Human fibrosarcoma HT1080 cells (KCLB no. 10121) and human prostate adenocarcinoma PC-3 cells (KCLB no. 21435) were purchased in 2012 from Korean Cell Line Bank (Seoul, Korea), where they were characterized by DNA fingerprinting. Cells were cultured in RPMI-1640 (Lonza, Walkersville, MD, USA) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS; Gibco/Invitrogen, Carlsbad, CA, USA) and 100 U/ml penicillin/100 μg/ml streptomycin (Gibco/Invitrogen) at 37°C in a humidified 5% CO₂ incubator and routinely screened for mycoplasma contamination. Human umbilical vein endothelial cells (HUVECs) purchased from Innopharmascreen (Asan, Korea) in October, 2012 were maintained in Endothelial Growth Medium-2 (EGM-2; PromoCell, Heidelberg, Germany) and used at passages 3 to 8 in the experiments. Murine melanoma B16F10 cells (KCLB no. 80008), which are metastatic to the lungs of C57BL/6J mice, were obtained in 2012 from Korean Cell Line Bank and cultured in complete Dulbecco’s modified Eagle’s medium (DMEM; Lonza). Specific pathogen-free female C57BL/6J mice and athymic nude mice were purchased from Taconic Farms Inc. (Samtako Bio Korea, Osan, Korea) and Nara Biotech (Seoul, Korea), respectively, and maintained in our animal facility for 1 week before use. Mice were housed in a barrier facility with 12-h light-dark cycles at 24±1°C and 55±5% humidity under specific pathogen-free conditions. All animal experimental procedures were approved by Korea Institute of Oriental Medicine Care and Use Committee with a reference number of #13-67 and #13-94, and performed according to the guidelines established by the Korea Institute of Oriental Medicine Care and Use Committee.

Type A gelatin from porcine skin, casein from bovine milk, type I collagen from calf skin, 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT), phorbol 12-myristate 13-acetate (PMA), mitomycin C, and curcumin were purchased from Sigma Chemical Co. (St. L ouis, MO, USA). Growth factor-reduced Matrigel Basement Membrane Matrix was obtained from BD Biosciences (Bedford, MA, USA), and recombinant mouse basic fibroblast growth factor (rMu bFGF) and recombinant murine vascular endothelial growth factor 165 (rMu VEGF₁₆₅) were purchased from PromoKine (Heidelberg, Germany). SB203580, PD98059 and SP600125 were obtained from Calbiochem (San Diego, CA, USA). Antibodies against IκBα, pIκBα, c-jun, p-c-jun, c-fos, p-c-fos, NF-κB p65, p38, p-p38, ERK, p-ERK, JNK, p-JNK and tubulin were obtained from Cell Signaling Technology (Danvers, MA, USA).

Dried ASF was purchased from Korea Medicine Herbs Association (Yeongcheon, Korea) and deposited in the herbal bank of Korea Institute of Oriental Medicine (KIOM; Daejeon, Korea) after verifying by Professor Ki Hwan Bae of the College of Pharmacy, Chungnam National University (Daejeon, Korea). To prepare aqueous extract of ASF, dried ASF (50 g) were placed in 1 liter of distilled water, and then heat-extracted at 115°C for 3 h in a Cosmos-600 Extractor (Gyeonseo Co., Inchon, Korea). Aqueous extract of ASF was filtered using standard testing sieves (150 μm, Retsch, Haan, Germany), and then concentrated by lyophilization. Freeze-dried ASF powder (50 mg) dissolved in 1 ml of DW was kept at −20°C prior to use after filtration through a 0.22-μm disk filter.

Five-week-old female C57BL/6J mice (n=3 per group) were daily administered vehicle (saline) or 50 mg/kg ASF. During 15-day experimental period, gross appearance and behavior were carefully monitored and body weight was daily measured. After sacrifice, organs including heart, liver, lung, spleen and kidneys were weighed. Hematological and serological parameters were measured using ADVIA 2120i hematology system (Siemens Healthcare Diagnostics, Tarrytown, NY, USA) and XL 200 (Erba Diagnostics, Mannheim, Germany), respectively.

