We prepared Lico B with 95% purity according to the method previously reported (18). Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), trypsin, penicillin and streptomycin (P/S) and phosphate-buffered saline (PBS) were purchased from Thermo Scientific (Logan, UT, USA). Antibodies against cyclin D1, p21, p27, actin, CHOP, DR4, DR5, cytochrome c, α-tubulin, MTCO1, Apaf-1, survivin, Bid, Bax, Bcl-_xl, myeloid cell leukemia-1 (Mcl-1) and poly (ADP-ribose) polymerase (PARP) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). A 4′,6-diamidino-2-phenylindole (DAPI) was obtained from Sigma-Aldrich, Inc. (St. Louis, MO, USA).

HN22 and HSC4 cells are human oral squamous cancer cell lines. HN22 cells were provided by Dankook University (Cheonan, Korea) and HSC4 cells were provided by Hokkaido University (Hokkaido, Japan) (19). Both cell types were maintained in DMEM supplemented with 10% FBS and 100 U/ml each of P/S at 37ºC in a 5% CO₂ incubator.

HN22 (6×10⁴/6-well plate) and HSC4 (7.5×10⁴/6-well plate) cells were grown for 24 h, and then treated with various concentrations of Lico B (0, 10, 20 and 30 μM). After treatment with Lico B for 24 and 48 h, cells were trypsinized and resuspended in complete medium. Each sample was mixed with trypan blue solution. Colored (non-viable) and dye-excluding (viable) cells were counted on a haemocytometer.

HN22 (6×10⁴/6-well plate) and HSC4 (7.5×10⁴/6-well plate) cells were seeded and allowed to incubate overnight in 10% FBS-containing DMEM. At 48 h after treatment with different concentration of Lico B (0, 10, 20 and 30 μM), cells were harvested by trypsinization and analyzed for the detection of early/late apoptosis and cell death using Annexin V and Dead Cell kit (EMD Millipore Corp., Billerica, MA, USA) according to the manufacturer's instructions. The percentage of apoptotic cells was calculated from triplicate samples by statistical analysis of dot plot using Muse Cell Analyzer (Merck Millipore).

After treatment with Lico B, the HN22 and HSC4 cells were harvested by trypsin digestion. The cells were washed in ice-cold PBS and fixed with 100% methanol at room temperature (RT) for 1 h. Fixed cells were stained with DAPI solution (2 μg/ml) and deposited on slides at RT in the dark. DAPI-stained cells were observed through a FluoView confocal laser microscope (Flouview FV10i, Olympus Corp., Tokyo, Japan).

HN22 (6×10⁴/6-well plate) and HSC4 (7.5×10⁴/6-well plate) cells were seeded and exposed to Lico B (0, 10, 20 and 30 μM) for 48 h. The harvested cells were washed in PBS and 200 μl of Muse cell cycle reagent (EMD Millipore Corp.) was added. The cells were incubated for 30 min at RT in the dark. Cell cycle distribution was analyzed by Muse Cell Analyzer (Merck Millipore).

The cell lysates were prepared using PRO-PREP™ Protein Extraction Solution (iNtRON Biotechnology, Korea) and then supernatants were collected by centrifugation. Samples containing equal amounts of protein were separated by SDS-PAGE and transferred to polyvinylidene difluoride membrane. The membranes were blocked with 5% skim milk in PBS with 0.1% Tween-20 at RT. Membranes were probed with primary antibodies at 4ºC overnight with mild shaking and then washed with PBST for a total of 30 min. Finally, membranes were incubated with the secondary antibodies for 2 h at RT. Antibody-bound protein bands were visualized using ECL Plus Western Blotting Detection system (Santa Cruz Biotechnology, Inc.).

Muse Oxidative Stress kit (EMD Millipore Corp.) was used to examine the effect of Lico B on the generation of ROS. Briefly, cells from each treatment were harvested, washed with PBS, and then resuspended in Muse Oxidative Stress Reagent working solution for 30 min at 37ºC in the dark. Intracellular reactive oxygen species was measured by flow cytometry using Muse Cell Analyzer (Merck Millipore).

