Extracellular vesicles derived from bone marrow mesenchymal stem cells loaded on magnetic nanoparticles delay the progression of diabetic osteoporosis via delivery of miR-150-5p

Extracellular vesicles derived from bone marrow mesenchymal stem cells (BMSC-EVs) are emerged as carriers of therapeutic targets against bone disorders, yet its isolation and purification are limited with recent techniques. Magnetic nanoparticles (MNPs) can load EVs with a unique targeted drug delivery system. We constructed gold-coated magnetic nanoparticles (GMNPs) by decorating the surface of the Fe3O4@SiO2 core and a silica shell with poly(ethylene glycol) (PEG)-aldehyde (CHO) and examined the role of BMSC-EVs loaded on GMNPs in diabetic osteoporosis (DO). The osteoporosis-related differentially expressed miR-150-5p was singled out by microarray analysis. DO models were then established in Sprague–Dawley rats by streptozotocin injection, where poor expression of miR-150-5p was validated in the bone tissues. Next, GMNPE was prepared by combining GMNPs with anti-CD63, after which osteoblasts were co-cultured with the GMNPE-BMSC-EVs. The re-expression of miR-150-5p facilitated osteogenesis in osteoblasts. GMNPE could promote the enrichment of EVs in the bone tissues of DO rats. BMSC-EVs delivered miR-150-5p to osteoblasts, where miR-150-5p targeted MMP14 and consequently activated Wnt/β-catenin pathway. This effect contributed to the enhancement of osteoblast proliferation and maturation. Furthermore, GMNPE enhanced the EV-based delivery of miR-150-5p to regulate the MMP14/Wnt/β-catenin axis, resulting in promotion of osteogenesis. Overall, our findings suggest the potential of GMNP-BMSC-EVs to strengthen osteoblast proliferation and maturation in DO, showing promise as an appealing drug delivery strategy against DO. Graphical abstract 1. GMNPs-BMSCs-EVs-miR-150-5p promotes the osteogenesis of DO rats. 2. miR-150-5p induces osteoblast proliferation and maturation by targeting MMP14. 3. Inhibition of MMP14 activates Wnt/β-catenin and increases osteogenesis. 4. miR-150-5p activates the Wnt/β-catenin pathway by downregulating MMP14. Supplementary Information The online version contains supplementary material available at 10.1007/s10565-022-09744-y.


Introduction
Diabetes mellitus often impairs the skeletal system (Schacter and Leslie 2017). The risk of developing bone disorders, including osteoporosis, is thus greatly increased (Cortet et al. 2019;Rehling et al. 2019). Available biological agents to treat diabetic osteoporosis (DO) consist of anti-diabetic drugs, anti-osteoporosis drugs, and anti-resorptive osteoporosis drugs (Mohsin et al. 2019). The advent of new and original molecular biology techniques has shed a light on the novel molecular mechanisms that underlie the development of individual treatment plans.
Extracellular vesicles (EVs) derived from mesenchymal stem (stromal) cells (MSCs) (MSC-EVs) can deliver therapeutic targets for various diseases, showing a promise as an alternative cell-free therapy (Rani et al. 2015). Despite this, the isolation and detection of EVs still exhibit several technical drawbacks such as limited sensitivity and time consumption (Oliveira-Rodriguez et al. 2017). Magnetic nanoparticles (MNPs) with surface decorated with stimuli-responsive polymers can enable efficient and rapid EV separation (Jauregui et al. 2018). MNPs are a class of nanomaterials, showing diverse applications, such as magnetic separation, magnetic imaging, magnetic biosensing (diagnostics), and drug and gene delivery as they , highlighting a promise of MNP-loaded EVs for disease treatment.
MicroRNAs (miRNAs) exhibit therapeutic use by controlling gene expression of target mRNAs (Liu et al. 2021a). It should be noted that EVs contain miRNAs, which have been reported to be essential for the mediation of intercellular communication and as potential therapeutic candidates for disease (Mori et al. 2019). miR-150 can trigger the osteoblastic phenotype related to osteoblast function and bone mineralization, which shows therapeutic potential of promoting bone formation in some diseases, especially in osteoporosis (Dong et al. 2015). miR-150-5p has been identified to target matrix metalloproteinase 14 (MMP14) and downregulated its expression in hepatoma cells . Pharmacologic inhibition of MMP14 can reduce bone resorption and potentiate bone gain (Delgado-Calle et al. 2018). Activation of the human MMP14 gene promoter links to enhanced nuclear translocation of β-catenin (Lu et al. 2014). Of note, activation of the Wnt/β-catenin pathways in human bone marrow MSCs (BMSCs) results in promotion of osteogenesis (Pang et al. 2022). In addition, activation of this pathway can promote osteoblastic differentiation, thus preventing the development of osteoporosis (Sharma and Nam 2019;Yang et al. 2020).
In light of these described above, MSC-EVs containing miR-150-5p may play a critical role in the progression of DO via the MMP14/Wnt/β-catenin axis. In this study, we utilized nanoparticles consisted of Fe 3 O 4 , SiO 2 , poly(ethylene glycol) (PEG), and aldehyde (CHO) (GMNPs) to load the EVs isolated from BMSCs and carried out co-culture assays to address this hypothesis.

