K. formosana branches, following the shedding of their leaves, were purchased from local herb stores in Taichung, Taiwan. K. formosana peduncle extracts were prepared by initial condensation, followed by lyophilization, as previously described (13). In brief, 100 g air-dried branch was boiled with 50% ethanol at 70°C for 24 h. The solvent was removed from the combined extract with a vacuum rotary evaporator. The filtrate was then lyophilized and stored at −20°C.

The total phenolic content was measured by photometric assay using a Folin-Ciocalteu reagent (Fluka, Buchs, Switzerland) (19). Folin-Ciocalteu reagent was added to each sample and sodium carbonate (Sigma, St. Louis, MO, USA) was added. The specific absorbance was measured immediately at 760 nm using a spectrophotometer (Thermo Biomate 5; Thermo Electron Corporation, San Jose, CA). Gallic acid (Sigma) was used as a standard phenolic compound for the calibration curve (y=0.7758x+0.0582, R2=0.9982). Total phenolic content was expressed as milligrams of gallic acid equivalents per gram of dry weight of plant (mg GA/g DW).

HPLC analysis (Waters 600 with a 2998 photodiode array detector; Waters Corp., Milford, MA, USA) was conducted with a LiChroCART RP-18 reversed phase column (200×4 mM; 5 μm Merck KGaA, Parmstadt, Germany), the mobile phase consisted of water/acetic acid (0.05%, v/v) (solvent A) and acetic acid/water/acetonitrile (0.05%, v/v; Sigma) (solvent B). Elution was carried out in a programmed gradient elution as follows: 0–30 min with 0–100% B.

To establish the first generation of xenograft-derived 786-O cell lines, 786-O cells were injected subcutaneously into the dorsal flank of five-week-old nude BALB/c nu/nu mice. The size of each tumor was measured as described previously (21). Solid tissue from the tumor was then mechanically and enzymatically disaggregated using collagenase (0.7 mg/ml) in serum-free medium (Sigma) into a single-cell suspension and was designated as 786-O-SI1. A second-generation xenograft-derived 786-O cell line was established following an identical protocol, using cells from the 786-O-SI1 cell tumor; this single-cell suspension was designated as 786-O-SI2. The third-generation xenograft-derived 786-O cell line was established from the tumor produced by the 786-O-SI2 cells and was designated as 786-O-SI3.

RCC cell line 786-O-SI3 and human kidney-2 (HK-2; a human proximal tubule epithelial cell line) cells were purchased from the Food Industry Research and Development Institute (Hsinchu, Taiwan). The RCC cell line 786-O-SI3 was cultured in RPMI 1640 medium (Gibco-BRL, St. Louis, MO, USA) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 mg/ml streptomycin and 100 units/ml penicillin (Sigma). HK-2 was cultured in a 1:1 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 medium (Gibco-BRL) containing 10% fetal bovine serum (FBS; Gibco-BRL). The cell cultures were maintained at 37°C in a humidified atmosphere of 5% CO₂. KFEE treatment was added to the culture medium at different concentrations (25, 50, 75 and 100 μg/ml) and was incubated with cells for 24 h. DMSO (final concentration 0.1%) without KFEE was used as a blank reagent.

To evaluate the cytotoxicity of KFEE, cell viability was determined using an MTT colorimetric assay (Sigma) (22). Cells were seeded in 24-well plates at a density of 3×10⁴ cells/well and were treated with concentrations of KFEE of 0–100 μg/ml at 37°C for 24 h. Following incubation for 24 h, the media were removed and the cells were washed with phosphate-buffered saline (PBS; Sigma). The medium was then changed and the cells were incubated with 20 ml MTT (5 mg/ml) for 4 h. The viable cell number/dish was directly proportional to formazan production, which, following solubilization with isopropanol (Sigma), was measured by a Hitachi U-1900 spectrophotometer (Hitachi, Tokyo, Japan) at 563 nm.

Wounds were introduced to the confluent monolayer of cells with culture-inserts (Ibidi GmbH, Martinsried, Germany) to create a cleared line. The medium was removed and replaced with RPMI 1640 medium containing 1% FBS, and KFEE was then added. The cells were incubated at 37°C and cell migration into the wound area was documented at 0, 6 and 24 h by a microscope (CKX41: Olympus, Tokyo, Japan).

The 786-O-SI3 cells were pre-treated with KFEE at indicated concentrations (0, 25, 50, 75 and 100 μg/ml) for 24 h. The cells were then harvested and seeded in a Boyden chamber (Neuro Probe, Cabin John, MD, USA) with 10⁴ cells/well in serum-free medium and were incubated at 37°C for 12 h. For the invasion assay, 10 ml Matrigel^® (25 mg/50 ml; BD Biosciences, Bedford, MA, USA) was applied to polycarbonate membrane filters (Neuro Probe, Cabin John, MD, USA) with a pore size of 8 mM and the bottom chamber of the apparatus contained a standard medium. Following incubation, the filters were air dried in a laminar flow hood for 5 h. The invaded cells were fixed with methanol and stained with Giemsa (Sigma). Cell numbers were counted with a light microscope (CKX41; Olympus) and the migration assay was carried out as described for the invasion assay, with no coating of Matrigel (23).

