GLP-2-carrying exosomes alleviate osteoporosis by upregulating miR-378a-3p to inhibit osteoclastic differentiation and NF-κB-MAPK pathway

Abstract

Background

Osteoporosis, characterized by excessive osteoclast activity, remains a major health challenge. This study investigated the therapeutic potential of exosomes derived from glucagon-like peptide-2 (GLP-2)-overexpressing macrophages in osteoporosis, with specific focus on miR-378a-3p-mediated regulation of osteoclastogenesis and inflammatory signaling pathways.

Methods

GLP-2 was overexpressed in mouse RAW264.7 cells. Exosomes were extracted and characterized. In vitro, effects on osteoclast differentiation were assessed using Cell Counting Kit-8, tartrate-resistant acid phosphatase staining, and quantitative real-time polymerase chain reaction. Differentially expressed miRNAs were screened via bioinformatics analysis. In vivo, an osteoporosis rat model was induced by bilateral ovariectomy. Histopathological evaluation of femoral tissue was performed via Hematoxylin-Eosin and tartrate resistant acid phosphatase staining. Bone microarchitecture, inflammatory cytokines, and signaling pathways were evaluated using micro-CT, enzyme-linked immunosorbent assay, and western blotting. Moreover, miR-378a-3p antagonist was applied to investigate the role of miR-378a-3p in osteoporosis.

Results

GLP-2-carrying exosomes inhibited osteoclast differentiation in vitro and improved bone microarchitecture in vivo. Bioinformatics analysis identified miR-378a-3p as a hub miRNA and miR-378a-3p was downregulated in osteoporosis. In vivo, GLP-2-carrying exosomes promoted miR-378a-3p level and inhibited osteoclast differentiation. MiR-378a-3p mimics inhibited the NF-κB/MAPK signaling, and the direct binding specificity between miR-378a-3p and TRAF6 was validated via dual-luciferase reporter assay. Mechanistically, GLP-2-carrying exosomes enhanced bone parameters, suppressed inflammation and NF-κB/MAPK signaling pathway, which were reversed by miR-378a-3p antagonist.

Conclusion

GLP-2-carrying exosomes showed promise in treating osteoporosis by inhibiting osteoclast differentiation and modulating NF-κB/MAPK pathways via miR-378a-3p regulation. This novel axis presents a promising dual-targeted strategy for osteoporosis treatment, with engineered exosomes offering significant translational potential for bone regenerative therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-025-06119-x.

Introduction

Osteoporosis, a systemic skeletal disorder characterized by diminished bone mineral density and deterioration of bone microarchitecture, significantly elevates fracture susceptibility, particularly in postmenopausal women and aging population [1–3]. The pathogenesis of this disease primarily stems from an uncoupling of bone remodeling processes, wherein osteoclast-mediated bone resorption exceeds osteoblast-dependent bone formation [4]. This imbalance is driven by multifactorial etiological mechanisms, including estrogen deficiency (accelerated bone loss in menopause), genetic polymorphisms, and modifiable risk factors [5, 6]. Globally, osteoporosis imposes a significant burden on public health, leading to disability, death, and socioeconomic costs [7]. Current treatment strategies include pharmacotherapy (such as bisphosphonates and hormone therapy), physiotherapy, and lifestyle modifications [8, 9]. Furthermore, novel therapies are continuously being explored with the aim of reducing fracture risks and improving patient outcomes [10]. Thus, it is necessary to seek other potential therapeutic targets.

Exosomes are small vesicles secreted by cells, carrying bioactive molecules such as proteins and nucleic acids, and are involved in intercellular communication and material exchange [11], which also exert beneficial effects in the fields of regenerative medicine and disease therapeutics [12, 13]. Studies have found that exosomes can carry signaling molecules derived from osteoblasts or osteoclasts, regulating the proliferation, differentiation, and function of bone cells, thereby affecting bone metabolism balance. Exosome-like nanovesicles derived from Pueraria lobata alleviate osteoporosis by enhancing autophagy [14]. Exosome-mediated targeted delivery of Nrf2 to chondrocytes can reduce degeneration of the cartilage endplate [15]. C2C12-derived exosomes can promote the proliferation and osteogenic differentiation of bone marrow mesenchymal stem cells through the miR-92a-3p/PTEN/AKT signaling pathway, alleviating the progression of osteoporosis in mice [16]. Therefore, exosomes play a significant role in osteoporosis.

Glucagon-like peptide-2 (GLP-2), an enteroendocrine peptide hormone primarily secreted by intestinal L-cells, is well-characterized for its trophic effects on intestinal epithelial proliferation and enhancement of mucosal barrier integrity via upregulating tight junction proteins [17]. Specific single-nucleotide polymorphisms (SNPs) within the GLP-1R gene locus demonstrate robust association with elevated susceptibility to postmenopausal osteoporosis [18]. Beyond its canonical gastrointestinal functions, emerging preclinical studies have illuminated the pleiotropic role of GLP-2 in skeletal homeostasis, positioning it as a promising therapeutic candidate for metabolic bone disorders, particularly postmenopausal osteoporosis [19]. Notably, in ovariectomized (OVX) rats, GLP-2 has been shown to improve trabecular microarchitecture, reduce inflammatory cytokines, and balance bone turnover markers by inhibiting osteoclast activity and promoting osteoblast function [20]. Mechanistically, GLP-2 can induce osteoclast apoptosis via the TGF-β-Smad2/3 signaling pathway and inhibit the activation of NF-κB/MAPK, highlighting its dual anti-resorptive and anti-inflammatory properties [20]. However, the molecular mechanisms underlying the role of GLP-2 in osteoporosis have not been fully elucidated.

