Isolation and characterization of hucMSCs
Umbilical cords were obtained from healthy newborn fetuses, and Wharton's jelly tissues were isolated under sterile conditions. Tissues were cut to 1–2 mm3 size tissue pieces and inoculated in culture flasks containing 10% fetal bovine serum (FBS) in a cell incubator at 37 °C and 5% CO2 saturated humidity. When cells grew to 80% confluence, tissue blocks were removed, digested, and subcultured with 0.25% trypsin (Gibco, Grand Island, NY, USA) containing EDTA. HucMSCs of Passage 3 were taken and the cell concentration was adjusted to 1 ´ 106/ml. Antibodies labeled with fluorescein isothiocyanate (FITC), Peridinin chlorophyll protein complex (PerCP)-cyanine (Cy)ä5.5, Allophycocyanin (APC), and Phycoerythrin (PE) were used (FITC-CD90, PerCP-Cyä5.5-CD105, APC-CD73, PE-CD45, PE-CD34, PECD11b, PE-CD19, and PE-HLA-DR) (BD Stemflow hMSC Analysis Kit). The cells were incubated with the antibodies at room temperature for 30 min, washed with phosphate-buffered saline (PBS), centrifuged at 300×g for 5 min, and the supernatant discarded. The cells were resuspended in PBS and analyzed using flow cytometry (BD Biosciences, San Jose, CA, USA). HucMSCs of Passage 3 were seeded in 24-well cell culture plates (Corning, NY, USA). When the cells grew to 80–90% confluence, they were replaced with osteogenic differentiation induction conditioned medium, adipose differentiation conditioned medium, or cartilage differentiation conditioned medium (Cyagen Biosciences, Guangzhou, China). After 21 days of induced differentiation, osteogenic differentiation was detected using with Alizarin red staining, and adipose differentiation was detected using oil red staining.
Isolation and characterization of exosomes derived from HucMSCs
HucMSC-Exosome (Exo1): HucMSCs of passage 3 (P3) were cultured with Dulbecco’s modified Eagle’s medium (DMEM)/F12 conditioned medium containing 10% exosome-free FBS (System Biosciences, Palo Alto, CA, USA), and the cell supernatant was collected after 48 h of culture. Osteogenic differentiation-Exosome (Exo2): HucMSCs of P3 were cultured to 80% confluence, washed twice with PBS, and replaced with osteogenic induction differentiation conditioned medium containing 10% exosome-free FBS, with 50 M vitamin C, 10 mM beta-phosphoglycerol, and 0.1 M dexamethasone (Sigma, St. louis, MO, USA). Cell supernatants were collected after 48 h of culture.
Collected cell supernatants were subjected to 2000 ´ g gradient centrifugation (Eppendorf, Hamburg, Germany) at 4 °C for 20 min and the pellet was discarded. The supernatant was removed to a new centrifuge tube and centrifuged at 10,000 ´ g at 4 °C for 40 min. The pellet was discarded and the supernatant was removed to a new centrifuge tube. The pellet was collected by centrifugation at 100,000´ g for 60 min, resuspended by PBS and centrifuged at 100,000 ´ g (Beckman, USA) for 60 min at 4 °C. The pellets were resuspended in PBS, and then filtered and sterilized through 0.22 μm sterile filter membrane, and stored in a freezer (Panasonic, Osaka, Japan) at -80 °C
After adjusting the exosome suspension to the appropriate concentration with PBS, the exosomes were dripped on special carbon-film copper mesh for electron microscopy observation. The exosomes on the mesh were stained with 2% phosphotungstic acid for 2–3 min, air-dried naturally, and photographed using transmission electron microscopy (Hitachi, Tokyo, Japan). Western blotting was used to detect three exosome-derived proteins, CD9, CD81, and HSP70 (ProteinTech, Rosemont, IL, USA). Vesicle diameter in the exosome suspensions was measured using particle size analyzer (Malvern Nanosight NS300, Malvern, UK).
