Exosomes Derived from Bone Marrow Mesenchymal Stem Cells Promote Angiogenesis in Ischemic Stroke Mice via Upregulation of MiR-21-5p

Exosomes derived from bone mesenchymal stem cells (BMSC-Exos) are one of the main factors responsible for the therapeutic effects of BMSCs. The study aimed to investigate whether BMSC-Exos could promote angiogenesis in ischemic stroke mice via miR-21-5p. In ischemic stroke mice, the therapeutic effects of BMSC-Exos were evaluated by neurological functions and infarct volume. Microvessel density was detected by BrdU/vWF immunofluorescence staining. In in vitro experiments, the proangiogenic effects of BMSC-Exos were assessed via proliferation, migration, and tube formation of human umbilical vein endothelial cells (HUVECs). The miR-21-5p inhibitor was transfected into BMSCs using Lipofectamine 2000. miR-21-5p expression was detected by qRT-PCR. The expression levels of VEGF, VEGFR2, Ang-1, and Tie-2 were determined by Western blot. BMSC-Exos significantly improved neurological functions and reduced infarct volume, upregulated microvessel density, and miR-21-5p expression after cerebral ischemia. In vitro assays revealed that BMSC-Exos enhanced HUVECs functions including proliferation, migration, and tube formation. BMSC-Exos increased the expression levels of VEGF, VEGFR2, Ang-1, and Tie-2. However, the proangiogenic effects of BMSC-Exos on HUVECs were reversed by the miR-21-5p inhibitor. These results suggest that BMSC-Exos could promote angiogenesis via miR-21-5p upregulation, making them an attractive treatment strategy for stroke recovery.


Introduction
Stroke is one of the leading causes of death and disability worldwide [1]. Up to now, the only recommended drug for the treatment of acute cerebral ischemia is tissue-type plasminogen activator (tPA), which works as a thrombolytic agent. However, the narrow therapeutic time window and potentially fatal hemorrhage severely limit its clinical application [2]. Therefore, novel effective strategies designed to improve functional recovery after ischemic stroke are urgently needed.
A growing number of studies have indicated that stem cell-based therapy represents a new approach to the treatment of ischemic stroke [3]. Mesenchymal stem cells (MSCs) can be obtained easily and expanded rapidly in vitro, which makes them an ideal candidate for cell-based therapy [4]. A variety of studies have demonstrated that transplanted MSCs could promote neurogenesis and angiogenesis, thus facilitating stroke recovery [5,6]. As for the underlying mechanism of these functions, it was initially believed that MSCs could be home to ischemic tissues and differentiated to replace injured cells [7]. However, subsequent

Focal Cerebral Ischemia Model in Mice
Focal cerebral ischemia was induced by middle cerebral artery occlusion (MCAO) as described previously [25]. Briefly, mice were anesthetized with 2% isoflurane, and right common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA) were carefully isolated. Then, a tip blunted and poly-L-lysine coated 6-0 nylon monofilament suture was inserted through right ECA and was advanced approximately 10 mm distal to the right carotid bifurcation to occlude the origin of MCA. After 60 min of ischemia, the suture was carefully withdrawn. Sham-operated mice received identical surgery without suture insertion. During the surgery, rectal temperature of mice was maintained at 37 • C with a homeothermic pad.

Neurological Function Evaluation
In all animals, behavioral tests were performed before MCAO and at 1, 3, 7, and 14 days after MCAO by an investigator who was blinded to experimental group design. Neurological deficit score was evaluated according to Zea Longa score [26]: 0, no deficit; 1, failure to fully extend left forepaw; 2, circling to the left; 3, paresis to the left; 4, depressed level of consciousness and no spontaneous walking. Corner test was carried out as described [27]. Briefly, two 30 cm × 20 cm × 1 cm boards were attached to each other at an angle of 30 • and with a small opening between the two boards. A mouse was placed into the central square facing the corner. When both sides of the vibrissae were stimulated by boards, the mouse then reared forward and upward, after which it turned back to face the open end. Each mouse was tested for ten trials, and the selected turning sides were recorded.

Infarct Volume Assessment
Mice were sacrificed 3 and 14 days after MCAO, and brains were removed and frozen immediately at −20 • C for approximately 5 min and then dissected into 1 mm-thick coronal slices. The slices were stained with 2% 2,3,5-triphenyltetrazolium chloride (TTC) for 15 min and fixed with 4% paraformaldehyde for 24 h. Infarct volume was evaluated by Image J software, and infarct volume percentage was calculated as follows: infarct volume (%) = [left hemisphere volume − (right hemisphere volume − infarct volume)]/left hemisphere volume × 100%.

