2.1 Culture of BM-MSCs

The use of human BM-MSCs (lot no. TL281098; Lonza, Basel, Switzerland) was authorised by the provider. Experiments using human cells were approved by The Tokyo Medical University Medical Ethics Committee and the Regulations Concerning Medical Research (D23-036). Briefly, cells were seeded at a density of 2.0 × 105 per 10-cm dish in 8 mL of MesenPro (Thermo Fisher Scientific, MA, USA) medium supplemented with 1% GlutaMAX™ (Thermo Fisher Scientific) and incubated at 37 °C and 5% CO2. BM-MSCs from the fourth passage were used for the experiments.

2.2 Generation of BM-MSCs overexpressing angio-miRNAs

A lentiviral system was used to overexpress miRNA-126, -135b, and -210 in MSCs. Briefly, lentiviral stocks containing human miRNA precursors were obtained from BioSettia Inc. (CA, USA). MSCs were plated on 6-well plates at a density of 1.0 × 105 cells/well and incubated for 24 h. Subsequently, the spent medium was replaced with a medium containing TransDux™ (System Biosciences, CA, USA), and MSCs were incubated with the lentiviral particles for 24 h. MSCs in the lentiviral vector group were transfected with lentiviral vectors containing miRNA-126, -135b, and -210 at a multiplicity of infection (MOI) of 1. MSCs in the control vector group were infected with an empty control vector at the same MOI. After transfection, the cell pools were selected using a medium containing 2.0 μg/mL puromycin for 3 days. The resulting genetically modified cells were named MSC-miRNA126, MSC-miRNA135b, MSC-miRNA210, and MSC-miRNA control.

2.3 miRNA extraction and quantitative polymerase chain reaction (qPCR)

Total miRNAs of cultured MSCs were extracted using an miRNeasy Kit for miRNA Purification (Qiagen, Venlo, Netherlands) following the manufacturer’s protocol. cDNA synthesis was performed using a TaqMan™ MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific). Real-time qPCR was performed using primers specific for miRNA-126, -135b, and -210. U6 was used as an internal control.

2.4 Separation and characterisation of EVs

The separation of EVs was conducted as follows. Briefly, 7 × 105 BM-MSCs were passaged in a 15-cm dish containing 20 mL of StemPro (Thermo Fisher Scientific). After 24 h, the spent BM-MSC culture medium was collected and centrifuged at 2000×g for 10 min. The supernatant was then filtered through a 0.22-μm pore membrane filter and poured into 13.2-mL Open-Top Thinwall Ultra-Clear Tubes (Beckman Coulter, CA, USA). The tubes were ultracentrifuged using an SW 41 Ti rotor (Beckman Coulter) at 210,000×g and 4 °C for 70 min. The supernatant was decanted, and the precipitated EVs were diluted with phosphate-buffered saline (PBS(–)). The EV solution was then ultracentrifuged under the same conditions, the supernatant was decanted, and the pellets were collected. Each type of EV isolated was named after the vector transfected into its BM-MSC source (EVcontrol, -126, -135b, and -210); EVs derived from unmodified BM-MSCs were named EVNative. The concentration of EV-associated proteins was measured using a Qubit Fluorometer (Thermo Fisher Scientific). The particle number and size distribution of EVs were analysed by dynamic light scattering using a NanoSight instrument (Spectris, London, UK) following the manufacturer’s instructions. The ExoScreen assay was performed as previously described to capture CD63 [16]. These EV analyses were performed following the minimal information available for studies on EVs (MISEV2023) [17].

2.5 Tube formation assay

Human umbilical vein endothelial cells (HUVECs; Lonza) were seeded at a density of 2.0 × 105 cells per 10-cm dish in 8 mL of EBM-2 medium (Lonza) and incubated at 37 °C and 5% CO2. Cells from passages five to nine were used for the tube formation assay. HUVECs and EVs from BM-MSCs (0.5, 1, or 10 μg/mL) suspended in EBM-2 medium were seeded on top of a Matrigel layer (Corning, NY, USA) in each well of a 24-well plate at a density of 1.5 × 104 cells and incubated at 37 °C for 24 h. In the control group, HUVECs were seeded onto Matrigel in PBS(–). Images were captured using a BZ-X 800 phase contrast microscope (Keyence, Osaka, Japan), and tube formation was analysed using the Angiogenesis Analyzer in ImageJ software (version 1.54d; https://imagej.net/; National Institutes of Health, Bethesda, MD, USA). The number of junctions, total length of segments, and total mesh areas counted by the software were compared.

