Exogenous C8-Ceramide Induces Apoptosis by Overproduction of ROS and the Switch of Superoxide Dismutases SOD1 to SOD2 in Human Lung Cancer Cells

Ceramides, abundant sphingolipids on the cell membrane, can act as signaling molecules to regulate cellular functions including cell viability. Exogenous ceramide has been shown to exert potent anti-proliferative effects against cancer cells, but little is known about how it affects reactive oxygen species (ROS) in lung cancer cells. In this study, we investigated the effect of N-octanoyl-D-erythro-sphingosine (C8-ceramide) on human non-small-cell lung cancer H1299 cells. Flow cytometry-based assays indicated that C8-ceramide increased the level of endogenous ROS in H1299 cells. Interestingly, the ratio of superoxide dismutases (SODs) SOD1 and SOD2 seem to be regulated by C8-ceramide treatment. Furthermore, the accumulation of cell cycle G1 phase and apoptotic populations in C8-ceramide-treated H1299 cells was observed. The results of the Western blot showed that C8-ceramide causes a dramatically increased protein level of cyclin D1, a critical regulator of cell cycle G1/S transition. These results suggest that C8-ceramide acts as a potent chemotherapeutic agent and may increase the endogenous ROS level by regulating the switch of SOD1 and SOD2, causing the anti-proliferation, and consequently triggering the apoptosis of NSCLC H1299 cells. Accordingly, our works may give a promising strategy for lung cancer treatment in the future.


C 8 -ceramide Exerts the Anti-Proliferation Potential against H1299 Lung Cancer Cells.
H1299 lung cancer cells were treated with increasing concentrations of C 8 -ceramide for 24 h, and proliferation rates were determined using Trypan blue assay. The rate of cellular proliferation of C 8 -ceramide-treated H1299 lung cancer cells decreased in a dose-dependent manner. The 50% inhibitory concentration (IC 50, 24 h) of C 8 -ceramide for H1299 cells was 22.9 µM ( Figure 1B). The quantitative analysis of cell proliferation assay showed the inhibition of growth in a dose-dependent manner (* p < 0.05, ** p < 0.001 for C 8 -ceramide treatment versus respective control).

C 8 -Ceramide May Cause the G1 Arrest of H1299 Cells
The cell cycle profiles of C 8 -ceramide treated H1299 lung cancer cells was examined using flow cytometry. Cells accumulated in the G1 phase when treated with 20 and 30 µM C 8 -ceramide. An increased sub-G1 population was observed at 30 and 50 µM C 8 -ceramide treated cells ( Figure 2). In Figure 3A, the profiles of Annexin V/PI -positive percentages were shown for the treatments with vehicle control (0.5% DMSO) or indicated concentrations (from 10 to 50 µM) of C 8 -ceramide for 48 h respectively. After 48 h of the C 8 -ceramide treatment, the Annexin V-positive percentages of H1299 cells rose in a dose-dependent manner, and the level of cleaved caspase-3 was shown ( Figure 3B,C).

The Detection of Endogenous ROS in C 8 -Ceramide-Treated H1299 Cells
To explore whether C 8 -ceramide affects the endogenous ROS level of H1299 cells, we analyzed ROS generation of C 8 -ceramide-treated H1299 cells using flow cytometer-based 2 ,7 -dichlorofluorescein diacetate (DCFDA) staining. The changes in endogenous ROS level by C 8 -ceramide treatment for 24 h were shown ( Figure 4A). The levels of endogenous ROS were significantly increased in H1299 cells in a dose-dependent manner (* p < 0.05 and ** p < 0.001) following C 8 -ceramide treatment (** p < 0.001) ( Figure 4B). Data presented as mean ± S.D. in triplicate. Asterisks indicated statistically significant differences compared with those of the control (* p < 0.05 and ** p < 0.001 for control versus C 8 -ceramide treatment respectively). (B) The quantitative analysis. Data presented as mean ± S.D. in triplicates. Five µM of camptothecin (CPT) as a positive control. Asterisks indicated statistically significant differences compared with those of the control (** p < 0.001 for C 8 -ceramide treatment versus respective control in 6 and 12 h).

Assessment of Migration in C8-ceramide-treated H1299 cells
To examine whether C 8 -ceramide affects the cellular migration, a critical index of cancer metastasis, the wound healing assay was conducted. Image panel shows the results of wound healing assay and Boyden's transwell assay ( Figure 5). As shown in Figure 5A,B, the results showed the moderately inhibitory effect of C 8 -ceramide on the migration of H1299 cells, whereas the no significant changes were observed when we further assessed the anti-migration effect of C 8 -ceramide, showing that sub-IC 50 dose (below 20 µM) of C 8 -ceramide is ineffective to suppress the invasion of H1299 lung cancer cells ( Figure 5C,D). Therefore, the results suggesting that C 8 -ceramide induces anti-proliferation and apoptosis rather than anti-migration and anti-invasion in NSCLC cancer cells.

