Tenuifolide B from Cinnamomum tenuifolium Stem Selectively Inhibits Proliferation of Oral Cancer Cells via Apoptosis, ROS Generation, Mitochondrial Depolarization, and DNA Damage

The development of drugs that selectively kill oral cancer cells but are less harmful to normal cells still provide several challenges. In this study, the antioral cancer effects of tenuifolide B (TFB), extracted from the stem of the plant Cinnamomum tenuifolium are evaluated in terms of their effects on cancer cell viability, cell cycle analysis, apoptosis, oxidative stress, and DNA damage. Cell viability of oral cancer cells (Ca9-22 and CAL 27) was found to be significantly inhibited by TFB in a dose-responsive manner in terms of ATP assay, yielding IC50 = 4.67 and 7.05 μM (24 h), but are less lethal to normal oral cells (HGF-1). Dose-responsive increases in subG1 populations as well as the intensities of flow cytometry-based annexin V/propidium iodide (PI) analysis and pancaspase activity suggested that apoptosis was inducible by TFB in these two types of oral cancer cells. Pretreatment with the apoptosis inhibitor (Z-VAD-FMK) reduced the annexin V intensity of these two TFB-treated oral cancer cells, suggesting that TFB induced apoptosis-mediated cell death to oral cancer cells. Cleaved-poly (ADP-ribose) polymerase (PARP) and cleaved-caspases 3, 8, and 9 were upregulated in these two TFB-treated oral cancer cells over time but less harmful for normal oral HGF-1 cells. Dose-responsive and time-dependent increases in reactive oxygen species (ROS) and decreases in mitochondrial membrane potential (MitoMP) in these two TFB-treated oral cancer cells suggest that TFB may generate oxidative stress as measured by flow cytometry. N-acetylcysteine (NAC) pretreatment reduced the TFB-induced ROS generation and further validated that ROS was relevant to TFB-induced cell death. Both flow cytometry and Western blotting demonstrated that the DNA double strand marker γH2AX dose-responsively increased in TFB-treated Ca9-22 cells and time-dependently increased in two TFB-treated oral cancer cells. Taken together, we infer that TFB can selectively inhibit cell proliferation of oral cancer cells through apoptosis, ROS generation, mitochondrial membrane depolarization, and DNA damage.


Introduction
Oral cancer is the sixth most common type of cancer globally [1] and is especially prevalent in areas that feature a high frequency of betel nut, alcohol, and cigarette consumption [2][3][4]. Oral cancer is likely ignored by patients in early stages and is commonly detected at a later stage. Late detection, combined with poor chemotherapy outcomes, leads to high morbidity and mortality rates of oral cancer [5]. While several drugs have proven effective at killing cancer cells, they are also toxic to normal tissue cells, and the need for selective antioral cancer drugs remains urgent.
Alternatively, we previously used the stems of C. tenuifolium [22] to identify several novel compounds, including tenuifolide A, isotenuifolide A, tenuifolide B (TFB), secotenuifolide A, and tenuifolin, along with some known compounds. Secotenuifolide A was found to provide the best antiproliferative effect against two human prostate cancer cells (DU145 and LNCaP) with IC 50 values < 7 µM after 24 h of treatment. For TFB (3-(1-methoxyeicosyl)-5-methylene-5H-furan-2-one), its IC 50 values were 246 and 22.2 µM for DU145 and LNCaP cancer cells after 24 h of treatment [22]. However, the biological effect of TFB against the oral cancer cells was not addressed as yet.
The current study first evaluates the possible selectively antiproliferactive effect and the mechanism of C. tenuifolium stem-derived TFB on oral cancer cells by analyzing cell viability, cell cycle progression, apoptosis, reactive oxygen species (ROS) induction, mitochondrial depolarization, and DNA damage.

Cell Viability and ATP Cellular Content
ATP content has been widely used to measure cell viability [24,25]. Figure 1 shows the ATP assay of cell viability after 24 h of treatment with TFB (0, 5, 10, and 15 µM). The viability of TFB-treated oral

Cell Cycle Progression
To examine whether the cell cycle was affected by TFB, the cell cycle progression was examined.

