged using a Fusion SOLO X imaging system (Vilber, France).

2.11. In vivo toxicity evaluation

Female C57BL/6 N mice (6 weeks old) were randomly divided into two groups: control and treatment groups that received PBS and RGNVs, respectively. The RGNVs (1 mg/kg) were administered orally once daily for 7 d Throughout the treatment period, mice were monitored for signs of distress or adverse effects. On Day 7, the mice were euthanized using carbon dioxide (CO₂), and major organs (heart, lungs, liver, kidney, and brain) and blood were collected. The organs were fixed in 4% paraformaldehyde and processed for histological analysis. Blood was centrifuged at 12,000 rpm for 10 min to isolate the serum, which was stored at −80 °C until analysis. Serum biochemical markers of liver function (alanine aminotransferase [ALT], aspartate aminotransferase [AST], and triglycerides [TG], high-density lipoprotein cholesterol [HDL-C], total bilirubin [TBIL]) and kidney function (blood urea nitrogen [BUN], creatinine) were measured using an automated analyzer (FUJI DRI-CHEM SLIDES) according to the manufacturer’s instructions. Serum biomarkers of the pro-inflammatory response (TNF-α and interleukin-1β [IL-1β]) were analyzed using an ELISA kit (R&D Biosystems, Minneapolis, MN, USA). Data from RGNV-treated and control mice were compared to detect any significant differences indicative of toxicity.

2.12. In vivo biodistribution of RGNVs

The RGNVs were labeled with the near-infrared fluorescent dye, DiR (Invitrogen, Waltham, MA, USA), for in vivo tracking. RGNVs (1 mg/ml) were incubated with DiR (10 mM final concentration) at room temperature for 1 h. Unbound DiR was removed by ultracentrifugation (100,000 × g, 4 °C), and DiR-labeled RGNV pellet was washed and resuspended in PBS. Female C57BL/6 N mice (6 weeks old) were orally administered DiR-bound RGNVs (1 mg/kg). At 12, 24 and 48 h post-administration, the mice were euthanized, and major organs/tissues (small intestine, large intestine, liver, spleen, kidneys, lungs, heart, brain, lymph nodes and femur bones) were harvested. The fluorescence distribution was immediately imaged using an in vivo imaging system (IVIS; NEWTON 7.0; Vilber, Collégien, France) to detect DiR signal in each tissue sample. The fluorescence intensity in each organ was quantified using an Kuant2.5 software (Vilber, Collégien, France) to determine the relative accumulation of RGNVs over time.

2.13. OVX-induced osteoporosis mouse model

A postmenopausal osteoporosis model was established by bilateral OVX in 6-week-old female C57BL/6 N mice. The mice were anesthetized (ketamine/xylazine) and ovaries were surgically removed via a small ventral incision. One week after surgery, OVX mice were randomly assigned to three groups using a computer-generated randomization schedule to minimize selection bias; (1) OVX groups (vehicle control, n = 10); (2) OVX mice + 17β-estradiol (E2, positive control, 0.1 mg/kg, n = 10); and (3) OVX + RGNVs (1 mg/kg, n = 10). E2 and RGNVs were orally administered once daily for 4 weeks. At the end of the 4-week treatment period, the mice were euthanized by CO₂ inhalation. Blood was collected, and femur samples were excised for analysis. The investigators administering the treatments were different from those performing the outcome assessments to maintain allocation concealment.

Excised femurs were fixed in 4% paraformaldehyde for 24 h at room temperature, then decalcified in 10% EDTA at 4 °C for 2 weeks. Decalcified femurs were embedded in paraffin and then cut into 5 µm thick sections for histological staining. Hematoxylin and eosin (H&E) staining was performed on the femur sections to assess bone morphology. The sections were stained with hematoxylin for 8 min, rinsed in water, counterstained with eosin for 2 min, dehydrated, and mounted for microscopic examination.

Tartrate-resistant acid phosphatase (TRAP) staining was performed on adjacent femur sections to identify the osteoclasts. After deparaffinization and rehydration, the sections were stained using a leukocyte acid phosphatase kit (Sigma-Aldrich) according to the manufacturer’s instructions. Sections were incubated with the TRAP staining solution at 37 °C for 1 h, rinsed, and counterstained with hematoxylin. TRAP-positive multinucleated cells (containing ≥ 3 nuclei) along bone surfaces were counted as osteoclasts under a light microscope.

2.14. Micro-CT analysis

Bilateral femurs isolated from female OVX mice treated with or without RGNVs were analyzed using micro-computed tomography (CT) to evaluate bone microarchitecture. Femurs were scanned using a Skyscan 1276 micro-CT system (Bruker, Belgium) at a resolution of ∼9.7 µm per slice. Approximately 120 cross-sectional slices were acquired, encompassing the distal femoral metaphysis (the region proximal to the growth plate). The reconstructed 3D images of the trabecular (cancellous) bone region (Tb) were generated using standard reconstruction algorithms. Bone morphometric parameters were calculated from the 3D region of interest, such as bone volume (BV, mm3); bone volume fraction (BV/TV, % of tissue volume); trabecular number (Tb.N, per mm); and trabecular separation (Tb.Sp, mm). Threshold and segmentation (sigma = 1.2, support = 2, threshold (Th) = 180) were applied uniformly to all samples to distinguish mineralized bone from marrow/soft tissue. The resulting indices were compared between the groups to quantify bone loss or preservation. The scanning and analysis of micro-CT were performed at the Chronic and Metabolic Diseases Research Center, Sookmyung Women’s University, South Korea.

2.15. Serum biochemical analysis (bone turnover markers)

Serum ALP and calcium (Ca) levels were measured to assess systemic indicators of bone metabolism. Serum samples from OVX control and OVX+RGNV mice were analyzed. ALP activity (a marker of bone formation and turnover) was measured using a colorimetric assay (similar to the ALP method described in Section 2.7) by monitoring the conversion of pNPP to pNP at 405 nm. The serum Ca concentration was determined using a commercially available assay slide (FUJI DRI-CHEM SLIDE) according to the manufacturer’s protocol. Elevated ALP and Ca serum levels in OVX mice relative to those in sham-operated control mice (CON) indicated high bone turnover associated with osteoporosis. Type I collagen (Novus Biologicals, Centennial, CO, USA) levels were quantified using a commercially available ELISA kit according to the manufacturer’s instructions.

