DMEM (cat. no. 30-2002) was obtained from the American Type Culture Collection. Antibiotics, FBS and 1X PBS were procured from GE Healthcare Life Sciences. Specific antibodies against focal adhesion kinase (FAK, cat. no. 3285), p-FAK (cat. no. 3281), steroid receptor coactivator (Src, cat. no. 2109), p-Src (cat. no. 6943), Akt (cat. no. 9272), p-Akt (cat. no. 9271), ERK (cat. no. 4695), p-ERK (cat. no. 9101), VEGF-R2 (cat. no. 9698) and p-VEGF-R2 (cat. no. 3817) were purchased from Cell Signaling Technology, Inc. A specific antibody against β-actin (sc-1616) was obtained from Santa Cruz Biotechnology, Inc. VEGF was procured from R&D Systems. Matrigel was purchased from Becton Dickinson (BD Biosciences). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), hemoglobin assay kit, and all other chemicals were obtained from Sigma-Aldrich (Merck KGaA). All other chemicals were of analytical grade.

The MGE and FGE were prepared following a previous method (6). Briefly, voucher specimens of mountain ginseng and farm-cultivated ginseng were deposited in the National Institute for Korean Medicine Development (NIKOM, Gyeongsan, Korea) after identification by Dr H. Lee, a herbalist. Dried fragments of the herbs were boiled in 30% ethanol solution for about 3 h, and deposited at 4°C after cooling. Next day, the separated supernatant was filtered through a 0.45 µm filter, and the filtrate was concentrated and then lyophilized. The dried pellets were stored at −20°C until use. The lyophilized powders of MGE and FGE were dissolved in 4% ethanol solution for in vitro and in vivo studies.

Ginsenoside contents in MGE and FGE were analyzed according to a previous method (6).

Male C57BL/6 mice (6 weeks and 19–21 g) were obtained from Koatech Co., and housed in cages (5 mice per cage) under specific pathogen-free conditions (21–24°C and 40–60% relative humidity) with a 12 h light/dark cycle. They were given free access to standard rodent food (Envigo) and water. All animal experiments were approved by the Committee of Animal Care and Experiment of NIKOM with a reference number (NIKOM-2020-5). Animal studies were performed according to the guidelines of the Animal Care and Use Committee at NIKOM.

Matrigel plug assay was performed following a modification of a previous protocol (9). Matrigel including VEGF (400 ng/ml) was mixed with MGE (0–200 µg/ml) or FGE (200 µg/ml). After adaptation, 200 µl of the Matrigel mixtures was subcutaneously injected into the flank of mice. After 2 weeks, mice were euthanized by CO₂ (in 30%/min). Matrigel plugs were isolated from sacrificed mice, and then washed with cold 1X PBS twice. The washed Matrigel plugs were used for hemoglobin content assay and histological analysis.

Hemoglobin content in Matrigel plugs was detected by a hemoglobin assay kit in accordance with the manufacturer's instructions.

Histological analysis was carried out according to a modified method, previously reported (10). To observe aspect of red blood cells in Matrigel plugs, briefly, deparafinized Matrigel plug on slices was stained with hematoxylin-eosin. The stained tissue slices were embedded with the Mounting solution. Histological features in Matrigel plugs were observed under a light microscope with 100× magnification.

SVEC4-10 cells, a murine endothelial cell line (11), were purchased from the American Type Culture Collection. The cells were cultured in DMEM medium containing 10% (v/v) FBS and antibiotics at 37°C in a humidified atmosphere of 5% CO₂. All in vitro tests contained a vehicle control group (0.016% ethanol).

Cell viability was determined following a modification of a method previously reported (12). Briefly, SVEC4-10 cells were seeded on a 96-well plate (1×10⁴ cells/well). Next day, the cells were incubated with MGE or FGE (0–200 µg/ml) for 24 h, and then further incubated with 100 µl culture media containing 300 µg/ml MTT reagent for 2 h. Next 100 µl of dimethyl sulfoxide was added to the plate after removal of the supernatant, and then the plate was incubated for 15 min. Cell viability was determined at 570 nm using a microplate reader (Tecan Sunrise).

Wound healing assay and Transwell migration assay were performed in accordance with a modified method, previously reported (13). SVEC4-10 cells were seeded on a 6-well plate (1×10⁵ cells/well), and then the cells were incubated until 95% confluence in 10% FBS medium. Next, the cells were wounded with a p20 pipette tip under serum free condition, and then incubated with MGE (0–200 µg/ml) or FGE (200 µg/ml) containing FBS overnight. Wound closure was observed under a light microscope with 100× magnification. To confirm cell migration using the Transwell migration assay, the cells were seeded on a 24-well Transwell insert (1×10³ cells/insert) in serum-free DMEM, and then incubated with MGE containing FBS (lower chamber) overnight. The insert insides were swabbed and then stained with 0.2% crystal violet in a 20% methanol solution. The stained cells were observed under a light microscope with 100× magnification. Wound closure and migrated cells were measured using ImageJ software (version 1.51j8 for Windows; National Institutes of Health).

