Induction of Bex genes by curcumin is associated with apoptosis and activation of p53 in N2a neuroblastoma cells

Brain expressed X-linked (Bex) genes are newer group of pro-apoptotic genes. Role of any Bex gene in neuroblastoma and Bex4 and Bex6 in any cancer is completely unknown. Re-expression of all endogenous Bex genes by any nutraceutical is also unknown. Therefore, we investigated the induction of all endogenous Bex genes and associated mechanisms by curcumin using N2a, an aggressive neuroblastoma cell line. Curcumin induced all endogenous Bex genes prior to apoptosis in N2a cells in a dose- and time-dependent manner. Wortmannin (PI-3Kinases inhibitor), SP600125 (JNK inhibitor) and pifithrin-α (p53 inhibitor) abrogated curcumin-mediated induction of Bex genes. Inhibition of curcumin-mediated induction of Bex genes by pifithrin-α also inhibited N2a cells apoptosis suggesting, a direct role of Bex genes in N2a cells apoptosis and involvement of p53 in Bex genes induction. Curcumin treatment activated p53 through hyperphosphorylation at serine 15 before Bex genes induction indicating Bex genes are novel downstream targets of p53. Collectively, curcumin, a safe nutraceutical has the potential to induce all endogenous Bex genes to harness their anti-cancer properties in neuroblastoma cells. Re-expression of Bex genes by curcumin acts as tumor suppressors and may provide alternate strategy to treat neuroblastomas and other cancers with silenced Bex genes.

. Curcumin induces apoptosis in N2a neuroblastoma cells. (a) Approximately, 80% N2a neuroblastoma cells were serum starved for 2 hours, treated with 50 μ M of curcumin in serum free media for 24 hours and bright field imaging was done from three random fields. DMSO treated cells were used as controls. Images show increased cell deaths with membrane blebbing (yellow arrow) and nuclear condensation (red arrow) in curcumin treated N2a cells than controls. (b) MTT assays (see materials and methods) were performed on N2a cells treated with 10, 25, 50 μ M of curcumin or with DMSO as controls for 24 hours. Amount Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420 any neuroblastoma cells. It is highly impractical to re-express all Bex genes employing gene therapy in a variety of cancers and in various tissues simultaneously. Therefore, manipulating tumor cells genome by nutraceutical/s or pharmaceutical/s to re-express Bex genes can be of great importance in controlling cancer cells growth and death. Until today, there is no report on utilization of any small molecule or phytochemical to induce all the endogenous Bex genes.
Curcumin (diferuloylmethane), the principal curcuminoid of turmeric (Curcuma longa), has a wide range of applications in several biological processes including induction of apoptosis in various cancer cell types 16,17 . Curcumin has been known to kill N2a neuroblastoma cells through apoptosis 18 . However, use of curcumin to induce Bex genes is not reported. Therefore, we hypothesized that curcumin-mediated neuroblastoma cell death might induce Bex genes.
In the present study, induction of all endogenous Bex genes was explored using curcumin-mediated apoptosis in N2a neuroblastoma cells. Cell signaling inhibitors were employed to investigate the possible molecular mechanisms behind curcumin-mediated induction of Bex genes and to associate the expression of Bex genes with apoptotic neuroblastoma cells death. Collectively, our studies for the first time suggest that all the Bex genes can be induced specifically by curcumin to harness their tumor suppressor functions by inhibiting cell proliferation and activating apoptotic factors in N2a neuroblastoma cells.