Isolation of osteoblasts
Osteoblasts (OBs) were derived from the skulls of suckling SD rats. OBs in the skull were isolated using an appropriate amount of 0.2% type II collagenase (Gibco，NY，USA) and inoculated into T75 culture flasks (Corning, NY, USA) containing DMEM medium (Gibco，NY，USA) (containing 10% FBS, penicillin 100 U/ml, streptomycin 0.1 mg/ml).
Cell Counting Kit-8 (CCK-8)
The isolated OBs were seeded into 96-well plates at a concentration of 3 ´ 103/ml, and 100 μL of cell suspension was added into each well. After 12 h, Exo1 and Exo2 were co-cultured with the OBs at three concentration of 0, 0.05, 0.1, and 0.2 mg/ml. Four control duplicate wells were set at each concentration. After incubation for 36 and 60 h, 10 μl of CCK-8 (MCE, Monmouth Junction, NJ, USA) reagent was added to each well, and after incubation at 37 °C for 2 h in the dark, the absorbance at 450 nm wavelength was detected using a microplate reader (Tecan, Männedorf, Switzerland).
Alkaline phosphatase (ALP) and alizarin red staining
The isolated OBs were seeded in 24-well plates, and after 24–48 h, when the cells had grown to 80% confluence, osteogenic induction was performed using osteogenic induction medium. Co-cultures with Exo1 and Exo2, at 0.2, 0.1, and 0 mg/ml were set, respectively. Each concentration gradient had four duplex holes, and the solution was changed every 3 days. Alkaline phosphatase kit(Beyotime, Shanghai, China) was used for analysis after 10 days. After 21 days, alizarin red (Solarbio, Beijing, China) staining was performed according to the manufacturer’s instructions. The results for the groups were compared and analyzed according to the depth of staining.
OBs were co-cultured with Exo1 and Exo2 in osteogenic induction medium conditions, respectively. Three concentration of 0.1, 0.2, and 0 mg/ml were set. After 7 days of culture, the cells were collected and 150 μL of Radioimmunoprecipitation assay (RIPA) Lysis Buffer (Beyotime, Shanghai, China) was added to each well of the 6-well plate for lysis. All protein concentrations were determined using a bicinchoninic acid (BCA) protein concentration determination kit (Beyotime). The protein concentrations in the samples were adjusted so that the protein content was the same in the same volume. The SDS-PAGE gel preparation kit (Beyotime) was used for constant voltage protein electrophoresis at 80 V for 40 min, followed by 120 V for 30 min. The separated proteins were transferred to PVDF membranes (0.2 μm; Thermo Scientific, Rockford, IL. USA) at 300 mA for 80 min. The membranes were the incubated in 10% skimmed milk (BD) for 1 h at room temperature, incubated overnight with primary antibodies at 4 °C, followed by incubation with horseradish peroxidase-labeled secondary antibody (Beyotime) at room temperature for 1 h. Immunoreactive proteins on the membranes were visualized using a chemiluminescent imager (Bio-Rad, Hercules, CA, USA) using extremely hypersensitive ECL luminescent reagents. The mouse-derived primary antibodies recognized target proteins of osteogenic differentiation including RUNX family transcription factor 2 (RUNX2), osteopontin (OST), collagen type I alpha 1 Chain (COL1A1), and ALP (Abcam, Cambridge, MA, USA), and the internal reference was detected using mouse-derived anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies (ProteinTech). The protein expression intensity was compared and analyzed according to the ratio of the gray value of target protein to the gray value of the internal reference protein.
Animal model of osteoporosis
This experimental study was approved by the Medical Laboratory Animal Ethics Committee, and all experimental procedures were in accordance with the procedures of the Medical Laboratory Animal Center of Shantou University Medical College. All C57BL/6J mice were purchased from Beijing Weitong Lihua Laboratory Animal Technology Co., Ltd (Beijing, China). Forty 10-week-old C57 female mice were divided into five groups with eight mice in each group: Sham group, ovariectomized (OVX) group, OVX+Exo1 group, OVX+Exo2 group, and OVX+E2 (estradiol) group. Referring to experimental studies such as Zhen Lian , all mice in the experimental group were anesthetized with 2% pentobarbital sodium. Except for the sham group, ovariectomy was performed to simulate the osteoporotic disease model caused by estrogen deficiency. The sham group underwent a sham operation as the control group. A small incision was made on the back of mice to remove the ovaries, including part of the fallopian tube, and the incision was sutured with 5-0 artificial absorbable sutures. After 1 week of recovery time, the experiment was carried out. The OVX+Exo1 group, OVX+Exo2 group, and OVX+E2 group were intraperitoneally injected with Exo1 (0.5 mg/kg), Exo2 (0.5 mg/kg), and E2 (0.15 mg/kg), respectively. The OVX group was injected with the same volume of PBS. Injection every 3 days for 6 weeks.
