Human bone marrow derived mesenchymal stem cells (MSCs) and human umbilical vein endothelial cells (HUVECs) were used in this study. MSCs were cultured in Dulbecco's modified Eagle's medium (DMEM; Life Technologies, NY, USA) containing a low glucose concentration supplemented with 10% fetal bovine serum (FBS; Life Technologies) and 1% gentamycin. HUVECs were grown in EGM™-2 growth medium (Lonza Biosciences, ML, USA) in 1% gelatin coated culture dishes. All cells were maintained in 5% CO₂ humidified incubator at 37℃.

Adult (7 week-old) male Balb/c mice (weight 21~26 gm) were maintained in a temperature-controlled room with access to food and water ad libitum. All experiments were approved by the Institute Animal Care and Use Committee of Lee Gil Ya Cancer and Diabetes Institute of Gachon University and were conducted humanely (approval number; LCDI-2013-0002). Mice were anesthetized with Zoletil (25 mg/kg) and Rompun (5~10 mg/kg) by intra-peritoneal injection. After anesthesia induction, thrombus of the left femoral artery was induced by topically applying a small strip of filter paper soaked with 10% ferric chloride (Sigma-Aldrich, MO, USA) for 3 mins. After 3 mins, MSCs (10⁶ or 10⁵ cells; high or low doses) were mixed with 80 µl of Plasma Solution A (CJ Cheil Jedang Corp., South Korea), and then injected into the ischemic skeletal muscle area in the single injection model. For control model, the same volume (80 µl) of Plasma Solution A (saline) was intramuscularly injected into ischemic areas. In the double injection model, MSCs or saline were injected once in 2 consecutive weeks in the same ischemic area. Protein and mRNA samples of skeletal muscles were obtained 4 weeks after MSCs first implantation.

After antigen retrieval using sodium citrate solution (ph. 7.2), tissue sections were incubated in blocking solution with normal horse serum to block non-specific binding. The tissues were incubated with specific antibodies (detailed in Supplemental Table 1) overnight at 4℃ and rinsed with phosphate buffer saline (PBS, Life technologies). The tissues were incubated with fluorescence conjugated secondary antibodies for 1 hr (Supplemental Table 1) rewashed with PBS. Nuclei were counterstained with 4', 6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich) at room temperature for 10 mins. Fluorescence signals were detected using a confocal microscope (LSM 710, Carl Zeiss) and fluorescence intensities and blood vessel sizes were measured using Zeiss Zen 2012 software (Carl Zeiss).

MSCs were supplied by FCB-Pharmicell (Sungnam, South Korea). These cells were extracted from healthy bone marrow of 51 years old man and passaged 5~6 times. To confirm their identities, the well-known MSCs specific stem cell markers, CD73 (+), CD105 (+), CD34 (–), and CD45 (–), were checked by fluorescence-activated cell sorting (FACS) analysis. The MSCs differentiated into adipocytes (adipogenic differentiation) or osteocytes (osteogenic differentiation) for confirmation of multipotent and each MSCs derived cells were especially stained with Oil Red O or alkaline phosphatase respectively.

MSCs (2×10⁵ cells) were labeled with CD105, CD34, and CD45 conjugated with Fluorescein isothiocyanate (FITC) and CD73 conjugated with phycoerythrin (PE) for 20 mins at room temperature in the dark. To check negative control, isotype controls was labeled with same FITC and PE. After incubation with antibodies, washed with PBS and stained cells were measured with FACS (Navios™ flow cytometer, Beckman Coultyer) and data analyzed by kaluza software (Beckman Coultyer).