Annexin V-Based Apoptosis
To validate the role of apoptosis in the increase in the subG1 population in TFB-treated Ca9-22 and CAL 27 cells, the annexin V/propidium iodide (PI) staining method was used. Figures 3A,B respectively show the patterns of dose response changes of annexin V/PI staining profiles of TFBtreated Ca9-22 and CAL 27 cells. By calculating the percentages of annexin V positive (%), the apoptosis level ( Figures 3C,D) show a significant increase in a dose-responsive manner in TFB-treated Ca9-22 and CAL 27 cells (p < 0.001). When the Ca9-22 and CAL 27 cells were pretreated with apoptosis inhibitor Z-VAD-FMK, apoptosis induced by different doses of TFB was decreased.

Cell Cycle Progression
To examine whether the cell cycle was affected by TFB, the cell cycle progression was examined. Figure 2A

Cell Cycle Progression
To examine whether the cell cycle was affected by TFB, the cell cycle progression was examined.

Annexin V-Based Apoptosis
To validate the role of apoptosis in the increase in the subG1 population in TFB-treated Ca9-22 and CAL 27 cells, the annexin V/propidium iodide (PI) staining method was used. Figures

Annexin V-Based Apoptosis
To validate the role of apoptosis in the increase in the subG1 population in TFB-treated Ca9-22 and CAL 27 cells, the annexin V/propidium iodide (PI) staining method was used. Figure 3A

Caspases-Based Apoptosis
To validate the role of apoptosis in the increase in annexin V intensity in TFB-treated Ca9-22 and CAL 27 cells, the pancaspase activity assay was used. Figures

Caspases-Based Apoptosis
To validate the role of apoptosis in the increase in annexin V intensity in TFB-treated Ca9-22 and CAL 27 cells, the pancaspase activity assay was used. Figure 4A Figure 4I shows the protein expressions of apoptosis signaling proteins, such as cleaved-poly (ADP-ribose) polymerase (PARP) and cleaved-caspases 3 and 8 gradually increased over 3-24 h and cleaved-caspase 9 was detected at 24 h in TFB-treated Ca9-22 cells. In TFB-treated CAL 27 cells, the protein expressions of cleaved-PARP and cleaved-caspases 3, 8, and 9 gradually increased from 3 to 6 h, moderately increased at 12 h, and declined at 24 h. In contrast, these apoptosis signaling proteins in TFB-treated HGF-1 were weak.  Figure 4I shows the protein expressions of apoptosis signaling proteins, such as cleaved-poly (ADP-ribose) polymerase (PARP) and cleaved-caspases 3 and 8 gradually increased over 3-24 h and cleaved-caspase 9 was detected at 24 h in TFB-treated Ca9-22 cells. In TFB-treated CAL 27 cells, the protein expressions of cleaved-PARP and cleaved-caspases 3, 8, and 9 gradually increased from 3 to 6 h, moderately increased at 12 h, and declined at 24 h. In contrast, these apoptosis signaling proteins in TFB-treated HGF-1 were weak.

ROS
To determine why TFB may inhibit cancer cell proliferation and induce apoptosis, the cellular ROS level was examined. Figure 5A

ROS
To determine why TFB may inhibit cancer cell proliferation and induce apoptosis, the cellular ROS level was examined. Figures 5A,B

Mitochondrial Membrane Potentials (MitoMP)
DiOC 2 (3)-based MitoMP detection assay was performed to evaluate the impact of TFB (0, 5, 10, and 15 µM)-induced ROS generation. Figure 6A,B show the MitoMP profiles for TFB-treated oral cancer Ca9-22 and CAL 27 cells, respectively, after a 24 h of treatment. By calculating the percentages of DiOC 2 (3)-negative in Figure 6A,B and comparing with the control, it was found that the MitoMP-negative (%) was gradually increased in TFB-treated Ca9-22 and CAL 27 cells in a dose-responsive manner (p < 0.001) ( Figure 6C,D). Therefore, the MitoMP level of Ca9-22 and CAL 27 cells was significantly decreased after TFB treatment. Figure 6E,

γH2AX Expression
To examine the role of DNA damage in TFB-induced antiproliferation of Ca9-22 oral cancer cells, the expression of DNA double strand break marker γH2AX was analyzed via both flow cytometry and Western blotting. Figure 7A shows that the flow cytometry-based γH2AX/PI staining profiles of