Moreover, an increasing body of evidence indicates that epigenetic regulators such as microRNAs (miRNAs) play crucial roles in the mechanisms underlying osteoporosis [21]. It has been reported that miR-378a-3p represses adipogenic differentiation while promoting osteogenesis in skeletal stem/progenitor cells [22]. Herein, we investigated the impact of exosomes overexpressing GLP2 on osteoporosis through in vivo and in vitro experiments, and the connection between miR-378a-3p and GLP-2-carrying exosomes was explored, elucidating the regulatory mechanism in osteoporosis and providing new therapeutic targets.

Materials and methods

Cell grouping and treatment

The murine monocyte RAW264.7 cell line was obtained from iCell (Shanghai, China) and cultured in ‌Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum. Cells were split into three groups: Control group (untreated cells, baseline RANKL response), LV-NC group (transfected with an empty lentiviral vector pSLenti, empty vector control) and LV-GLP-2 group (transfected with the recombinant plasmid pSLenti-GLP-2 for GLP-2 overexpression) (n ≥ 3). Empty lentiviral vector pSLenti and recombinant plasmid pSLenti-GLP-2 were purchased from OBiO Tech, Inc (Shanghai, China). All groups were co-cultured with complete medium containing receptor activator of nuclear factor-κB ligand (RANKL, 4 ng/mL, Sigma-Aldrich, St. Louis, MO, USA) for 5 days. Culture supernatants were collected for exosome isolation.

Exosome isolation

Supernatants from each group were centrifuged and treated with Exosome Concentration Solution (MedChemExpress, Shanghai, China). After vortexing and incubation at 2–8 °C for 2 h, samples were ultracentrifuged at 10,000 ×g for 60 min. The pellet containing exosomes was resuspended in PBS and centrifuged again at 12,000 ×g at 4 °C for 2 min. The final supernatant was enriched with exosomes.

Exosome size distribution analysis

Nanoparticle Tracking Analysis (NTA) was employed to determine the hydrodynamic diameter and concentration of exosomes. Prior to measurement, the sample chamber was sequentially rinsed with deionized water and PBS buffer to minimize contamination. The instrument was calibrated using 110 nm polystyrene microspheres (PS beads) to ensure measurement accuracy. Exosome samples were diluted in PBS to avoid signal saturation. The ZetaView system, maintained at a controlled temperature range of 23–30 °C during analysis, recorded and analyzed NTA measurements across 11 distinct positions to ensure statistical robustness and minimize positional bias.

Transmission electron microscopy (TEM)

Exosomes were visualized using negative staining TEM to assess their ultrastructural integrity. Briefly, purified exosomes were mixed with 4% paraformaldehyde and deposited onto Formvar-coated copper grids (film side facing downward). Grids were rinsed with PBS to remove unbound particles, followed by sequential fixation with 1% glutaraldehyde and extensive washing with ddH₂O. For contrast enhancement, grids were subjected to uranyl oxalate negative staining on ice, then embedded in methylcellulose to stabilize the sample. Excess liquid was blotted using filter paper, and grids were air-dried overnight. Imaging was performed under a transmission electron microscope at an accelerating voltage of 80 kV, capturing characteristic exosomal morphology.

Cell counting Kit-8 (CCK-8) assay

CCK-8 (Beyotime, Shanghai, China) was applied to measure cell proliferation (n ≥ 3). RAW264.7 cells were cultured in 96-well plates, treated with exosomes from each group, and RANKL was added to induce osteoclast differentiation. Three replicates were set up for each group, with blanks on the periphery to prevent water loss. After 5 days, 10 µL CCK-8 reagent was added to each well, followed by incubation at 37 °C and 5% CO₂ for 2 h. Absorbance was measured at 450 nm using a microplate reader.

Tartrate resistant acid phosphatase (TRAP) staining

TRAP staining was performed using a TRAP Staining Kit (Servicebio, Wuhan, China) (n ≥ 3). RAW264.7 cells cultured for 5 days were rinsed with the fixative solution from the kit for 30 s, followed by thorough washing with deionized water to avoid excessive drying. The TRAP working solution was prepared by mixing acetate buffer, hexazotized pararosaniline (a 1:1 mixture of sodium nitrite and pararosaniline solution), and naphthol AS-BI phosphate solution. TRAP staining solution was then added to the wells (2 mL/well) in a 6-well plate. The plate was wrapped in aluminum foil and incubated at 37 °C for 1 h in the dark. After incubation, the plate was rinsed three times with ultrapure water and air-dried before microscopic observation.

