The human NSCLC cell line NCI-H460 was purchased from the American Type Culture Collection, ATCC (Manassas, VA, USA) and was cultured as monolayers in Roswell Park Memorial Institute-1640 medium (RPMI) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine and 100 U/ml penicillin and streptomycin. Cell cultures were incubated at 37°C in a humidified incubator filled with 5% CO₂. Cells were subcultured routinely with 0.25% (w/v) trypsin in 0.53 mM ethylenediamine tetraacetic acid (EDTA) and were seeded according to the manufacturer's instructions at ~70% confluence.

Vanillin (Fig. 1), as USP-secondary standard purity grade (>99.0% purity) was purchased from Sigma-Aldrich (St. Louis, MO, USA). The vanillin stock solution was freshly diluted with RPMI medium, and the treatment solution was prepared by dilution of the stock solution with culture media to the desired concentrations. Cell culture media were also used as the control solvent for the experiments. Glutamate, penicillin-streptomycin antibiotics, and phosphate-buffered saline (PBS) were purchased from Gibco (Grand Island, NY, USA). Methanol, dimethyl sulfoxide (DMSO) and mouse monoclonal antibodies to Oct4 (1:1,000) and ubiquitin (1:1,000) were purchased from Sigma-Aldrich. Rabbit monoclonal antibodies to ABCG2 (1:1,000), β-catenin (1:1,000), Nanog (1:1,000) and phosphorylated Akt (P-Akt, (1:1,000) in Ser473 and Thr308 were from Cell Signaling Technology, Inc. (MA, USA). A goat monoclonal antibody against ALDH1A1 (1:1,000) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and rabbit monoclonal antibody against CD133 (1:1,000) was purchased from United States Biological (MA, USA).

Cell viability and proliferation assays were performed using the surrogate 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay. Briefly, NCI-H460 cells were seeded onto 96-well plates at 10,000 cells/well and 2,500 cells/well for the cell viability and proliferation assays, respectively. After cell attachment, various concentrations of vanillin (0–400 µM) were added for 24 h for the cell viability assay. Cells pretreated with the same concentration of vanillin for 1 and 3 days were subcultured and incubated for 24, 48 and 72 h for the cell proliferation assays. At the end of each incubation time, the medium was removed and replaced with MTT solution (Life Technologies, Carlsbad, CA, USA) and incubated for 3 h at 37°C. The MTT solution was then substituted with 100 µl of DMSO to dissolve the formazan crystals, and then the absorbance was measured at 570 nm using a microplate reader. The absorbance was calculated and represented as the % viability and % relative proliferation, respectively, compared with the untreated (control) cells that were set at 100%.

Hoechst 33342 and propidium iodide (PI) (Sigma-Aldrich) were used to stain the nuclei to identify necrotic and apoptotic cell death, respectively. Cells were treated with vanillin for 24 h and then were washed with PBS, followed by incubation with either 5 µM PI or 10 µM Hoechst 33342 at 37°C for 30 min. Cells were visualized and captured under a fluorescence microscope at ×10 magnification (Olympus IX5; 10X with DP70 digital camera system; Olympus, Tokyo, Japan). PI-positive necrotic cells and nuclear condensation of apoptotic cells were scored and reported as the % of all cells viewed.

The NCI-H460 cells were pretreated with vanillin (0–50 µM) for 1 and 3 days before being subjected to the soft agar colony-formation assay to determine the anchorage-independent growth property of cancer stemness. Soft agar was prepared using a combination of complete media and 1% (w/v) agarose at a 1:1 ratio. This mixture was allowed to solidify in 24-well plates as a bottom layer. Next, pretreated cells suspended in RPMI complete media were mixed with 0.3% (w/v) agarose, added onto the prepared bottom layer and then incubated at 37°C. Further complete media were applied every 2 days to prevent the soft agar drying. After 2 weeks, the colony was examined under a phase-contrast microscope at ×4 magnification (Olympus IX5; 4X with DP70 digital camera system; Olympus). The relative colony number and size were analyzed compared with those of the control group.

