SDF-1 secreted by mesenchymal stem cells promotes the migration of endothelial progenitor cells via PI3K/Akt pathway


 Background: Cell-based therapeutics bring great hope in areas of unmet medical needs. Mesenchymal stem cells (MSCs) has been suggested to facilitate neovascularization mainly by paracrine action, and endothelial progenitor cells (EPCs) can differentiate into mature endothelial cells. Studies have demonstrated that a combination cell therapy that includes MSCs and EPCs has a favorable effect on ischemic limbs. However, the mechanism of combination cell therapy remains unclear. Herein, we investigate whether stromal cell-derived factor (SDF)-1 secreted by MSCs contributes to. Furthermore, we examined whether SDF-1 affects EPC migration via Phosphoinositide 3-Kinases (PI3K)/protein kinase B (termed as Akt) signaling pathway.Methods: First, intramuscular MSC injections were supplemented with intravenous EPC injections in the mouse model of hind limb ischemia. The incorporation of Qdot® 525 labeled-EPC into the vasculature and capillary density was evaluated by CD31 immunohistochemistry and immunofluorescence, respectively. Then, the concentration of SDF-1 secreted by MSCs was detected via quantitative immunoassay. Flow cytometry was performed to quantify CXC chemokine receptor (CXCR) 4-positive EPCs. The effect of MSCs on EPC migration was measured by a transwell system and a tube-like structure formation on Matrigel. The SDF-1 antagonist AMD3100 and the PI3K inhibitor wortmannin were separately used to determine the participation of CXCR4 and PI3K into EPC migration. Finally, western blot assay was performed to detect the effect of SDF-1 secreted by MSCs on Akt phosphorylation in EPCs.Results: The combination delivery of MSCs and EPCs via a “dual-administration” approach enhanced the incorporation of EPCs into the vasculature and increased the capillary density in mouse ischemic hind limb. The SDF-1 concentration secreted by MSCs was 2.61 ng/ml after 48 h. CXCR4-positive EPCs increased after incubation with MSC-conditioned medium (CM). MSCs contributed to EPC migration and tube-like structure formation, both of which were suppressed by AMD3100 and wortmannin. Phospho-Akt induced by MSC-CM was attenuated when EPCs were pretreated with AMD3100 and wortmannin.Conclusions: The paracrine action of MSCs contributes to EPC migration. Furthermore, SDF-1 secreted by MSCs induces EPC migration. The mechanism of this migration is related to the activation of the Akt pathway


Unilateral Hind Limb Ischemia (HLI) Model and Cell Transplantation
The femoral arteries of 6-month-old (30-35 g) male mice were ligated to induce left hind limb ischemia as described previously [14]. To track EPC incorporation at early time point after transplantation (on day 3), we seeded passage 3 cells in EGM in a 60 mm culture dish. When 80-90% confluence was reached, the cells were labeled using the Qtracker® 525 Cell Labeling Kit (Life technology, Carlsbad, CA, USA) according to the manufacturer's protocol. At 24 h after operation, the mice with ischemic limbs were randomly allocated into four groups as follows: phosphate-buffered saline (PBS) group; MSC (1 × 10 6 ) group; Qdot® 525 labeled-EPCs (1 × 10 6 ) group; and a combination of MSCs (1 × 10 6 ) and Qdot® 525 labeled-EPCs (1 × 10 6 ) group. Cells or PBS were administrated as above. After 3 days, the adductor muscles of ischemic and healthy limbs were immediately harvested for frozen section samples.
We also investigated the combination effect of MSCs and EPCs on neovascularization at late time point after transplantation (on day 14). At 24 h after operation, the mice were randomly allocated into four groups that received the following injections: PBS; MSCs (1 × 10 6 ); EPCs (1 × 10 6 ); or a combination of MSCs (1 × 10 6 ) and EPCs (1 × 10 6 ). MSCs suspended in 50 µl of PBS or PBS alone were infused to the gracilis muscles at four sites.
Then, EPCs suspended in 20 µl of PBS or PBS alone were injected via the tail vein. After 14 days, the adductor muscles of ischemic and healthy limbs were harvested for paraffin section samples.