Microarray dataset analysis

The gene expression profiles analyzed in this study were downloaded from the ‌Gene Expression Omnibus database of the National Center for Biotechnology Information in the United States (https://www.ncbi.nih.gov/geo/). Two microarray datasets related to “Osteoporosis”, GSE74209 and GSE201543, were identified and selected. For subsequent analyses, the gene dataset GSE74209, which included six normal samples (Normal) and six osteoporotic samples (Osteoporosis), and the gene dataset GSE15209, which contained four normal samples (Normal) and six osteoporotic samples (Osteoporosis), were chosen. The Gene Expression Omnibus Repository (GEO2R, http://www.ncbi.nlm.nih.gov/geo/geo2r) was utilized to analyze the datasets. Differentially expressed genes (DEGs; Tumor vs. Control) were determined based on a p-value cutoff of < 0.05 and an absolute log₂ fold change (|log₂FC|) of ≥ 1. These DEGs were visualized using heatmaps. A Venn analysis was performed on the differentially expressed miRNAs from the two datasets to identify common miRNAs.

MiR-378a-3p mimics and antagonist transfection

RAW264.7 cells were separated into Control, Mimics NC, miR-378a-3p Mimics, antagonist NC and miR-378a-3p antagonist groups (n ≥ 3). Mimics NC, miR-378a-3p mimics, antagonist NC and miR-378a-3p antagonist (Ribobio, Guangzhou, China) were transfected into RANKL-induced RAW264.7 cells. Transfection was carried out according to the instructions using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA), and cells were collected 48 h later for subsequent experiments.

Animal model construction and grouping

Healthy female Sprague Dawley rats aged 6–8 weeks and weighing 180–200 g were obtained from SPF Biotechnology (Beijing, China) and housed for one week for adaptation. All rats were housed under specific pathogen-free (SPF) conditions, free access to food and water, and maintained in a temperature-controlled room at 22 ± 2 °C and 55% humidity with a 12-hour light/dark cycle. Rats were divided into four groups: Control, OP Model, OP + GLP-2 and OP + GLP-2 + miR-378a-3p antagonist groups (n ≥ 6). Each rat constituted an independent experimental unit. The OP rats were induced by bilateral ovariectomy [23]. To overexpress GLP-2, OP rats were treated with GLP-2 overexpressing exosomes (750 µg) via tail vein injection twice weekly for 8 weeks [24, 25]. In OP + GLP-2 + miR-378a-3p antagonist group, OP rats were treated with GLP-2 overexpressing exosomes plus intratibial injection of miR-378a-3p antagonist (40 µL every two weeks) [26]. Mice were anesthetized by isoflurane inhalation and then euthanized. All animal experiments were approved by the local ethical committee.

Micro-CT reconstruction

Rat femoral specimens were fixed vertically along their long axis in the sample holder and scanned at 0°rotation using the NEMO Micro CTviva CT (NMC-200) (n ≥ 6). The region of interest (ROI) of the cancellous bone was selected for three-dimensional reconstruction. Reconstructed images were quantitatively analyzed using the Micro-CT software. Physical parameters assessed included bone volume fraction (BV/TV), trabecular number (Tb.N), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th) and tissue mineral density (TMD) [27].

Histological staining

The tissue was preserved in 4% paraformaldehyde, underwent decalcification in 15% ethylenediaminetetraacetic acid, embedded in paraffin wax, and subsequently sectioned at a thickness of 4 μm. For hematoxylin and eosin (HE) staining, the paraffin-embedded sections were dewaxed, rehydrated, and then stained with hematoxylin followed by eosin. Furthermore, the sections were subjected to TRAP staining.

Enzyme-linked immunosorbent assay (ELISA)

Rat serum was collected (n ≥ 6), and inflammatory cytokines, tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), interleukin-6 (IL-6) and osteoporosis indicators such as serum calcium and alkaline phosphatase (ALP) levels were measured using corresponding ELISA kits (Esebio, Shanghai, China). Absorbance values were measured at 450 nm using a microplate reader, and sample concentrations were calculated via standard curves.

Quantitative real-time polymerase chain reaction‌ (qRT-PCR)

Total RNA was extracted from cells using TRIZOL reagent (Invitrogen, USA) (n ≥ 3). cDNA was synthesized, and RT-PCR was performed on an ABI7500 quantitative PCR instrument (Applied Biosystems, USA) by SYBR Green Master Mix (Thermo Fisher Scientific). The reaction conditions were as follows: 95 °C for 30s, 95 °C for 10s, 60 °C for 30s for 40 cycles. Ct values were analyzed using the 2^−ΔΔCt method. β-actin was chosen as internal reference. Primers were synthesized by Sangon (Shanghai, China), and primer sequences were shown in Table 1.

Table 1.