Establishment of rat models of DO
Seventy male SD rats (7-8 weeks old; 180 ± 20 g) were housed with 60-70% relative humidity at 25 ± 2 °C under a 12-h light/dark cycle. They were free access to food and water. Following 7 days of acclimatization, 6 rats were injected with normal saline and used as control and 64 rats were randomly selected for DO modeling by intraperitoneal injection with streptozotocin (STZ; 60 mg/kg, Sigma-Aldrich, St Louis, MO) (Wang-Fischer and Garyantes 2018;Cao et al. 2021;An et al. 2019). The study, approved by the Animal Ethics Committee of the First Affiliated Hospital of Bengbu Medical College (approval no.: 2021-245), was conducted according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.
After 72 h, random blood glucose > 16.7 mmol/L was indicative of successful modeling. Bone mineral density (BMD) and bone mineral content (BMC) were assessed every 4 weeks using dual-energy X-ray absorptiometry (DXA; Hologic, Marlborough, MA). The significant statistical difference in BMD and BMC between the two groups denoted successful establishment of DO ( Supplementary Fig. 1).

Isolation and culture of primary BMSCs and osteoblasts
BMSCs were isolated as previously described (Huang et al. 2021). The bilateral tibia and femur of rats were collected, and the bone marrow was washed with DMEM/F12 to obtain a mixed cell suspension. The suspension was subsequently centrifuged (1000 r/ min, 5 min) and a precipitate containing BMSCs was harvested, then re-suspended in DMEM/F12 and cultured with 5% CO 2 at 37 °C. For the characterization of BMSCs (Duan et al. 2020;Dominici et al. 2006), alizarin red S staining was performed to test the calcium deposition at the 3rd week of osteogenic differentiation induction, so as to examine the osteogenesis of BMSCs, and oil red O staining was applied to observe the lipid droplets at the 3rd week of adipogenic differentiation induction to assess the adipogenesis of BMSCs. Cell morphology and growth were observed and identified under an inverted microscope ( Supplementary Fig. 2).
The osteoblasts were isolated from trabecular bone fragments by sequential enzymatic digestion (Huang et al. 2021;Zhao et al. 2010). The morphology of cells was finally observed with an inverted microscope ( Supplementary Fig. 3).

RT-qPCR
Total RNA was extracted with TRIzol method and reverse-transcribed into cDNA with PolyA Tailing Reverse Transcription Kit (B532451, Sangon; for miRNA detection) or RevertAid RT Reverse Transcription Kit (K1691, Thermo Fisher Scientific; for mRNA detection). RT-qPCR was followed using Applied Biosystems QuantStudio System (4,489,084, Thermo Fisher Scientific). As normalized to U6 or GAPDH, the fold changes were calculated using the 2 −ΔΔCt method (Livak and Schmittgen 2001) (Supplementary Table 2).

Isolation and identification of EVs from BMSCs
The EVs were isolated from BMSCs by differential ultracentrifugation-based method as previously described (Liao et al. 2019).
Co-culture of BMSC-EVs with osteoblasts and uptake of EVs by osteoblasts BMSCs were transiently transfected with Cy3-labeled miR-150-5p, and then co-cultured with osteoblasts using a Transwell co-culture system (Guan et al. 2020;Zhu et al. 2019). After nucleus staining with Hoechst 33,342, uptake of the labeled EVs by osteoblasts was visualized by a confocal laser scanning microscope (ZEISS LSM 800).
Detection of osteoblast viability, mineralized nodules, and ALP activity For viability assay of osteoblasts, CCK-8 kit (10 μL, C0038, Beyotime, Shanghai, China) was applied to assess cell viability (Ge et al. 2019). Osteoblasts were fixed and then stained with 1% alizarin red or ALP staining solution to observe the number of mineralized nodules or ALP activity respectively ).