BMSC-Exos Uptake by HUVECs
To label Exos with green fluorescent dye, Exos were firstly resuspended with PKH67 dye (Sigma-Aldrich, Munich, Germany) and incubated for 5 min, then terminated by 2% bovine serum albumin. Next, Exos were ultracentrifuged at 100,000× g for 1 h to remove unbound dyes. HUVECs were incubated with PKH67-labelled Exos for 12 h, followed by 4% paraformaldehyde fixation for 15 min and DAPI stain for 5 min. The internalization of PKH67-labeled Exos by HUVECs was observed using a fluorescence microscope (Leica, Wetzlar, Germany).

Transwell Migration Assay
Transwell migration assay was carried out using 24-well chambers (8 µm, Corning, Corning, NY, USA). RPMI-1640 of 500 µL containing 1% FBS was added into the lower chamber, and HUVECs (6 × 10 4 cells/well) suspended in 100 µL FBS-free medium were seeded in the upper chamber with or without Exos. After 8 h of migration, nonmigratory cells were removed from the top of the insert membrane using humidified cotton swabs. The migrated cells at the bottom surface of membrane were fixed in 4% paraformaldehyde and stained with 0.1% crystal violet. The migrated cells were imaged and counted at 5 random fields. Each group was triplicated.

Tube Formation Assay
After thawed overnight at 4 • C, 50 µL per well matrigel matrix (Corning, Bedford, MA, USA) was added into precooled 96-well plates and incubated at 37 • C to polymerize for 30 min. Next, HUVECs (2 × 10 4 cells/well) in FBS-free RPMI-1640 containing Exos were seeded onto matrigel-coated plates. Capillary-like tubular structures were captured after 6 h incubation. Total tube lengths from five random microscopic fields were calculated using Angiogenesis Analyzer Image J software. Each group was triplicated.

Quantitative Reverse Transcription-Polymerase Chain Reaction (qRT-PCR)
Total RNA from brain tissues of ischemic boundary zone or cells was extracted using miRNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was produced from the total RNA using Mir-XTM miRNA First-strand Synthesis Kit (TaKaRa, Dalian, China). Subsequently, the product from reverse transcription was amplified with the SYBR Premix Ex Taq Kit (TaKaRa, Dalian, China) on an iQ5 real-time PCR detection system (Bio-Rad, Hercules, CA, USA). Relative expression levels of miRNA were calculated by 2 −∆∆Ct and were normalized to U6. Each sample was repeated 3 times and at least three samples obtained from independent experiments were examined. All primers used in this study are listed in Table 1.

Western Blot Analysis
Total protein from brain tissues of ischemic boundary zone or cells was extracted with RIPA lysis buffer containing protease inhibitor PMSF (Beyotime, Shanghai, China), and quantified using BCA protein assay kit (Beyotime, Shanghai, China). Firstly, equal amounts of protein were separated by 10% sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) and transferred to PVDF membranes (Millpore, Billerica, CA, USA). After being blocked with 5% skim milk in TBST, the membranes were incubated with following primary antibodies overnight at 4 goat anti-Ang-1 (1:1000; Santa Cruz, Santa Cruz, CA, USA). Then, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies of goat anti-rabbit IgG (Cell Signaling Technology, Danvers, MA, USA), goat anti-mouse IgG (Thermo Fisher Scientific, Waltham, MA, USA) or donkey anti-goat IgG (Santa Cruz, CA, USA) at room temperature for 1 h. Signals were visualized by enhanced chemiluminescence detection kit (Millpore, CA, USA). Relative expression levels were normalized to GAPDH.

Statistical Analysis
Data were analyzed by SPSS software (version 25.0, SPSS, Chicago, IL, USA) and presented as mean ± standard error of mean (SEM) unless indicated otherwise. Neurological deficit and corner test data were analyzed by nonparametric Kruskal-Wallis H test. All other data were analyzed by one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls post-hoc test. p < 0.05 was considered statistically significant.