Matrigel was added to a 96-well plate (50 µl/well), which was then incubated for 30 min at 37°C. SVEC4-10 cells, pretreated with MGE or FGE for 30 min, were seeded on the Matrigel-coated 96-well plate. After 3 h, the cell morphology was observed under a light microscope with 100× magnification. The tube lengths and areas were measured using ImageJ software (National Institutes of Health).

Immunoblotting analysis was evaluated following a method previously reported (14). Briefly, the blotted proteins on PVDF membrane were visualized using a chemiluminescent reaction (Immnobilon Western; Millipore Corporation) with an Imaging system (ImageQuant LAS 4000, GE Healthcare Life Sciences). The level of target proteins was compared to that of a loading control (non-phosphorylated proteins or β-actin), and the results were expressed as a ratio of density of each protein identified by a protein standard size marker (BIOFACT Co., Ltd.). The density of each inverted band was measured using ImageJ software (National Institutes of Health).

The experimental results were listed as means ± SD for in vitro studies or SEM for in vivo studies One-way analysis of variance (ANOVA) was used for multiple comparisons (GraphPad Prism version 5.03 for Windows; GraphPad Software, Inc.). We applied the Dunnetts test or Tukey's test for one-way ANOVA for significant variations between treated groups. Differences at the *P<0.05 and **P<0.01 levels were considered statistically significant.

Comparison between MGE and FGE. To compare the effects of MGE and FGE on cell migration and tube formation in vascular endothelial cells, we investigated the effects of MGE and FGE on cell migration of SVEC4-10 cells. As presented in Fig. 1, MGE and FGE did not show cytotoxicity within the concentration ranges (0–200 μg/ml). Subsequently, when SVEC4-10 cells were stimulated by FBS, the cells completely covered a wound area (Fig. 2A). In contrast, MGE dose-dependently inhibited the cell migration of FBS-activated SVEC4-10 cells, and the inhibitory effect of MGE at 200 µg/ml was more potent than that of FGE at the same concentration (Fig. 2A). Based on the cell migration results, we examined whether MGE and FGE could suppress the formation of capillary-like network of FBS-activated SVEC4-10 cells. The inclusion of MGE significantly reduced the formation of capillary-like network in the cells, and the inhibitory effect of MGE at maximum dose was stronger than that of FGE at the same concentration (Fig. 2B). The results suggest that MGE and FGE have anti-angiogenic properties within non-cytotoxic ranges in vascular endothelial cells. In addition, the anti-angiogenic effect of MGE is stronger than that of FGE.

As we found that MGE was able to inhibit the cell migration and capillary-like network formation of vascular endothelial cells on a 2-D structure, we tried to confirm whether MGE was capable of suppressing the cell migration of FBS-stimulated SVEC4-10 cells on a 3-D structure. Consistent with the cell migration and capillary-like network formation on the 2-D structure, MGE concentration-dependently attenuated Transwell cell migration of the cells (Fig. 3). This finding suggests that MGE may be able to attenuate the formation of neo blood vessels in a whole body system.

After we found that MGE was able to reduce the cell migration and tube formation of vascular endothelial cells, we were interested in the effect of MGE on the VEGF-R2 signaling pathway because angiogenic factor-exposed vascular endothelial cells are capable of formatting neo blood vessels through activation of the VEGF-R2 signaling pathway (15). MGE not only reduced both phosphorylation and expression of VEGF-R2 (Fig. 4A) but also attenuated the phosphorylation of FAK, Src, Akt and ERK, intermediate proteins in the VEGF-R2 signaling pathway (16), in SVEC4-10 cells activated by FBS (Fig. 4B). Also, MGE at 50 µg/ml almost blocked the phosphorylation of FAK, one of the intermediate proteins of the VEGF-R2 signaling pathway (Fig. 4B). These results suggest that MGE has anti-angiogenic properties through inhibition of activation of the VEGF-R2 signaling cascade in vascular endothelial cells.

Finally, because we found that MGE possesses anti-angiogenic properties through inhibiting the phosphorylation of proteins related to the VEGF-R2 signaling cascade in vascular endothelial cells, we tried to confirm the anti-angiogenic action of MGE on a whole body system. In histological observation of a Matrigel plug assay in vivo, MGE dose-dependently attenuated accumulation of erythrocytes induced by VEGF in Matrigels. Also, the inhibitory effect of MGE at 200 µg/ml was stronger than that of FGE (Fig. 5). Consistent with the histological results, MGE dramatically reduced the level of hemoglobin, a biomarker of blood vessels, in Matrigel including VEGF (400 ng/ml) in mice (Fig. 5). Surprisingly, MGE at 200 µg/ml almost suppressed the level of hemoglobin in Matrigels to a level similar to that in Matrigels of the control group. In addition, the inhibitory effect of MGE was more potent than that of FGE at the same concentration (Fig. 5). These findings suggest that MGE is able to inhibit neo blood vessels in a whole body system. Overall, MGE may be a possible herbal drug candidate or adjuvant for therapy for cancer patients.