2.16. In vitro osteoclast TRAP assay

To confirm osteoclast differentiation of mHSCs in vitro, TRAP staining was performed on cultured cells. mHSCs were induced toward the osteoclast lineage with M-CSF and RANKL (as described in Section 2.4) in the with or without of RGNVs. Cells were fixed and TRAP stained using acid phosphatase kit. TRAP-positive multinucleated cells (≥3 nuclei) were identified as osteoclasts and counted under a microscope (Leica Microsystems). To ensure unbiased quantification, fields were chosen using a systematic random sampling method using a coordinate grid overlay, and the observer was blinded to the experimental groups during counting. For TRAP enzymatic activity, culture supernatants were incubated with pNPP substrate solution, the reaction was stopped using 1 N NaOH, and the absorbance was measured at 405 nm (TECAN, Männedorf, Switzerland) to quantify TRAP activity released by osteoclasts. These vitro assays allowed the quantification of osteoclastogenesis and its inhibition by RGNVs.

2.17. Bone resorption assay

Bone resorption activity was evaluated using bone slice cultures. Differentiated osteoclasts were seeded onto bone slices in 24-well plates and cultured in the presence of RANKL (100 ng/ml), M-CSF (30 ng/ml), and RGNVs (10 µg) or vehicle control (0.1% DMSO). After 7 days of incubation, cells were removed from the bone surface by gentle mechanical agitation. Bone resorption was assessed by microscopic observation and 5% sodium hypochlorite.

2.18. F-actin ring staining

For actin ring analysis, osteoclasts differentiated in the presence or absence of RGNVs (10 µg/ml) were fixed with 4% paraformaldehyde for 15 min at room temperature. Cells were stained with rhodamine-conjugated phalloidin (R415; Thermo Fisher Scientific; 1:60 in PBS) for 20 min in the dark to visualize F-actin. After rinsing with PBS, fluorescence microscopy was used to examine actin ring formation.

2.19. Histomorphometry analysis

Histomorphometric analysis was conducted using femur bones from the treated mice. Femur bones were fixed in 4% paraformaldehyde (or 10% neutral formalin solution [Sigma-Aldrich, HT501128]) for 3 days, decalcified with 14% neutral-buffered EDTA (Sigma-Aldrich, ED-1) for 3 weeks, and embedded in paraffin. Osteoclasts were quantified in paraffin-embedded tissues stained with TRAP and methyl green (Tissue ProTech). Osteoclasts were identified as multinucleated TRAP-positive cells adjacent to the bone. Osteoblasts were quantified using paraffin-embedded tissues stained with hematoxylin (H08) and eosin Y (EY07) (Tissue ProTech). The terminology and units used for the histomorphometric analysis adhered to the recommendations of the Nomenclature Committee of the American Society for Bone and Mineral Research. Bone histomorphometric analysis was performed in a blinded manner, where the examiner was unaware of the treatment group identities during the quantification of osteoclasts and osteoblasts. Similarly, micro-CT reconstruction and analysis were conducted by a blinded operator to prevent observer bias, using a computerized semi-automated system (Osteomeasure Ver2.02, OsteoMetrics, TN, USA) with light microscopy. All tissue staining and histomorphometric analyses were performed at the Chronic and Metabolic Diseases Research Center at Sookmyung Women’s University, South Korea.

2.20. Statistical analysis

All experiments were performed at least in triplicate. For in vitro experiments, ‘n’ refers to the number of independent biological replicates. For in vivo experiments, ‘n’ refers to the number of individual animals. ‘n’ is denoted at each figure. Data are expressed as mean ± standard deviation (SD) for in vitro results or mean ± standard error of the mean (SEM) for in vivo animals. GraphPad Prism (GraphPad Software, San Diego, CA, USA) was used for statistical analyses. For comparisons among multiple groups, one-way or two-way analysis of variance (ANOVA), with Dunnett’s test was used. For two-group comparisons, Student’s t-test was used. A P value < 0.05 was considered statistically significant.

3. Results and discussion

3.1. Isolation and characterization of RGNVs

RGNVs were isolated from red ginseng (Fig. S1) using differential ultracentrifugation (Fig. 2A). TEM confirmed that the isolated RGNVs were spherical vesicles bound by a lipid bilayer membrane (Fig. 2B). The vesicles had a relatively uniform size distribution with an average diameter of ∼217 nm, as determined by NTA (Fig. 2C). TEM imaging established that the size and morphology of RGNVs were in line with those of plant extracellular vesicles reported in the literature [10]. Extracellular vesicle’s membrane proteins, namely CD81, CD63 and CD9, were found to be distributed on the lipid membrane of RGNVs as well (Fig. 2D). These characteristics are consistent with those of PENs. The yield of the RGNVs was approximately 27 mg vesicular protein per 1 g of starting red ginseng material (Fig. 2E).

Fig. 2.

Isolation and characterization of RGNVs. (A) Schematic representation of the ultracentrifugation-based isolation process of RGNVs; (B) TEM image of RGNVs, confirming their vesicular morphology. Scale bar: 500 nm; (C) Size distribution and particle concentration of RGNVs analyzed using NTA; (D) Distribution of extracellular vesicle’s marker proteins CD81 (purple), CD63 (yellow) and CD9 (blue) visualized by confocal microscopy; (E) Quantification of protein content in RGNVs extracted 1 g of red ginseng using the BCA protein assay; (F) SDS-PAGE analysis of RGNVs proteins, visualized by Coomassie blue staining; (G) KEGG pathway annotation of proteins identified in RGNVs; (H) Ginsenoside composition of RGNVs and GNVs determined using HPLC.

RGNVs are enriched in various proteins and ginsenosides derived from red ginseng. SDS-PAGE revealed distinct protein bands between 17 and 56 kDa (Fig. 2F), indicating a defined protein cargo profile. Comparative proteomic analysis of RGNVs and ginseng-derived vesicles (GNVs) identified a subset of proteins unique to RGNVs (40 proteins present only in RGNVs) (Table S4), and a larger set shared by both (114 proteins in common) (Table S5). These proteins spanned functional categories related to metabolism, stress response and cellular structure (Fig. S2A). Gene ontology (GO) analysis revealed that RGNV proteins are associated with catalytic activity and binding functions and contribute to cellular metabolic processes and structural components of cells (Fig. S2B). Notably, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis suggested that RGNV proteins are involved in stress-response signaling pathways (Fig. 2G), which could be relevant to their bioactivity in mammalian systems. High-performance liquid chromatography (HPLC) confirmed the presence of key ginsenosides such as such as Rg1, Rb1, Rc and Re, encapsulated within the RGNVs (Fig. 2H). However, RGNVs exhibited a markedly higher total ginsenoside content and uniquely contained characteristic red ginseng ginsenosides, including Rh1, RG3s, RG3r, Rk1 and Rg5 (Fig. 2H), reflecting a phytochemical profile enriched by the red ginseng steaming process. Interestingly, miRNA analysis revealed that while 2.57 µg miRNA was detected in GNVs, no measurable miRNA was found in RGNVs (Fig. S3), likely due to thermal degradation during red ginseng processing. Together, these findings suggest that RGNVs possess a distinct molecular composition that may underlie their therapeutic potential in modulating osteoblast and osteoclast activity.