Characterization of BMSCs and BMSC-Exos
BMSCs (P3) exhibited a typical spindle-shaped morphology ( Figure 1A). After differentiation induction, calcium deposits and intracytoplasmic lipid droplets appear that were stained by Alizarin Red S or Oil Red O, respectively ( Figure 1B,C), indicating that BMSCs had the potential for osteogenic and adipogenic differentiation. To characterize BMSCs phenotype, BMSCs surface markers were analyzed by flow cytometry. The results showed that these cells were strongly positive for CD29 (99.83%) and CD90 (99.88%) (mesenchymal stem-cell specific markers), whereas they were negative for CD34 (1.44%) and CD45 (0.30%) (hematopoietic cell-specific markers) ( Figure 1D).
As presented in Figure 2A, BMSC-Exos had a typical spherical or cup-shaped morphology. NTA revealed a bell-shaped curve of BMSC-Exos size distribution with a peak at approximately 108.4 nm, and the average concentration of the particles was 4.8 × 10 7 particles/mL ( Figure 2B). Additionally, isolated BMSC-Exos expressed exosomal markers CD9, CD63, and TSG101 without the expression of non-exosome marker protein GAPDH ( Figure 2C). These characteristics indicated that the BMSC-derived particles were Exos.

BMSC-Exos Ameliorated Ischemic Brain Injury in Mice
To determine whether BMSC-Exos improved neurological function recovery, the neurological deficit score and the corner test were performed before MCAO and at 1, 3, 7, and 14 days after MCAO in mice ( Figure 3A). The results demonstrated that BMSC-Exos decreased the neurological deficit score and right-turn number in a dose-related manner ( Figure 3B,C). Compared with the MCAO group, BMSC-Exos significantly reduced infarction volume three days after ischemia. At 14 days after ischemia, apparent atrophy of the ischemic hemisphere was observed in all groups of mice, except the sham-operated group. The cerebral atrophy in the BMSC-Exos group was markedly reduced compared to the MCAO group. In addition, compared with the 25 µg BMSC-Exos group, the infarct volume was significantly decreased in the 50 µg BMSC-Exos group at 3 and 14 days, respectively ( Figure 3D,E). As presented in Figure 2A, BMSC-Exos had a typical spherical or cup-shaped morphology. NTA revealed a bell-shaped curve of BMSC-Exos size distribution with a peak at approximately 108.4 nm, and the average concentration of the particles was 4.8 × 10 7 particles/mL ( Figure 2B). Additionally, isolated BMSC-Exos expressed exosomal markers CD9, CD63, and TSG101 without the expression of non-exosome marker protein GAPDH ( Figure 2C). These characteristics indicated that the BMSC-derived particles were Exos.   As presented in Figure 2A, BMSC-Exos had a typical spherical or cup-shaped morphology. NTA revealed a bell-shaped curve of BMSC-Exos size distribution with a peak at approximately 108.4 nm, and the average concentration of the particles was 4.8 × 10 7 particles/mL ( Figure 2B). Additionally, isolated BMSC-Exos expressed exosomal markers CD9, CD63, and TSG101 without the expression of non-exosome marker protein GAPDH ( Figure 2C). These characteristics indicated that the BMSC-derived particles were Exos.  ( Figure 3B,C). Compared with the MCAO group, BMSC-Exos significantly reduced infarction volume three days after ischemia. At 14 days after ischemia, apparent atrophy of the ischemic hemisphere was observed in all groups of mice, except the sham-operated group. The cerebral atrophy in the BMSC-Exos group was markedly reduced compared to the MCAO group. In addition, compared with the 25 μg BMSC-Exos group, the infarct volume was significantly decreased in the 50 μg BMSC-Exos group at 3 and 14 days, respectively ( Figure 3D,E).

BMSC-Exos Promoted Angiogenesis in the Ischemic Boundary Zone
Proliferative microvessel density was estimated by BrdU/vWF immunofluorescent staining on day 14 after MCAO. The number of BrdU + /vWF + cells was significantly increased in the MCAO and BMSC-Exos groups compared with the sham group. Furthermore, microvessel density in the BMSC-Exos group was significantly higher than that in the MCAO group ( Figure 4A,B).

BMSC-Exos Promoted Angiogenesis in the Ischemic Boundary Zone
Proliferative microvessel density was estimated by BrdU/vWF immunofluorescent staining on day 14 after MCAO. The number of BrdU + /vWF + cells was significantly increased in the MCAO and BMSC-Exos groups compared with the sham group. Furthermore, microvessel density in the BMSC-Exos group was significantly higher than that in the MCAO group ( Figure 4A,B).