Synthesis and characterization of nanoparticles
Various chemical substances were fixed on the surface of MNPs (102,138, XFNANO, Nanjing, Jiangsu, China) to produce GMNPs or RhB-labeled GMNPs (Supplementary Table 3). The Fe 3 O 4 @SiO 2 -PEG-CHO (GMNPs or RhB-labeled GMNPs) was obtained according to previous report .
The intermediate products and final products were analyzed by SEM (S-4800, Hitachi, Shanghai Fulai Optical Technology Co., Ltd., Shanghai, China) with energy-dispersive spectrometer (IQLAAHGABM-FAAWMACL, Thermo Fisher Scientific) and Fourier transform infrared spectrometer (912A0770, Thermo Fisher Scientific). DLS (Nano ZS90, Malvern) was applied for size and size distribution measurement.
Preparation and biocompatibility of GMNPE GMNP E was generated with GMNPs (Fe 3 O 4 @SiO 2 -PEG-CHO) bound by the rabbit antibody to CD63 (PA5-92,370, Thermo Fisher Scientific). In short, 1 mg/ mL GMNP solution was purified by magnetic separation, and incubated with anti-CD63 antibody overnight at 4 °C to obtain GMNP E . Fluorescent GMNP E was prepared using the same method, with the aforementioned GMNPs replaced by RhB-labeled GMNPs.
In order to assess drug carrying capacity of GMNP E , thermogravimetric analysis (TGA; STA 6000, PerkinElmer) was conducted for analyzing MNPs, GMNPs, and GMNP E . The adsorption capacity and stability of HSA were measured with FITC-labeled HSA by fluorescence spectroscopy.
In vitro identification of GMNPE-EV binding and targeting efficiency Purified EVs were labeled with PKH67 (green) (MINI67, Sigma). After magnetic separation, GMNP E -EVs were obtained, with the morphology observed by a high-resolution TEM (200 kV, CM200, Philips). Furthermore, RhB and PKH67 were separately used to label GMNP E and EVs to confirm this binding .
Uptake of GMNPE-EVs by osteoblasts GMNP E -EVs (RhB-labeled GMNP E , PKH67-labeled EVs) were incubated with osteoblasts for 3 h at 37 °C and fixed with PFA. Cytoskeleton staining was performed to observe the localization of GMNP E -EVs in the cells, and the nuclei were counterstained with Hoechst 33,342.
Flow cytometry FITC-labeled GMNP E or PKH67-labeled GMNP E -EVs containing different ratios of anti-CD63 were co-cultured with osteoblasts in PBS at 4 °C for 30 min. These samples were analyzed using a FC500 flow cytometer (Morey Biosciences, Inc., Shanghai, China). The FlowJo v.10 (Tree Star) software was applied for data processing.
Immunofluorescence detection of GMNPE-EV uptake in rat femoral tissues GMNP E -EVs (RhB-labeled GMNP E , PKH67-labeled EVs) were injected into DO rats via tail vein. An N52 neodymium magnet was used to provide an external magnetic field in the rat femur, and the flowing GMNP E nanoparticles were collected to the femur. Next, the rats were euthanized and the femoral, brain, liver, and kidney tissues were removed to prepare longitudinal frozen sections, which were subjected to immunofluorescence staining with OCN (10 µg/mL, mouse antibody, 33-5400, Thermo Fisher Scientific) and 80i fluorescence microscopy.

Histomorphometric analysis
On days 10 and 3 before euthanasia, the rats were intraperitoneally injected with 0.1% calcein (10 mg/ kg body weight, C0875, Sigma-Aldrich) dissolved in PBS. After euthanasia, the femurs were removed from the rats. The samples were fixed, dehydrated, embedded in methyl methacrylate, and cut into 60-μm-thick sections. The double labeling of calcein was observed under a fluorescence microscope and the Image-Pro Plus 6 software was applied to measure mineral apposition rate (MAR) of trabecular bones.