To examine cytotoxicity of ASF, cells (5×10³ cells/well/96-well plate) were treated with various concentrations of ASF (10 to 1,000 μg/ml) for 48 h, and then MTT assay was performed as described previously (17). Relative cell viability was represented as percentage compared with that of untreated control cells.

After HT1080 cells (5×10⁴) were suspended in 2 ml RPMI medium with specified concentrations of ASF, 0.3% agar and 10% FBS, they were applied to the bottom agar pre-solidified with 3 ml RPMI medium containing 0.6% agar and 10% FBS. During incubation for 2 weeks, colony formation was daily observed under a phase-contrast microscope and photographed.

In vitro wound healing assay was performed as described previously (17). In brief, HT1080 cells with 90% confluence were pre-incubated with mitomycin C (25 μg/ml) for 30 min and scratched by scraping across the cell monolayer. After washing with PBS to eliminate floating cell debris, cells were permitted to migrate for 36 h in the presence of ASF and wound healing was observed under a phase-contrast microscope and photographed.

Using a Transwell chamber with polycarbonate membrane (10 mm diameter, 8 μm pore size), in vitro migration and invasion assays were performed. Transwell chamber coated with 20 μl 1:8 mixture of Matrigel: RPMI as the intervening barrier was used for invasion assay. Cells were pre-incubated with indicated concentrations of ASF for 12 h. After filling the lower chamber with 600 μl 10% FBS/RPMI, cells (1×10⁵) suspended in 100 μl serum-free RPMI were added to upper chamber, and incubated for 12 or 24 h for migration and invasion, respectively. For migration of HUVECs, experimental conditions are specified in the Figures. After removal of the cells remaining in the upper surface of the chamber, migrated and invaded cells were stained with 0.2% crystal violet/20% methanol (w/v) solution.

The secretions of MMP-2 and MMP-9 in the culture supernatants were measured by gelatin zymography. In brief, cells pre-incubated with ASF in serum-free medium for 12 h were stimulated with 5 nM PMA for an additional 24 h. Collected serum-free conditioned medium (CM) were electrophoresed on an 8% SDS-PAGE containing 0.1% gelatin and gels were washed with washing buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 2.5% Triton X-100) and incubated in reaction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM CaCl₂, 0.02% NaN₃, 1 μM ZnCl₂) at 37°C. Gels were stained with Coomassie Brilliant Blue R-250 staining solution (Bio-Rad Laboratories, Hercules, CA, USA) and destained with 10% isopropanol/10% acetic acid (v/v) solution. The secretion of uPA was examined by casein-plasminogen zymography. Conditioned medium were electrophoresed on a 10% SDS-PAGE containing casein (1 mg/ml) and human plasminogen (15 μg/ml), and gels were washed, incubated and stained as described earlier. For the detection of collagenase activity, conditioned medium were electrophoresed on a 10% SDS-PAGE containing collagen (4 mg/ml). MMP-2, MMP-9 and uPA were detected as clear bands against blue background at the size of 72/64, 92 and 55/52 kDa, respectively.

Total RNA was isolated using PureHelix™ RNA extraction solution and 3 μg RNA was reverse transcribed to cDNA using HelixCript™ 1st Strand cDNA Synthesis kit (NanoHelix Co., Daejeon, Korea) according to the manufacturer’s instructions. The cDNA aliquots were analyzed by semi-quantitative PCR using specific primers (Table I) and PCR products were electrophoresed on 1% agarose gels and stained with EcoDye™ (Solgent Co., Daejeon, Korea). Band intensity was quantitatively analyzed using ImageJ software (National Institute of Health, USA) for each gene.

Whole cell lysates were extracted using M-PER Mammalian Protein Extraction Reagent (Thermo Scientific, Rockford, IL, USA), and protein concentrations were determined by Bicinchoninic Acid (BCA) kit (Sigma). After immunoblot analysis, proteins were detected using a PowerOpti-ECL Western blotting detection reagent (Animal Genetics, Inc. Korea) and an ImageQuant LAS 4000 mini (GE Healthcare, Piscataway, NJ, USA), and band intensity was analyzed using ImageJ software.

