In order to investigate the effect of cedrol on growth of human cancer cells, we treated a panel of human cancer cell lines and assayed the cell viability. Cedrol suppressed the growth of each cell line in a dose-dependent manner (Supplementary Table S1). Human leukemia cell K562 and colon cancer cell HT29 were found to be most sensitive among them with growth inhibitory (GI50) concentrations 179.5 and 185.5 μM, respectively. Cedrol drastically suppressed the growth of K562 and HT29 cells at 100 and 200 μM concentrations (Figure 1A). Next, we examined the apoptosis-inducing activity of cedrol in K562 cells by annexin V/7-7-amino-actinomycin (AAD) double staining and analyzing DNA contents by flow cytometry. Annexin V is a calcium-dependent phospholipid-binding protein with high-affinity phosphatidylserine, which makes it useful for analyzing apoptotic cells that are exposed with phosphatidylserine, while 7-AAD is a standard probe for cell viability that distinguishes between viable from non-viable cells, thus preferably used in combination with annexin V. Cedrol treatment caused high annexin V staining of cells, considered as apoptotic cells (Figure 1B). Results also show that cedrol partially induced necrosis in K562 cells by 2% (100 μM) and 5.1% (200 μM) (Figure 1B). The pro-apoptotic activity of cedrol in K562 cells was also evident from an early apoptosis marker, DNA strand fragmentation. The band pattern on agarose gel indicates the remarkable increase in DNA fragmentation (∼200 bp) especially at 200 μM cedrol (Figure 1C). The effect of cedrol on cell cycle progression in K562 cells was also observed (Supplementary Figure S1). Propidium iodide (PI) staining of K562 cells showed increased accumulation of cells in the sub-G1 phase with increasing concentrations of cedrol by 5% (100 μM) and 18.4% (200 μM). Although this result supports the pro-apoptosis effect of cedrol, the effect on cell cycle progression was not clear. Collectively, these results demonstrate that cedrol suppresses growth of human cancer cells by inducing apoptosis.

Next, we performed experiments to study the mechanisms by which cedrol induced cell death. Firstly, we examined the effect of cedrol on proteins in apoptosis cascade in the presence and absence of z-VAD-fmk (pan-caspase inhibitor). Treatment of K562 and HT29 cells by cedrol showed activation of caspase-3 in a dose-dependent manner. Addition of z-VAD-fmk to cells inhibited the caspase-3-activating effect of cedrol (Figure 2A). We also assessed the viability of cells treated with cedrol in combination with z-VAD-fmk. Addition of z-VAD-fmk restored the cedrol-mediated suppression of cell viability in K562 and HT29 cells but not completely (Figure 2B). The reason behind this observation could be partial necrosis induced by cedrol especially at 200 μM, as shown in Figure 1B. These results suggest that cedrol suppresses growth of cancer cells through induction of apoptosis as a mechanism, but other growth regulatory pathways may also be associated. Secondly, we determined the apoptosis pathway (intrinsic or extrinsic) activated by cedrol. As seen in Figure 2C, the level of BID (BH3-interacting-domain death agonist) was dramatically suppressed by cedrol. BID is a pro-apoptotic member of the Bcl-2 protein family which is known to initiate the mitochondrial intrinsic pathway of apoptosis. Cedrol further caused the activation of caspase-9 and subsequently caspase-3 in a dose-dependent manner. Caspase-3 then executes cell death by cleavage of key nuclear substrate PARP (poly ADP ribose polymerase) (Figure 2C). These results confirm that cedrol induced mitochondrial intrinsic apoptosis in K562 and HT29 cells. Studies using sesquiterpenes like resveratrol, costunolide, and tehranolide have shown anticancer activities through the Fas/CD95-caspase-8-dependent extrinsic pathway of apoptosis (Reis-Sobreiro et al., 2009; Choi et al., 2012; Noori and Hassan, 2012). Resveratrol induced a co-clustering of the raft protein contents and death-induced silencing complex (DISC) constituents: Fas/CD95, FADD (Fas-associated protein with death domain), and procaspase-8 (Reis-Sobreiro et al., 2009). Unlike other sesquiterpenes, cedrol inhibited the levels of FADD in K562 and HT29 cells, which rules out its possibility to be recruited in the DISC. However, pro-caspase-8 was inhibited by cedrol especially at a higher concentration (200 μM) which may be a non-specific activation, which often occurs in the intrinsic apoptosis pathway.