Statistical analysis
Unpaired t-test, one-way ANOVA and two-way ANOVA, or repeated measures ANOVA with Tukey's tests were utilized to calculate statistical significance for two-group, multi-group, and time-based data. All results processed using the SPSS 21.0 software were presented as mean ± SD. p < 0.05 suggests statistically significant difference.
The above-said results indicate that miR-150-5p restores the osteogenesis of DO rats.

BMSC-EV-miR-150-5p facilitates the proliferation and maturation of osteoblasts
The EVmiRNA website predicted that miR-150-5p was enriched in MSC-EVs ( Fig. 2A). Under a TEM, the isolated vesicles sourced from BMSCs exhibited a representative double-layer membrane structure, cup-shaped or flat-shaped, with diameters of about 100 nm (Fig. 2B, C). Immunoblotting results showed expression of CD63, HSP70, and TSG101 in EVs but the absence of GM130 in the isolated vesicles ( Fig. 2D), demonstrating the typical characteristics of EVs. In addition, expression of miR-150-5p was notably higher in BMSCs compared with osteoblasts isolated from DO rats (Fig. 2E).
The BMSCs transiently transfected with Cy3labeled miR-150-5p mimic were co-cultured with osteoblasts. After 72 h, the osteoblasts showed obvious Cy3 fluorescence, which became disappeared in response to the addition of EV release inhibitor GW4869. Green fluorescence was also detected in the osteoblasts co-cultured with EVs from the Cy3labeled miR-150-5p mimic-transfected BMSCs (Fig. 2F). Accordingly, miR-150-5p could be transmitted into osteoblasts through BMSC-EVs.
Synthesis and characterization of GMNPs SEM images showed increased size of spherical Fe 3 O 4 and Fe 3 O 4 @SiO 2 (MNPs) with rough surface changed into relatively smooth after PEG modification at 200 nm ( Supplementary Fig. 6A). The energydispersive spectrum analysis of GMNPs showed the existence of Fe 3 O 4 , SiO 2 , PEG, and hydrazine bond in GMNPs ( Supplementary Fig. 6B). Infrared spectroscopy revealed the expected characteristic peaks of the functional groups during each synthesis step, including the modification of Fe 3 O 4 by SiO 2 , hydrazone, and PEG-CHO ( Supplementary Fig. 6C).
Zeta potential measurement data presented that the surface charge gradually changed according to the electronic properties of varied functional groups ( Supplementary Fig. 6D). The magnetic measurement results showed that these nanoparticles were superparamagnetic without evident hysteresis loop before and after modification ( Supplementary Fig. 6E), indicating that the effect of surface modification on the core structure could be negligible. X-ray diffraction pattern results also supported this finding (Supplementary Fig. 6F). Compared with anti-CD63 antibody that directly bound to MNPs without PEG modification, namely Fe 3 O 4 @SiO 2 @anti-CD63 antibody (MNP E ), the stability of GMNP E was improved  Fig. 1 The promoting effect of miR-150-5p on the osteogenesis in DO rats. A Differential expression of miR-150-5p in the GSE142872 dataset (3 serum samples from osteoporosis patients and 3 serum samples from healthy individuals) from GEO database (https:// www. ncbi. nlm. nih. gov/ gds). Differential analysis was conducted using R language "limma" package with p < 0.05 as the threshold to identify differentially expressed genes. B miR-150-5p expression determined by RT-qPCR in the bone tissues of normal and DO rats. C miR-150-5p expression determined by RT-qPCR in the bone tissues of miR-150-5p mimic or miR-150-5p inhibitor-treated DO rats. D mRNA expression of RUNX2 and BSP deter-mined by RT-qPCR in the bone tissues of miR-150-5p mimic or miR-150-5p inhibitor-treated DO rats. E Immunoblotting of RUNX2, BSP, OPN, and OCN proteins in the bone tissues of miR-150-5p mimic or miR-150-5p inhibitor-treated DO rats. F Serum levels of CTX-I and TRAP5b in miR-150-5p mimic or miR-150-5p inhibitor-treated DO rats determined by ELISA. G Immunohistochemistry analysis of the rate of positive osteoblasts in miR-150-5p mimic or miR-150-5p inhibitortreated DO rats. H Immunofluorescence staining analysis (calcein double labeling) for the bone mineralization rate on the trabecular bone surface in miR-150-5p mimic or miR-150-5p inhibitor-treated DO rats. * p < 0.05. n = 6 ( Supplementary Fig. 6G). Hence, the aforesaid results indicate the successful synthesis of GMNPs.