3.2. RGNVs enhance osteogenic differentiation in vitro

We first examined whether RGNVs could promote osteoblast proliferation and differentiation in vitro using the MC3T3-E1 preosteoblast cell line (Fig. 3A). The fluorescently labeled RGNVs were efficiently internalized by MC3T3-E1 cells (Fig. S4A). Quantitative flow cytometry analysis confirmed that this uptake occurred in a concentration-dependent manner, showing a significant increase in mean fluorescence intensity (Fig. S4B). These results indicate the effective cellular uptake of RGNVs for intracellular cargo delivery. To elucidate the endocytic pathway involved in RGNV uptake, cells were pretreated with methyl-β-cyclodextrin (MβCD), an inhibitor of cholesterol/lipid raft-mediated endocytosis or chlorpromazine (CPZ), an inhibitor of clathrin-mediated endocytosis. While RGNV uptake was maintained in the MβCD-treated group, it was markedly reduced upon CPZ treatment, indicating that RGNVs are predominantly internalized via clathrin-mediated endocytosis (Fig. 3B). The MTT assay demonstrated that RGNV treatment had no cytotoxic effects on the MC3T3-E1 cells (Fig. 3C). In particular, low doses (1 and 5 µg/ml) tended to increase cell proliferation on Day 3 compared to untreated controls, although this increase was significant only at certain time points (Fig. 3C). These results indicate that RGNVs are well tolerated by osteoblasts and may modestly stimulate their growth.

Fig. 3.

RGNVs-mediated enhanced osteoblast differentiation in MC3T3-E1 cells. (A) Schematic representation of the experimental setup for MC3T3-E1 cells treated with RGNVs (1, 5 and 10 µg/ml) for 3 and 7 d to evaluate osteoblast activity; (B) Fluorescence microscopy images of MC3T3-E1 cells pretreated with MβCD or CPZ prior to the treatment of DID-labeled RGNVs; (C) MTT assay results assessing the cytotoxicity of RGNVs at different concentrations after 3 and 7 d of treatment; (D, E) Cellular and medium ALP activities after 3 and 7 d of RGNVs treatment; (F) Von Kossa staining of MC3T3-E1 cells treated with different concentrations of RGNVs for 3 and 7 d; (G) Alizarin Red staining of MC3T3-E1 cells following RGNVs treatment at various concentrations for 3 and 7 d; (H) Quantification of eluted Alizarin Red staining to measure mineralization levels. Data are presented as mean ± SD. Statistical significance was assessed using one-way ANOVA with Dunnett’s test (*P < 0.05, **P < 0.01, ***P < 0.001 vs. control; n = 3 per group).

To evaluate the osteogenic activity, MC3T3-E1 cells were cultured in osteogenic differentiation medium with or without RGNVs. ALP activity, an early marker of osteoblast differentiation [45], was measured in the cell lysates and culture supernatants. RGNV treatment significantly elevated the ALP activity in a dose- and time-dependent manner (Fig. 3D and 3E). After 7 d, the highest dose of RGNVs (10 µg/ml) increased intracellular ALP levels compared to those in controls (Fig. 3D), and also raised ALP levels released into the medium (Fig. 3E). These results indicate that RGNV exposure accelerated osteoblastic differentiation.

Next, we assessed matrix mineralization, a hallmark of late-stage osteoblast differentiation (Fig. 3F–3H). Von Kossa staining revealed that RGNV-treated cultures showed enhanced deposition of phosphate minerals (Fig. 3F). By Day 7, MC3T3-E1 cells treated with 5 or 10 µg/ml RGNVs showed markedly darker Von Kossa staining compared to untreated cells, reflecting greater mineralized matrix accumulation (Fig. 3F). Similarly, Alizarin Red S staining of calcium deposits demonstrated a dose-dependent increase in mineralized nodule formation following RGNV treatment (Fig. 3G). Qualitatively, RGNV-treated osteoblast cultures contained more and larger mineralized nodules (red-stained areas) than controls over the 7-day differentiation period (Fig. 3G). Quantitative extraction of Alizarin Red dye confirmed the significantly higher calcium deposition in the extracellular matrix in the RGNV group (Fig. 3H). Cells treated with native RGNVs exhibited the most pronounced mineralization. Notably, mineral deposition was also observed, albeit to a lesser extent, in the group treated with heat-inactivated RGNVs (Fig. S5). To examine this possibility, the osteogenic effects of ginsenoside Rg3 were directly compared with those of intact RGNVs, with Rg3 administered at a concentration equivalent to its content in the RGNV treatment condition. Although Rg3 treatment promoted osteoblast differentiation and mineral deposition relative to the control group, RGNV treatment induced a greater increase in ALP activity and mineralized nodule formation (Fig. S6A and S6B). Consistently, Von Kossa and Alizarin Red S staining revealed more robust matrix mineralization in RGNV-treated cultures than in Rg3-treated cultures (Fig. S6C–S6E). These results suggest that both heat-stable components (ginsenosides) and heat-labile components (proteins) of RGNVs may contribute to osteoblast.

To validate these findings in a more physiologically relevant system, we investigated the effects of RGNVs on primary osteoblasts derived from mHSCs, which were differentiated into osteoblasts with or without RGNVs (10 µg/ml) (Fig. 4A). Consistent with the MC3T3-E1 results, RGNVs were internalized in the primary mouse osteoblasts (Fig. S7), and significantly enhanced the osteogenic differentiation of the primary cells (Fig. 4B–4F). After 7 d, RGNV-treated primary osteoblasts exhibited higher intracellular and secreted ALP activities than the untreated cells (Fig. 4B and 4C). Mineralization assays mirrored these results (Fig. 4D–4F). Primary osteoblast cultures exposed to RGNVs showed more extensive calcium phosphate deposition by Von Kossa staining, as well as intensified Alizarin Red staining for calcium, relative to the controls (Fig. 4D and 4E). The amount of minerals (Alizarin Red-bound) was significantly higher in RGNV-treated cultures (Fig. 4F).