We also performed tube formation analysis of HUVECs cultured with single or combined EVs (1 μg/mL) derived from transfected BM-MSCs. Combined EVs refer to a mixture of two or three EVs (EV126 and EV135b; EV135b and EV210; EV126 and EV210; and a combination of EV126, EV135b, and EV210). To analyse the angiogenic capacity of the BM-MSC culture medium, a non-contact co-culture model was established by placing a Transwell insert (Corning) in each well. Briefly, 1.0 × 104 BM-MSCs were seeded in Transwell plates containing MesenPro medium. Tube formation by HUVECs was analysed after they were cultured separately from BM-MSCs in Transwell plates for 24 h. The number of junctions, total length of segments, and total mesh areas were compared between the native EVs, transfected EVs, and co-culture models.

2.6 Mouse hindlimb ischaemia model

All animal experiments were approved by the Institutional Animal Care and Use Committee of Tokyo Medical University (R6-020). All animal studies were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. This work followed the ARRIVE guidelines 2.0. After the experimental procedures, the mice were euthanised via inhalation of pure CO2 gas in a euthanasia chamber.

Briefly, 10-week-old male BALB/c mice were purchased from Japan SLC, Inc. (Hamamatsu, Japan). The mice were allocated to 11 groups with 4 mice per group: sham surgery, PBS(–), native EVs, control EVs, single EVs (EV126, EV135b, or EV210), double EVs (EV126 and EV135b, EV135b and EV210, or EV126 and EV210), and triple EVs (EV126, EV135b, and EV210). The mice were then anaesthetised via intraperitoneal injection of a mixture of midazolam (0.4 mg/mL), butorphanol (0.5 mg/mL), and medetomidine (0.075 mg/mL). Both hindlimbs were fixed in the supine position with extension and abduction. An oblique incision was made on the left groin, and the subcutaneous fat pad was then dissected. The femoral artery, vein, and nerve were identified and separated gently. The common femoral artery was ligated at two points using 6–0 polyvinylidene fluoride sutures (Kono Seisakusho, Ichikawa, Japan).

Unilateral hindlimb ischaemia was induced by excising the left common femoral artery between the two ligations. After the excision of the common femoral artery, 50 μL of EVs containing 4 μg protein were injected into the extensor and flexor muscles of the thigh from the surgical site. PBS(−) was injected into the mice in the control group. The incision was closed using 4–0 silk sutures (Kono Seisakusho). The right hindlimb was used as the internal control. Mice were placed on a heating pad maintained at 37 °C during the procedure. Four mice underwent a sham operation in which the left common femoral artery was exposed but not ligated.

2.7 Evaluation of limbs

The severity of the ischaemic injury was evaluated based on the grade of limb necrosis according to a previous study [18]: grade 0, normal limb without necrosis; grade I, black toenails with necrosis limited to the toes; grade II, necrosis extending to the foot; grade III, necrosis extending to the knee; and grade IV, necrosis extending to the hip or loss of the entire limb.

Blood flow was monitored using a MoorFLPI-2 instrument (Moor Instruments, Devon, UK) on days 0 (before and after the procedure) and 7. To evaluate the alterations in limb perfusion, the blood flow ratio of the left to the right foot was analysed. The region of interest was set at the ankle, which reflected peripheral blood perfusion.

2.8 Histological and immunohistochemical analyses

Limb muscle tissues were immunohistochemically stained with haematoxylin–eosin and anti-CD31 antibodies (Cat. no. GTX130274, GeneTex International Corporation, CA, USA). Briefly, the left hamstrings were harvested from the mice in each group on day 14 (n = 1 per group) and fixed with a 10% formalin solution (Muto Pure Chemicals, Tokyo, Japan). The samples were then embedded in paraffin, and 3-μm-thick serial sections were processed for haematoxylin–eosin staining.

For immunostaining to detect CD31, sections were incubated with anti-CD31 antibodies (1:4000) overnight at 4 °C and subsequently incubated with the secondary antibody Histofine Simple Stain Mouse MAX-PO® (1:1; Nichirei, Tokyo, Japan) for 30 min at 25 °C. To evaluate angiogenesis in ischaemic muscle, we compared the ratio of the area occupied by capillary endothelial cells in ischaemic muscle between mice injected with EVs and PBS(–). The areas of the muscle and capillary endothelial cells were calculated using ImageJ software.

2.9 Statistical analyses

The results are expressed as the mean ± standard error of the mean. Differences between groups were determined using analysis of variance (ANOVA). All data were analysed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was set at p < 0.05.