The Modulation of SOD1 and SOD2 in C 8 -Ceramide Treated H1299 Cells
The C 8 -ceramide-induced treatment modulated the levels of SOD1 and cyclin D1 in H1299 lung cancer cells on a protein level, which was examined by Western blotting in the present study. Both SOD1 (decreased) and cyclin D1 (increased) levels in C 8 -ceramide-treated H1299 cells were significantly changed at the concentration of 20 and 30 µM ( Figure 6). In contrast, the protein levels of SOD2 were upregulated dramatically ( Figure 6).

Discussion
The modulations of ceramides as the strategy for lung cancer therapies have been reported [12,14,41]. Both exogenous and endogenous ceramides have been reported to play essential roles in the apoptotic death of cancer cells induced by ionizing radiation [42] or chemotherapeutic agents [43,44]. However, little is known regarding the effects of long carbon-chain ceramides [45]. Among the subtype of NSCLC cells, large cell carcinoma has been reported to exhibit higher invasiveness, and it is difficult for treatment [46,47]. Therefore, we first used H1299 cells, a cell line of large cell carcinoma to examine the anti-lung cancer activity of C 8 -ceramide in the study. We also observed the correlations of ROS and SOD expression in lung cancer H1299 cells following exogenous C 8 -ceramide treatment.
The modulation of endogenous ROS are essential for cellular survival and proliferation [48] and increased ROS level could provide cancer cells with advantages of survival and growth [49]. However, excessive oxidative stresses may cause cell death. A previous study showed that superoxide dismutase SOD1 acts as an endogenous ROS scavenger and a potential contributor to the survival of the cancer cell under conditions of high oxidative stress [50]. In the study, our results showed that the C 8 -ceramide treatment causes the decreased expression of SOD1 concomitantly increased the level of ROS stress.
Cyclin D1 has been thought to be oncogenic [51], and cyclin D1 overexpression was frequently observed in cancers. In addition, the upregulated expression of cyclin D1 promotes the proliferation of HT29 human colon cancer cells [52]. Crebanine, an aporphine alkaloid, was reported to exert an anti-cancer activity through down-regulating cyclin D1 expression in lung adenocarcinoma A549 [53]. On the contrary, cyclin D1 overexpression was also reported to be correlated with apoptosis under specific concomitant signals of arrest such as serum starvation or proliferation arrest [38], which may due to the feedback loop between cyclin D1 and tumor suppressors. Recently, Sun's work suggested that cyclin D1 is required for miRNA let-7-induced cancer repression and the cell death [54].
Papa and Manfredi showed that cancer cells have elevated levels of reactive oxygen species (ROS), which are generated by modulating superoxide dismutase (SOD) as an essential antioxidant enzyme [25]. Furthermore, the activity of SOD2 was decreased to 87% of breast cancer cases, whereas the levels of SOD1 and ROS were upregulated. The ratio of SOD2 to SOD1 seems to be critical for maintaining endogenous ROS in cells. Therefore, the switch mechanism of SOD1 to SOD2 may play a significant role in cellular physiology, such as invasion or proliferation of breast cancer cells. Consistently, the decrease of SOD1 and SOD2 ratio was also observed in H1299 cells following C 8 -ceramide treatment, suggesting that the anti-lung cancer effects of C 8 -ceramide may be closely correlated with the mechanism of SOD1 to SOD2 switch.
Recent studies regarding cell apoptosis showed the biological correlation between cyclin D1 and SOD1 in various human diseases, including genetic diseases and cancers [28,31,55]. The cell proliferation and cell cycle are regulated by directly interacting with cyclin D1. Cyclin D1 was up-regulated when it was involved in neurodegenerative processes related to SOD1 [30]. Furthermore, recent evidence strongly suggested the relationship between SOD and cyclin D, for example, a cell cycle signaling that is cyclin D at the neuronal G1-S checkpoint may be critical for the neuronal cell death caused by mutant SOD1 [28].
However, according to the results of those above studies, likewise, our findings suggest that C 8 -ceramide causes a high level of ROS that was contributed by down-regulated SOD1 and up-regulated cyclin D to promote cell cycle G1 arrest, the growth inhibition, and apoptosis of H1299 cells. Our results indicate that a cell cycle signaling changed by the up-regulation of ROS may reconstitute a critical step in the cell death pathway caused by SOD1 and cyclin D1, which was treated by exogenous long carbon-chain C 8 -ceramide.
Importantly, the results of our study showed that C 8 -ceramide induces apoptosis of H1299 cells, indicating the potential of C 8 -ceramide against human lung cancer cells. Menon et al.'s work investigated the regulatory role of SOD2 (MnSOD) and cyclin D1 in N-acetyl-L-cysteine (NAC)-induced G1 phase arrest in mouse fibroblast [56], and their results suggested that NAC upregulates endogenous O 2 pathway and induces G1-arrest through both increasing SOD2 activity and decreasing cyclin D1 [56]. However, our results showed both SOD2 and cyclin D1 are significantly accumulated in C 8 -ceramide-induced G1-arrest, suggesting the cell types and drug treatments cause the different mechanism from Menon's work.
In the study, C 8 -ceramide treatment may cause the switch of SOD1/SOD2 expression and the up-regulation of cyclin D1, which could sensitize NSCLC cells towards proliferation inhibition and apoptosis, and we herein presented a proposed model of the pathway accordingly (Figure 7). This model illustrated that the down-regulation of SOD1 was controlled by ROS, which resulted in negative feedback to SOD1, causing the excessive level of cyclin D1 in the H1299 cells. The up-regulation of ROS level and SOD2 were increased, and while the expression of SOD1 was decreased by the switching mechanism of SOD1/SOD2, which resulted in the up-regulation of cyclin D1 and the arrest of cell cycle G1 phase. Eventually, the cellular apoptosis is initiated, and the effector caspase-3 is activated, causing apoptosis of H1299 cells. The results of our study suggested that the anti-growth potential of exogenous long carbon-chain C 8 -ceramide against human non-small-cell lung cancer cells may occur through the modulation of the ratio of SOD1 and SOD2. Accordingly, the study suggested that C 8 -ceramide could be used for chemoprevention or chemotherapeutics of lung cancer treatment in future applications. Figure 7. A proposed model of C 8 -ceramide-induced apoptosis and anti-proliferation of non-small-cell lung cancer cells by modulating the differential expressions of superoxide dismutases and cell cycle G1 arrest. After C 8 -ceramide treatment, the SOD1 to SOD2 switch stimulated by ROS induces the excess accumulation of cyclin D1, a feedback loop, which in turn causes cell cycle G1 arrest. Eventually, C 8 -ceramide induces the growth arrest and the apoptotic cell death in lung cancer H1299 cells. The upwards arrows and downwards arrows indicate the upregulation and downregulation respectively.