γH2AX Expression
To examine the role of DNA damage in TFB-induced antiproliferation of Ca9-22 oral cancer cells, the expression of DNA double strand break marker γH2AX was analyzed via both flow cytometry and Western blotting. Figure 7A shows that the flow cytometry-based γH2AX/PI staining profiles of TFB-treated Ca9-22 cells after 24 h of treatment. Figure 7B shows the γH2AX-positive intensity of TFB (0, 5, 10, and 15 µM)-treated Ca9-22 cells increased in a dose-responsive manner (p < 0.001). Moreover,  Figure 7C shows that the γH2AX expression by Western blotting of TFB-treated Ca9-22 cells after 24 h of treatment with indicated doses were dramatically increased at higher doses.
TFB-treated Ca9-22 cells after 24 h of treatment. Figure 7B shows the γH2AX-positive intensity of TFB (0, 5, 10, and 15 μM)-treated Ca9-22 cells increased in a dose-responsive manner (p < 0.001). Moreover, Figure 7C shows that the γH2AX expression by Western blotting of TFB-treated Ca9-22 cells after 24 h of treatment with indicated doses were dramatically increased at higher doses.
For the time course experiments, Figure 7D shows that γH2AX expression of TFB-treated Ca9-22, CAL 27, and HGF-1 cells were increased in a time-dependent manner. The TFB-induced γH2AX levels were dramatically induced in both oral cancer cells (Ca9- 22 and CAL 27). In contrast, TFBinduced γH2AX levels increased only slightly at 12 h in oral normal HGF-1 cells.
Based on ATP content assays, the current study found that the IC50 values of TFB were 4.67 and 7.05 μM after 24 h of treatment in oral cancer cells (Ca9-22 and CAL 27), respectively. In general, the sensitivity of TFB to oral cancer cells (Ca9-22 and CAL 27) was higher than that of other Cinnamomum plants to other types of cancer cells. Moreover, TFB was less cytotoxic to human prostate cancer DU145 and LNCaP cells [22]. These results suggest that TFB may have a cancer cell type-specific antiproliferation effect.   Figure 7D shows that γH2AX expression of TFB-treated Ca9-22, CAL 27, and HGF-1 cells were increased in a time-dependent manner. The TFB-induced γH2AX levels were dramatically induced in both oral cancer cells (Ca9-22 and CAL 27). In contrast, TFB-induced γH2AX levels increased only slightly at 12 h in oral normal HGF-1 cells.
Based on ATP content assays, the current study found that the IC 50 values of TFB were 4.67 and 7.05 µM after 24 h of treatment in oral cancer cells (Ca9-22 and CAL 27), respectively. In general, the sensitivity of TFB to oral cancer cells (Ca9-22 and CAL 27) was higher than that of other Cinnamomum plants to other types of cancer cells. Moreover, TFB was less cytotoxic to human prostate cancer DU145 and LNCaP cells [22]. These results suggest that TFB may have a cancer cell type-specific antiproliferation effect.
Following 24 h of treatment, the cytotoxicity (IC 50 ) of taxol in human prostate cancer DU145 and LNCaP cells is 4.84 and 6.32 µM, respectively [22]. The IC 50 of cisplatin in oral cancer Ca9-22 cells is 10.2 µM (data not shown). Therefore, our developed TFB (IC 50 = 4.67 and 7.05 µM) has similar sensitivity to these clinical drugs in oral cancer cells (Ca9-22 and CAL 27). Moreover, we found that TFB is less harmful to normal oral HGF-1 cells (Figure 1), suggesting that TFB selectively kills oral cancer cells and may prevent side effects of oral cancer therapy. Similarly, Cinnamomum stem bark extract has been reported to selectively kill other types of cancer cells. In Cinnamomum burmannii Blume stem bark extract after 24 h of treatment, the IC 50 values of nasopharyngeal carcinoma cells (HK1 and C666-1) were 108.32 and 224.32 µg/mL, respectively, whereas the IC 50 of immortalized human skin keratinocyte HaCaT cells was 320.29 µg/mL [27]. However, our study only tested one normal oral cell line, and further study is needed to confirm these findings, including more normal oral cell lines to further show the lack of possible side-effects of TFB.
Secotenuifolide A, also isolated from the same material of this study (C. tenuifolium stems), has been reported to inhibit cell proliferation, increase the subG1 population, induce apoptosis and ROS generation, and decrease mitochondrial membrane potential in human prostate cancer cells, DU145 [22]. Secotenuifolide A also exhibited a release of cytochrome c from mitochondria and the activation of caspase-9/caspase-3 [22]. Similarly, the TFB from C. tenuifolium stems showed the same effect of oxidative stress (ROS induction and MitoMP depletion) on oral cancer Ca9-22 and CAL 27 cells (Figures 5 and 6). In contrast, the TFB-induced ROS generation in HGF-1 cells was only slightly induced. These results suggest that TFB exhibited selective ROS induction in oral cancer cells (Ca9-22 and CAL 27), but less induction in HGF-1 cells. NAC pretreatment experiments ( Figure 5A,B) validated that ROS was relevant to TFB-induced cell death because the ROS generation in the two types of oral cancer cells was reduced by NAC pretreatment.
Annexin V and pancaspase results (Figures 3 and 4) support that TFB is apoptosis-inducible in oral cancer cells. Moreover, the TFB-induced apoptosis was reduced by Z-VAD-FMK pretreatment ( Figure 3A,B), suggesting that apoptosis was involved in selective killing by TFB. However, the role of apoptosis signaling in TFB-induced apoptosis was not addressed specifically. Caspases 8 and 9 involved in intrinsic and extrinsic apoptotic pathways, respectively. Both converge in activating the executioner caspases 3 and 7 [28]. PARP is also involved in apoptosis [29,30]. To address the role of apoptosis signaling, we found that both TFB-treated oral cancer cells (Ca9-22 and CAL 27) induced activation of PARP and caspases 3, 8, and 9 by cleavage. However, these oral cancer cells displayed a differential expression of these TFB-induced apoptosis proteins. For example, cleaved-caspase 8 was mainly or early induced during 3-12 h in TFB-treated Ca9-22 cells, but cleaved-caspase 9 induction showed a later response at 24 h ( Figure 4I). For CAL 27 cells, cleaved-PARP and cleaved-caspases 3, 8 and 9 were upregulated early after 3 h. These caspases peaked at 6 h, gradually declining by 24 h. Similarly, other drug-induced apoptosis also showed a similar tendency for cleaved-caspase expression. For example, cleaved-caspase 3 increased at 6-12 h and declined at 24 h in 0.2 µM staurosporine-treated human endothelial cornea cells [31]. After treatment of 0.2 µM staurosporine, cleaved-caspase 3 was also increased at 2-8 h and declined at 12-24 h for human cervical cancer HeLa cells. Cleaved-caspase 3 increased at 0.5-1 h and declined at 2-24 h in cervical cancer C-33A cells [32].