Fig. 4.

RGNVs promote osteoblast differentiation and inhibit osteoclast formation in vitro. (A) Schematic representation of the experimental design for RGNVs treatment (10 µg/ml) in primary mouse osteoblast; (B, C) Cellular and medium ALP activities after 7 d of RGNVs treatment; (D) Von Kossa staining of mHSCs treated with RGNVs for 3 and 7 d; (E) Alizarin Red staining of mHSCs treated with RGNVs for 3 and 7 d; (F) Quantification of eluted Alizarin Red staining to assess mineralization; (G) mRNA expression levels of osteoblast differentiation markers measured by qRT-PCR after 7 d of RGNVs treatment. GAPDH was used as an internal control; (H) Western blot analysis of osteoblast differentiation marker proteins with β-actin as the loading control; (I) Schematic representation of the experimental procedure for evaluating the effect of RGNVs on osteoclast differentiation. Primary mouse osteoclasts were treated with RGNVs at various concentrations (0, 1, 5, 10 µg/ml); (J) TRAP staining of primary osteoclast cultures following RGNVs treatment to assess osteoclast differentiation. Scale bar: 100 µm; (K) Quantification of TRAP activity released from TRAP+ multinucleated cells (MNCs) after RGNVs treatment; (L) Images of osteoclast resorption pits following RGNV treatment. Scale bar: 200 µm; (M) Fluorescence images of F-actin rings in osteoclasts following RGNV treatment. Scale bar: 200 µm. Data are presented as mean ± SD. Statistical significance was assessed using one-way ANOVA with Dunnett’s test (*P < 0.05, **P < 0.01, ***P < 0.001 vs. control; n = 3–5 per group).

Importantly, RGNV treatment also upregulated the expression of key osteogenic genes and proteins in primary osteoblasts (Fig. 4G and 4H). The mRNA levels of the master osteoblast transcription factor Runx2, as well as the major bone matrix proteins ALP, OPN and ProCOL1, were increased in RGNV-treated cells compared to controls (Fig. 4G). Consistently, the protein levels of Runx2, ALP, OPN and ProCOL1 were higher in RGNV-treated osteoblasts, as shown by Western blot analysis (Fig. 4H). These molecular changes corroborate the results of functional assays, indicating that RGNVs drive osteoblastic cells toward a mature, bone-forming phenotype. Taken together, our in vitro results demonstrated that RGNVs can enhance osteoblast differentiation and mineralized matrix formation, suggesting a pro-anabolic effect on bone tissue.

3.3. RGNVs inhibit osteoclastogenesis in vitro

Next, we examined the effects of RGNVs on osteoclast differentiation. Primary osteoclast precursors (obtained by inducing mHSCs with M-CSF and RANKL) were treated with RGNVs at various concentrations during their differentiation into osteoclasts. (Fig. 4I). The formation of multinucleated osteoclasts was evaluated using TRAP staining (Fig. 4J). In control cultures (with RANKL but without RGNVs), numerous TRAP-positive multinucleated osteoclasts were formed, as expected (Fig. 4J). In contrast, RGNV treatment reduced osteoclast formation in a dose-dependent manner. Cultures treated with 5 or 10 µg/ml RGNVs showed markedly fewer and smaller TRAP-positive multinucleated cells compared to untreated osteoclast cultures (Fig. 4J). Quantitative analysis confirmed that the RGNVs significantly suppressed osteoclastogenesis (Fig. 4J). In line with these observations, the enzymatic activity of TRAP released by osteoclasts into the medium was also significantly decreased by RGNV treatment in a concentration-dependent manner (Fig. 4K). In addition to suppressing osteoclast differentiation, RGNVs impaired osteoclast function. Resorption pit formation was visibly reduced following RGNV treatment, indicating attenuated bone-resorptive activity (Figs. 4L and S8). Furthermore, RGNV-treated cells exhibited disrupted and diminished F-actin ring formation compared with the RANKL-treated group, suggesting defective sealing zone organization in mature osteoclasts (Fig. 4M). These findings indicate that RGNVs not only inhibit osteoclast differentiation but also impair the cytoskeletal organization and resorptive activity of mature osteoclasts. These results indicated that RGNVs interfered with the differentiation of osteoclast precursors into bone-resorbing osteoclasts.

3.4. In vivo biodistribution of orally administered RGNVs

For a therapy to be effective in treating osteoporosis, it should ideally reach bone tissue after administration. To evaluate the in vivo biodistribution of RGNVs, we first assessed their stability under GI conditions. In simulated gastric fluid (pH 2.0) and intestinal fluid (pH 6.5), RGNVs exhibited an average diameter of 233.8 ± 4.7 nm and 216.2 ± 1.3 nm with a concentration of 2.05 × 10⁸ l and 3.01 × 10⁸ particles/ml, respectively, which is slight changes in size and reduced concentrations compared to the native RGNVs (229.5 ± 6.3 nm and 9.48 × 10⁸ particles/ml) (Fig. S9A and S9B). Nevertheless, RGNVs yielded after incubated under both conditions retained mineralization-promoting effects (Fig. S9C and S9D), confirming RGNVs physicochemical stability in GI environments.

Next, we examined the tissue distribution of RGNVs after oral administration. RGNVs labeled with the DiR fluorescent tracer were orally administered to healthy mice, and the major organs were imaged at 12-, 24- and 48-h post-administration (Fig. 5A). At 12 h, strong DiR fluorescence was detected in the GI tract (stomach and intestine), indicating that RGNVs survived gastric exposure and were in the process of being absorbed through the gut (Fig. 5A). After 24 h, fluorescence signals became evident in systemic organs, such as the liver, spleen, lung, kidney and notably bone (Fig. 5A). At 48 h, the RGNVs showed a broad distribution throughout the body, with particularly high accumulation in the bone tissue (Fig. 5A). Quantitative analysis of the DiR signal confirmed that the bone showed peak RGNV accumulation at 48 h, which was significantly higher than that in soft organs (Fig. 5B). The liver and spleen also showed uptakes, which was expected given their role in filtering nanoparticles, however, bone targeting was prominent. In particular, the sustained liver signal and the fluctuating accumulation in the small intestine (Fig. 5B) suggests the involvement of enterohepatic circulation [46,47]. This physiological recycling, well-documented for ginsenosides [48], likely prolongs the systemic availability of RGNVs, facilitating their sustained accumulation in bone tissue up to 48 h post-administration.