Vanillin-pretreated cells at a cell density of 2,500 cells/well were cultured in 24-well ultralow attachment plates in RPMI serum-free medium for 7 days. Next, the spheroids were disaggregated by trypsinization, resuspended in RPMI serum-free medium as a single cell suspension and then cultured onto 24-well ultralow attachment plates. After an additional 14 days, the presence of secondary spheroids was investigated by phase-contrast microscopy at ×4 magnification (Olympus 1X51 with DP70). The relative spheroid number and size were analyzed compared with those of the control group.

Cells were lysed using lysis buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM Na₃VO₄, 10% glycerol, 50 mM NaF, 100 mM phenylmethylsulfonyl fluoride (PMSF; Sigma-Aldrich), and a protease inhibitor cocktail (Merck Millipore, MA, USA), sonicated and incubated on an ice bath for 45 min. The protein content was measured using the BCA protein assay kit (Pierce™ Thermo Fisher Scientific Inc., IL, USA). An equal amount of denatured protein was separated by electrophoresis [10% (w/v) acrylamide resolving gel] and transferred to a 0.45-µm nitrocellulose membrane (Bio-Rad, Hercules, CA, USA). Each membrane was blocked with 5% (w/v) nonfat-milk in Tris-buffered saline solution containing 1% (v/v) Tween-20 (TBST) for at least 30 min and then incubated with the indicated primary antibody at 4°C overnight. Thereafter, the membranes were washed three times with TBST and then incubated with secondary antibody at room temperature for 2 h. The antigen-antibody complexes were detected, after washing with TBST as above, using chemiluminescent solution (Pierce Biotechnology) and were exposed to film (Carestream Health, Inc., Rochester, NY, USA). The densitometry of the target protein was measured using the NIH ImageJ program and quantified as the relative expression level to that of the control group.

Protein interaction was examined by immunoprecipitation. Cells were incubated with 10 µM lactacystin for 1 h prior to treatment with vanillin to inhibit proteasome function while preserving the ubiquitin-protein complex. Treated cells were collected and lysed in TMN buffer (50 mM Tris-HCl pH 7.5, 140 mM NaCl and 0.5 mM MgCl₂) containing 10% glycerol, a protease inhibitor cocktail, 1% nonylphenylpolyethylene glycol (NP-40) and 1% PMSF for 30 min, followed by centrifugation at 12,000 rpm at 4°C for 15 min. The supernatants were blocked with protein G agarose beads (GE Healthcare, Little Chalfont, UK) to remove non-specific binding. After centrifugation at 3,000 rpm at 4°C for 5 min, the protein contents were measured. The supernatants were collected and stored at 4°C as an input for immunoblotting where 300 µg of protein was incubated with anti-Akt antibody at 4°C overnight and then incubated with protein G agarose beads for 2 h at 4°C. The immunoprecipitates were collected, washed with TMN buffer and resuspended in 30 µl of 2X SDS sample buffer. The samples were boiled at 95°C for 5 min and then were subjected to western blot analysis using antibodies against ubiquitin.

The data were collected from at least four independent experiments, normalized by the control groups and presented as the means ± standard deviation (SD). The significant difference among the groups was analyzed using one-way ANOVA followed by the post hoc test. Statistical analysis was performed using SPSS software (IBM Inc., NY, USA) and statistical significance was accepted at the p<0.05 level.

To diminish the interference of the cytotoxic effect of vanillin on cancer stemness, a non-toxic concentration of vanillin was used on the NCI-H460 cells, as determined by the cytotoxicity assay. Cells were treated with various concentrations of vanillin for 24 h and were then subjected to the MTT cell viability assay. Fig. 2A demonstrates that 200 µM vanillin caused a significant (p=0.003) mortality with ~30% cell death, whereas at a concentration of <100 µM at least 80% viable cells remained. To confirm the cell death mechanism, cells were similarly treated with vanillin for 24 h and then were incubated with either Hoechst 33342 or PI, where apoptotic nuclei and necrotic bodies were clearly observed in the cells treated with a high (200 µM) dose of vanillin (Fig. 2B and C). This result is consistent with the cell viability testing, revealing that vanillin at concentrations of <100 µM are non-toxic to NCI-H460 cells, so these doses were used for subsequent experiments.