To check the osteogenic differentiation ability, MSCs were incubated with osteogenic differentiation medium which is mixed with 0.1 µM dexamethasone (Sigma-Aldrich), 10 mM β-glycerol phosphate (Sigma-Aldrich) and 50 µM L-ascorbic acid (Sigma-Aldrich) in MSCs growth medium including 10% FBS, 100 U/mL penicillin (Life Technologies), 100 mg/mL streptomycin (Life Technologies), high-glucose DMEM (Life Technologies) for 14 days at 37℃ and 5% CO₂. MSCs (3×10⁵) were planted into 6-well plate (BD Falcon) and the osteogenic differentiation medium was changed every third day. After PBS washed, differentiated cells were fixed with 100% methanol (Merck Millipore, Ohio, USA) in ice for 2 mins and washed with distilled water twice. The cells incubated with alkaline phosphatase substrate, 5-Bromo-4-Chloro-3-Indolyl Phosphate / Nitroblue Tetrazolium (BCIP/NBT) liquid substrate (Sigma-Aldrich), for 10 mins at room temperature. For validation of adipogenic differentiation potency, same number of MSCs was used. Adipogenic differentiation induction medium is prepared with 1 µM dexamethasone, 10 µg/mL insulin, 100 µM indomethacin, and 0.5 mM methyl isobutylzanthine in MSCs growth medium for 72 hrs. After PBS washed twice, differentiated cells were fixed with 4% paraformaldehyde for 4 hrs and stained with 60% oil red-O solution (Sigma-Aldrich) for 45 mins at room temperature. Used reagents for adipogenic differentiation were obtained from Sigma-Aldrich. All differentiated cells were observed by light microscope (Olympus CX21).

To determine the collagen formation in the ischemia area, we have performed Masson's Trichrome staining by using commercial kit (Sigma-Aldrich). To fix the skeletal muscle tissues slides, put slides in the bouin's solution (Sigma-Aldrich) for 15 mins at 56℃. Afterwards, tissues were stained with iron hematoxylin solution (Merck Millipore) for 5 mins to observe nucleus, then Biebrich Scarlet solution (Sigma-Aldrich) was used for 2 mins to stain acidophilic tissue components. These slides were treated with phospho-acids, phosphomolybdic-phosphotungstic acid solution (Sigma-Aldrich), for differentiating tissues. Lastly, tissues slides were washed with running water and collagen of skeletal muscle tissues slides was stained by aniline blue solution (Sigma-Aldrich).

All stained tissues slides were immersed in each absolute alcohol and xylene for 3 mins and these slides were covered with glass by using DPX mounting medium (Sigma-Aldrich) and they were analyzed under the light microscope.

To detect apoptotic cells after ferric chloride treatments, TUNEL was performed according to the instruction. Briefly, tissues sections were incubated with mixture of TUNEL reagent for 1 hr at 37℃ in the dark. Nuclei were counterstained by DAPI. After PBS wash, tissue sections were mounted with Vectashield mounting medium and coverslipped on glass slide and validated by confocal microscope.

Sectioned tissue slides were stained followed by immunofluorescence staining. Diameter of vWF positive blood vessels were measured by Zen 2009 software (Carl zeiss, Germany).

To investigate proteins secreted by MSCs, we measured angiopoietin-1 (ANG-1), hepatocyte growth factor (HGF) and vascular endothelial growth factor (VEGF) and in mesenchymal stem cells-conditioned medium (MSC-CM). Briefly, MSC-CM was incubated in 96-well plates at 37℃ for 90 mins and then each primary antibody for 2 hrs at room temperature and then washed with PBS containing 0.1% Tween 20 (Sigma-Aldrich). Secondary antibodies were administered for 2 hrs at room temperature. After developing with tetramethylbenzidine (TMB, Sigma-Aldrich) solution for 10 mins in the dark and stopping the reaction with 2 N H₂SO₄ (Sigma-Aldrich), absorbances were measured at 450 nm using a spectrophotometer.

To confirm the angiogenic capacity of MSCs, we prepared MSC-CM and implemented angiogenesis assays. When we used HUVECs for this study, all HUVECs starved by incubating them for 24 hrs in HUVECs basal medium which's components is growth medium containing 1% FBS without growth factors. MSCs was incubated for 2 days in normal cell culture condition and then, this MSC-CM was concentrated by Amicon Ultra Centrifusal Filter (Merck Millipore).

Wound migration assay was used for validation of chemotactic motility. A scratch wound of 100% confluent HUVECs was introduced with a sterile pipette tip, and treating cell with PBS or MSC-CM. After incubation for 20 hrs, numbers of cells that migrated into wound fields were counted using the Image J program (NIH).