Conclusions
TFB treatment induces apoptosis, ROS generation, mitochondrial depolarization, and DNA damage, which ultimately results in the antiproliferation of oral cancer Ca9-22 cells. This study also shows that TFB selectively kills the two oral cancer cell lines tested here and opts for its application in anti-oral cancer therapies.

Measurement of Cell Viability-Cellular ATP Content
Cellular ATP level was determined by using the ATP-lite Luminescence ATP Detection Assay System (PerkinElmer Life Sciences, Boston, MA, USA) according to the manufacturer's instructions with a slight modification [44]. Briefly, Ca9-22 cells were plated at 4000 cells/well in 96-well plates. After seeding overnight, cells were treated with vehicle control (DMSO) or with TFB at indicated concentrations (5, 10, and 15 µM) for 24 h. After removing the medium solution, 100 µL of serum-free medium and 50 µL of a mammalian cell lysis solution was added per well of a microplate with orbital shaking at 100 rpm for 5 min. Then, 100 µL of the cell lysates/well was transferred to white 96-well plates and reacted with 50 µL of substrate solution (D-Luciferin and luciferase) under orbital shaking at 100 rpm for 5 min and then left to stand in darkness for a further 10 min. Finally, the luminescence was assayed using a microplate luminometer (CentroPRO LB 962, Berthold, ND, USA).