Histological Assessment of Transplanted Mice
For frozen section samples, tissues were embedded in OCT compound and snap frozen in liquid nitrogen. Frozen sections with a thickness of 8 µm were mounted on silane-coated glass slides and air-dried for 1 h. The section samples were then washed for 5 min thrice with PBS and blocked with normal goat serum (Solarbio, Beijing, China) for 20 min at room temperature (RT). Subsequently, the sections were stained with rabbit antibody against mouse CD31 (1:100; Cell Signaling Technology, Danvers, MA, USA) overnight at 4 °C. After three washes in PBS for a total of 30 min, a secondary TRITC-conjugated goat anti-rabbit IgG antibody (1:50; ZSGB-BIO, Beijing, China) was added for 30 min at RT. Any excess liquid was removed from the specimen, and one drop of the mounting reagent (glycerinum: PBS = 9:1) was applied (Thermo Fisher Scientific) to the specimens.
Photographs were taken using a Zeiss LSM 510 META laser confocal microscope (Zeiss, Germany).
For paraffin section samples, tissues were fixed, dehydrated, and paraffin embedded.
Paraffin sections with a thickness of 5 µm were prepared. Vascular density was determined by quantifying the CD31-positive vessels/mm 2 present in the peri-infarct region. The sections were incubated with primary rabbit anti-mouse CD31 (1:200; Abcam) and then with secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (1:10000; Abcam). After rinsing with PBS thrice, a DAB working solution was added for 5 min. The sections were counterstained with hematoxylin for 10 s and mounted with neutral balsam. Photographs were taken using an inverted microscope (Olympus, Japan). CD31-positive staining was measured in two sections of four distinct views of each specimen by using the Image-Pro Plus 6.0 software.

Migration Assays
To investigate the effect of SDF-1 secreted by MSCs on EPC migration, we performed the assay using the transwell assembly (Corning) with 6.5 mm diameter inserts (8 µm pore size) as described previously [15]. Briefly, EPCs were pretreated with 50 ng/ml AMD3100 for 2 h [7] and 0.1 µM wortmannin for 1 h [16]. AMD3100 is a highly selective antagonist of SDF-1 that binds to its receptor, CXCR4. By comparison, wortmannin is a PI3K inhibitor.
Non-pretreated EPCs and pretreated EPCs (1 × 10 5 ) were harvested and suspended in 100 µl of EBM-2 supplemented with 2% FBS and then reseeded in the upper compartment.
MSCs (5 × 10 4 ) were suspended in 600 µl EBM-2 supplemented with 2% FBS and replated in the lower compartment of the transwell chamber. After incubation at 37 °C for 18 h, the cells on the filters were stained with 0.1% crystal violet. Thereafter, the filters were washed with 33% acetic acid, and the OD 570 nm value of the eluate was detected using a spectrophotometer (Thermo Fisher Scientific).

Preparation of MSC-conditioned Media (MSC-CM)
MSCs were seeded in DMEM supplemented with 10% FBS until 90% confluence. After washing the cells with PBS, the medium was changed into EBM supplemented with 1% BSA and conditioned at 37 °C and 5% CO 2 . After 24 and 48 h, the medium was collected and centrifuged at 300 g for 10 min to remove cell debris and then filtered (0.22 µm pore size; Merck Millipore, Billerica, MA, USA). The control medium comprised EBM and 1% BSA in the absence of cell procedure.

Flow Cytometry Analysis of CXCR4 Expression on EPCs Induced by MSC-CM
We investigated whether SDF-1 secreted by MSCs affects CXCR4 expression in EPCs. First, quantitative immunoassay was used to assess the ability of MSCs to produce SDF-1 according to the manufacturer's protocol (Elabscience, Wuhan, China). The MSC-CM obtained above was detected. Data were acquired using a spectrophotometer, and measurement wavelength was 450 nm.