GMNPE promotes the enrichment of EVs in the bone tissues of DO rats
We combined the anti-CD63 antibody targeting EVs to aldehyde-functionalized magnetic nanoparticles to form a Schiff base between the aldehyde group and the primary amine of the antibody, and finally negatively charged GMNP E nanoparticles were formed. Hence, negative charge of GMNP E is conducive to repelling the negatively charged proteins during blood circulation. Immunofluorescence results showed the presence of anti-CD63 antibody on the surface of GMNP E (Supplementary Fig. 7A).
In addition, TGA showed antibodies accounting for 54% of GMNP E (Supplementary Fig. 7B). GMNP E were subsequently co-cultured with EVs. TEM observation results showed an obvious life. hust. edu. cn/ EVmiR NA#!/). B Morphological characterization of the isolated EVs observed using a TEM. C The size distribution of the isolated EVs analyzed by nanoparticle tracking analysis. D Immunoblotting of EV surface makers CD63, HSP70, and TSG101 along with negative EV marker GM130 in the isolated EVs. E miR-150-5p expression in osteoblasts isolated from DO rats and BMSCs determined by RT-qPCR. F Cy3-labeled green fluorescence signal observed under a fluorescence microscope. G Osteoblast viability of osteoblasts co-cultured with EV-miR-150-5p mimic or EV-miR-150-5p inhibitor measured by CCK-8 assay. H Alizarin red S staining of mineralized nodules in osteoblasts co-cultured with EV-miR-150-5p mimic or EV-miR-150-5p inhibitor. I ALP staining of ALP activity in osteoblasts co-cultured with EV-miR-150-5p mimic or EV-miR-150-5p inhibitor. J mRNA expression of RUNX2 and BSP in osteoblasts co-cultured with EV-miR-150-5p mimic or EV-miR-150-5p inhibitor determined by RT-qPCR. K Immunoblotting of RUNX2, BSP, OPN, and OCN proteins in osteoblasts co-cultured with EV-miR-150-5p mimic or EV-miR-150-5p inhibitor. * p < 0.05. The cell experiment was repeated 3 times independently core-shell-corona structure of GMNP E -EVs, accompanied by obvious membrane EVs with vesicle-like structure ( Supplementary Fig. 7C). Fluorescence confocal microscopic observation further confirmed the ability of GMNP E (RhB-labeled GMNP E and PKH67labeled EVs) to capture EVs, wherein EV-related fluorescence (green) and GMNP E -related fluorescence (red) were co-localized ( Supplementary Fig. 7D).
Next, the EV binding capacity of GMNPs containing anti-CD63 was assessed when substituted with IgG isotype control antibody (GMNP N ) as a control. Compared with GMNP N , the capture of EVs by GMNP E was significantly enhanced (Supplementary Fig. 7E).
In addition, immunoblotting results showed the presence of CD63 in GMNP E -EV samples, which was same as the positive control, whereas CD63 was absent in GMNP N -EV samples (Supplementary Fig. 7F). These results confirmed that GMNP E could capture EVs. Next, EVs or GMNP E -EVs were co-cultured with osteoblasts, and the results suggested that GMNP E considerably raised the number of EVs in osteoblasts ( Supplementary Fig. 7G). Furthermore, confocal microscopic images showed that injection with GMNP E -EVs led to more EVs enriched in the osteoblasts of femoral tissues of DO rats than injection with EVs alone ( Supplementary  Fig. 7H). In the other tissues such as brain, heart, lung, liver, spleen, and kidney, GMNP E -EV signal was barely detected (Supplementary Fig. 7I).
Therefore, GMNP E may potentiate the enrichment of EVs in the bone tissues of DO rats.
The aforementioned findings highlighted the pro-proliferative and pro-mature effects of GMNP E -EV-miR-150-5p on osteoblasts.