Primer sequences in this study

Name	Sequences (5’-3’)
GAPDH-F		CCGGGAAACTGTGGCGTGATGG
GAPDH-R		AGGTGGAGGAGTGGGTGTCGCTGTT
TRAP-F		ATGACGCCAATGACAAGAGGTTCC
TRAP-R		TTGTGCCGAGACATTGCCAAGG
c-Fos-F		TTGGAGCCGGTCAAGAACATTAGC
c-Fos-R		CCAGTCTGCTGCATAGAAGGAACC
NFATc1-F		GCTGTGAAGGCTTCTGCTGGAG
NFATc1-R		TCGTGGCTGGTGGTGGAGAC
OSCAR -F		GCCGCTACCGCAAGACAGAC
OSCAR -R		CTGATCCGTTACCAGCAGTTCCAG
Cathepsin K-F		AGCAGTACAACAGCAAGGTGGATG
Cathepsin K-R		CACTTCTTCGCTGGTCATGTCTC
U6-F		GCTTCGGCAGCACATATACTAA
U6-R		CGAATTTGCGTGTCATCCTT
miR-378a-3p		ACUGGACUUGGAGUCAGAAGG

Dual-luciferase reporter assay

A mutant TRAF6 3’-UTR sequence containing point mutations within the miR-378a-3p binding region was constructed using the Phusion™ Site-Directed Mutagenesis Kit (Thermo Fisher Scientific). Cells were seeded in 96-well culture plates and co-transfected with the reporter gene constructs and the Renilla luciferase expression vector pRL-TK using Lipofectamine™ 2000. At 48 h post-transfection, luciferase activity was quantified using the Dual-Luciferase^® Reporter Assay System.

Western blotting (WB)

Protein samples were extracted from collected cellular (n ≥ 3) and femoral tissue (n ≥ 6) groups. Subsequently, sodium dodecyl sulfate polyacrylamide gel electrophoresis was performed. Following electrophoresis completion, proteins were transferred onto a polyvinylidene fluoride membrane at 4 °C for 2 h. The membrane was then blocked with 5% skimmed milk-Tris buffered saline with Tween under constant agitation at room temperature for 2 h. After blocking, the membrane was briefly rinsed with washing buffer and incubated overnight at 4 °C with primary antibodies (1/1000, Abcam, Cambridge, Britain) against NF-κB p65 (ab32536), P-NF-κB p65 (ab76302), p38 (ab170099), P-p38 (ab178867), JNK (ab179461), p-JNK (ab124956), ERK (ab184699), and p-ERK (ab201015). The membrane underwent three 10-min TBST washes followed by incubation with corresponding goat anti-rabbit secondary antibodies (1/5000, Abcam) at room temperature for 1 h. Protein bands were visualized through chemiluminescent detection system exposure, followed by standard photographic development and fixation processes. Quantitative data analysis was conducted using Image J software.

Statistical analysis

Data were analyzed using GraphPad Prism 8.0. Measurement data were presented as mean ± standard deviation. T-tests were used for comparison between two groups, and One-Way ‌Analysis of Variance‌ followed by Tukey’s multiple comparisons test was used for comparison among multiple groups. This study assessed the normality of the dataset using the Shapiro-Wilk test. For datasets violating the normality assumption, non-parametric alternatives-specifically the Mann-Whitney U test-were employed to calculate statistical significance (P-values). P < 0.05 was considered statistically significant.

Results

GLP-2-carrying exosomes inhibits osteoclast differentiation

Exosomes were extracted from the Control, LV-NC, and LV-GLP-2 treated with RANKL-induced RAW264.7 cells. The particle size distribution of exosomes was analyzed, which revealed that all three groups exhibited comparable profiles in terms of particle concentration distribution. Specifically, exosomal particles within the size range of 80–110 nm accounted for 92.8% of the total in the Control group, 92.3% in the LV-NC group, and 95.3% in the LV-GLP2 group (Fig. 1A). As shown in Fig. 1B, TEM analysis confirmed that exosomes derived from the Control, LV-NC, and LV-GLP2 groups displayed the characteristic bilayered lipid membrane structure typical of exosomes, presenting a predominantly round or oval. Furthermore, WB demonstrated that exosomes from all three groups were positive for the exosomal marker proteins CD63 and CD81, confirming their identity (Fig. 1C). Notably, while CCK-8 assays revealed no significant differences in proliferation among Control, LV-NC, and LV-GLP-2 groups, indicating that the observed inhibition of osteoclastogenesis was not attributable to reduced cellular viability or proliferative capacity (Fig. 1D), TRAP staining and qRT-PCR analysis demonstrated pronounced suppression of osteoclast differentiation markers (TRAP, c-Fos, NFATc1, OSCAR and Cathepsin K) in LV-GLP-2-treated cells (all p < 0.01, Figs. 1E–F). This indicated that GLP-2-carrying exosomes specifically targeted differentiation pathways in osteoclast precursors without altering cell viability or expansion—suggesting a mechanism focused on modulating functional maturation.

Fig. 1.