MMP was assessed using Muse MitoPotential kit (EMD Millipore Corp.) according to the manufacturer's protocol. Briefly, HN22 and HSC4 cells treated with Lico B were harvested to observe quantitatively MMP. The cells were incubated with Muse MitoPotential working solution at 37ºC for 20 min. Then, 5 μl of 7-aminoactinomycin D (7-AAD) was added and samples were further incubated at RT for 5 min. Muse Cell Analyzer (Merk Millipore) was employed to detect MMP.

The activity of caspase-1, -3, -4, -5, -6, -7, -8 and -9 was measured using a Muse Multi-caspase kit (MCH100109, Merck Millipore) according to the manufacturer's instructions. HN22 and HSC4 cells were incubated with various concentrations of Lico B (0, 10, 20 and 30 μM) for 48 h. Cells were harvested and incubated with 5 μl of Muse Multi-Caspase reagent working solution at 37ºC for 30 min. Then, 150 μl of Muse Caspase 7-AAD working solution was added to each sample. Multi-caspase assay was performed with Muse Cell Analyzer (Merck Millipore).

Using Student's t-test, the statistical significance was assessed. The result with p-value <0.05, indicates statistical significance of data.

Trypan blue staining was performed to determine the viability of HN22 and HSC4 cells treated with Lico B (0, 10, 20 and 30 μM) for 24 and 48 h, and then live cells were counted using a haemocytometer. As shown in Fig. 1A, cell viability was significantly decreased in a concentration- and time-dependent manner by Lico B (Fig. 1A). The IC₅₀ value was 15 μM for HN22 cells and 13 μM for HSC4 cells at 48 h. Cell morphology was observed under an inverted phase contrast microscope after the cells were treated with Lico B for 48 h, and images were obtained (Fig. 1B). Significant cell shrinkage, formation of plasma membrane blebs and a decreased cellular attachment rate were observed in the Lico B-treated groups. These results suggest that Lico B has an inhibitory effect on the cell growth of HN22 and HSC4 cells.

Phosphatidylserine (PS) of normal cells are mainly distributed on the inner layer of the membrane. However, early apoptotic cells transpose PS on the outer layer of the membrane (20). Annexin V can detect PS on the outer membrane, designating early apoptotic cells while 7-AAD can permeate into plasma membrane and bind to DNA when the membrane of cells lose integrity. Annexin V and 7-AAD absorb laser with different wavelengths so that both fluorophores can be analyzed at the same time (20). To determine whether the Lico B could induce apoptosis in HN22 and HSC4 cells, we stained cells with Annexin V/7-AAD for analysis by flow cytometry. As shown in Fig. 1C, Lico B caused apoptosis of OSCC cells in a concentration-dependent manner. The apoptotic cells of HN22 were 12.62±1.22, 47.27±6.83, 79.84±2.99% at 10, 20 and 30 μM of Lico B, respectively (Fig. 1C, left). In HSC4, population of apoptotic cells was 36.63±3.33, 60.92±3.01, 83.36±1.43% at 10, 20 and 30 μM of Lico B, respectively (Fig. 1C, right). Additionally, the characteristic appearance of apoptosis was analyzed by DAPI staining. HN22 and HSC4 cells treated with Lico B presented nuclear shrinkage, chromatin condensation and fragmentation of nuclear bodies compared to control (Fig. 2A). The cell cycle distribution was assessed after PI staining by Muse Cell Analyzer. As shown in Fig. 2B, the treatment of cells with Lico B at 30 μM caused 36.83±3.85 or 39.37±0.6% induction of sub-G₁ phase in HN22 (Fig. 2B, left) and HSC4 (Fig. 2B, right), respectively. Overall, these findings suggest that Lico B can induce apoptosis in HN22 and HSC4 cells.

To determine the effect of Lico B on the cell cycle progression, HN22 and HSC4 cells were treated with different concentrations of Lico B (0, 10, 20 and 30 μM) for 48 h and the cell cycle distribution was assessed by Muse Cell Analyzer. Fig. 2B showed that HN22 and HSC4 cells accumulated in G₁ phase after treatment of Lico B (Fig. 2B). To confirm G₁ block in cell cycle progression, the level of G₁ cell cycle components was monitored by western blot analysis. Fig. 2C showed that Lico B increased the protein levels of p21 and p27, and decreased the levels of cyclin D1 (Fig. 2C). These results support that Lico B causes G₁ phase arrest in OSCC cells.