Tube-like Structure Formation Assays
The tubule formation assays were performed with a thick Matrigel (BD Biosciences) according to the manufacturer's instruction. Briefly, a pre-cooled 48-well plate was coated with the Matrigel, which was melted into liquid at 4 °C overnight. The plate was placed at 37 °C and 5% CO 2 for 45 min to allow polymerization of the Matrigel. Meanwhile, EPCs were pretreated with 50 ng/ml AMD3100 for 2 h and 0.1 µM wortmannin for 1 h. The nonpretreated EPCs and pretreated EPCs (2 × 10 4 ) were suspended in 350 µl of MSC-CM or EBM supplemented with 2% FBS and then inoculated on top of the Matrigel. After incubating for 6-8 h at 37 °C and 5% CO 2 , all wells were photographed (× 5 amplification) using an inverted microscope (Zeiss). Tubule formation was quantified with the Angiogenesis analyzer from Image-Pro Plus 6.0.

Western Blot Assays
To investigate the migration signaling, we pre-incubated EPCs with 50 ng/ml AMD3100 for

Neovasculature
To study the effect of MSCs on recruitment of EPCs from the systemic circulation, we measured the incorporation of injected EPCs into the microvasculature in the ischemic hind limb. Transplanted EPCs labeled with Q-tracker were identified in tissue sections by green fluorescence, whereas the native mouse vasculature stained by anti-CD31 antibody was identified by red fluorescence in the same tissue sections. Three days after cell administration, the incorporation of EPCs into vasculature increased in the combined group compared with that in the EPC group (Fig. 1).

Local Delivery of MSCs Increase Vascular Ratio per Area
Angiogenesis promoted by MSCs results from the paracrine effect; hence, we evaluated its beneficial effect with the combined cell therapy. The nuclei of muscle cells were on the edge, and they moved inward when muscle cells degenerated. At 14 days after cell delivery, hematoxylin and eosin (HE) staining showed that a small number of nuclei were located in the center of muscle cells in the combination group of MSCs and EPCs in contrast to the PBS and EPC groups.
The vascular ratio per area was represented by CD31-positive staining. It was significantly higher in the combined group than in the PBS and MSC groups. Although the vascular ratio increased in the EPC group, no statistically significant difference was observed compared with the combined group (Fig. 2).

SDF-1 Production by MSCs and CXCR4 Expression by EPCs
We performed ELISA assay with MSC-CM.

Effect of SDF-1 Secreted by MSCs on EPC Migration
EPC migration is a critical step in neovascularization. Transwell migration assay revealed that EPCs, which migrated to the lower surface of the inserts, significantly increased in the MSC-CM group compared with the EBM group. We examined the mechanism of SDF-1 secreted by MSCs on EPC migration with AMD3100 and wortmannin. Notably, the preincubation of AMD3100 or wortmannin resulted in significantly less EPC migration than that seen in the MSC-CM group (Figs. 4A and B). The results demonstrated that blockade of CXCR4 remarkably suppressed EPC migration induced by MSCs similar to the blockade of PI3K.

Effect of SDF-1 in MSC-CM on Tube-like Structure Formation of EPCs
Cell migration was also included in the process of tube-like structure formation. Compared with EBM, EPCs migrated, assembled, and formed complete tube-like structures under MSC-CM induction. The pre-incubation of AMD3100 or wortmannin resulted in incomplete tubule formation than that seen in no pre-incubation group containing MSC-CM. These structures were quantified with total mesh area and total segment length. The data indicated that the number of total mesh area and total segment length significantly decreased with AMD3100 or wortmannin pre-incubation. The blockade of CXCR4 has been suggested to suppress the formation of a tube-like structure similar to the blockade of PI3K.
(A) Tube-like structures were formed by EPCs under the indicated conditions (n = 3). Scale bar, 200 μm. (B) Quantification of total mesh area and total segment length in the tubelike structure. The experiment was repeated thrice. **p < 0.01, ***p < 0.001.