GLP-2-carrying exosomes suppressed osteoclast differentiation in RAW264.7 cells. (A) Particle size distribution profile of isolated exosomes as determined by nanoparticle tracking analysis. (B) Morphological characterization of exosomes via transmission electron microscopy. (C) Western blotting (WB) of exosomal surface markers (CD9/CD63/CD81). (D) Cell Counting Kit-8 (CCK-8) assay for cell viability detection. (E) Quantitative real-time polymerase chain reaction‌ (qRT-PCR) analysis of osteoclast-related gene expression. (F) Tartrate resistant acid phosphatase (TRAP) staining was carried out to assess osteoclast differentiation. ^**P < 0.01 vs. LV-NC group

GLP-2-carrying exosomes suppresses NF-κB/MAPK signaling pathways

After treating RAW264.7 cells with exosomes from the Control, LV-NC, and LV-GLP-2 groups for 5 days, WB analysis showed that phosphorylation levels of NF-κB and MAPK signaling pathway proteins (p-NF-κB, p-JNK, p-ERK, p-p38) in the LV-GLP-2 group were reduced (all p < 0.01, Fig. 2).

Fig. 2.

GLP-2-carrying exosomes inhibited NF-κB and MAPK signaling pathways. WB analysis of NF-κB and MAPK signaling pathway-related proteins (NF-κB, p-NF-κB, JNK, ERK, p38, p-JNK, p-ERK, p-p38). ^**P < 0.01 vs. LV-NC group

MiR-378a-3p exerts inhibitory effects on osteoclast differentiation

Two microarray datasets (GSE74209 and GSE201543) related to “Osteoporosis” were selected. Volcano plots showed DEGs in both datasets (Fig. 3A). Using GEO2R, 170 DEGs (59 upregulated and 111 downregulated) were identified in GSE74209 dataset and 226 DEGs (71 upregulated and 155 downregulated) in GSE201543 dataset (Fig. 3B). Venn analysis identified 15 common miRNAs in both datasets (Fig. 3C). Emerging evidence indicates that miR-378a-3p potentiates osteogenic differentiation [22], therefore, miR-378a-3p was selected for further investigation. qRT-PCR analysis showed that miR-378a-3p expression was significantly increased in GLP-2 overexpressing exosomes (p < 0.01, Fig. 3D).

Fig. 3.

miR-378a-3p mimics inhibited osteoclast differentiation. A-B. Volcano plots of differentially expressed miRNAs in datasets GSE74209 and GSE201543; red denotes upregulated miRNAs, blue denotes downregulated miRNAs. C. Venn diagram of common differentially expressed miRNAs, with overlapping regions indicating common miRNAs. D. QRT-PCR validation of miR-378a-3p expression in exosomes. E. QRT-PCR was carried out to verify the transfection efficiency of miR-378a-3p. F. QRT-PCR analysis of osteoclast differentiation-related gene expression. G. TRAP staining was performed to evaluate osteoclast differentiation. ^**P < 0.01 vs. LV-NC/Mimics NC group

In RANKL-induced RAW264.7 cells, transfection with miR-378a-3p mimics was performed, and miR-378a-3p expression levels were quantified via qRT-PCR. The results demonstrated no significant differences between the Control and Mimic NC groups, whereas transfection with miR-378a-3p mimics led to a marked upregulation of miR-378a-3p expression (p < 0.01, Fig. 3E), confirming successful transfection efficacy. To assess osteoclast differentiation, qRT-PCR analysis of osteoclast-associated genes demonstrated that transfection with miR-378a-3p mimics resulted in significant downregulation of TRAP, c-Fos, NFATc1, OSCAR, and Cathepsin K mRNA expression levels (all p < 0.01, Fig. 3F). Furthermore, TRAP staining was conducted in RANKL-induced RAW264.7 cells, which revealed that miR-378a-3p mimics significantly suppressed osteoclast differentiation compared to the Mimics NC group (Fig. 3G). These findings collectively indicated that miR-378a-3p exerted inhibitory effects on osteoclast differentiation.

MiR-378a-3p suppresses the NF-κB/MAPK signaling pathway

In RANKL-induced RAW264.7 cells, miR-378a-3p mimics and miR-378a-3p antagonist were applied. WB analysis was performed to evaluate the expression of NF-κB and MAPK signaling pathway-related proteins, which indicated that compared with Mimics NC, transfection with miR-378a-3p mimics significantly reduced the phosphorylation levels of key proteins in the NF-κB and MAPK signaling pathways, including p-IκB, p-NF-κB, p-JNK, p-ERK, and p-p38 (all p < 0.01, Fig. 4A). Conversely, when compared to the antagonist NC, the miR-378a-3p antagonist markedly promoted the phosphorylation levels of these key proteins in the NF-κB and MAPK signaling pathways (all p < 0.01, Fig. 4B).

Fig. 4.

miR-378a-3p attenuated NF-κB/MAPK pathway activation. A-B. WB analysis of the regulatory effects of miR-378a-3p Mimics (A) and miR-378a-3p antagonist (B) on NF-κB and MAPK pathways. ^**P < 0.01 vs. Mimics NC/antagonist NC group

Through TargetScan-based predictive analysis, potential binding sites between miR-378a-3p and TRAF6, a gene associated with the NF-κB/MAPK pathway, were identified, as illustrated in Fig. 5A. Furthermore, the direct binding specificity between miR-378a-3p and TRAF6 was validated via dual-luciferase reporter assays, and a significant reduction in luciferase activity was observed exclusively with the wild-type reporter (Fig. 5B; p < 0.01). Both qRT-PCR and WB results demonstrated that miR-378a-3p can suppress TRAF6 expression (p < 0.01, Fig. 5C-D), mechanistically linking direct targeting to downstream NF-κB/MAPK pathway inhibition. Therefore, these findings suggested that miR-378a-3p exerted its inhibitory effects on osteoclast differentiation, at least in part, by attenuating the activation of the NF-κB and MAPK signaling cascades.