Tube formation was estimated using a matrigel assay. In brief, growth factor-reduced matrigel (BD Biosciences, San Diego, USA) was spread onto a 4-well plate (BD Falcon) and incubated for 40 mins at 37℃. After HUVECs starvation, HUVECs were seeded (5×10⁴ cells/well) on the matrigel and then, treated with PBS or MSC-CM. After 20 hrs incubation, images were captured by light inverted microscope (Olympus CKX31).

Endothelial cells proliferation was determined using a tetrazolium / phenazine methosulfate (MTS/PMS, Promega) assay. Briefly, HUVECs (1×10⁴ cells/well) seeded in a 96-well plate (BD Falcon) for FBS starvation. Cells were then treated with PBS or MSC-CM at 37℃ for 48 hrs. Freshly prepared MTS/PMS solution (20 µl/well) was mixed HUVECs basal medium (100 µl/well) in the plates and then cells were incubated for 4 hrs at 37℃ in the dark. Absorbance was read at 490 nm using a spectrophotometer (PerkinElmer). All experiments were performed three times in triplicate.

HUVECs were incubated in basal medium (Basal-M) or MSC-CM for 2 days. Whole cell lysates containing equal amount of proteins (30 µg protein/lane) were resolved by 4~12% NuPAGE Bis-Tris gel (Thermo Scientific, NH, USA) and transferred to polyvinylidene fluoride (PVDF) membrane using semi-dry (25 voltage, 10 mins; ATTO, Japan). For protein isolation, cells were collected in RIPA buffer (TAKARA, Japan) with protease inhibitor and phosphatase inhibitors (TAKARA). The PVDF membranes were incubated with 5% skim milk as a blocking solution, followed by proper primary antibodies (Supplemental Table 1) in blocking solution overnight at 4℃.

The membranes were washed with Tris-buffered saline with 0.1% tween 20 (TTBS) three times and incubated with appropriate secondary antibodies (Supplemental Table 1) for 1 hr at room temperature. The blotting membranes were developed with enhanced chemiluminescence (ATTO) on LAS-4000 (GE healthcare).

To validate injected MSCs viability, human specific markers including CD29 and CD44 genes were used. After RNA extraction, complementary DNA (cDNA) was synthesized by using the PrimeScript 1st strand cDNA Synthesis Kit (TAKARA) followed by manufacturer's manual. One hundred nano gram of each cDNA, 5 µl SYBR Premix (TAKARA) and 0.4 µM forward and same volume of reverse primers (listed in Supplemental Table 2) were mixed, then the level of interesting gene expression was detected on a CFX384 Touch™ Real-Time PCR Detection System (Bio-Rad, CA, USA).

p-values were calculated and analyzed using SPSS software version 22 (SPSS, IL, USA). The Kruskal-Wallis test and the Mann-Whitney U test were used to determine the significances of differences between MSCs viability, blood vessel sizes, M2 macrophage infiltrations, and histology findings. p-values of <0.05 were considered significant.

Extracted MSCs showed typical surface markers and differentiation potentials. Analysis of MSCs surface markers by FACS revealed 99.99% and 99.98% were CD73 or CD105 positive, respectively, and that 0.78% and 0.63% were CD34 or CD45 positive (Supplemental Fig. 1A). MSCs also differentiated into adipocytes or osteocytes depending on the differentiation method used. To assess adipogenesis of MSCs, lipid droplets were stained with Oil Red O.

Differentiated osteocytes showed positivity to alkaline phosphatase, in areas of mineral deposition (Supplemental Fig. 1B). To assess the effects of injection frequency and of cell dose on MSCs survival, we analyzed protein and mRNA levels of injected MSCs in ischemic areas. CD44 levels were confirmed by immunofluorescence analysis. Double high dose of MSCs injection in ischemic muscles resulted in the largest number of CD44 expressing cells, and double low dose of MSCs injection resulted in 1.59 fold increase in the number of CD44 expressing MSCs than a single high dose of MSCs injection (p<0.05; Fig. 1A). qRT-PCR results for CD29 and CD44 showed that the survival rate of injected MSCs resulting from double high dose of MSCs injection was greatest. Interestingly, expression level of CD29 in double low dose injected tissues was 2.33-fold higher than those of single high dose injected tissues. CD44 expression pattern was similar with CD29. The level of CD44 in double low dose of MSCs injections was 1.69-fold higher than in single high dose of MSCs injection (p<0.05; Fig. 1B). Based on these results, double injections of either high or low dose of MSCs provided better cell survival than a single high dose injection.