miR-150-5p stimulates the proliferation and maturation of osteoblasts by targeting MMP14
We then intended to determine the mechanism of miR-150-5p in osteoblasts. miRBD website showed 123 downstream genes of miR-150-5p, which were then submitted to Venn diagram analysis to attain the intersection genes with the osteogenesis-related genes predicted by the GeneCards database. Six genes were found at the intersection, namely, TNF, MMP14, CDK4, IL7, KMT2A, and SYP (Fig. 4A). More importantly, binding sites were observed between MMP14 and hsa-miR-150-5p as well as rno-miR-150-5p (Fig. 4B). Therefore, we selected MMP14 as the target for the subsequent experiments.
In addition, higher expression of MMP14 was observed at the mRNA and protein in the bone tissues of DO rats versus normal rats (Fig. 4C, D). Luciferase analysis revealed a decline in the luciferase activity of MMP14-WT but that of MMP14-MUT was not affected following miR-150-5p mimic (Fig. 4E).
Additionally, miR-150-5p overexpression-induced elevation in the mRNA expression of RUNX2 and BSP along with protein expression of RUNX2, BSP, OPN, and OCN were also reversed by re-expression of MMP14 (Fig. 4J, K, Supplementary Fig. 4D).
Overall, miR-150-5p targets MMP14 and inhibits its expression, augmenting the proliferation and maturation of osteoblasts.
Inhibition of MMP14 activates Wnt/β-catenin and promotes proliferation and maturation of osteoblasts Next, we proceeded to elucidate the molecular mechanism of MMP14 involved in osteogenesis. The GeneMANIA website predicted interaction between MMP14, Wnt1, and CTNNB1 (Fig. 5A). At the same time, the top 50 genes related to osteogenesis were screened by the GeneCards website. A protein-protein interaction (PPI) network of MMP14, Wnt1, and CTNNB1 and osteogenesis-related genes was constructed by STRING database, where MMP14, Wnt1, and CTNNB1 were found to be closely related to osteogenesis (Fig. 5B). KEGG results showed that MMP14 can mediate the Wnt/β-catenin pathway (Fig. 5C). However, it is unknown whether this MMP14-mediated pathway functioned in the osteogenesis process.
Together, suppression of MMP14 may contribute to Wnt/β-catenin activation, thus augmenting the proliferation and maturation of osteoblasts.
GMNPE-EV-miR-150-5p promotes the osteogenesis in DO rats through the MMP14/Wnt/β-catenin axis Finally, we intended to characterize the effect of GMNP E -EV-miR-150-5p on the osteogenesis in DO rats through the MMP14/Wnt/β-catenin axis. Of note, lower levels of miR-150-5p and β-catenin yet higher MMP14   Fig. 4 miR-150-5p strengthens the proliferation and maturation of osteoblasts by binding to MMP14. A Venn plot (http:// bioin forma tics. psb. ugent. be/ webto ols/ Venn/) of downstream genes of miR-150-5p predicted by the miRBD website (http:// mirdb. org/) and the osteogenesis-related genes predicted by the GeneCards website (relevance score ≥ 20; https:// www. genec ards. org/). B Binding sites between MMP14 and hsa-miR-150-5p as well as rno-miR-150-5p. C MMP14 mRNA expression in the bone tissues of normal and DO rats determined by RT-qPCR. D Immunoblotting of MMP14 protein in the bone tissues of normal and DO rats. n = 6. E Binding of miR-150-5p to MMP14 verified by dual-luciferase reporter assay in HEK-293 T cells. F miR-150-5p expression and MMP14 mRNA expression in osteoblasts in the presence of with miR-150-5p mimic alone or combined with oe-MMP14 determined by RT-qPCR. G Osteoblast viability in the presence of with miR-150-5p mimic alone or combined with oe-MMP14 measured by CCK-8 assay. H Alizarin red S staining of mineralized nodules in osteoblasts in the presence of with miR-150-5p mimic alone or combined with oe-MMP14. I ALP staining of ALP activity in osteoblasts. J mRNA expression of RUNX2 and BSP in osteoblasts in the presence of with miR-150-5p mimic alone or combined with oe-MMP14 determined by RT-qPCR. K Immunoblotting of RUNX2, BSP, OPN, and OCN proteins in osteoblasts in the presence of with miR-150-5p mimic alone or combined with oe-MMP14. * p < 0.05. The cell experiment was repeated 3 times independently level were observed in the bone tissues of DO rats treated with either PBS or GMNP E -EV-mimic NC + sh-NC than normal rats. miR-150-5p and β-catenin expression was elevated while MMP14 expression was diminished in the bone tissues of DO rats upon treatment with GMNP E -EV-miR-150-5p mimic. Conversely, the elevation in β-catenin levels by GMNP E -EV-miR-150-5p mimic was offset by sh-β-catenin (Fig. 6A).
In addition, consistent protein levels were obtained as shown by immunoblotting results (Fig. 6B, Supplementary Fig. 4 J). The number of OCN-positive osteoblasts and the bone mineralization rate on the trabecular bone surface were reduced in the DO rats treated with either PBS or GMNP E -EV-mimic NC + sh-NC as compared to normal rats. However, GMNP E -EV-miR-150-5p mimic resulted in elevated number of OCN-positive osteoblasts and bone mineralization rate. Further silencing of β-catenin led to elevations in OCN-positive osteoblasts and bone mineralization rate (Fig. 6C, D). As illustrated in Fig. 6E, ELISA data exhibited increased CTX-I and TRAP5b serum levels in the DO rats treated with either PBS or GMNP E -EV-mimic NC + sh-NC as compared to normal rats, but reduced CTX-I and TRAP5b serum levels in the DO rats treated with GMNP E -EV-miR-150-5p mimic. Loss of β-catenin further caused a decrease in the CTX-I and TRAP5b levels.
Furthermore, micro-CT analysis results suggested a decline in the Tb.Th, BV/TV, and Tb.N and an increase in Tb.Sp in the DO rats treated with PBS or GMNP E -EV-mimic NC + sh-NC. In the DO rats treated with GMNP E -EV-miR-150-5p mimic + sh-NC, Tb.Th, BV/TV, and Tb.N were increased but Tb.Sp was decreased. However, further silencing of β-catenin led to opposite results (Fig. 6F).