Fig. 5.

(A) TargetScan prediction analysis of the binding site between miR-378a-3p and TRAF6. (B) Dual-luciferase reporter assay validating the interaction between miR-378a-3p and TRAF6. C-D. QRT-PCR and WB analysis of the effect of miR-378a-3p on TRAF6 expression. ^**P < 0.01 vs. Mimics NC

GLP-2-carrying exosomes ameliorate osteoporosis by upregulating miR-378a-3p in vivo

In OP rats, administration of GLP-2-carrying exosomes and miR-378a-3p antagonist was performed, and miR-378a-3p expression was analyzed by qRT-PCR. The results demonstrated that miR-378a-3p levels were significantly reduced in the OP model group compared to the Control group (p < 0.01). However, treatment with GLP-2-carrying exosomes markedly restored miR-378a-3p expression in the OP model group (p < 0.01), whereas co-administration of miR-378a-3p antagonist resulted in a significant reduction of miR-378a-3p levels (p < 0.01, Fig. 6A).

Fig. 6.

GLP-2-carrying exosomes restored bone microarchitecture via miR-378a-3p in osteoporosis (OP) model rats. (A) QRT-PCR detection of miR-378a-3p expression in femoral tissue. (B) Micro-CT reconstruction of bone structure. (C) Quantitative assessment of bone parameters. BV/TV: bone volume fraction, Tb.N: trabecular number, Tb.Sp: trabecular separation, Tb.Th: trabecular thickness. TMD: tissue mineral density. ^**P < 0.01 vs. Control group; ^##P < 0.01 vs. OP Model group; ^$$P < 0.01 vs. OP + GLP-2 group

Micro-CT reconstruction revealed distinct trabecular bone structural differences among groups. In the Control group, femoral trabeculae exhibited uniform alignment, dense architecture, and intact continuity with abundant interconnected networks. In contrast, the OP group displayed sparse, fragmented, and disorganized trabeculae, accompanied by extensive bone marrow regions devoid of trabeculae, indicative of severe microstructural deterioration. The OP + GLP-2 group showed partial restoration of trabecular architecture, characterized by increased density and structural regularity. Notably, miR-378a-3p antagonist reversed these improvements, leading to trabecular loss (Fig. 6B). Quantitative analysis of bone morphometric parameters revealed that the OP model group exhibited significant reductions in BV/TV, Tb.Th, Tb.N and TMD, alongside elevated Tb.Sp (all p < 0.01) compared to the Control group, and GLP-2-carrying exosomes ameliorated these deficits, increasing BV/TV, Tb.Th, Tb.N and TMD while reducing Tb.Sp (all p < 0.01). However, co-administration of miR-378a-3p antagonist with GLP-2 exosomes partially abrogated these benefits, resulting in decreased BV/TV, Tb.Th, Tb.N, and TMD, and increased Tb.Sp compared to the OP + GLP-2 group (all p < 0.01, Fig. 6C). These findings indicated that the therapeutic effects of GLP-2 exosomes on osteoporosis could be effectively reversed by miR-378a-3p antagonist, underscoring miR-378a-3p as a critical mediator in this process.

HE staining revealed significant bone loss and a reduced bone tissue area in the OP rat model group compared to the Control group. Treatment with GLP-2-carrying exosomes (OP + GLP-2 group) resulted in increased bone mass and bone tissue area. However, the beneficial effects of GLP-2-carrying exosomes on osteoporosis were reversed upon transfection with miR-378a-3p antagonist (Fig. 7A). TRAP staining demonstrated enlarged and more deeply stained areas in the OP model group compared to the Control group, indicating an increased osteoclast-positive reaction area and enhanced osteoclast activity. In contrast, the OP + GLP-2 group exhibited a significant reduction in both staining area and depth; miR-378a-3p antagonist led to a resurgence in osteoclast activity (Fig. 7A).

Fig. 7.

Histopathological and biochemical effects of GLP-2-carrying exosomes in OP rats. (A) Histopathological evaluation of femoral tissue via Hematoxylin-Eosin and TRAP staining (magnification: 200×, scale bar: 50 μm). (B) Enzyme-Linked Immunosorbent Assay (ELISA) quantification of serum calcium and alkaline phosphatase levels. (C) ELISA measurement of serum inflammatory cytokines. (D) qRT-PCR analysis of osteoclast-related gene expression in femoral tissue. ^**P < 0.01 vs. Control group; ^##P < 0.01 vs. OP Model group; ^$$P < 0.01 vs. OP + GLP-2 group

