Exosome isolation

Under aseptic conditions, 10 cm of umbilical cords of healthy newborns were taken from the operating room and placed in a sterile, wide-mouth bottle containing a mixture of 300 ml PBS and double antibody. After the pretreatment in the ultra-clean studio, the umbilical cord was removed and placed in a sterile petri dish containing PBS on the ultra-clean workbench. Wash until PBS is clear, cut the Walton glue separated from umbilical cord to 1 cc fragments with tissue shears, put 1 cc into T-25 cm2 inoculation flask, add 6 ml of medium containing FBS, and culture in a 5%C02 cell incubator at 37 ℃. When the fusion of cells reached 70–80%, the hUCMSCs of the P4 generation with good growth status were selected and cultured in the high-glucose DMEM/F12 medium containing 10% FBS. When the fusion of cells reached 60–80%, the serum-free medium was changed for culture. After 72 h, the sample medium was collected. The obtained supernatant was then passed through a 0.22 μm disposable filter and collected. When the hUCMSCs had reached 60–80% confluence, exosomes were isolated from the collected medium by hypercentrifugation. Supernatant collected from cultured mesenchymal stem cells was centrifuged at 2000 g for 20 min to remove large fragments and dead cells (Optima L-100 XP Ultracentrifuge, Backman, USA), The upper liquid was centrifuged at 10,000 g for 30 min, and the filtered liquid was centrifuged at 100000 g for 80 min. Finally, the precipitated exosomes were collected. The exosomes were stored at − 80 °C. To quantify the isolated exosomes, The protein content of exosomes was examined by BCA assay. Western blotting was used to examine specific exosome surface markers, including CD9 (ab92726; Abcam), CD63(ab13404;Abcam) and TSG101(ab125011; Abcam).

Transmission electron microscopy (TEM) analysis of hUCMSCs-Exo

The exosomes samples were dropped on the carbon supporting membrane mesh for 3–5 min, and then the excess of liquid was removed with filter paper. Drop 2% phosphotungstic acid (Servicebio) on the copper grid of the carbon supporting membrane for 2–3 min, then used a piece of filter paper to wipe off excess solution. Exosome morphology was observed and the images were acquired and analyzed under TEM (HT7700, hitachi).

Particle size analysis of exosomes

Take 5 μl of exosomes separated by hypercentrifugation and dilute them to 1 mL. Pipette the samples gently for thorough mixture, then transfer them to designated sample tank, place the tank into NS300 (Malvern company) nanometer particle size analyzer to test the particle size range of the exosomes. The parameters as follows: particle size range of 50–200 nm, molecular weight range of 1000–20107, Dalton temperature range of 25 ℃, 4.0 mW He–Ne laser, wavelength at 633 nm. Each sample was analyzed continuously at least three times (3–5 times). Data was analyzed using Brookhaven Instruments’ Nanosight NTA data analysis software.

Western blotting analysis

The protein concentration of the exosomes sample was measured by BCA method (Beyotime Biotechnology), 20 μg protein was used for SDS-PAGE (Beyotime Biotechnology). 10 μl RIPA lysate (Beyotime Biotechnology) was added to the same volume of PBS resuspended UCMSCs exosome. The supernatant was extracted after centrifugation at 10000 g at 4 ℃ for 5 min. The supernatant was mixed with Loading Buffer at a volume ratio of 5:1, and the mixture was boiled at 100 ℃ for 15 min in a constant temperature metal bath. Protein samples and protein Maker (Prestained).

Color Protein Ladder15–120 kD, Beyotime Biotechnology) were added into the 12% SDS-PAGE gel hole in sequence. Run the glue at 80 V until the sample runs out of the concentrated glue, then transfer the glue at 100 V to the bromophenol blue to the rubber bottom (EPS-300, Tianneng Shanghai). The membrane was transferred at 200 mA, and the time of membrane transfer was determined according to the size of the target protein. The general rule was 1 min for 1kd. After the transfer, the PVDF membrane was taken out and soaked in 5% skimmed milk TBST, face up, for 1.5 h. Dilute the primary antibody with TBST diluent buffer containing 5% BSA according to the instructions, Then the film was incubated overnight at 4 °C and washed with TBST for 15 min (3 times in total). Incubate at room temperature in secondary antibody dilution (using 5% skimmed milk powder TBST) for 1 h, then wash the membrane with TBST for 10 min (3 times). ECL reagent (A: B = 40:1) was then used to incubate the membrane, and imaging was detected by Typhoon scan (Typhoon parameters: wavelength 473 nm, voltage 485 V).