Discussion
The data collected from this study supported the proproliferative and pro-mature effect of miR-150-5p in BMSC-EV-loaded GMNPs on the osteoblasts via activation of MMP14-mediated Wnt/β-catenin pathway.
We illustrated decreased miR-150-5p in the bone tissues of DO rats while its re-expression could enhance the osteogenesis in DO rats. In line with these findings, a recent study has demonstrated decreased miR-150-3p expression in osteoporosis . Notably, our findings displayed that miR-150-5p could elevate the expression of osteogenesisrelated genes (RUNX2, BSP, OPN, and OCN) (Wang et al. 2017;Tang et al. 2020), and increased the proportion of OCN-positive osteoblasts and bone mineralization rate, contributing to its pro-mature effect. Meanwhile, it has been confirmed that forced expression of miR-150 in mouse osteoblast cell line MC3T3-E1 leads to increased expression of OC, ALP, type I collagen, and OPN, leading to promotion of bone formation (Dong et al. 2015). These lines of evidence support the promoting property of miR-150-5p in the osteogenesis in DO.
The current study further revealed that miR-150-5p enriched in BMSC-EVs which can carry miR-150-5p to promote the proliferative and mature functions of osteoblasts. Osteoblast maturation can be tested with ALP and markers of osteoblast differentiation (Im et al. 2004;Crous and Abrahamse 2021). A previous study has elaborated the pro-proliferative and prodifferentiation roles of BMSC-EVs in the osteoblasts (Qin et al. 2016). Exosomal miR-150-3p can further strengthen the role of BMSC-derived exosomes in augmenting the proliferating and differentiating potentials of osteoblasts , which is consistent with our findings. Considering that, we constructed GMNP E loaded with EV-miR-150-5p and determined the optimal structure of GMNP E to capture EVs. Subsequent results of this study further demonstrated that GMNP E promoted the enrichment of EVs in the bone tissues of DO rats whereby enhancing the proliferative and differentiating properties of osteoblasts. In partial agreement with our results, hydroxyapatite-coated magnetite (Fe 3 O 4 ) nanoparticles have been demonstrated to enhance the proliferation and differentiation of osteoblasts by increasing secretion of ALP, collagen, and calcium deposition, thus preventing A PPI among MMP14, Wnt1, and CTNNB1 predicted by GeneMANIA website (http:// genem ania. org/). B PPI network of MMP14, Wnt1, CTNNB1, and osteogenesis-related genes constructed by STRING database (https:// string-db. org/) with the Cytoscape 3.6.0 software (https:// cytos cape. org/) applied for visualization. C KEGG enrichment analysis of the candidate genes using R "ClusterProfiler" package. D Immunoblotting of Wnt1 and β-catenin proteins in the bone tissues of normal and DO rats. n = 6. E Immunoblotting to testify the silencing efficiency of sh-MMP14 in osteoblasts. F Immunoblotting to testify the silencing efficiency of sh-β-catenin in osteoblasts. G Immunoblotting of MMP14, Wnt1, and β-catenin proteins in osteoblasts in the presence of sh-MMP14 alone or combined with sh-β-catenin. H Osteoblast viability in the presence of sh-MMP14 alone or combined with sh-β-catenin assessed by CCK-8 assay. I Alizarin red S staining of mineralized nodules in osteoblasts in the presence of sh-MMP14 alone or combined with sh-β-catenin. J ALP staining of ALP activity in osteoblasts in the presence of sh-MMP14 alone or combined with sh-β-catenin. K mRNA expression of RUNX2 and BSP in osteoblasts in the presence of sh-MMP14 alone or combined with sh-β-catenin determined by RT-qPCR. L Immunoblotting of RUNX2, BSP, OPN, and OCN proteins in osteoblasts in the presence of sh-MMP14 alone or combined with sh-β-catenin. * p < 0.05. The cell experiment was repeated 3 times independently ◂ osteoporosis (Tran et al. 2012;Tran and Webster 2013). These findings were further substantiated in the DO rat models, supporting the enhanced osteoblast functions induced by GMNP E -EV-miR-150-5p.