Furthermore, the OP model group showed significantly elevated serum levels of osteoporosis biomarkers calcium and ALP compared to the Control group (all p < 0.01). GLP-2-carrying exosomes reduced these levels (all p < 0.01), but the reductions were reversed upon transfection with miR-378a-3p antagonist (all p < 0.01, Fig. 7B). ELISA revealed marked increases in TNF-α, IL-6, and IL-1β levels in the OP model group compared to the Control group (all p < 0.01). GLP-2-carrying exosomes significantly decreased these cytokine levels (all p < 0.01), but the decrease was mitigated by miR-378a-3p antagonist transfection (all p < 0.01, Fig. 7C). qRT-PCR analysis of osteoclast-related gene expression in rat femoral tissues showed significantly higher levels of TRAP, c-Fos, NFATc1, OSCAR, and Cathepsin K in the OP model group compared to the Control group (all p < 0.01). GLP-2-carrying exosome treatment significantly downregulated these genes (all p < 0.01), but this downregulation was reversed following miR-378a-3p antagonist transfection (all p < 0.01, Fig. 7D).

GLP-2-carrying exosomes inhibit the NF-κB and MAPK pathways through up-regulating miR-378a-3p

WB analysis was conducted to assess the expression of proteins associated with the NF-κB and MAPK signaling pathways. Compared to the Control group, the OP model group exhibited elevated phosphorylation levels of NF-κB and MAPK pathway-related proteins (p-NF-κB, p-JNK, p-ERK, p-p38) (all p < 0.01). Treatment with GLP-2-carrying exosomes in the OP model group led to a significant reduction in the expression levels of p-NF-κB, p-JNK, p-ERK, and p-p38 (all p < 0.01). However, the decrease in the expression of these phosphorylated proteins following GLP-2 exosome treatment was reversed upon transfection with miR-378a-3p antagonist (all p < 0.01, Fig. 8).

Fig. 8.

GLP-2-carrying exosomes modulated NF-κB and MAPK pathways in vivo. WB analysis of NF-κB and MAPK signaling pathway-related protein expression. ^**P < 0.01 vs. Control group; ^##P < 0.01 vs. OP Model group; ^$$P < 0.01 vs. OP + GLP-2 group

Discussion

Osteoporosis is a prevalent disease that significantly increases the risk of fractures [28]. Therefore, it is imperative for identifying novel therapeutic targets to address its pathophysiological mechanisms. In this study, we demonstrated that GLP-2-carrying exosomes effectively ameliorated osteoporosis by suppressing osteoclast differentiation and inhibiting the NF-κB/MAPK signaling pathway through upregulation of miR-378a-3p (Figure S1). In vitro, GLP-2-carrying exosomes inhibited osteoclastogenesis and the NF-κB/MAPK pathway. In vivo assay revealed that GLP-2-carrying exosomes restored bone microarchitecture and attenuated inflammatory cytokines in osteoporotic rats. Crucially, administration of a miR-378a-3p antagonist substantially reversed these therapeutic outcomes, unequivocally establishing the pivotal regulatory role of miR-378a-3p in mediating the bone-protective effects of GLP-2-carrying exosomes.

Functioning as natural delivery vehicles for diverse bioactive payloads, exosomes demonstrate considerable translational promise in osteoporosis intervention strategies [29]. GLP-2 has been established as a critical regulator of human bone turnover, synergizing with the gut-derived hormone glucose-dependent insulinotropic polypeptide to improve bone metabolic homeostasis, thereby emerging as a promising therapeutic candidate for osteoporosis [30]. In aging mouse models such as Senescence-Accelerated Mouse Prone 6, GLP-2 ameliorates bone loss and intestinal barrier dysfunction, highlighting its dual efficacy in enhancing osteogenesis and counteracting senile osteoporosis [31]. In the present study, both in vivo and in vitro experiments demonstrated that GLP-2-carrying exosomes attenuated osteoclastogenesis by downregulating key differentiation markers. Concurrently, these exosomes restored trabecular microarchitecture, evidenced by increased BV/TV, Tb.Th, and reduced Tb.Sp. Furthermore, GLP-2 exosomes significantly suppressed pro-inflammatory cytokines, demonstrating its therapeutic potential in ameliorating osteoporosis through multi-target modulation of bone resorption, inflammation, and structural remodeling.

Non-coding RNAs also play significant roles in musculoskeletal disorders. Current research indicates that specific miRNAs are implicated in the proliferation and differentiation of stromal cell lineages, demonstrating considerable therapeutic potential for tendon injuries [32], and are further involved in the diagnosis and treatment of osteoarthritis [33]. Furthermore, small interfering RNAs (siRNAs) influence the dysregulation of genes encoding growth factors and tendon structural proteins [34], potentially serving as therapeutic targets for rheumatoid arthritis and osteoporosis [35, 36]. MiR-378a is an emerging metabolic miRNA that has been shown to ameliorate metabolic disorders in diseases such as diabetes and obesity [37]. Wang et al. discover that the circHIPK3/miR-378a-3p/HDAC4 axis plays a potential regulatory role in osteoporotic fractures [38]. Low-intensity vibration promotes osteogenic differentiation of bone marrow mesenchymal stem cells by upregulating the miR-378a-3p/Grb2 pathway, enhancing bone density and formation in aged rats [39]. MiR-378a-3p promotes chondrogenic differentiation of synovial-derived mesenchymal stem cells by activating the BMP2-Smad pathway, representing a potential therapeutic target for osteoarthritis [40]. LncRNA MALAT1 regulates tenogenic differentiation of tendon-derived stem cells by modulating the miR-378a-3p/MAPK1 axis [41]. Herein, bioinformatics analysis identified miR-378a-3p as the hub miRNA in osteoporosis, and we found that it was significantly down-regulated in model mice, while overexpression of GLP increased the expression of miR-378a-3p. Additionally, miR-378a-3p antagonist reversed the inhibitory effect of GLP-2-carrying exosomes on osteoclastogenesis and inflammation, indicating that GLP-2-carrying exosomes alleviated osteoporosis progression by up-regulating miR-378a-3p.