Six weeks later, the tibia of mice were taken, and five groups were subjected to micro computed tomography (microCT) scanning, to detect and analysis various data indicators: Bone volume (BV), relative bone volume (BV/trabecular volume (TV)), cortical bone area (Ct.Ar), cortical bone thickness (Ct.Th), bone surface (BS), ratio of bone surface area to bone volume (BS/TV), bone mineral content (BMC), bone mineral density (BMD), trabecular number (Tb.N), trabecular separation/Spacing(Tb.Sp), and trabecular thickness (Tb.Th). After microCT (SCANCO, Wangen-Brüttisellen, Switzerland) scanning, mouse tibias were fixed with 4% paraformaldehyde and decalcified with 10% EDTA for 4 weeks, and paraffin-embedded sectioning was performed. After hematoxylin and eosin (HE) staining, the sections were observed and analyzed under a microscope.
MicroRNA Sequencing Analysis
When the P3 generation HucMSCs grew to 80% confluence, they were divided into three groups, and each group had three samples. The control group was replaced with conditioned medium without exosome serum, and the supernatant was collected after 48 h of culture; osteogenic group 3 and osteogenic group 7 were replaced with osteogenic induction differentiation medium containing 10% exosome-free serum, and the supernatant was collected after 48 h and 7 days of culture, respectively. The cell supernatants of the three groups were separated by ultracentrifugation. RNA sequencing uses next generation sequencing (NGS) technology was used to obtain the sequences of miRNAs (18–30 nt or 18–40 nucleotides) in exosomes. Sequencing data were then compared with databases to identify and analyze the small RNA sequences.
The sequencing results of three groups of miRNAs analyzed statistically for differentially expressed miRNAs. The p-value was corrected by multiple hypothesis tests using the Q value. Genes with two or more coincidence differences and a Q-value less than or equal to 0.001 were considered as significant differentially expressed genes. The domain value of p-value was determined by controlling the FDR (False Discovery Rate). The FDR value of the difference test was obtained, and the multiple of differential expression of a gene between different samples was calculated according to the expression amount of the gene, as calculated using the FPKM value (Fragments Per Kilobase of transcript per Million mapped reads). The smaller the FDR value, the greater the difference multiples, indicating more significant expression differences. Differentially expressed genes were defined as those with an FDR < 0.001 and > 2-fold expression difference.
RNAhybrid , miRanda , and TargetScan  were used to predict the potential target genes of the miRNAs, and then the intersections of the results predicted by three software were noted. According to the results of differential miRNA detection, hierarchical clustering analysis was performed using the pheatmap function in the R software to form a clustering heat map of differentially expressed miRNAs between the groups. The biological functions of genes were investigated using Kyoto Encyclopedia of Genes and Genome (KEGG) Pathway analysis. According to KEGG Pathway  public database, pathway significance enrichment analysis was carried out to identify those pathways that were significantly enriched in candidate genes compared with the whole genome background. Pathways with a Q value < 0.05 were defined as pathways that were significantly enriched in differentially expressed genes. Pathway significant enrichment can identify the most important biochemical metabolic pathways and signal transduction pathways in which the candidate genes participate.
Statistical analysis was performed using the software Statistical Product and Service Solutions (SPSS)19.0 (IBM Corp., Armonk, NY, USA). GraphPad Prism 8.0 (GraphPad. Inc, La Jolla, CA, USA) and ImageJ Launcher software (NIH, Bethesda, MD, USA) were used for image editing and gray value analysis, respectively. The two groups were compared and analyzed using t-tests. Differences were considered significant at P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).