For tumor-induced angiogenesis, mixture containing PC-3 cells (3×10⁶/100 μl serum-free medium), heparin (50 U/ml), and Matrigel (500 μl) with or without ASF was subcutaneously injected into the abdomen of athymic nude mice. At days 10–14 after implantation, Matrigel plugs were carefully removed and hemoglobin content of the gel plug was measured using a Drabkin’s reagent kit (Sigma) as described previously (18). The concentration of hemoglobin was calculated based on the set of hemoglobin standard (Sigma).

Using an ECMatrix assay kit (Millipore, Temecula, CA, USA), HUVEC tube formation assay was performed according to the manufacturer’s instructions. In brief, 10 μl chilled ECMatrix was transferred to each well of a pre-cooled 1 μ-Slide Angiogenesis ibiTreat chamber (ibidi GmbH, Germany) and allowed to solidify for at least 1 h at 37°C. A mixture of HUVECs (5×10³ cells), bFGF (1 μg/ml), VEGF₁₆₅ (100 ng/ml), and/or ASF in endothelial cell basal medium-2 (EBM-2; PromoCell, Heidelberg, Germany), or HUVECs in ASF-treated CM was seeded onto the surface of polymerized ECMatrix, and incubated for 4 to 24 h at 37°C. For the collection of CM, cells were treated with ASF for 24 h in complete media conditions, washed twice with 0.5% serum media, and then incubated for another 24 h in 0.5% serum media. Cellular network-like structures were examined under phase-contrast inverted light microscopy and photographed. Tube formation was quantified by counting tube number and represented as the average of five fields of view in each well.

Cells were treated with indicated concentrations of ASF for 48 h. The levels of VEGF₁₆₅, bFGF and PDGF in culture media were determined with the Human ELISA Development kit (Peprotech, Rocky Hill, NJ, USA) according to the manufacturer’s instructions.

After injection of B16F10 cells (3×10⁵ cells/200 μl PBS) via the tail veins, C57BL/6J mice were randomly divided into 2 groups (n=4 for each group). Administration of vehicle (saline) or 50 mg/kg ASF (day 0), for 17 days, mice received daily oral administration and were then sacrificed. Lungs were fixed in Bouin’s solution (Sigma) and black colonies were counted macroscopically.

The values are presented as means ± standard deviation (SD). Statistical significance of the difference between groups was analyzed by Two-way analysis of variance (ANOVA) and Student’s t-test with SigmaPlot 8.0 software. P-value <0.05 was considered significantly different.

To examine the anti-metastatic effect of ASF in vivo, we first examined whether repeated administration of ASF (50 mg/kg) is non-toxic systemically. Over 15 days, corresponding to the in vivo experimental conditions, ASF intake did not influence the body weight of mice (Table II) or cause abnormal behavior or death. Organ weights were not significantly different between control and ASF-treated mice (Table III). In serological analyses, the aspartate aminotransferase/alanine aminotransferase (AST/ALT) and blood urea nitrogen/creatinine (BUN/CRE) ratios were not significantly affected by ASF intake, indicating that ASF did not cause hepatic or renal damage, respectively (Table IV). In hematological analyses, the number of red blood cells (RBCs) and the hemoglobin (Hb) level, an indicator of RBC balance and anemia, were not altered by ASF intake. In addition, the number of white blood cells (WBCs) and other hematologic parameters were also similar to those of control mice (Table V). Next, we investigated the inhibitory effect of ASF on in vivo tumor metastasis by comparing the ability of B16F10 cells to colonize the lungs of C57BL/6J mice after tail vein injection. As shown in Fig. 1, B16F10 cells in control mice metastasized to the lungs and formed a considerable number of black colonies (1273.5±157.1). In ASF-administered mice, the number of black colony in lungs were significantly decreased (402.6±160.6) to approximately 30% of control mice. These data indicate that ASF intake effectively inhibited pulmonary metastasis of intravenously injected B16F10 cells with no systemic side effects during treatment.