The NF-κB signaling pathway serves as a master regulator of inflammatory cascades and disease pathogenesis [42, 43]. Emerging pharmacological interventions targeting this pathway demonstrate therapeutic efficacy in osteoporosis management: Aconine counteracts osteoclast-mediated bone resorption and ferroptosis via NF-κB inhibition [44], while ellagic acid suppresses osteoclastogenesis and postmenopausal bone loss by blocking NF-κB-driven pyroptosis [45]. Notably, Nrf2 activators have been shown to mitigate bone resorption through STING-dependent NF-κB suppression [46]. This pathway further exhibits extensive crosstalk with key regulators of bone homeostasis, including PI3K/AKT, WNT, and MAPK cascades [47]. Mechanistic studies reveal multi-pathway modulation strategies—StemRegenin 1 impedes osteoclast differentiation by disrupting AhR-c-src-NF-κB/p-ERK MAPK-NFATc1 signaling [48], whereas puerarin alleviates bone loss through TRAF6/ROS-dependent MAPK/NF-κB axis inhibition [49]. N, N-Dimethylformamide exerts anti-osteoporotic effects in hyperglycemic conditions through dual suppression of MAPK and NF-κB signaling cascades [50]. Our findings demonstrated that GLP-2-carrying exosomes markedly suppressed NF-κB/MAPK pathway phosphorylation in osteoporotic mice, an effect significantly attenuated by miR-378a-3p antagonism. While prior studies, such as those utilizing C2C12-derived exosomes carrying miR-92a-3p, demonstrated therapeutic effects on osteoporosis by broadly modulating osteoblast-osteoclast balance [16]. Unlike miR-92a-3p, which primarily regulates osteogenesis, miR-378a-3p exhibits multi-pathway inhibitory effects on osteoclast differentiation by suppressing TRAF6/NF-κB and MAPK signaling cascades. Moreover, GLP-2 is known to improve nutrient absorption and calcium homeostasis [51, 52], synergizing with miR-378a-3p’s anti-resorptive effects.

While this work established the miR-378a-3p/NF-κB-MAPK axis as a key mediator of GLP-2-carrying exosome efficacy in osteoporosis, several limitations warrant mention. First, the precise mechanism by which GLP-2 upregulates miR-378a-3p—whether via transcriptional regulation, exosomal sorting, or miRNA stabilization—remains unresolved and will be explored using RIP-seq and promoter-reporter assays. Second, the exclusive use of an OVX model limits extrapolation to non-estrogen-deficient osteoporosis. Future studies will employ aged and glucocorticoid-induced models to broaden therapeutic validation. Furthermore, current methods for exosome isolation, such as ultracentrifugation, present limitations in terms of yield, purity, and scalability for clinical-grade production. The development of stable exosome formulations and optimized storage conditions is critical for clinical translation. Variations in individual immune responses to exosomes may occur, potentially affecting therapeutic efficacy and safety. Consequently, future clinical studies should incorporate close monitoring of participants’ immune reactions and explore personalized treatment regimens.

In summary, this study elucidated that GLP-2-carrying exosomes exerted potent anti-osteoporotic effects by suppressing osteoclast differentiation and NF-κB/MAPK pathway activation through upregulation of miR-378a-3p in vivo and in vitro. These findings highlight a novel mechanism by which exosome-miRNA axis regulates bone remodeling, offering a dual-targeted approach to simultaneously inhibit osteoclast activity and inflammatory signaling. This mechanistically positions exosome-based delivery as a promising targeted therapeutic strategy for human osteoporosis, potentially overcoming limitations of systemic drug administration (e.g., bisphosphonate toxicity). Future clinical translation requires validation in human bone-marrow-derived cells, safety/efficacy assessment of humanized exosomes in non-human primates.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Jiping Shen: Conceptualization; Formal analysis; Methodology; Writing - original draft; Validation; Resources; Kefen Wu, Yi Lu and Kan Xu: Formal analysis; Methodology; Validation; Editing; Yanling Huang and Yu Hu: Data curation; Investigation; Software; Review & editing; All authors have read and approved the manuscript.

Funding

Not applicable.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The experiments conformed to the Guide for the Care and Use of Laboratory Animals. Animal study has been approved by the Animal Ethics Committee of Zhongshan Hospital. All methods are reported in accordance with ARRIVE guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.