Overwhelming ER stress resulted from accumulation of ROS has been known to trigger apoptotic signaling (15,21). As shown in Fig. 3A, ROS production from HN22 and HSC4 cells treated with Lico B was significantly increased. CHOP is one of the components of ER stress-mediated apoptosis pathway (16). To assess the effect of Lico B on CHOP expression in HN22 and HSC4 cells, western blotting was performed. Fig. 3B showed that the protein levels of CHOP were increased in a dose-dependent manner by Lico B. CHOP is suggested to induce apoptosis through upregulation of DR4 and DR5 expression (16). We also observed that Lico B treatment increased dose-dependently DR4 and DR5 in HN22 and HSC4 cells (Fig. 3B). These results showed that treatment of OSCC cells with Lico B induced ER stress, which, in turn, upregulates CHOP, DR4 and DR5.

Dynamic changes in Bcl-2 family proteins were associated with death receptor-mediated apoptosis (22). Bcl-_xl functions as an anti-apoptotic protein, whereas Bax is pro-apoptotic protein (23). Bid is then truncated to induce apoptosis. Downregulation of Mcl-1 protein may be required to initiate a cascade of apoptotic signals leading to release of cytochrome c (5). To investigate whether expression of apoptosis regulatory proteins may be modulated by Lico B, the protein levels of Bid, Bax, Bcl-_xl and Mcl-1 in HN22 and HSC4 cells treated with Lico B were analyzed by western blotting. Fig. 3C showed that the levels of pro-apoptotic Bax were increased and the expression levels of Bid, Bcl-_xl and Mcl-1 were decreased in a concentration-dependent manner.

Death receptor induces apoptosis through the disruption of the mitochondrial membrane potential (MMP) and sequential release of cytochrome c from mitochondria into the cytosol (24). The MMP was measured by Muse Cell Analyzer with cationic, lipophilic dye. The dye permeates mitochondrial membrane of intact cells and accumulates within inner membrane. Thus, cells with depolarized mitochondria show a decrease in fluorescence (25). The data indicated that treatment of cells with Lico B induced a loss of MMP in HN22 and HSC4 cells (Fig. 4A). The loss of MMP leads to the release of cytochrome c into the cytosol then, triggering the downstream processes in the apoptotic cascade (26). To examine whether Lico B could cause the release of cytochrome c into the cytosol, subcellular distribution of cytochrome c in cytosol and mitochondria fraction were analyzed by western blotting. The results showed that cytochrome c in cytosol fraction of HN22 and HSC4 cells treated with Lico B (0, 10, 20 and 30 μM) was concentration-dependently increased (Fig. 4B). These results indicated that Lico B facilitated mitochondrial membrane depolarization and release of cytochrome c into the cytosol. Additionally, to investigate downstream target proteins including Apaf-1, survivin and PARP in mitochondria-mediated apoptosis, their expression was analyzed by western blotting with specific antibodies. As shown in Fig. 4B, the expression levels of survivin and PARP were diminished, but Apaf-1 and C-PARP were increased in a concentration-dependent manner by Lico B.

To examine whether Lico B-induced apoptosis would be associated with the activation of caspases, the expression levels and activity of caspases such as caspase-1, -3, -4, -5, -6, -7, -8 and -9 in the Lico B-treated OSCC cells were assessed using Muse Cell Analyzer. As shown in Fig. 5, HN22 and HSC4 cells treated with Lico B exhibited enhanced multi-caspase activity in a concentration-dependent manner. The multi-caspase activity of HN22 cells (Fig. 5, left) was 4.33±0.76, 34.09±2.07 and 81.1±2.86%, respectively, at 10, 20 and 30 μM of Lico B compared with control. In HSC4 cells (Fig. 5, right), multi-caspase activity was 10.08±0.85, 38.5±2.25 and 74.29±2.81% at 10, 20 and 30 μM of Lico B, respectively.