Although cedrol treatment could sensitize K562 and HT29 cells to apoptosis and partial necrosis, the pan-caspase inhibitor (z-VAD-fmk) could not fully abrogate the effects of cedrol. Therefore, we attempted to search other mechanistic possibilities associated with growth inhibitory effects of cedrol in K562 and HT29 cell lines. Analysis of proteins in cell survival signaling pathways shows that cedrol inhibited the levels of pERK 1/2 and pAkt and pmTOR proteins in a dose-dependent manner (Figure 3A). Cedrol further inhibited the levels of anti-apoptosis proteins of the Bcl-2 family (Bcl-2 and Bcl-X_L) as well as XIAP. It is also noteworthy that cedrol remarkably inhibited the levels of cyclin D1 and c-Myc in both cell lines (Figure 3A). The expression of Bcl-2 family proteins is under the transcriptional control of both the ERK 1/2 and NF-κB signaling pathways. Thus, we examined the levels of NF-κB protein in K562 and HT29 cells treated with cedrol. We found that cedrol inhibited the levels of the p65 subunit of NF-κB protein in cytoplasmic as well as nuclear lysates in a dose-dependent manner (Figure 3B). The effect of cedrol on downregulation of NF-κB was also confirmed by analyzing the nuclear translocation of NF-κB proteins. Cedrol effectively inhibited the nuclear levels of p65 protein as observed by a dramatic reduction at 200 μM. These findings suggest that cedrol interferes with PI3K/AKT/ERK and NF-kB signaling as additional mechanism in growth inhibition of cancer cells.

In order to explore the mechanism underlying the effects of cedrol, we assessed the effect on the organization of cell membranes. K562 cells were treated with Bodipy-labeled sphingomyelin and cholesterol followed by confocal fluorescence microscopy imaging (Figure 4A). The control-treated cells displayed a preferential localization of both cholesterol and sphingomyelin in certain areas of the plasma membrane. Cedrol treatment (200 μM) caused a redistribution of both lipid contents across the membrane surface. These observations suggest that cedrol alters the distribution of cholesterol and sphingomyelin contents in the plasma membrane. Since cholesterol and sphingomyelin are key components in the membrane raft domains (Edidin, 2001; van Deurs et al., 2003), these alterations by cedrol could be associated with the integrity of lipid raft domains of the membrane. To test this possibility, we examined the effect of cedrol in combination with a known lipid raft-disrupting agent MβCD (2.5 mg/ml). Addition of MβCD caused about 14% more cell death in K562 cells as compared to cedrol alone (Figure 4B). In HT29 cells, MβCD caused 16.6% and 15% more cell death as compared to cedrol alone at 100 and 200 μM, respectively (Figure 4C). These results showed the additive combination effects of cedrol and MβCD on growth inhibition. We also examined the levels of proteins in apoptosis cascade and AKT/ERK signaling when treated in combination with cedrol and MβCD. The combination of cedrol and MβCD showed a more exaggerated response as compared to cedrol alone. Apoptosis-related caspases (9 and 3) and cell survival-related proteins (pAkt and pErk) were highly inhibited in combination as compared to cedrol alone (Figure 4C). In addition to the lipid raft disruption approach, we also utilized the cholesterol replenishing approach using the cholesterol:MβCD complex prepared by the previously described method (Christian et al., 1997). Our results show that treatment of cells with the cholesterol complex (10 μg/ml) dramatically attenuated the effects of cedrol (Figure 4C). K562 and HT29 cells treated with the cholesterol complex showed lesser inhibition of pro-caspases (9 and 3) and pAkt and pErk as compared to cedrol alone (Figure 4C). It is thus evident that lipid raft disruption exaggerates, and cholesterol replenishment abrogates, the effects of cedrol. These results indicate two clear possibilities that effects of cedrol are highly dependent on the integrity of membrane lipid raft and that apoptosis induction and suppression of pro-survival signaling act as parallel mechanisms.