Effect of SDF-1 in MSC-CM onAkt Phosphorylation in EPCs
To investigate whether SDF-1 in MSC-CM affects Akt phosphorylation, we evaluated Akt Ser473 phosphorylation in EPCs stimulated with MSC-CM for 30 min. Western blot analysis revealed that the level of phospho-Akt remarkably increased, and the increase was significantly suppressed by AMD3100 or wortmannin (Fig. 6).  [20]. In the present study, we used late EPCs to identify their integration into endothelium because early EPCs generate high levels of angiogenic cytokines, and late EPCs have the potential to form blood vessels [21].
MSCs are non-hemopoietic stromal cells, which are characterized by the multilineage differentiation potential, paracrine action, and low immunogenicity. MSCs make an important contribution to postnatal vasculogenesis, especially during tissue ischemia [22].
Mounting evidence shows that paracrine action probably underlie the vascular effects of MSCs [23][24][25]. Moreover, the conditioned medium of MSC cultures induces fibroblast, keratinocyte, and endothelial cell migration and promotes the formation of capillary-like structures by HUVECs [26].
In combined cell therapy, we critically consider the delivery methods of cells aside from the selection of cell types. To maximize the local concentration of EPCs in the ischemic area and increase cell invasion, we chose to implement a "dual-administration" approach, which was developed by Franz and Bartsch to treat patients with arterial occlusive disease [27][28][29]. In this approach of the present study, intramuscular MSC injections were supplemented with intravenous EPC injections in contrast to previous works in which the subjects received a mixture of MSCs and EPCs [30,31]. The theoretical foundation of the design is that cytokines are the crucial stimulating factor for stem cell homing, and they build an attractive gradient, forming migratory route and guiding EPC migration to the region to be vascularized [32,33]. To generate as many chemoattractants as possible, we envisioned that a vast array of chemoattractants secreted by intramuscular-administrated MSCs are locally deposited in ischemic muscle, attracting the homing of intravenousadministrated EPCs. Our results showed that Qdot® 525 labeled-EPCs in vasculature increased, and vessel numbers also increased when the combination of MSCs and EPCs was infused into the hind limb of ischemic mice. Thus, we conclude that chemoattractants secreted by MSCs promote EPC migration to the neovascularization sites.
We explored the molecular mechanism involved in EPC migration via in vitro experiments.
Previous studies have suggested that SDF-1 induces EPC migration after binding to CXCR4, which is highly expressed on EPCs [34,35]. In the present work, SDF-1 concentration increased in MSC culture supernatants of different time spans. In addition, the level of CXCR4 protein in EPCs increased after incubation in MSC-CM. The data indicated that SDF-1 produced by MSCs promotes CXCR4 expression in EPCs.
Akt, a multifunctional serine/threonine protein kinase, is the downstream of class I PI3K and various receptors. PI3K-Akt signaling pathway also has a crucial effect on multiple processes, including cell proliferation, cell survival, cell migration, activation of integrins, MMP, and angiogenesis. Using both chemical inhibitors to detect the role of PI3K/Akt signaling, we found that CXCR4 and PI3K participate in EPC migration. Furthermore, Akt phosphorylation results in cytoskeleton changes in many cells [36]. In the present study, Akt phosphorylation induced by MSC-CM was inhibited by both AMD3100 and wortmannin, indicating that phospho-Akt is the downstream of SDF-1/CXCR4/PI3K. Consistent with our results, Yu et al. revealed that SDF-1/CXCR4 mediates the migration of BMSCs toward heart MI through the activation of PI3K/Akt [37]. Dimova et al. reported that SDF-1 treatment of cardiac stem/progenitor cells increase Akt phosphorylation [38].

Conclusion
In a word, we argue that in combined cell therapy, MSCs facilitate the migration of circulating EPCs to neovascularization sites via the SDF-1/CXCR4/PI3K/Akt signaling pathway. Herein, we provided new insights into the mechanisms underlying the effects of combined cell therapy. These novel findings suggest that the modulation of the homing mechanism may be used as a therapeutic strategy to improve the efficacy of stem cell therapy [39].       (B) Quantification of total mesh area and total segment length in the tube-like structure. The experiment was repeated thrice. **p < 0.01, ***p < 0.001. (B) Bar graphs show the ratio of the densitometry measurement of phosphor-Akt to that of β-actin and to that of Akt. Data were obtained from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, NS: no significance.