Mechanistic investigations indicated that loss of MMP14, a target gene of miR-150-5p, conferred the pro-proliferative and pro-mature effects of miR-150-5p on osteoblasts. Consistently, miR-150-5p is able to target MMP14 by binding to its 3′-UTR in fibroblast-like synoviocytes in the context of rheumatoid arthritis, thus causing downregulation of its expression (Chen et al. 2018 of MMPs can cause a battery of physiological conditions, associated with imbalanced bone remodeling, leading to bone osteolysis or bone formation (Paiva and Granjeiro 2017). Further investigation explained that inhibition of MMP14 activated the Wnt/β-catenin pathway, thus accelerating osteoblast proliferation and maturation. Wnt pathway has been reported to act as an inducer of osteogenic differentiation of MSCs (Yuan et al. 2016). In addition, osteogenesis of human BMSCs can be stimulated following activation of Wnt/β-catenin pathway . Recent study has proposed that Wnt/βcatenin pathway activation accelerates osteoblast activity, corresponding to increased expression of osteogenesis-related genes RUNX2 and OCN in MC3T3-E1 cells or rat primary osteoblasts (Wang et al. , 2017. Meanwhile, activating the Wnt/ β-catenin pathway delays osteoporosis in ovariectomized rats (Liu and Guo 2020). Thus, disruption of the MMP14-mediated blockade of Wnt/β-catenin pathway may contribute to therapeutic potentials in promoting osteogenesis and inhibiting the DO progression.
Overall, our study indicates that miR-150-5p delivered by BMSC-EV-loaded GMNPs can enhance the osteogenesis through modulating the MMP14-Wnt/ β-catenin axis, thus attenuating DO (Fig. 7). Thus, BMSC-EV-mediated transfer of miR-150-5p may act as a potentially effective strategy for diagnosing and monitoring the progression of DO. However, miR-150-5p exerts its therapeutic effect on DO by inhibiting the inflammatory response in DO rats and its specific molecular mechanism require further exploration in future studies. In addition, Wnt/β-catenin regulation is mainly induced by the two Wnt pathway antagonists: SOST and DKK1 (Rossini et al. 2013). Therefore, in addition to targeting MMP14 to activate Wnt/β-catenin, miR-150-5p may also promote Fig. 7 Schematic diagram of the mechanism by which BMSC-EV-loaded GMNPs affect the progression of DO. GMNPs loaded with BMSC-EVs can transfer miR-150-5p to osteoblasts where miR-150-5p binds to MMP14 and results in decreased MMP14 expression, leading to activation of the Wnt/β-catenin pathway. By this mechanism, proliferation and maturation of osteoblasts are enhanced and the DO progression is delayed. The black upward and downward arrows in the figure represent the upregulation and downregulation of gene expression, respectively miR-150-5p ----

MMP14
β-catenin Wnt1 osteoblast proliferation and maturation by targeting SOST and DKK1 to activate Wnt/β-catenin, thus alleviating DO. Further investigation is required for validation of this speculation.

Data availability
The data that supports the findings of this study are available in the manuscript and supplementary materials.
Code availability Not applicable.

Declarations
Ethics approval The study, approved by the Animal Ethics Committee of the First Affiliated Hospital of Bengbu Medical College (approval no.: 2021-245), was conducted according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.