Prior to evaluation of the in vitro anti-metastatic potential of ASF, we first determined its cytotoxicity in HT1080 cells in the absence or presence of serum using MTT assays. At concentrations up to 250 μg/ml, ASF had no cytotoxic effects; however, 500 μg/ml ASF exhibited a ~10% decrease in cell viability (Fig. 2A). Therefore, we treated cells with 25, 50 or 100 μg/ml ASF in all subsequent experiments. Metastatic cancer cells possess the ability to form colonies in semi-solid agar, which is termed anchorage-independent growth. As shown in Fig. 2B, untreated control HT1080 cells formed sizable colonies from a single cell, whereas at concentrations of 25–100 μg/ml, ASF remarkably suppressed anchorage-independent growth by approximately 30–63%. In wound healing assays, control HT1080 cells rapidly migrated across the wound region, leading to approximately 28, 50 and 70% healing at 12, 24 and 36 h, respectively. ASF treatments of 25, 50 and 100 μg/ml dose-dependently suppressed wound migration to approximately 45, 40 and 23% healing, respectively, at 36 h (Fig. 2C). Using the Transwell culture system, serum-induced migration and invasion were significantly reduced in a dose-dependent manner in ASF-treated HT1080 cells by approximately 20–70 and 30–90%, respectively, compared with untreated control cells (Fig. 2D).

In order to elucidate the anti-metastatic effects of ASF, we first examined the transcriptional levels of various MMPs using semi-quantitative RT-PCR. Under resting conditions, HT1080 cells expressed substantial amounts of MMPs, including MMP-1, -2, -3, -9, -13, MT1-MMP and uPA. PMA stimulation significantly increased MMP-1, -3, -9, -13, MT1-MMP and uPA levels but decreased PAI-1. MMP-2, TIMP-1 and TIMP-2 levels were essentially unchanged by PMA stimulation. However, as shown in Fig. 3A, ASF treatment efficiently inhibited the PMA-induced increase in MMP-9, -13, MT1-MMP and uPA levels and decreased PAI-1 levels dose-dependently. Next, to examine whether ASF influences proteolytic activity of HT1080 cells, we performed zymography using gelatin, collagen type I, and casein-plasminogen as substrates (Fig. 3B). In gelatin zymography, HT1080 cells constitutively secreted MMP-2 and MMP-9. PMA stimulation partially converted proMMP-2 to active MMP-2 and increased MMP-9 activity. ASF treatment dose-dependently suppressed MMP-9 activity in response to PMA stimulation, but it did not prevent PMA-induced conversion to active MMP-2. Collagenase activity in ASF-treated cells was also lower compared with control cells. Furthermore, in casein-plasminogen zymography, uPA activity after PMA-stimulation was remarkably reduced by ASF treatment in a dose-dependent manner. These results indicate that ASF suppresses the metastatic potential of HT1080 cells via downregulation of proteolytic activities. In U937 human leukemic monocyte lymphoma cells and PC-3 human prostate cancer cells, ASF significantly decreased MMP-9 gelatinolytic activity (Fig. 3C), further supporting the inhibitory effect of ASF on proteolytic degradation.