Fig. 5.

In vivo distribution of orally administered RGNVs. (A) Fluorescence images showing the biodistribution of DiR-labeled RGNVs in major organs, including the small intestine, colon, brain, lung, heart, liver, spleen, kidney, lymph node and bone, at 12-, 24- and 48-h post-administration in C57BL/6 mice; (B) Quantification of DiR fluorescence intensity in each tissue using Living Image 3.1 software. Data are presented as mean ± SEM (n = 3 per group). Statistical significance was assessed using a one-way ANOVA with Dunnett’s test. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, comparing the control group (CON; PBS) with the RGNVs-treated group (1 mg/kg) at different time points.

These results suggest that orally administered RGNVs can cross the GI barrier, enter the circulation, and preferentially localize to the bone. The GI absorption and bone deposition of RGNVs likely follow from the natural stability of plant nanovesicles in the digestive tract and possibly intrinsic bone-targeting properties of their components. The efficient homing of RGNVs to the bone tissue in vivo supports their potential use as bone-targeted delivery nanocarriers for therapeutic agents or active biological treatments for skeletal diseases.

3.5. RGNVs prevent bone loss in an OVX-induced osteoporosis mouse model

We evaluated the therapeutic efficacy of RGNVs in vivo using estrogen-deficient osteoporosis model. OVX female mice, which mimic postmenopausal osteoporosis, were orally administered RGNVs or control treatment for 4 weeks. The low-dose 17β-estradiol (E2) group served as positive control for anti-osteoporotic effect. One-week post-OVX, the mice received either vehicle (OVX group), E2 (0.1 mg/kg, three times weekly), or RGNVs (1 mg/kg, three times weekly) as previous study [14] for 4 weeks, after which bone analyses were performed (Fig. 6A). Based on this biodistribution profile, in which the DiR fluorescence in bone peaked at 48 h and gradually declined after 3 days, RGNVs were administered three times per week in the subsequent in vivo experiments to maintain effective vesicle exposure [14].

Fig. 6.

Anti-osteoporotic effects of RGNVs in an OVX-induced osteoporosis model. (A) Schematic representation of the experimental timeline for RGNVs administration in the OVX-induced osteoporosis mouse model. RGNVs were orally administered three times per week for 4 weeks, followed by sample collection at 9 weeks; (B) Micro-CT images of the metaphyseal trabecular bone structure in the distal femur across different treatment groups. Scale bar: 1 mm; (C) Quantification of key bone parameters, including BV, BV/TV, Tb.N and Tb.Sp in femur tissues; (D) Histological analysis of femoral sections using H&E staining to assess trabecular bone integrity. Scale bars: 500 µm (top), 200 µm (bottom); (E) Quantification of osteoblast activity, including the N.Ob/B.Pm and Ob.S/BS; (F) Measurement of serum ALP and calcium levels across different groups; (G) TRAP staining of femoral sections from OVX-induced osteoporosis mice treated with RGNVs. Scale bars: 100 µm; (H) Quantification of osteoclast activity in femoral sections, including N.Oc/B.Pm and Oc.S/BS; (I) Serum CTX-1 levels measured by ELISA. Data are presented as mean ± SEM (n = 3–8 per group). Statistical significance was assessed using one-way ANOVA with Dunnett’s test. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, comparing between the control (CON) and RGNVs-treated groups.

As expected, OVX-only mice exhibited significant bone loss compared to sham-operated controls. In contrast, RGNV-treated OVX mice showed preserved bone structure and mass (Fig. 6B and 6C). Micro-CT analysis of the distal femurs revealed that OVX caused substantial deterioration of trabecular bone architecture. Compared to CON groups, OVX mice had lower trabecular bone volume and connectivity, characterized by a reduced BV/TV, decreased Tb.N, and increased Tb.Sp (Fig. 6B and 6C). Interestingly, the RGNV administration prevented these osteoporotic changes. RGNV-treated OVX femurs showed markedly higher trabecular bone volume and density than untreated OVX femurs. Quantitatively, RGNV treatment increased cancellous bone volume (BV and BV/TV) and Tb.N to levels comparable to or slightly lower than those in the healthy CON group, and was significantly greater than that in the OVX group (Fig. 6B and 6C). Concurrently, Tb.Sp was significantly reduced in RGNV-treated bones (indicating denser trabecular packing) compared to that in the OVX controls. These micro-CT results demonstrate that RGNV therapy mitigated OVX-induced bone loss and preserved trabecular bone microarchitecture.

Histological analysis further supported the bone-protective effects of RGNVs in vivo (Fig. 6D). H&E-stained sections of the femoral trabecular bone showed that OVX control mice had sparse, thin trabeculae with few osteoblasts lining the bone surfaces (Fig. 6D). In contrast, RGNV-treated OVX mice retained a greater number of trabeculae, with more osteoblastic cells visible on the bone surfaces (Fig. 6D). Osteoblast parameters were quantified by histomorphometry (Fig. 6E). The number of osteoblasts per bone perimeter (N.Ob/B.Pm) and the osteoblast surface relative to the bone surface (Ob.S/BS) were significantly higher in RGNV-treated mice than in OVX controls (Fig. 6E). These data indicated that RGNV treatment enhanced the presence and activity of osteoblasts on bone surfaces in vivo, which is consistent with our in vitro findings.

RGNV-mediated improvements in bone morphology were accompanied by normalization of serum bone turnover markers [49,50]. Serum ALP activity and calcium levels were elevated in OVX mice relative to those in CON mice, reflecting high bone turnover and loss after OVX (Fig. 6F). RGNV treatment significantly reduced these levels (Fig. 6F). RGNV-treated OVX mice showed lower serum ALP and calcium levels than untreated OVX mice, which were similar to the levels observed in non-OVX controls (Fig. 6F). Specifically, ALP, which was high in OVX rats due to excessive bone remodeling, was reduced by RGNV administration, and serum calcium, which tends to increase due to osteoclastic bone resorption, was also reduced. This suggests that RGNVs helped restore a balanced bone turnover rate.