BMSC-Exos Increased VEGF/VEGFR2 and Ang-1/Tie-2 Protein Expressions after MCAO in Mice
To test whether BMSC-Exos promotes angiogenesis via upregulating the expression of VEGF, VEGFR2, Ang-1, and Tie2, the expression levels of these proteins were detected by western blot analysis. Compared with the sham group, the expression levels of VEGF, VEGFR2, Ang-1, and Tie-2 were increased in the MCAO and BMSC-Exos group. Furthermore, the expression levels of these proteins were increased in the BMSC-Exos group compared with the MCAO group ( Figure 5A,B). Meanwhile, double-labeling immunofluorescence indicated that VEGF, VEGFR2, Ang-1, and Tie-2 were co-localized with brain endothelial cells in the ischemic boundary zone ( Figure 5C).

BMSC-Exos Increased VEGF/VEGFR2 and Ang-1/Tie-2 Protein Expressions after MCAO in Mice
To test whether BMSC-Exos promotes angiogenesis via upregulating the expression of VEGF, VEGFR2, Ang-1, and Tie2, the expression levels of these proteins were detected by western blot analysis. Compared with the sham group, the expression levels of VEGF, VEGFR2, Ang-1, and Tie-2 were increased in the MCAO and BMSC-Exos group. Furthermore, the expression levels of these proteins were increased in the BMSC-Exos group compared with the MCAO group ( Figure 5A,B). Meanwhile, double-labeling immunofluorescence indicated that VEGF, VEGFR2, Ang-1, and Tie-2 were co-localized with brain endothelial cells in the ischemic boundary zone ( Figure 5C).

BMSC-Exos Increased miR-21-5p Expression after MCAO in Mice
To study the changes in specific miRs expression levels after stroke, the expressions of several candidate miRs, including let-7i-5p, miR-21-5p, miR-22-3p, and miR-486, were detected by qRT-PCR in the ischemic boundary zone at day 14 after MCAO. The expression levels of these miRs were increased in the BMSC-Exos group compared with the MCAO group ( Figure 6). In particular, miR-21-5p expression levels were increased approximately 22-fold ( Figure 6).  To study the changes in specific miRs expression levels after stroke, the expressions of several candidate miRs, including let-7i-5p, miR-21-5p, miR-22-3p, and miR-486, were detected by qRT-PCR in the ischemic boundary zone at day 14 after MCAO. The expression levels of these miRs were increased in the BMSC-Exos group compared with the MCAO group ( Figure 6). In particular, miR-21-5p expression levels were increased approximately 22-fold ( Figure 6).

HUVECs Uptake BMSC-Exos
To confirm that BMSC-Exos could be internalized into HUVECs, BMSC-Exos were labeled with PKH67 and then added to HUVECs for 12-h incubation. Fluorescence images showed that PKH67-labeled BMSC-Exos were located in the cytoplasm of HUVECs ( Figure  7A). This result revealed that our purified BMSC-Exos could be taken up by HUVECs. Figure 6. The expression levels of miRs in ischemic brains were determined by qRT-PCR. n = 3/group. * p < 0.05. ** p < 0.01, ns, no significance; BMSC-Exos, bone marrow mesenchymal stem cell-derived exosomes; MCAO, middle cerebral artery occlusion.

HUVECs Uptake BMSC-Exos
To confirm that BMSC-Exos could be internalized into HUVECs, BMSC-Exos were labeled with PKH67 and then added to HUVECs for 12-h incubation. Fluorescence images showed that PKH67-labeled BMSC-Exos were located in the cytoplasm of HUVECs ( Figure 7A). This result revealed that our purified BMSC-Exos could be taken up by HUVECs.

BMSC-Exos Promoted HUVECs Angiogenesis by Transferring miR-21-5p
Since endothelial cell proliferation, migration, and sprouting are critical in angiogenesis, the effects of BMSC-Exos on HUVECs proliferation, migration, and tube formation were studied. The results indicated that BMSC-Exos could enhance the proliferation, migration, and tube formation of HUVECs ( Figure 7B-H). Therefore, 50 µg/mL BMSC-Exos was selected for the following experiments.