It has been reported that the transcription factors NF-κB and AP-1 are involved in regulating MMP expression and invasion in various cancer cell types (19–21). To examine whether the anti-metastatic effect of ASF is linked to the suppression of NF-κB and AP-1 activities, we examined the levels of p-IκBα, IκBα, p-c-jun, and p-c-fos in response to PMA stimulation using western blot analyses in control cells and ASF-treated cells. As shown in Fig. 4A, IκBα phosphorylation and degradation were markedly increased by PMA stimulation in control cells. However, in ASF-treated cells, the extent of IκBα phosphorylation and degradation was insignificant. Phosphorylation of c-jun and c-fos was also increased by PMA stimulation in control cells, whereas c-fos phosphorylation was almost completely blocked by ASF treatment. In addition, ASF-mediated suppression of phosphorylation of both IκBα and c-fos was dose-dependent (Fig. 4B). In control cells, the p65 subunit translocated from the cytosol to the nucleus upon PMA stimulation. In ASF-treated cells, however, p65 nuclear translocation was significantly inhibited in a dose-dependent manner (Fig. 4C). It has been reported that activation of MAPKs, including p38, ERK and JNK, plays a pivotal role in increasing MMP-9 activity and expression (22). As shown in Fig. 4D, phosphorylation of p38, ERK and JNK was rapidly increased by PMA stimulation in control cells. In ASF-treated cells, PMA-induced p38 phosphorylation was almost completely blocked, but ERK and JNK phosphorylation was only partially inhibited. To confirm the involvement of these signal transduction pathways in PMA-stimulated MMP-9 expression and ASF-mediated inhibition, we used gelatin zymography to analyze the effects of specific inhibitors on PMA-induced MMP-9 activity in HT1080 cells. As reported previously, we found that PMA-induced MMP-9 activity was almost completely inhibited by the p38 inhibitor (SB203580) and was significantly inhibited by inhibitors of ERK (PD98059), JNK (SP600125), and AP-1 (curcumin) (22). In addition, PMA-induced MMP-9 activity was significantly reduced by transient expression of dominant negative IκB kinase, dnIKKα or dnIKKβ (data not shown) (18). These results indicate that specific inhibition of p38, ERK, JNK, NF-κB and AP-1 contributes to suppression of PMA-induced MMP-9 activity.

Angiogenesis, regulated by angiogenic factors produced by tumor cells, is critical for growth, invasion and metastasis of solid tumors (8). To examine the effect of ASF on tumor-induced angiogenesis in vivo, we performed Matrigel plug assays in athymic nude mice. As reported previously, PC-3 cells remarkably induced formation of new blood vessels in the Matrigel plug, and the plugs exhibited a bright red color (Fig. 5A) (23). In the presence of ASF, Matrigel plugs were light red/yellow in color, indicating inhibition of angiogenesis. The hemoglobin content in the plug with the PC-3 cells was 15.6 mg/g. ASF treatment significantly decreased hemoglobin levels by approximately 50%, suggesting that ASF can inhibit tumor-induced angiogenesis in vivo. The levels of pro-angiogenic factors, including VEGF-α, bFGF and PDGF, in conditioned media (CM) from ASF-untreated and -treated PC-3 cells were determined using ELISA analyses. We found that ASF treatment significantly decreased the production of angiogenic factors (Fig. 5B). In addition, we observed that CM from control PC-3 cells strongly induced tube formation in HUVECs. However, in HUVECs plated with CM from ASF-treated PC-3, tube formation was considerably diminished in terms of tubular network and tube number (Fig. 5C). During neovascularization, cancer cells attract endothelial cells by secretion of pro-angiogenic factors. To examine whether ASF could inhibit endothelial cell migration toward PC-3 cells, we used Transwell migration assays and seeded HUVECs in the upper chamber and ASF-treated PC-3 cells in the lower chamber. As shown in Fig. 5D, PC-3 cells in the lower chambers induced migration of HUVECs. ASF treatment of PC-3 cells reduced HUVEC migration toward PC-3 cells in a dose-dependent manner, indicating that ASF suppresses tumor cell-derived chemotactic motility of endothelial cells. In ASF-treated HT1080 cells, we also observed strong inhibitory effects on tube formation (Fig. 5E) and HUVEC migration (Fig. 5F). To determine the effects of ASF on endothelial cells, HUVECs in EGM-2 media were treated for 24 h with 25, 50 and 100 μg/ml ASF. Cell viability was analyzed using MTT assays. As shown in Fig. 6A, ASF had no effect on HUVEC proliferation. The degree of EGM-2-, bFGF- and VEGF-α-stimulated tube formation in HUVECs was similar between control and ASF-treated HUVECs (Fig. 6B). In addition, EGM-2-mediated migration of HUVECs was not inhibited by ASF treatment (Fig. 6C). These data suggest that ASF exerts its anti-angiogenic influence by targeting cancer cells but not endothelial cells.