The activities of MMP-2 and MMP-9 on the condition medium were measured by gelatin-zymogram protease assays, as previously described (24). Samples were prepared with a standard SDS gel-loading buffer containing 0.01% SDS without β-mercaptoethanol and were not boiled prior to loading. The prepared samples were then subjected to 8% SDS-PAGE (0.75-mm; acrylamide/bis-acrylamide 30/1.2; containing 0.1% gelatin; Sigma). Electrophoresis was performed at 150 V in an OWL P-1 apparatus (Alpha Multiservices, Inc., Contoe, TX, USA) for 3 h. Following electrophoresis, the gels were washed twice with 100 ml distilled water containing 2% Triton X-100 (Sigma) on a gyratory shaker at room temperature for 30 min to remove the SDS. The gel was then incubated in 100 ml reaction buffer (40 mM Tris-HCl, pH 8.0, 10 mM CaCl₂, 0.02% NaN₃; Sigma) at 37°C for 12 h, stained with Coomassie brilliant blue R-250 and de-stained with methanol/acetic acid/water (50/75/875, v/v/v; Sigma). u-PA activity was visualized as previously described (24). 2% w/v casein and 20 mg/ml plasminogen (Sigma) were added to 8% SDS-PAGE gels. Electrophoresis and zymography were then performed for gelatin zymography.

786-O-SI3 cells grown on glass coverslips were fixed in 4% paraformaldehyde (Sigma) at room temperature for 12 min. Following washing in PBS, the cells were blocked and permeabilized in PBS containing 4% bovine serum albumin and 0.1% Triton X-100 at room temperature for 90 min. Filamentous actin was detected by incubating coverslips with rhodamine-conjugated phalloidin (1:200) at 4°C overnight and nuclei were counterstained with DAPI (Sigma) at room temperature for 1 h. The cells were viewed and photographed with a ZEISS Axioskop2 upright fluorescence microscope (Carl Zeiss AG, Oberkochen, Germany).

Samples of cell lysates or nuclear fractions were separated by 12.5% SDS-PAGE and transferred onto a nitrocellulose membrane (GE Heakthcare, Taipei, Taiwan), as previously described (18). The blot was subsequently treated using standard procedures and probed with the following antibodies: c-Jun (cat. #3270), c-Fos (cat. #5348), phospho-FAK Tyr 925 (cat. #3284), phospho-MEK1/2 (cat. #9121), total-MEK1/2 (cat. #9122), phospho-MLC-2 (cat. #3671), total-MLC-2 (cat. #3672), phospho-ERK1/2 (cat. #9101), total-ERK1/2 (cat. #9107) (1:1,000 dilution, monoclonal, Cell Signaling Technology, Inc., Danvers, MA, USA) and total-paxillin (cat. #sc-136297), MMP-2 (cat. #sc-13594), β-actin (cat. #sc-10731), total-FAK (cat. #sc-932) and C23 (cat. #sc-17826) (1:1,000 dilution; monoclonal; Santa Cruz Biotechnology, Inc., Santa Cruz CA, USA) and phospho-paxillin Ser 178 (cat. #44–1026; 1:1,000 dilution; monoclonal; Biosource, Camarillo, CA, USA). Protein expression was detected by chemiluminescence with an Enhanced Chemiluminescence Plus detection kit (Amersham Life Sciences, Inc., Piscataway, NJ, USA).

Five-week-old male C57BL/6 mice (National Taiwan University Animal Center, Taiwan) were housed under a regular 12-h light/dark cycle and with ad libitum access to a standard rodent diet (Laboratory Rodent Diet 5001; LabDiet, St. Louis, MO). Cells (1×10⁶) suspended in 0.1 ml PBS were injected into the tail vein of the C57BL/6 mice. On the following day (day 1), the mice were randomly divided into three groups (n=3 for each group) to be fed by oral gavage with saline (control) or KFEE (0.1 and 0.2 g/kg of body weight daily). Three untreated mice were used as wild-type controls. Following 31 days, the animals were sacrificed using CO₂. The lungs were isolated and weighed, and metastatic nodules on the surface of the lungs were counted under a microscope (Axioskop 2 Plus, Carl Zeiss, Inc., Oberkochen, Germany). The lungs were fixed in neutral buffered 5% formalin (Sigma), and sections were prepared and stained with hematoxylin and eosin (Sigma) for morphological studies (25).

Statistically significant differences throughout this study were calculated by Student’s t-test (SigmaStat 2.0; Jandel Scientific, San Rafael, CA). P<0.05 was considered to indicate a statistically significant difference between values.