Notably, RGNV treatment reduced the osteoclast activity in vivo. TRAP staining of femur sections revealed an abundance of osteoclasts on the trabecular surfaces of OVX control mice, whereas fewer TRAP-positive osteoclasts were observed in the RGNV-treated group (Fig. 6G). The histomorphometric indices of osteoclasts were greatly improved by RGNV therapy (Fig. 6H). The N.Oc/B.Pm and Oc.S/BS in RGNV-treated mice were significantly reduced compared to those in OVX controls (Fig. 6H). Moreover, the eroded surface (ES/BS), a measure of the bone resorption area, was significantly lower in RGNV-treated bones (Fig. S10). In addition, serum CTX-1 levels, a biochemical marker of bone resorption, were elevated in OVX mice but were reduced by RGNVs administration, indicating suppression of osteoclast- mediated bone resorption in vivo (Fig. 6I). The eroded surface (ES/BS), bone area (B. Ar), and bone perimeter (B.Pm) were also significantly restored to normal levels compared with the reduced values in the OVX group (Fig. S10). These findings confirmed that RGNVs suppressed osteoclast-mediated bone resorption in vivo, complementing their stimulation of osteoblast-mediated bone formation.

Collectively, the in vivo results demonstrated that the oral administration of RGNVs effectively prevented estrogen deficiency-induced bone loss. RGNV-treated osteoporotic mice maintained greater bone mass and microarchitectural integrity, with increased osteoblast and decreased osteoclast indices, relative to untreated OVX mice. The therapeutic effects of RGNVs were comparable to those of low-dose estrogen in this model. Notably, the oral administration of RGNVs achieved these benefits, which is advantageous for clinical translation. Thus, RGNVs exert dual bone-protective actions in vivo, making them promising candidates for osteoporosis therapy.

3.6. Mechanistic insights: RGNVs modulate osteogenic and osteoclastogenic signaling pathways

To elucidate the molecular mechanisms underlying the dual effects of RGNVs on bone cells, we performed additional analyses of osteogenic and osteoclastogenic signaling pathways in the presence of RGNVs. Given the pronounced effect of RGNVs on osteoblast differentiation, we investigated whether RGNVs could activate known bone anabolic signaling cascades. RNA-seq profiling of RGNV-treated osteoblasts (MC3T3-E1) and untreated cells was performed to identify differentially expressed proteins. A scatter plot of gene expression revealed numerous genes that were significantly upregulated or downregulated by RGNV treatment (Fig. 7A). GO classification of these genes showed enrichment in categories related to bone formation. Notably, RGNV-treated cells had increased levels of genes associated with the positive regulation of bone mineralization and osteoblast differentiation (GO terms), supporting the concept that RGNVs trigger pro-osteogenic pathways (Figs. 7B and S11). Pathway analysis further indicated that RGNV exposure influenced signaling pathways such as PI3K-Akt and MAPK (Figs. S11 and S12) that are involved in cell growth and differentiation, respectively.

Fig. 7.

BMP-2/Smad signaling in the RGNVs-mediated enhanced osteoblast differentiation in MC3T3-E1 cells. (A) Scatter plot illustrating log2-transformed gene expression data for the control (CON) group (X-axis) and the 10 µg/ml RGNVs-treated group (Y-axis). Genes with significantly increased expression are marked in red, while those with decreased expression are in blue; (B) GO enrichment analysis of RGNV-targeted genes using DAVID bioinformatics tool. The top 10 GO terms in the biological process category are presented as a bar chart ranked by –log(P-value). (I: Positive regulation of transcription by RNA polymerase II; II: Response to endoplasmic reticulum stress; III: Cellular response to lipopolysaccharide; IV: Positive regulation of gene expression; V: Bone mineralization; VI: Positive regulation of osteoblast differentiation; VII: Negative regulation of transcription by RNA polymerase II; VIII: Angiogenesis; IX: Positive regulation of cell migration; X: Positive regulation of cell population proliferation); (C) mRNA expression levels of BMP-2/Smad signaling markers in MC3T3-E1 cells treated with RGNVs at different concentrations (1, 5 and 10 µg/ml) for 3 and 7 d, as measured by qRT-PCR. GAPDH was used as an internal control; (D) Protein expression levels of BMP-2/Smad signaling markers in MC3T3-E1 cells following RGNVs treatment at different concentrations (1, 5 and 10 µg/ml) for 3 and 7 d, analyzed by Western blot. β-actin was used as the loading control; (E–F) Cellular ALP activity and culture medium ALP activity in osteoblasts treated with RGNVs in the presence or absence of a BMP pathway inhibitor; (G-H) Representative images of Alizarin Red S and Von Kossa staining in osteoblasts treated with RGNVs in the presence or absence of a BMP-2 inhibitor. Data are presented as mean ± SD. Statistical significance was assessed using one-way or two-way ANOVA with Dunnett’s test (*P < 0.05, **P < 0.01, ***P < 0.001 vs. control; n = 3 per group).

Focusing on osteogenic growth factors (Figs. S11 and S12), we found that RGNVs induced the BMP-2 pathway (Fig. 7), a key driver of osteoblast differentiation. In MC3T3-E1 osteoblasts, RGNV treatment led to a significant increase in BMP-2 gene expression on Day 3 and 7 compared to that in untreated cells (Fig. 7C). Concurrently, the mRNA levels of the downstream transcription factors Smad1, Smad5 and Smad8 were elevated in RGNV-treated cells (Fig. 7C). These Smad proteins are intracellular mediators of BMP signaling that promote Runx2 activation and osteogenesis. Western blotting confirmed that RGNV treatment upregulated BMP-2 protein expression and enhanced the activation (phosphorylation) of Smad1/5/9 in osteoblasts relative to the controls (Fig. 7D). To determine whether this signaling axis is required for RGNV-mediated osteogenesis, a BMP pathway inhibitor was applied during osteoblast differentiation. Inhibition of BMP-2 markedly suppressed the RGNV-induced increase in cellular and secreted ALP activity (Fig. 7E and 7F). Consistently, RGNV-enhanced mineralization, visualized by Alizarin red and Von Kossa staining, was also diminished in the presence of the inhibitor (Fig. 7G and 7H). Together, these findings support that RGNVs promote osteoblast differentiation and matrix mineralization through the BMP-2/Smad pathway. This resulted in the upregulation of Runx2, ALP, OPN and ProCOL I at both mRNA and protein levels (Fig. 8A and 8B). In addition, RGNVs modulated MAPK-related signaling in both MC3T3-E1 cells and primary osteoblasts, as evidenced by changes in the phosphorylation of MAPK-associated proteins (Fig. S13A). Consistent with these signaling changes, RGNV treatment also increased the mRNA expression of osteogenesis-related genes in both cell types (Fig. S13B and S13C). These findings further support that RGNVs promote osteoblastic differentiation through coordinated regulation of multiple osteogenic signaling pathways. A similar trend was observed in primary osteoblast cultures. RGNVs increased the BMP-2/4 and phospho-Smad1/5/8 levels (Fig. S14), which correlated with elevated Runx2 and other differentiation markers expressions (Fig. 4G and 4H). These results indicate that RGNVs activate the BMP-2-Smad-Runx2 signaling pathway in osteoblasts, explaining, at least in part, the mechanism by which they enhance osteogenic differentiation. This was evidenced by increased BMP-2 expression and Smad1/5/8 phosphorylation upon RGNV treatment, agreeing with previous studies, which showed that certain ginsenosides can stimulate osteoblast activity via BMP-2 signaling [28,51,52]. For example, ginsenosides Rg1 and Rd have been reported to enhance osteoblast differentiation by upregulating BMP-2 and downstream osteogenic genes​ [26]. Our findings align with and extend these reports by showing that the collective cargo of red ginseng vesicles can elicit similar pro-osteogenic effects. The vesicular form may improve the stability and delivery of these compounds to the target cells.