To investigate ability of MSCs to induce angiogenesis, we measured the migration, tube formation, and proliferation ability in HUVECs. Mean migration ability and mean tube length formed by HUVECs after MSC-CM treatment were 9.30-fold (Fig. 2A) and 3.89-fold higher (Fig. 2B), respectively, than in PBS. Moreover, HUVECs proliferation induced by MSC-CM was 2.28-fold greater than that induced by PBS alone (Fig. 2C). ANG-1, HGF, and VEGF are important growth factors during angiogenesis, and we used ELISA to assess levels of ANG-1, HGF, and VEGF proteins excreted into MSC-CM. ELISA results showed that amounts of secreted ANG-1, HGF and VEGF were 1.77, 1.98, and 1.59-fold higher, respectively, in MSC-CM than in basal medium (Fig. 2D). To study the angiogenesis mechanism in vitro, we investigated changes in the expressions of pAKT and MAPK in MSC-CM treated HUVECs. Immunoblot analysis showed MSC-CM treatment increased pERK1/2 and pAKT expressions. However, the expression levels of AKT, ERK1/2, p38, pp38, SAPK/JNK, and pSAPK/JNK did not change after MSC-CM treatment (Fig. 2E). In vivo, injected MSCs were found to promote angiogenesis. The size of vWF positive blood vessels after double high dose of MSCs injections were significantly greater (by 2.63 and 3.25-fold) than single high dose or double low dose of MSCs injection (Fig. 2F). Interestingly, no significant differences were observed between vWF positive vessel sizes after double low dose injections or single high dose of MSCs injection, and single low dose injection resulted in larger vessels that single high dose of MSCs injection. These results indicate that injection frequency is more important than the dose administered with respect to the promotion of angiogenesis by MSCs in limb ischemia.

In order to confirm the effect of MSCs on M2 macrophage modulation, we used immunofluorescence analysis and a M2 specific marker arginase 1 protein (Arg 1). The number of Iba1 positive cells (a marker of macrophage activation) was higher in saline injected mice than in MSCs treated mice. However, it was changed in all MSCs injected mice except after a single low dose injection and these difference among high dose of MSCs and double low dose of MSCs injection was not stand out (Fig. 3A, B).

Although the number of Iba1 positive cells decreased after MSCs injection, the proportion of Arg 1 expressing Iba1 cells was increased by MSCs injection. In particular, the number of Arg 1 expressing Iba1 cells was significantly higher after single and double high dose of MSCs injections than after double low dose injections (Fig. 3B). These results show the higher MSCs doses than minimum number (10⁵ MSCs) of MSCs better modulated macrophage and M2 macrophage infiltrations than increasing injection frequency.

Limb ischemia induced fibrosis and apoptosis were dramatically reduced in 4 weeks after single or double high dose of MSCs injections, but not after single low dose of MSCs injection (Fig. 4A, B). Double high and low dose of MSCs injections reduced fibrosis areas by 12.92 and 10.97-fold, respectively, versus double injections of saline (Fig. 4A). Numbers of apoptotic cells (as indicated by TUNEL staining) were reduced by 4.09, 5.95-fold after double high or low dose of MSCs injections, respectively, versus double injections of saline (Fig. 4B). Interestingly, no statistical difference was observed between animals administered double high or low dose of MSCs injections.

Finally, limb conditions in all MSCs injected groups were better than in the saline group (Fig. 4C). Among the all MSCs injected groups, limb salvage ratio was highest in the double high dose of MSCs injection group (85.71%) followed by the double low dose of MSCs injection group (81.82%), which both had greater limb salvage ratios than the single high or low dose of MSCs injection groups. Necrosis ratio in the double high dose of MSCs injection group was lowest (14.29%) and the double low dose of MSCs injection group was next lowest (18.18%). These results indicate injection frequency is more important than dosage in terms of preventing limb loss in limb ischemia (Fig. 4C).