Quantitative analysis of phytochemicals revealed the presence of 38.8±0.4 mg/g polyphenols in KFEE. Reference compounds included gallic acid, gallocatechin, catechin, caffeic acid, rutin, naringin, quercetin, kaempferol, anthraquinone, myricetin, morin, apigenin and galangin (Fig. 1A and B). To evaluate the bioactive compounds in KFEE, K. formosana was successively extracted with 50% ethanol. HPLC analysis of KFEE (Fig. 1C) and 13 standard compounds showed peaks corresponding to the retention time chromatography and absorbance at 254 nm. Gallic acid and caffeic acid were included in the composition of KFEE (Fig. 1D).

To determine whether the serial xenotransplantation of a cancer cell line in nude mice enriches the malignant subpopulation, tumor-forming 786-O cells were injected subcutaneously into the dorsal flank of five-week-old nude BALB/c nu/nu mice. Three xenograft-derived 786-O cell lines, namely 786-O-SI1, 786-O-SI2 and 786-O-SI3, were established from the xenograft tumors (Fig. 2A). 786-O-SI3 cells significantly exhibited increased MMP-2 activity compared to that of the control 786-O cell line (P<0.001; Fig. 2B), indicating that the successive implantation of 786-O cells in nude mice increased the ability of the tumor to metastasize.

A previous study reported no cytotoxic effects of KFEE (≤100 μg/ml) on RCC 786-O-SI3 cells (17). In the present study, RCC 786-O-SI3 and HK-2 cells were viable in the presence of 0, 25, 50, 75 and 100 μg/ml KFEE (Fig. 3A and B) (P>0.05). In all subsequent experiments, this concentration range was used for KFEE.

Incubation of 786-O-SI3 with 1% FBS produced marked cell migration into the wounded area at 6 and 24 h following wounding, whereas wounds treated with KFEE showed significantly delayed wound healing under identical conditions (Fig. 3C and D). In addition, KFEE had a concentration-dependent inhibitory effect on cell migration, with 87.5±2.5% cell migration compared to that of the control (P<0.001) (Fig. 3E and F). Under identical conditions, it was also observed that KFEE reduced the invasion of RCC 786-O-SI3 cells in a concentration-dependent manner, which was demonstrated using a cell invasion assay using a Matrigel-coated Boyden chamber. Following treatment with 100 μg/ml KFEE, invasion capability was decreased to 89.8±4.0% compared to that of the control (P<0.001) (Fig. 3E and G).

To determine the involvement of MMPs and u-PA in the KFEE-induced inhibition of invasion and migration of 786-O-SI3 RCC cells, the effects of KFEE on MMP and u-PA activity were investigated using gelatin and casein zymography, respectively, under serum-starved conditions. KFEE reduced the activity of MMP-2 (P<0.001) in gelatin zymography (Fig. 3H) and that of u-PA (P<0.001) in casein zymography assays (Fig. 3I). Furthermore, KFEE decreased the migratory and invasive capacities of 786-O-SI3 cells and the cell shape changed from rounded to spindle-shaped, indicating that KFEE may regulate metastasis by controlling cytoskeletal events essential for motility (Fig. 4A).

Given that KFEE inhibited the invasion, migration and transcriptional levels of MMP-2 and u-PA activities in 786-O-SI3 cells, the effects of KFEE on the expression of the Erk1/2/mitogen-activated protein kinase (MAPK) pathways were investigated by western blot analysis to elucidate the underlying mechanisms. KFEE significantly inhibited phosphorylation of FAK Tyr925, Paxillin Ser178, MEK1/2 and Erk1/2 (Figs. 4B and C) in RCC 786-O-SI3 cells. Therefore, inhibition of the FAK Tyr925 and Erk1/2 signaling pathways may be involved in the reduction of MMP-2 and u-PA activity, as well as tumor cell invasion. In contrast, the phosphorylation of myosin light chain (MLC)2 (Fig. 4C), the cell actin dynamic regulatory protein, was decreased.

Western blot analysis revealed that KFEE significantly inhibited c-Jun and c-Fos levels in the nuclear extracts of RCC 786-O-SI3 cells (P<0.001) (Fig. 4D).

786-O-SI3 cells primarily form lung tumors. C57BL/6 mice were injected with 786-O-SI3 cells via the tail vein and administered either KFEE or saline via oral gavage. KFEE reduced 785-O-SI3 cell pulmonary metastasis formation. Within 31 d of injection, the average body weight of KFEE-treated mice was lower than that of the control group (Fig. 5A). The mean lung weights of animals receiving 0.1 g/kg/d KFEE (0.4503±0.0838 g; P<0.001) and 0.2 g/kg/d KFEE (0.3288±0.1208 g; P<0.001) were significantly lower than those of the control animals (0.7154±0.0.1601 g; Fig. 5B). The lungs of the control mice were visibly riddled with metastatic tumor nodules compared with those of the KFEE-treated mice (Fig. 5C). Histopathology of the lung also showed marked reduction in tumor mass in the lungs of the KFEE-treated animals (Fig. 5D).