Cell Proliferation Assay
The proliferation rate was determined by a Trypan blue dye exclusion assay as previously described [57,58]. In brief, 1 × 10 5 H1299 cells were seeded onto a 12-well culture plate. The cells were treated with C 8 -ceramide at indicated concentrations for 24 h respectively. Afterward, 0.2% Trypan blue was added to the wells. Finally, the viable cells were counted by Countess ® Automated Cell Counter (Invitrogen, San Diego, CA, USA). The assay was performed in triplicates, and the IC 50 was calculated from the slope and intercept accordingly to the concentrations of C 8 -ceramide between the half-maximal proliferative inhibitions.

Cell Cycle Distribution
Propidium iodide (PI) (Sigma-Aldrich St. Louis, MO, USA) was used for assessing DNA content. In brief, cells were treated with indicated concentrations of C 8 -ceramide for 24 h respectively. After harvest, cells were fixed with 70% ethanol and washed with phosphate-buffered saline (PBS). After centrifugation, cells were incubated with 10 µg/mL RNase A and 10 µg/mL propidium iodide in PBS for 15 min at room temperature in the dark. The distributions of the cell cycle were analyzed using a BD LSRII flow cytometer (Becton Dickinson, San Jose, CA, USA).

Flow Cytometry-based ROS assessment
The changes in endogenous ROS levels were assessed using the redox-sensitive fluorescence indicator 2 ,7 -dichlorofluorescein diacetate (DCFDA) (Sigma-Aldrich, St. Louis, MO, USA). The ROS assessment was described previously [56]. Briefly, cells were treated with or without C 8 -ceramide for 24 h respectively and then harvested and stained with the oxidative dye 100 nM DCFDA in PBS for 30 min at 37 • C. The measurement wavelengths for excitation and emission were 485 and 530 nm, respectively.

Wound Healing Assay
The cells were seeded and grown on a 12-well plate for overnight. Then cells were scratched by a 200 µL tip to generate a wound area. Cells were further incubated with medium containing 8% FBS media at 37 • C for 12 h for reconstructing the wound area. The wound areas were measured by a software TScratch (http://www.cse-lab.ethz.ch) [59].

Boyden's Transwell Assay
The invasion ability was performed by a 12-well plate combined with inserts with polycarbonate filters (8-µm pore size). Briefly, the lower well contained 800 µL of medium containing 10% FBS. One hundred thousand H1299 cells in serum-free medium were seeded onto a transwell insert (Greiner) and were incubated for 16 h. Invaded cells were fixed with 4% paraformaldehyde and then stained with Giemsa (Merck). All stained cells were counted under a microscope (TE2000-U; Nikon, Tokyo, Japan)

Western Blotting Assay
In brief, cells were collected for lysate preparation. Forty µg of protein lysates were loaded and electrophoresed on 10% SDS-polyacrylamide gel and then transferred to nitrocellulose membranes, and then were blocked with 5% nonfat milk. Subsequently, the membranes were reacted with primary antibodies against SOD1 (GeneTex Co.