Human NSCLC cell lines (A549 and H1299) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and were grown in 10% fetal bovine serum-Dulbecco's modified Eagle's medium (FBS-DMEM) (HyClone Laboratories, Logan, UT, USA). Human umbilical vein endothelial cells (HUVECs) were purchased from Lonza (Walkersville, MD, USA) and used between passages 4 and 6 for all experiments. Cells were cultured in EGM-2^® BulletKit media, according to the manufacturer's instructions (Lonza).

Marmesin was isolated in an ethyl acetate fraction partitioned from the ethanolic extract of B. kazinoki. The following pharmacological agents and antibodies were purchased from commercial sources: anti-phospho-Src (Y416), anti-Src, anti-phospho-MEK (S217/S221), anti-MEK, anti-phospho-ERK (T202/Y204), anti-phospho-Akt (S473), anti-phospho-p70^S6K (T421/S424) and phospho-pRb (S780) (Cell Signaling Technology, Beverly, MA, USA); anti-ERK, anti-Akt, anti-p70^S6K, anti-VEGFR-2, anti-integrin α3, anti-integrin β1, anti-ILK, anti-Cdk4, anti-Cdk2, anti-actin antibodies, and mouse and rabbit IgG-horseradish peroxidase conjugates (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Subconfluent A549 and H1299 cells, plated on 6-well plates (5×10⁴ cells/well; SPL Life Sciences, Gyeonggi-do, Korea), were serum-starved for 24 h in basal DMEM to synchronize cells in the G₁/G₀ phase of the cell cycle, and pretreated with marmesin (0.1–10 µM) for 30 min prior to 10% FBS stimulation for 24 h. In some experiments, quiescent cells were pretreated with marmesin (10 µM) for 30 min, followed by 10% FBS stimulation for 12 h. After stimulation, cells were thoroughly rinsed with phosphate-buffered saline (PBS; pH 7.4) to remove any residual marmesin, and further incubated with 10% FBS for another 12 h until the end of the 24 h time point. Cell viability was determined by a Muse™ Cell Analyzer using cell count and viability assay kit (Merck Millipore, Billerica, MA, USA), and the cell proliferation was quantified as previously described (19). The results from triplicate determinations (mean ± standard deviation) are presented as the percentage of viable cells of total cell count or the fold-increase of the untreated controls.

Quiescent cells were pretreated with marmesin for 30 min, followed by 10% FBS stimulation for different time points, as indicated. Cells were rinsed twice with ice-cold PBS and lysed by incubation in 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 100 µg/ml AEBSF, 10 µg/ml aprotinin, 1 µg/ml pepstatin A, 0.5 µg/ml leupeptin, 80 mM β-glycerophosphate, 25 mM sodium fluoride and 1 mM sodium orthovanadate for 30 min at 4°C. Cell lysates were clarified at 13,000 × g for 20 min at 4°C, and the supernatants were subjected to western blotting as previously described (20). All western blot analyses are representative of at least three independent experiments. Bands of interest were integrated and quantified by the use of National Institutes of Health (NIH) ImageJ version 1.34s software.

The upper side of the Transwell insert (6.5-mm diameter insert, 8-µm pore size) (Corning Costar, Inc., Corning, NY, USA) was coated with 50 µl of 1 mg/ml Matrigel^® (BD Biosciences, Bedford, MA, USA) diluted in serum-free DMEM. Aliquots (100 µl) of cells (5×10⁵ cells/ml) resuspended in serum-free DMEM were added to the upper compartment of the Matrigel-coated Transwell and 600 µl of serum-free DMEM was added to the lower compartment. After serum starvation for 2 h, the cells were pretreated with marmesin (10 µM) for 30 min, followed by 10% FBS stimulation for 16 h. The inserts were fixed with methanol and using a cotton-tipped swab the non-invasive cells were removed from the top of the membrane. After staining with 0.04% Giemsa staining solution (Sigma-Aldrich Co., St. Louis, MO, USA), the numbers of invasive cells (mean ± standard deviation) were determined from six different fields using ×200 objective magnification (21,22).