The effect of vanillin on NCI-H460 cell proliferation was also investigated. The cells were pre-treated with vanillin for 1 and 3 days, prior to measuring their proliferation after a further 24, 48 and 72 h. The proliferation of the cells pretreated with 100 µM vanillin for 3 days was significantly reduced as early as 24 h, whereas lower doses (<100 µM) had no significant effect at any detection time (Fig. 2E). Furthermore, pretreatment with vanillin for 1 day showed non-detectable changes compared with that in the control group (Fig. 2D). These results indicated that a low dose of vanillin (0–50 µM) neither induced cell death nor inhibited cell proliferation in the NCI-H460 cells. Cell growth and survival in the absence of extracellular matrix is one of the crucial behaviors of the cancer stem-like phenotype. Thus, to avoid any anti-proliferative effect of vanillin on cancer stem cell growth, vanillin doses that showed no such effect on NCI-H460 cells under the attachment condition were then used for the cancer stem-like phenotype characterization.

Because self-renewal and tumorigenicity are distinctive behaviors for the cancer stem-like phenotype, anchorage-independent growth and spheroid formation have been used to characterize these features. To investigate the suppressive effect of vanillin on colony and spheroid formation, NCI-H460 cells were pretreated with vanillin (0–50 µM) for 1 and 3 days, followed by examination of their anchorage-independent growth and spheroid formation. Fig. 3A shows representative images in which both the colony size and number of cells pretreated with vanillin for 1 day were not different from those of untreated control cells. However, a significant and vanillin dose-dependent reduction in the cell size and number was observed when pre-incubated with vanillin for 3 days (Fig. 3B and C). Approximately 80 and 90% reduction in the colony size and number, respectively, were observed in the NCI-H460 cells pretreated with 50 µM vanillin for 3 days (p=0.027).

The effect of vanillin on the self-renewal property was evaluated by the spheroid formation assay. Primary spheroids, consisting of several types of cells, including progenitors, mature cells and stem cells (29), were first initiated for 7 days and then dissociated to single cells. The single cells were allowed to grow under detachment conditions to form secondary spheroids over 14 days, reflecting the pluripotency of CSCs. Interestingly, pretreatment with vanillin for 3 days markedly attenuated the number of secondary spheres formed under detachment conditions, with only ~0.2-fold of the number of sphere remaining after pretreatment with 50 µM vanillin for 3 days (p=0.022). In contrast, this effect was not found with a 1-day pretreatment with vanillin (Fig. 3D–F). These results suggest that vanillin can abolish the cancer stem-like phenotype of CSCs, in terms of both spheroid formation and anchorage-independent growth.

Cancer stem cell markers are differentially expressed in different types of cancer, with CD133, ABCG2 and ALDH1A1 being distinctly expressed in lung cancer stem cells (3). Given the inhibitory effect of vanillin on the CSC phenotype, its effect on stemness cell markers was then evaluated in the NCI-H460 cells. NCI-H460 cells were incubated with vanillin (0–50 µM) for 1 and 3 days, and then the expression of CSCs markers was determined. Based on western blot analysis (Fig. 4), CSC protein marker expression was not significantly changed in response to the treatment with vanillin for 1 day (50 µM, p=0.124). However, the level of CSC markers, including those of CD133, ABCG2 and ALDH1A1, were markedly attenuated in a dose-dependent manner after treatment with vanillin for 3 days (Fig. 4B). Quantitative analysis demonstrated that a reduction in the expression level of these proteins was observed in cells treated with vanillin for 3 days at concentrations as low as 10 µM (p=0.021). Thus, vanillin suppresses not only the CSC phenotype but also CSC markers in NCI-H460 cells.