Nox2 is a membrane-bound member of the NADPH oxidase enzyme complex, an important source of non-mitochondrial intracellular ROS generation (Leto et al., 2009). We utilized DCFH-DA assay in combination with phorbol 12-myristate 13-acetate (PMA), specific for detection of Nox2 activity-dependent ROS generation (Figure 5). PMA, a diester of phorbol, functions as a tumor promoter via activation of PKC signaling. PMA is a known and specific activator of Nox2, which stimulates cells for ROS generation through phosphorylation of p47phox sites of Nox2 (Leto et al., 2009; Jeon et al., 2010). Recently, it has been demonstrated that the potential cytosolic tail of Nox2 was phosphorylated during PMA activation by a PKC-dependent mechanism (Raad et al., 2009). Treatment of K562 cells with cedrol (200 μM) significantly inhibited the generation of ROS to 68% (P = 0.033) as compared to vehicle control (100%). PMA-treated K562 cells showed high fluorescence intensity (169%, P = 0.0004), indicating significantly enhanced production of ROS. Treatment of PMA-sensitized cells with cedrol inhibited the relative superoxide generation to near the level of control (111%). This inhibition of ROS generation was highly significant as compared to PMA-treated cells (P = 0.0081) but not significant to vehicle control (P = 0.20). Cedrol showed a reduction in ROS even without treatment with Nox2 activator PMA, yet with lesser statistical significance (P = 0.033). However, in K562 cells treated with Nox2 activator PMA, the reduction in ROS was statistically highly significant (P = 0.0081). This assay demonstrates that inhibition of ROS generation by cedrol is possibly mediated by the inhibition of Nox2 activity. Since Nox enzyme complexes are located in the lipid raft-rich region of the plasma membrane (Leto et al., 2009), the effect of cedrol appears to be dependent on the lipid raft-destabilizing property. Considering that high levels of ROS are functional features of cancer cell growth promotion, these properties of cedrol are likely to be associated with growth inhibition of cancer cells.

Ceramide production has been found in disrupted plasma membrane and then exerted a pro-apoptotic response in cancer cells (Ogretmen and Hannun, 2004). We further analyze the lipid raft-dependent pro-apoptotic mechanism of cedrol by quantifying the intracellular ceramide species in K562 cells by LC-MS/MS (Figure 6). Incubation of K562 cells with cedrol for 18 h led to a dose-dependent increase in intracellular accumulation of ceramide species (C16, C18, C24:1, and C24:0). Ceramide can be produced via multiple mechanisms including through the hydrolysis of sphingomyelin by acid and neutral sphingomyelinase or by a de novo synthesis through a pathway involving the ceramide synthase (Ogretmen and Hannun, 2004). Ceramide was reported to act as tumor-suppressor lipid through activation of apoptosis and suppression of cell survival (Pettus et al., 2002; Huang et al., 2011), especially limiting the PI3K signaling (Kitatani et al., 2016). Ceramide was also shown to affect the stress-induced ceramide on apoptosis through PI3K downregulation. Ceramide induction directly downregulated the levels of PI3K, which resulted in inhibition of Akt phosphorylation (Zundel and Giaccia, 1998). This suggested that ceramide levels could act as a general apoptotic rheostat and control cell survival through PI3K and Akt. This study showed similar observations with the treatment of cancer cells with cedrol, which caused the induction of the ceramide product and resulted in apoptosis activation and suppression of cell survival. These observations indicate that ceramide biosynthesis acts as a plausible mechanism for growth inhibition and that it was apparently dependent on membrane lipid raft destabilization.

K562 and HT29 cell lines purchased from the American Type Culture Collection (ATCC) were maintained in RPMI 1640 medium (Gibco Invitrogen, Grand Island, NY, United States) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, United States), 2 mM L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin (Gibco Invitrogen) at 37°C in humidified atmosphere of 5% CO₂. Cedrol was purchased from Sigma Chemicals.

K562 and HT29 cells (4000 cell/well) were seeded in 96-well plates and treated with cedrol for 48 h. Cell viability in K562 cells was assessed by using Ez-Cytox WST assay kit (Daeil Lab, Seoul, South Korea) according to the manufacturer’s instructions. MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Biosesang Inc., South Korea) was performed in HT29 cells, as previously described (Mishra et al., 2013a, b). Optical densities of WST and MTT assay products were measured at 460 and 570 nm using a microplate reader (Epoch Micro-Volume Spectrophotometer System, United States).

Apoptosis induction was analyzed by annexin V/7-amino-actinomycin (7-AAD) double staining using Annexin V-FITC Apoptosis Detection Kit II (BD Bioscience, San Jose, CA, United States) according to the manufacturer’s instructions and as described earlier (Mishra et al., 2013a, b). K562 cells (2 × 10⁶/ml) were treated with cedrol for 18 h followed by annexin V/7-AAD staining as per standard protocol and then flow cytometry analysis of apoptosis using FACSCalibur (BD Bioscience, San Jose, CA, United States).