Curcumin induces apoptosis in N2a neuroblastoma cells in a dose-dependent manner.
Bright field images exhibit curcumin-mediated N2a cells death. These images show membrane blebbing (yellow arrow) and nuclear condensation (red arrow) only in curcumin treated cells, which are commonly seen in apoptotic cell death (Fig. 1a). Results from MTT assay show, curcumin inhibited cell proliferation significantly in a dose-dependent manner, p < 0.001, n = 3 and H = 31.75 (Fig. 1b). Fluorescent images from LIVE/DEAD assays show a dose-dependent increased dead cells population in curcumin treated cells (Fig. 1c). These images also show increased membrane blebbing and fragmented nuclei in curcumin treated cells than control cells. Cell scoring analysis indicated 5 ± 2.3%, 11.4 ± 5% and 100% cell deaths in N2a cells treated with 10, 25 and 50 μ M of curcumin respectively and the difference between the treatments is highly significant, p < 0.001, n = 4, H = 43.49 (Fig. 1d). DNA fragmentation assay demonstrated increased DNA fragmentation in curcumin treated N2a cells than controls with maximum at 50 μ M (Fig. 1e). Densitometric analysis demonstrated approximately 1.8 ± 0.31fold increased DNA fragmentation (p < 0.05) in cells treated with 50 μ M of curcumin than control cells (Fig. 1f) suggesting curcumin induces large-scale genetic lesions during N2a cells death. TUNEL assay, a widely used method to detect apoptosis was also performed. Fluorescent images from TUNEL assay show dose-dependent increased TUNEL positive cells in curcumin treated N2a cells than controls (Fig. 1g). Cell count analysis indicated approximately 3.9 ± 4.2, 3.9 ± 2.3, 18.06 ± 8.56 and 80 ± 6.07% TUNEL positive cells in control, 10, 25 and 50 μ M of curcumin treated N2a cells respectively (Fig. 1h). These data suggest that curcumin-mediated N2a cells death involves apoptosis. To avoid 100% cell death in 50 μ M of curcumin treatment, 25 μ M of curcumin treatment was used in subsequent experiments.
Curcumin inhibits ERK1/2 activation but activates intrinsic apoptotic pathway in N2a cells. Activation of extracellular signal-regulated kinases (ERK1/2) through hyperphosphorylation is a key feature in almost all cancers 19 . Hence, inactivation of ERK1/2 along with activation of apoptotic factors is necessary in curcumin-mediated neuroblastoma cell death. Therefore, the phosphorylation status of ERK1/2 was studied and western blot analysis of ERK1/2 phosphorylation exhibited decreased ERK1/2 phosphorylation at all the time points studied with maximum at 30 minutes of curcumin treatment (Fig. 2a). Densitometric analysis of phosphorylated ERK1/2 normalized with ERK1/2 demonstrated approximately 90.1 ± 5.1%, 88.4 ± 11.1%, 68.3 ± 20.3% and 35.5 ± 24.8% decrease in ERK1/2 phosphorylation at 15, 30, 60 and 120 minutes of curcumin treatment respectively than controls (Fig. 2b), suggesting curcumin rapidly inhibits mitotic cell signals and prepares the cells towards apoptotic cell death. In order to confirm apoptotic cell deaths in curcumin treated N2a cells, activation of caspases were studied. Proteolytic activation of caspases are key features of apoptosis 20 . Western blot analysis indicated increased proteolytic activation of caspase-9 but not caspase-8 in curcumin treated cells than of formazan produced was calculated and plotted as histograms, which indicate curcumin reduces N2a cell viability in a dose-dependent manner. (c) N2a cells used for LIVE/DEAD assay were cultured and treated as mentioned above. After 24 hours, fluorescent images were captured using 40X objective lens from three random fields at same settings. Images clearly exhibit decreased live cells (green) but increased dead cells (red) with membrane blebbing (yellow arrow) and chromatin fragmentation (red arrow) in a dose-dependent manner. (d) Percentage of dead cells from LIVE/DEAD imaging was calculated and plotted as histograms, which show curcumin kills N2a cells in a dose-dependent manner. (e) DNA fragmentation assay on N2a cells treated with 10, 25, 50 μ M of curcumin or DMSO alone for 24 hours. Equal amount of DNA from all samples was resolved using agarose gel electrophoresis and imaged. The original gel image is shown in Supplementary Fig. S6. (f) Intensity of fragmented DNA to total DNA was measured and values were plotted as histograms showing increased oligonucleosomal DNA fragmentations in curcumin treated cells than controls. (g) N2a cells in culture were treated with curcumin similar to live-dead assay. After 24 hours, cells were collected and TUNEL assay was performed in suspension. Fluorescent images were captured from random fields at 40X magnification at same settings. (h) TUNEL positive cells was calculated in percentage and displayed as histograms. All histograms are displayed as mean ± standard deviation of three independent experiments and p-value displayed was calculated either using non-parametric one-way ANOVA or Student's t-test. * = p ≤ 0.05 is considered statistically significant.
Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420 Total cell lysates were obtained after indicated time points and western blot analysis was performed using anti-phospho-ERK1/2 antibody. Blots were stripped and re-immunoblotted with anti-ERK1/2 antibody. (b) Intensities of phosphorylated ERK1/2 bands were normalized with corresponding ERK1/2 band intensities and histograms were plotted as mean ± standard deviation of three independent experiments, which show curcumin significantly inhibited ERK1/2 phosphorylation with maximum at 30 minutes. (c) Approximately 60-70% confluent N2a cells were serum starved for 2 hours and treated with 25 μ M of curcumin for 24 hours. Equivalent amount of DMSO was treated in control cells. Western blot analysis was performed utilizing anticaspase-8, 9 or 3 antibodies indicating increased proteolysis of caspases-9 and -3 but not caspase-8 in curcumin treated N2a cells than controls at 24 hours. Blots were stripped and immunoblotted with anti-GAPDH antibody. Densitometric analysis of cleaved caspase-9 (d) and caspase-3 (e) are plotted as histograms. (f) Western blot analysis was performed on control and 25 μ M of curcumin treated N2a cells at 24 hours of treatment utilizing anti-PARP-1 antibody which indicates increased proteolysis of PARP-1 in curcumin treated N2a cells. (g) Level of cleaved PARP-1 (89 kDa) was calculated after normalizing its intensity to total PARP-1 and GAPDH. Values obtained are plotted as histograms of mean ± standard deviation from three independent experiments. P-values displayed were calculated using two tailed unpaired, Student's t-test. * = p ≤ 0.05 is considered statistically significant. The original immunoblot images are shown in Supplementary Fig. S7 (for a) and Supplementary Fig. S8 (for c,f).
Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420  and Bex6 (f) PCR products were obtained and normalized with corresponding GAPDH band intensity, and displayed as histograms of mean ± standard error of mean from three independent experiments. P-values displayed were calculated by using two-tailed, unpaired Student's t-test and * = p ≤ 0.05 is considered statistically significant.
Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420 Bex PCR products increased with increase in PCR cycle numbers suggesting the intensities are below saturation at 32 and 34 cycles but approaching saturation by 36 cycles, which justify quantification of Bex genes expression.