Total RNA was purified with PureHelix™ RNA extraction solution (Nanohelix Co., Daejeon, Korea). Integrity of RNA was checked by agarose gel electrophoresis and ethidium bromide staining. One microgram of RNA was used as template for each reverse transcriptase (RT)-mediated polymerase chain reaction (PCR) using First Strand cDNA Synthesis kit (BioAssay Co., Ltd., Daejeon, Korea). Primers for PCR were synthesized by Bioneer Corporation (Daejeon, Korea). Primer sequences were as follows: MMP-2 forward, 5′-GCTCAGATCCGTGGTGAGAT-3′ and reverse, 5′-GGTGCTGGCTGAGTAGATCC-3′; VEGFR-2 forward, 5′-TGCCTACCTCACCTGTTTCCT-3′ and reverse, 5′-TACACGGTGGTGTCTGTGTCA-3′; VEGF-A forward, 5′-TCGGGCCTCCGAAACCATGA-3′ and reverse, 5′-CCTGGTGAGAGATCTGGTTC-3′; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward, 5′-GAAGGTGAAGGTCGGAGTC-3′ and reverse, 5′-GAAGATGGTGATGGGATTTC-3′. Bands of interest were integrated and quantified by the use of NIH ImageJ version 1.34s software.

Quiescent cells were pretreated with marmesin (10 µM) for 30 min, followed by 10% FBS stimulation for 12 h. Cells were washed with PBS to remove any residual marmesin, and then further stimulated with 10% FBS for another 12 h. ELISA assay was performed to measure the concentration of VEGF in the conditioned media using VEGF ELISA kit (R&D Systems, Minneapolis, MN, USA), according to the manufacturer's instructions. Secreted VEGF levels were analyzed at 450 nm using BioTek Synergy Mx microplate reader (BioTek Instruments, Winooski, VT, USA).

Tube formation assay was performed to examine the ability of conditioned media from marmesin-treated NSCLC cells to regulate angiogenic responses in vitro. Quiescent NSCLC cells were pretreated with marmesin (10 µM) for 30 min, followed by 10% FBS stimulation for 12 h. Cells were washed with PBS to remove any residual marmesin, and then further stimulated with 10% FBS for another 12 h. After stimulation, conditioned media were collected. Quiescent HUVECs (4×10⁴ cells/ml) were added to Matrigel^®-coated plates and treated with conditioned media for 9 h. Tube formation was observed with an Olympus CKX41 inverted microscope (CAchN 10/0.25php objective) and ToupTek Toupview software (version ×36, 3.5.563; Hangzhou ToupTek Photonics Co., Zhejiang, China).

Statistical analysis was performed using Student's t-test, and was based on at least three different experiments. The results were considered to be statistically significant at P<0.05.

We first investigated the ability of marmesin to regulate cell proliferation in p53 wild-type A549 and p53-deficient H1299 NSCLC cells (Fig. 1A and B). Marmesin treatment inhibited mitogen-stimulated cell proliferation in a dose-dependent manner and did not alter cell viability and morphology at the highest concentration used in the present study, indicating that marmesin-mediated inhibition of cell proliferation was not mediated by induction of apoptosis or cytotoxicity. Based on these findings, we next analyzed the changes in the cell cycle regulatory proteins in the marmesin-treated NSCLC cells (Fig. 1C). Marmesin treatment markedly suppressed mitogen-induced expression of cyclin-dependent kinase 4 (Cdk4), but not Cdk2, to levels observed in the untreated controls, resulting in inhibition of pRb phosphorylation in both NSCLC cell lines. Marmesin has previously been reported to inhibit proliferation by downregulation of Cdk4, Cdk2 and cyclin D in VEGF-A-treated HUVECs (14). Although the molecular mechanism of marmesin in regulating cell cycle-related proteins appears slightly different in cell types, these findings demonstrate the antiproliferative activity of marmesin against various types of cells, independently of p53 expression status. In addition, the inhibitory effect of marmesin on NSCLC cell proliferation was not changed after the withdrawal of marmesin at the 12 h time point, and was sustained up to 24 h, suggesting that this effect may be irreversible until the end of this experiment (Fig. 2).