Grouping and treatment of experimental mice

8 week-old male DB/DB mice (weighing 210 ± 20 g, No. SCXK2014-0004) were purchased from south model company, and were fed at 26 ℃, 38.5% humidity, 12 h of light, and 12 h of darkness cycle (7:30 a.m. to 7:30 p.m.) in the SPF environment. The Animal Ethics Committee of Tongji University approved the protocol for mice. 13 mice were randomly tested for blood glucose and given insulin weekly. Then the DB/DB mice were randomly divided into two groups at 26 weeks: hUCMSC-exosome group (n = 8), injection with 100 μg (Protein concentration) hUCMSC-exosomes suspended in 0.1 ml PBS; PBS group (n = 5), injection with 0.1 ml PBS. 28 days after injection, all mice were sacrificed by cervical dislocation, and the calf tibialis anterior muscle group and gastrocnemius muscle group were completely isolated, which were divided into 3 parts and put into electron microscopy solution, 4% formaldehyde fixation solution and liquid nitrogen for subsequent experiments. These experiments include immunohistochemistry, transmission electron microscopy to assess microvascular morphological changes, 0.5% Evans blue immunofluorescence to assess microvascular permeability, and proteomics to detect differential protein expression changes (Fig. 7).

Immunofluorescence staining

The paraffin tissue sections were successively soaked in xylene, anhydrous ethanol, alcohol of different concentrations and distilled water. Then, the sections were soaked in EDTA antigenic repair solution (PH9.0), boiled and cooled naturally. After washing and sealing, the prepared primary antibody CD31 (1: 200 ab222783 Abcam) was dropped onto the sections. After incubating the secondary antibodies and washing, the tablets were sealed with anti-fluorescence quenching sealing tablets. The images were observed and collected under a Nikon inverted fluorescence microscope.

Microvascular permeability analysis

After injected with 0.5% EB (1.5 ml/kg) into the caudal vein for 1 h, the mice were anesthetized. The calf muscle tissue was collected and fixed in 4% paraformaldehyde solution for 12 h. After automatic dehydrating and embedding, paraffin pathological sections (anti-slide slides) were made. The slices were roasted for 2.5 h, dewaxed with gradient alcohol, hydrated, and sealed with direct glycerin. The distribution of EB in muscle tissue was immediately observed under excitation light of 540 nm using fluorescence microscope. The average fluorescence intensity was evaluated by Image J under the same exposure conditions.

Transmission electron microscopy

The muscle tissues in the electron microscope fixative solution were fixed in turn, rinsed and dehydrated. Pure 812 embedding agent was prepared, and the sample was inserted into the embedding plate. After oven at 37 ℃ for one night, the temperature was raised to 60 ℃ for 48 h. The samples were cut into 60–80 nm sections using an ultra-thin slicer, and the sections were stained and observed under a transmission electron microscope.

Proteomic analysisProtein extraction and labeling

SDT lysis method was used to extract protein from muscle tissue, and BCA (Bio-Rad, USA) method was used for protein quantification. Filter Aided Proteome Preparation (FASP) method was used for trypsin enzymolysis, and the peptides were desalted using the C18 Cartridge. After lyophilized, the peptides were added to 40μL 0.1% formic acid solution for resolution. 100 μg of peptides were taken from each sample and labeled by Thermo’s TMT labeling kit. (Thermo Scientific).

LC–MS/MS analysis

Easy NLC with nanositre flow rate and HPLC liquid phase system was used to separate the samples. After chromatographic separation, the samples were analyzed by Q-Exactive mass spectrometry. For identified and quantitative analysis, original data of mass spectrometry were transferred to Mascot2.2 and Proteome Discoverer1.4.

Bioinformatic analysis

For hierarchical clustering analysis, we used Cluster 3.0 and Java Treeview software. Briefly, the quantitative information of the target protein set is normalized to the interval of (− 1,1). As for the hierarchical clustering heatmap, we used the ComplexHeatMap R package (R Version 3.4) to simultaneously classify the expression levels of the samples and proteins (distance algorithm: Euclide, linkage: Average linkage). The sub-cellular location prediction was carried out using Cello (http://cello.life.nctu.edu.tw/), which was modeled by multi-class SVM classification system. We used Fisher's Exact Test to compare the distribution of the target protein set and the total protein set in GO classification or KEGG pathway. Blast2GO was used to annotate the target protein set, which can be summarized in four steps: Blast, Mapping, Annotation, and Annotation Augmentation at InterProScan. As for KEGG pathway, KAAS (KEGG Automatic Annotation Server) software was used to annotate the target protein set.

Statistical analysis

Image J was used to quantitatively evaluate the mean fluorescence intensity of Evans blue staining and CD31 immunofluorescence, and the analyzed data were presented in a bar chart. Images were taken from the calf muscles for each analysis group and were obtained using the same exposure Settings. The data were analyzed by SPSS 20.0 statistical software. Significant differences between the PBS and exosome groups were evaluated by independent T-test. All data are expressed in X̅ ± SD, P < 0.05 is considered to have significant difference.