The CSC phenotypes include not only an elevated expression of stemness markers, but also high levels of the related transcription factors, including Nanog and Oct4 (10–12) that endow a self-renewal property to the cells (30). To assess whether the negative regulation of CSCs properties by vanillin involved these transcription factors, NCI-H460 cells were treated with various concentrations of vanillin for 1 and 3 days, and then protein expression levels of Oct4 and Nanog were evaluated. As shown in Fig. 4B, the Oct4 and Nanog protein expression levels were significantly decreased in the cells treated with vanillin for 3 days, but not in the cells treated for 1 day (Fig. 4A). These results illustrate a good correlation with the other stemness phenotypes (Fig. 3 and 4), demonstrated by their suppression after treatment with vanillin for 3 days.

Because the expression levels of Nanog and Oct4, which contribute to the stemness behavior, are regulated directly by Akt (15), whether vanillin modulates Akt protein expression levels or its activity in terms of P-Akt expression was evaluated next. NCI-H460 cells were treated with vanillin for 3 days prior to determination of the Akt and P-Akt expression levels. Western blot analysis showed that the level of P-Akt (Ser473 and Thr308) or active Akt, was decreased (Fig. 5A), suggesting that the reduced CSCs phenotypes may result from an Akt-dependent mechanism. To verify whether the reduction of P-Akt by vanillin is a result of the suppression of P-Akt or total Akt expression levels, the total Akt expression level was also investigated. As shown in Fig. 5A, the total Akt level was clearly downregulated in a similar pattern. Densitometry analysis of P-Akt relative to total Akt was then analyzed to confirm the hypothesis that vanillin interferes with the Akt expression level. The quantitative results demonstrated that the ratio of P-Akt over its parental form are also changed, indicating the reduction of Akt expression may be, at least in part, a mechanism by which vanillin suppresses Akt activity.

To confirm the above finding that the downregulation of Nanog and Oct4 by vanillin is by an Akt-dependent pathway, perifosine (1,1-dimethylpiperidinium-4-yl octadecyl phosphate), an Akt inhibitor, was used as a positive control. H460 cells were treated with perifosine (0–5 µM) for 3 days, and the mentioned proteins were observed by western blotting as before. Interestingly, inactivation of Akt by perifosine caused a reduction of its downstream signaling, including Oct4 and Nanog (Fig. 5B), similar to that with vanillin (Fig. 4B). The CSC markers CD133, ABCG2 and ALDH1A1 were also downregulated corresponding to the suppression of Akt activity. Although the vanillin-blocked Akt activation occurs partly through downregulating the parental form, the total effect is similar to that by perifosine. These data indicated the Akt function that promotes cancer stemness property via Oct4 and Nanog, which was conversely inhibited by vanillin.

Emerging evidence has indicated that the existence of Akt is mainly governed by ubiquitin-proteasomal degradation via E3 ligase (31,32). The reduction of Akt by the enhancement of this mechanism results in the interference of its downstream signaling mediators, including Oct4 and Nanog (11,15,16,33,34). Accordingly, we next examined whether downregulation of Akt by vanillin occurred through the ubiquitin proteasomal pathway. Protein degradation by this pathway requires the protein to bind with ubiquitin prior to recognition by the proteasome (35). NCI-H460 cells were pretreated with 10 µM lactacystin (Lac), a proteasomal inhibitor, prior to vanillin treatment (50 µM) for 3 h, and the presence of Akt-ubiquitin complex was determined by immunoprecipitation. Fig. 5C shows that the level of the Akt-ubiquitin interaction was markedly increased ~2-fold in the vanillin-treated group, which indicated that suppression of CSCs by vanillin may be a consequence of vanillin promoting Akt degradation through the ubiquitin-proteasomal pathway.