We next examined the effect of marmesin on cell invasion in the A549 and H1299 cells. Marmesin treatment markedly blocked mitogen-stimulated cell invasion (Fig. 3A and B). The regulatory pattern of marmesin on NSCLC cell invasion was very similar to that on NSCLC cell migration (data not shown). Based on these findings, we analyzed the changes in expression of matrix metalloproteinases (MMPs) in the marmesin-treated NSCLC cells. Expression and activity of MMPs have been known to modulate cell migration, invasion and angiogenesis by degrading extracellular matrix components and cell surface molecules (23–26). As shown in Fig. 3C, marmesin treatment suppressed mitogen-induced expression of MMP-2 in both cell lines. In contrast, the levels of tissue inhibitor of metalloproteinases-2, an endogenous inhibitor of MMPs, were not altered in the mitogen- or marmesin-treated NSCLC cells (data not shown) (27–29). Collectively, these findings suggest that inhibition of cell invasion by marmesin is mediated, at least in part, through the suppression of MMP-2 expression (Fig. 3C).

To elucidate the molecular mechanisms and therapeutic targets of marmesin in regulating NSCLC cell proliferation and invasion, we examined the changes in activation of mitogen-stimulated signaling pathways including Src kinase, mitogen-activated protein kinase (MEK), extracellular signal-regulated kinase (ERK), Akt and p70S6 kinase (p70^S6K) in marmesin-treated NSCLC cells (14,30). As expected, mitogenic stimulation significantly increased the phosphorylation of MEK, ERK, Akt and p70^S6K, as compared with unstimulated controls (Fig. 4). In contrast, marmesin treatment markedly inhibited mitogen-stimulated phosphorylation of Src, MEK, ERK, Akt and p70^S6K in both NSCLC cells. Moreover, marmesin treatment markedly suppressed mitogen-induced expression of cell signaling molecules such as VEGFR-2, integrin β1 and integrin-linked kinase (ILK), and a key angiogenic factor VEGF, which play important roles in cancer growth and progression associated with angiogenesis (Fig. 5) (31–33). In addition to direct antitumor activity, these findings suggest the possibility that marmesin may regulate angiogenic responses through inhibition of VEGF expression and secretion in NSCLC cells.

In the tumor microenvironment cancer cells secrete a variety of biological molecules including cytokines and growth factors, which play important roles in cellular responses such as proliferation, migration, invasion and angiogenesis (5,34). Based on inhibitory effect of marmesin on VEGF expression, we thus analyzed the secreted levels of VEGF in conditioned media collected from marmesin-treated NSCLC cells. As shown in Fig. 6A, marmesin treatment significantly inhibited the secretion of VEGF from both NSCLC cell lines in response to mitogenic stimulation. H1299 cells were found to be more responsive to marmesin-mediated inhibition of VEGF secretion, as compared with A549 cells. These findings are similar to the patterns of VEGF transcription levels in marmesin-treated NSCLC cells (Fig. 5C), demonstrating that marmesin inhibits VEGF secretion through downregulation of VEGF mRNA expression. To determine whether secreted biomolecules including VEGF from marmesin-treated NSCLC cells affect the cellular fate of adjacent and/or other cells, we next performed in vitro angiogenesis assay using conditioned media from NSCLC cells treated with or without marmesin (Fig. 6B). The conditioned media from marmesin-treated NSCLC cells significantly inhibited the formation of capillary-like structures by HUVECs. Although the types and levels of biomolecules secreted from marmesin-treated NSCLC cells remain to be further identified, these observations suggest that marmesin-mediated inhibition of VEGF expression and secretion in NSCLC cells may be one of the major factors in the modulation of tumor-derived angiogenesis (Fig. 6B).