DNA fragmentation was assessed by electrophoresis of genomic DNA extracted from K562 cells as described previously (Muller et al., 1996). Briefly, K562 cells (2 × 10⁶/ml) treated with cedrol for 18 h were harvested and homogenized with 50 μl of lysis buffer (50 mM Tris–HCl, 1% NP-40, 20 mM EDTA). The supernatant collected from whole-cell lysate was treated with 5 μg RNase A and 1% SDS for RNA digestion at 56°C for 2 h, subsequently with 2.5 μg/μl proteinase K for protein digestion at 37°C for 2 h. DNA was precipitated using 10 M ammonium acetate and ethanol, dissolved in gel loading buffer, and electrophoresed on 1% agarose gel with ethidium bromide staining.

K562 and HT29 cells (2 × 10⁵/ml) were treated with cedrol for 18 h. Preparation of whole-cell lysates, protein quantification, gel electrophoresis, and Western blotting were performed as previously described (Mishra et al., 2013a, b). Nuclear protein lysates were prepared by homogenizing cells in lysis buffer mixed with Igepal CA-360 solution (0.6%). The cell homogenate was centrifuged, and the cytoplasmic fraction was collected as supernatant. The cell pellet was further treated with extraction buffer and centrifuged to collect the supernatant containing the nuclear fraction. Equivalent amounts of protein from cell and nuclear lysates were resolved by SDS-PAGE and immunoblotted with primary antibodies. GAPDH was detected as loading control. Bands were detected using ECL Western blotting detection reagents (AbFrontier, Seoul, South Korea).

Fluorescent Bodipy-labeled analogs of sphingolipids and cholesterol are specifically utilized to study the lipid microdomains of the cell membrane. Bodipy-labeled sphingomyelin and cholesterol can visualize the membrane lipid microdomains in living cells based on their fluorescence properties and help in examining the relationship between membrane domains at the cell surface (Marks et al., 2008). K562 cells were treated with vehicle control or cedrol for 20 h and collected by centrifugation at 2000 rpm for 3 min at 4°C. The pellet was resuspended in serum-free medium and then incubated with 20 μM of each Bodipy-labeled sphingomyelin and cholesterol (Invitrogen, United States) in serum-free medium for 30 min at 4°C. Cells were then washed with cold serum-free medium and observed under a confocal microscope.

Non-mitochondrial production of ROS was analyzed to assess the effect of cedrol on the plasma membrane-dependent enzyme activity of Nox2. K562 cells (2 × 10⁶/ml) were treated with cedrol for 18 h and stimulated with 0.3 μM phorbol-12-myristate 13-acetate (PMA) to induce generation of ROS. Five minutes later, cells were harvested and incubated with 10 mM of redox-sensitive dye 2’,7’-dichlorodihydrofluorescein diacetate (DCFH-DA) (Molecular Probes, Invitrogen, United States) for 30 min in the dark at 37°C. DCFHA fluorescence was measured by flow cytometry using FACSCalibur (BD Bioscience, San Jose, CA, United States).

Ceramide quantification was performed by LC-MS/MS technique on a Shimadzu Prominence UPLC system (Shimadzu, Kyoto, Japan) equipped with a thermostatted autosampler, binary pump, degasser, and thermostatted column compartment. The chromatographic separation was performed on an XTerra MS C18 column (2.1 × 50 mm, 5 μm particles; Waters Corporation, Milford, MA, United States) with a SecurityGuard C18 guard column (4 mm × 20 mm I.D.; Phenomenex, United States). MS analysis was performed on an Applied Biosystems API2000 triple-quadrupole mass spectrometer by multiple reaction monitoring (MRM) in positive and negative electrospray ionization mode. Quantification of each compound was achieved by comparison of the analyte/internal standard peak area ratios using C16, C18, C20, C24:1, and C24:0 ceramide mixtures (0–2000 pmol) as calibration standards. Acquisition and analysis of data were performed with Analyst^TM software (version 1.4.2, Applied Biosystems, Foster City, CA, United States). K562 cells treated with either vehicle control or cedrol were harvested and total 200 μg protein cell lysates were used for ceramide species quantification.

All values are expressed as mean with standard deviations (SD). Data were analyzed using SPSS v17 and Systat SigmaPlot 10.0 statistical software. Statistical significance was calculated by Student’s t-test and one-way analysis of variance (ANOVA) for multiple comparisons. P < 0.05 was considered statistically significant.