Fig. 8.

RGNVs enhance osteogenic gene and protein expression in MC3T3-E1 cells. (A) mRNA expression levels of osteoblast differentiation markers analyzed by qRT-PCR after treatment with RGNVs for 3 and 7 d GAPDH was used as an internal control; (B) Western blot analysis of osteoblast differentiation marker proteins with β-actin as a loading control after 3 and 7 d of RGNVs treatment. Data are presented as mean ± SD. Statistical significance was assessed using one-way ANOVA with Dunnett’s test (*P < 0.05, **P < 0.01, ***P < 0.001 vs. control; n = 3 per group).

To understand the mechanism of osteoclast inhibition, we assessed the expression of key osteoclastogenic markers and signaling molecules. Western blotting was performed using RANKL-stimulated osteoclast precursors, with or without RGNVs. Mechanistically, RANKL binding to its receptor RANK triggers cascades like NF-κB and MAPK, culminating in the activation of two transcription factors that are the master regulators of osteoclast differentiation, c-Fos and NFATc1, essential for osteoclast gene expression. Indeed, RGNV exposure attenuated the activation of the major signaling pathways involved in osteoclastogenesis (Fig. 9). Phosphorylation of ERK1/2, JNK and p38 MAPKs, as well as phosphorylation of IκBα (a readout of NF-κB pathway activation), were all lower in RGNV-treated cultures compared to RANKL-only controls (Fig. 9A). This resulted in diminished nuclear NF-κB activity. RGNV treatment markedly downregulated the protein expression of NFATc1 and c-Fos (Fig. 9B). NFATc1 auto-amplifies its expression during osteoclastogenesis, however, RGNVs appear to interrupt this auto-amplification loop. Cathepsin K, a late-stage osteoclast marker and bone-degrading enzyme, was also reduced in RGNV-treated cells (Fig. 9B), indicating a downstream effect on the attenuated NFATc1/c-Fos activity. Consistent with these molecular findings, RGNV treatment reduced the number of TRAP-positive multinucleated osteoclasts, whereas co-treatment with NF-κB/MAPK pathway activators partially reversed this inhibitory effect (Fig. 9C). Likewise, the decrease in TRAP activity induced by RGNVs was significantly attenuated by pathway activation (Fig. 9D), further supporting that RGNVs suppress osteoclast differentiation and function through inhibition of NF-κB/MAPK signaling. The suppression of NF-κB/MAPK signaling by RGNVs is consistent with prior studies on ginsenosides [53,54]. For instance, ginsenoside Rh2 is reported to inhibit osteoclast differentiation by down-regulating NF-κB, NFATc1, and c-Fos [44], and ginsenoside Rg3 has been shown to inhibit RANKL-induced JNK and p38 activation in osteoclast precursors​ [55]. RGNVs contain a mixture of saponins, which likely act synergistically to produce a broad inhibitory effect on osteoclast signaling networks. Our data suggest that RGNVs effectively recapitulate the anti-resorptive actions of individual ginsenosides, but potentially, with enhanced potency, owing to the multicomponent nature of the vesicles. In summary, the in vitro data demonstrated that RGNVs inhibited osteoclast formation by blocking the RANKL-driven differentiation process, as evidenced by fewer TRAP-positive cells and reduced expression of osteoclast-specific genes/proteins.

Fig. 9.

RGNVs-mediated inhibition of osteoclast activity mHSCs. (A) Western blot analysis of osteoclast differentiation markers after treatment with different concentrations of RGNVs, with Actin as a loading control; (B) Western blot analysis of osteoclast activation markers following RGNVs treatment for 0–3 d, with Actin as a loading control; (C) Images of TRAP staining in osteoclasts treated with RGNVs in the presence or absence of NF-κB/MAPK pathway activators. Scale bar: 200 µm; (D) TRAP activity in osteoclasts treated with RGNVs in the presence or absence of NF-κB/MAPK pathway activators. Data present as mean ± SD (n = 3–5 per group). Statistical significance was assessed using an one-way or two-way ANOVA with Dunnett’s test. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, comparing the control group (CON) with the RGNVs-treated group.

Collectively, RGNVs mechanistically promoted bone formation and inhibited bone resorption by modulating critical signaling pathways in both cell types. In osteoblasts, RGNVs activated BMP-2/Smad/Runx2 signaling, driving the expression of osteogenic genes and enhancing matrix mineralization. In osteoclast precursors, RGNVs inhibited the RANKL–NF–κB/MAPK signaling axis, thereby preventing the induction of c-Fos and NFATc1 essential for osteoclast differentiation. This dual modulation of anabolic and catabolic pathways in bone cells underlies the capacity of RGNVs to rebalance bone remodeling in favor of bone formation.

While our study identified BMP-2/Smad and NF-κB/MAPK as major pathways affected by RGNVs, the exact molecular components within RGNVs responsible for these actions remain unidentified. RGNVs contain a complex cargo, such as multiple ginsenosides, proteins, and oligonucleotides, and it would be informative to determine which specific cargo molecules or combinations drive osteogenic and anti-osteoclastogenic effects. Unpacking these questions requires further fractionation of the RGNV content and testing of the isolated fractions.

3.7. Biocompatibility and safety of RGNVs

An important consideration for any new therapeutic, especially one derived from natural sources, is its safety profile (Fig. 10A). In the short-term 7-d oral toxicity test, no adverse clinical signs were observed in RGNV-treated mice, and their body weights remained stable (data not shown). Histopathological examination of the major organs (heart, lung, liver, kidney and brain) did not reveal any abnormalities in the RGNV-treated animals (Fig. 10B). H&E-stained sections showed normal tissue architecture with no evidence of inflammation, necrosis, or other pathologies, comparable to those in the PBS-treated control mice. This indicated that the RGNVs did not cause overt organ damage at the administered dose.