Measurement of Cell Cycle Progression
The cellular DNA was stained with PI as previously described [45]. Cells were plated at 3 × 10 5 cells/2 mL cell culture medium on a 6-well plate. Briefly, cells were added with vehicle (DMSO only) or TFB. After collection for 70% ethanol fixation overnight, the centrifuged cell pellets were resuspended in 1 mL of PBS containing 50 µg/mL PI for 15 min at room temperature in darkness. Subsequently, these samples were examined using a FACSCalibur flow cytometer (Becton-Dickinson, Mansfield, MA, USA) (excitation: 488 nm and emission: 617 nm) and BD Accuri C6 software.

Measurement of Apoptosis by Annexin V Staining
Annexin V (Strong Biotech Corporation, Taipei, Taiwan)/PI (Sigma, St. Louis, MO, USA) double-staining for apoptosis analysis was performed as previously described [46]. Cells were plated at 3 × 10 5 cells/2 mL cell culture medium on a 6-well plate with or without 0.1 mM Z-VAD-FMK pretreatments for 2 h (Selleckchem.com; Houston, TX, USA). Briefly, cells were added with vehicle or TFB. The cells were then resuspended in the binding buffer containing 5 µg/mL of annexin V-fluorescein isothiocyanate and 50 µg/mL of PI and examined with a BD Accuri C6 flow cytometer (Becton-Dickinson, Mansfield, MA, USA) (excitation: 488 nm and emission: 525 nm and 617 nm for FITC and PI, respectively) and BD Accuri C6 software.

Measurement of Intracellular ROS
2 ,7 -Dichlorodihydrofluorescein diacetate (DCFH-DA) (Sigma Chemical Co., St. Louis, MO, USA) was used to detect intracellular ROS as previously described [33]. Cells were plated at 3 × 10 5 cells/2 mL cell culture medium on a 6 cm dish. Briefly, cells were added with vehicle or TFB with or without 2 mM NAC pretreatment for 1 h (Sigma; St. Louis, MO, USA). After the collection and PBS washing, cells were treated with 0.1 µM DCFH-DA in serum-free medium for 30 min at 37 • C in darkness. Cells were resuspended in PBS after centrifugation and examined with a BD Accuri C6 flow cytometer (excitation: 488 nm and emission: 525 nm) and BD Accuri C6 software.

Measurement of MitoMP
A MitoProbe™ DiOC 2 (3) assay kit (Invitrogen, San Diego, CA, USA) was used to detect mitochondrial membrane potential (MitoMP) as described previously [34]. Cells were plated at 3 × 10 5 cells/2 mL cell culture medium on a 6-well plate. Briefly, cells were added with vehicle or with TFB. The TFB-treated cells were washed in 1 mL of PBS/well, provided with 1 mL of medium/well, loaded with 10 µL of 10 µM DiOC 2 (3), and left to stand at 37 • C in 5% CO 2 for 20-30 min. After harvesting and washing, cells were resuspended in PBS and examined immediately using a FACSCalibur flow cytometer (excitation: 488 nm and emission: 525 nm) and BD Accuri C6 software.

Measurement of DNA Damage by γH2AX Expression
DNA double strand breaks were detected by both flow cytometry [48] and Western blotting [35] as described previously. For flow cytometry, TFB-treated cells were fixed, washed, and incubated at 4 • C for 1 h in 2 µg/mL of p-Histone H2AX (Ser 139) (γH2AX) monoclonal antibody (sc-101696; Santa Cruz Biotechnology, Santa Cruz, CA, USA). After washing, cells were suspended for 1 h in a secondary antibody (Jackson Laboratory, Bar Harbor, ME, USA) for 30 min at room temperature. Finally, the cells were resuspended in 20 µg/mL of PI for flow cytometry analysis (BD Accuri™ C6; Becton-Dickinson).
For Western blotting of γH2AX expression, 30 µg protein lysates were resolved in 10% SDS-PAGE. Except when p-Histone H2AX (Santa Cruz Biotechnology) was chosen for the primary antibody, procedures were the same as those employing Western blotting for apoptosis proteins, mentioned above.

Statistical Analysis
All data are shown as mean ± SD. The significant differences between test and control were analyzed with a Student t-test.