Discussion
In the present study, we firstly confirmed that systemic treatment of BMSC-Exos could improve cerebral ischemia injury and promote angiogenesis in mice. Simultaneously, BMSC-Exos increased proangiogenic protein expression, including VEGF, VEGFR2, Ang-1, and Tie-2, and upregulated miR-21-5p expression in ischemic boundary regions. Then, the miR-21-5p inhibitor was transfected into BMSCs, which further confirmed that BMSC-Exos improved HUVECs angiogenesis via miR-21-5p transfer in vitro.
Increasing evidence suggests that BMSCs exert their therapeutic effects through paracrine mechanisms, including Exos secretion [13,15]. BMSC-Exos have been reported as therapeutic agents for the treatment of central nervous diseases due to the ability to cross the blood-brain barrier [28,29]. More important, exosomes not only show the same effects as BMSCs but also show the advantages of targeted delivery, low immunogenicity, and high biocompatibility. Cross-species administration of MSC-EVs was used in a variety of in vivo experimental models, in which the majority demonstrated beneficial outcomes to reflect the immunocompatibility of EVs [30]. In this study, we first identified the characteristics of BMSCs and BMSC-Exos and further found that BMSC-Exos improved neurological function and reduced infarct volume after cerebral ischemia in mice. Ni et al. also reported that rat BMSC-Exos exerted a neuroprotective function in traumatic brain injury mice [31]. Thus, our research presents new evidence that BMSC-Exos possess cross-species therapeutic effects following cerebral ischemia.
Nowadays, accumulating evidence has shown that angiogenesis is essential in brain tissue repair following stroke, and the promotion of angiogenesis is broadly recognized as a promising therapeutic strategy [32]. Angiogenesis is commonly defined as a multi-step process, involving proliferation, sprouting, migration, and tube formation from pre-existing vasculature [33]. In the present study, the results showed that BMSC-Exos could improve angiogenesis in ischemic stroke mice and promote the proliferation, migration, and tube formation of HUVECs in vitro. Angiogenesis is regulated by multiple proangiogenic factors, among which VEGF, VEGFR2, Ang-1, and Tie-2 play the most important roles. VEGF binds to its receptor VEGFR2 and triggers a downstream angiogenic signaling pathway, which promotes endothelial cell proliferation and migration and ultimately forms a new vascular tube [34,35]. However, the newly formed endothelial cell tubes are unstable due to the lack of pericytes and the formation of the perivascular extracellular matrix. The Ang-1/Tie-2 interaction regulates the maturation of newly formed vasculature, which eventually results in complex vascular network formation [36]. In this study, we found that BMSC-Exos could dramatically improve protein expressions of VEGF, VEGFR2, Ang-1, and Tie-2 in the ischemic boundary zone of stroke in mice. Double immunofluorescence results showed that these molecules were expressed in brain endothelial cells. However, we could only find the improved expression of VEGF and VEGFR2 in HUVECs after BMSC-Exos administration.
qRT-PCR data indicated that BMSC-Exos upregulated the expression of these miRs, among which miR-21-5p showed the greatest effect. Then, the miR-21-5p inhibitor was transfected into BMSCs, and qRT-PCR results showed that the expression level of miR-21-5p was significantly decreased in BMSCs and BMSC-Exos. We also found that the expression level of miR-21-5p was significantly decreased in HUVECs treated with Exos derived from BMSCs transfected with miR-21-5p inhibitor. The proangiogenic effects of BMSC-Exos on HUVECs were greatly abolished by miR-21-5p inhibitor. Moreover, the expression levels of VEGF and VEGFR2 in HUVECs were also significantly reduced by Exos derived from BMSCs transfected with miR-21-5p inhibitor. These results suggest that the therapeutic effects of BMSC-Exos on angiogenesis of HUVECs may depend on miR-21-5p transportation.
This study also has some limitations. First, we did not detect miR expression profile of BMSC-Exos, nor the difference between BMSC-Exos and HUVECs. Although BMSCderived exosomal miR-21-5p was demonstrated to promote angiogenesis in HUVECs in vitro, we could not rule out the possibility that BMSC-Exos increased endogenous miR-21-5p expression. Moreover, we also did not validate the proangiogenic activity of exosomal miR-21-5p by loss-of-function experiments in vivo, which could not determine whether BMSC-Exos promoted angiogenesis by transferring miR-21-5p or upregulating endogenous miR-21-5p in ischemic stroke in mice. Second, the target genes of miR-21-5p need to be further investigated. TargetScan, miRWalk, miRDB, and PicTar databases were used to predict the target genes of miR-21-5p. The results showed that reversion-inducing cysteine-rich protein with kazal motifs (RECK) was one of the target genes of miR-21-5p (data not shown). Some studies have reported that RECK plays an essential role in brain angiogenesis [43,44]. Therefore, it is necessary to confirm whether miR-21-5p plays a pro-angiogenic role in ischemic stroke by targeting RECK in the future.

Conclusions
In conclusion, our study indicates that BMSC-Exos can promote angiogenesis by upregulating miR-21-5p after stroke in mice, which provides new insights into the mechanism of BMSC-Exos that might be used in the treatment of ischemic stroke.