Fig. 10.

Toxicity evaluation for RGNVs in vivo. (A) Schematic representation of the experimental design for toxicity assessment following oral administration of RGNVs for 7 d; (B) H&E staining images of major organs, including the heart, lung, liver, kidney and brain that are dissected from the control (CON) and RGNVs-treated groups. Scale bar: 200 µm; (C) Serum biochemical analysis of liver toxicity markers (ALT, AST and TG) and kidney toxicity markers (HDL-C, BUN, creatinine and TBIL) in the CON and RGNVs-treated groups; (D) Serum biochemical analysis of systemic pro-inflammatory markers (TNF-α and IL-1β). Data are presented as mean ± SEM (n = 3 per group). In the t-test, no significant differences were observed between groups.

We also used a panel of serum biochemical markers to detect subclinical toxicity. The levels of ALT, AST and TG in the RGNV-treated mice were not significantly different from those in the controls, suggesting no hepatotoxicity (Fig. 10C). Similarly, kidney function markers, such as BUN and creatinine, and other metabolic indicators, such as HDL-C and TBIL, showed no significant alterations following RGNV treatment (Fig. 10C). Furthermore, to assess potential immunotoxicity, serum inflammatory cytokines were quantified. RGNV administration did not induce elevated levels of TNF-α and IL-1β (Fig. 10D), confirming the absence of a systemic immune response. That is, all values in the RGNV group were within the normal range when observed in the control group. These findings were confirmed by quantitative comparisons, which showed no significant differences between the RGNV-treated and control animals on any of the measured parameters.

Overall, these evaluations indicated that RGNVs are biocompatible and do not elicit detectable toxicity in mice at therapeutically effective doses. The absence of systemic toxicity suggests that RGNVs have a favorable safety profile for potential translational use. This is consistent with reports of other plant-derived nanovesicles that have generally shown minimal toxicity or immunogenicity in animal models [[12], [13], [14],55].

The excellent biocompatibility of RGNVs can be attributed to their natural origin and the fact that humans have long consumed ginseng as food and medicines. Unlike synthetic nanoparticles, plant vesicles are composed of natural lipids and proteins that are likely to be recognized and processed by the body without eliciting harmful responses. This biocompatibility, combined with our efficacy data, suggests that the RGNVs are a promising and safe therapeutic modality for treating chronic osteoporosis.

Compared with conventional osteoporosis treatments, RGNVs offer several potential advantages. Current anti-resorptive drugs such as bisphosphonates and denosumab effectively reduce osteoclast activity, but do not promote new bone formation and can have side effects such as GI irritation, atypical femoral fractures or osteonecrosis of the jaw with long-term use [6]. Conversely, anabolic agents such as teriparatide and abaloparatide build bones but require daily injections and carry risks such as hypercalcemia [56]. While Romosozumab represents a clinical breakthrough as a dual-action agent, its associated cardiovascular risks limit its application in certain high-risk patient populations. Furthermore, the transition from anabolic to anti-resorptive therapy in sequential treatment regimens adds significant complexity to patient management. In contrast, RGNVs provide a natural dual-action approach using a single agent capable of simultaneously stimulating osteoblasts via the BMP-2/Smad pathway and suppressing osteoclasts via the RANKL-NF-κB/MAPK pathway. This could result in a net gain in the bone with a superior safety profile. Indeed, the RGNVs are derived from a natural source and demonstrate no detectable toxicity in our models, suggesting that they could be used long-term with a lower risk profile. Another important aspect of RGNVs is their oral efficacy. Many osteoporosis biologics currently require injections, whereas RGNVs can be formulated as an oral supplement or functional food ingredient, greatly improving patient compliance. While we observed no acute toxicity (4 weeks of treatment in mice), comprehensive long-term immunogenicity profiles and regulatory frameworks for RGNV require further establishment before clinical application.

Despite promising findings, the clinical translation of exogenous extracellular vesicles faces significant challenges. Key hurdles include the lack of standardized isolation methods, difficulties in large-scale production (yield), and potential source-to-batch heterogeneity. In this study, we addressed heterogeneity by utilizing standardized Red Ginseng roots processed according to ISO 19610. However, current isolation techniques like ultracentrifugation are labor-intensive and may limit scalability. Future studies should explore tangential flow filtration (TFF) or other scalable technologies to overcome these production barriers.

4. Conclusion

This study demonstrates that RGNVs enhance osteoblast differentiation and mineralization while simultaneously suppressing osteoclast differentiation and bone resorption. RGNVs exert these dual effects primarily by activating pro-osteogenic signaling (BMP-2/Smad/Runx2 pathway) and inhibiting osteoclastogenic signaling (RANKL-induced NF-κB/MAPK pathways). Orally administered RGNVs accumulated in bone tissue in vivo, significantly restoring the trabecular bone volume and microarchitecture in an OVX-induced osteoporosis mouse model, and exhibited no systemic toxicity. These findings suggest that RGNVs are a promising biocompatible therapeutic agent for osteoporosis, with a dual mechanism of action that promotes bone formation and inhibits bone resorption. Further research is warranted to optimize their bioavailability and evaluate their efficacy in long-term and clinical settings; although, RGNVs have a strong potential as a natural alternative or complement to conventional osteoporosis treatments.

CRediT authorship contribution statement

Dong-ha Kim: Writing – original draft, Visualization, Methodology, Data curation, Conceptualization. Se Hwan Mun: Writing – original draft, Visualization, Methodology, Investigation, Conceptualization. Jae-Hee Kwon: Methodology, Data curation. Bum-Joon Koo: Investigation. Hyun-Woo Kim: Methodology, Data curation. Sang-Hun Choi: Investigation, Data curation. Kwang-Won Seo: Writing – original draft, Supervision. Young Yang: Writing – original draft, Visualization, Supervision. Yong Yook Lee: Investigation. Jin Sun Jung: Investigation. Do-Kyun Kim: Writing – original draft, Supervision, Conceptualization. Moon-Chang Baek: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Jihoon Kim: Writing – review & editing, Writing – original draft, Visualization, Supervision. Young-Eun Cho: Writing – review & editing, Writing – original draft, Supervision, Project administration, Conceptualization.

Conflicts of interest

The authors declare that there is no conflicts of interest.