Curcumin induces Bex genes in N2a neuroblastoma cells in a time-dependent manner.
To establish optimal duration of curcumin stimulation on Bex genes, we performed time course experiments ranging from 2 to 24 hours using semi-quantitative RT-PCR analysis. Expression of Bex1, Bex2, Bex4, and Bex6 genes was upregulated at all the time points with a maximal induction at 4 hours. Intensity of GAPDH band did not change markedly between treated and control samples (Fig. 4a). Densitometric analysis at 4 hours of curcumin treatment exhibits Bex1 expression was upregulated by 8.8, 9.7 and 5.2-fold and Bex2 was overexpressed by 10.4, 5.6 and 4.1-fold at 32, 34 and 36 cycles respectively than controls ( Fig. 4b and c respectively). Similarly, Bex4 was up regulated by 106.5, 22.6 and 8.9-fold, and Bex6 was up regulated by 291.4, 135.9 and 71.5-fold at 32, 34 and 36 cycles respectively in curcumin treated cells than controls ( Fig. 4e and f respectively). Curcumin also induced Bex3 gene with maximum at 8 hours of treatment. Unlike other Bex genes, Bex3 induction by curcumin was not observed at 2 hours of treatment (Fig. 4a). Densitometric analysis demonstrated that curcumin upregulated Bex3 gene by 30.8, 14.2 and 196.7-fold at 32, 34 and 36 cycles respectively than control N2a cells (Fig. 4d). Taken together, curcumin has the potential to induce all the Bex genes over a long period.

Curcumin specifically induces Bex genes in N2a neuroblastoma cells. Whether induction of Bex
genes is specific to curcumin-mediated N2a cell apoptosis or a general consequence of apoptosis by any anti-cancer agent, we utilized both curcumin and 8-Methoxypyrimido [4′ ,5′ :4,5]thieno(2,3-b) Quinoline-4(3 H)-One (MPTQ) to induce Bex genes. MPTQ is a new anti-cancer agent. It induces apoptosis in different cancer cell lines 21 and through intrinsic apoptotic pathway in mouse and human neuroblastoma cells 22 . Both 10 μ M of MPTQ and and Bex6 (f) PCR products was performed and normalized with corresponding GAPDH band intensity. Values are displayed as histograms of mean ± standard deviation from three independent experiments, which indicates induction of all endogenous Bex genes are observed in curcumin but not in MPTQ treated N2a cells. P-values displayed were calculated by using two-tailed, unpaired Student's t-test and * = p ≤ 0.05 is considered statistically significant. 25 μ M of curcumin induce N2a neuroblastoma cells apoptosis to a similar amount, 15%, and 11% respectively. Therefore, we utilized 10 μ M of MPTQ in parallel with 25 μ M of curcumin to understand whether induction of Bex genes is associated with any type of apoptosis or specific to curcumin. Unlike curcumin, MPTQ failed to induce any Bex genes at various doses tested in N2a cells (Fig. 5a). Densitometric analysis showed 3.03 ± 1.4, 2.68 ± 1.4, 209 ± 67.7, 7.56 ± 3.2 and 70.9 ± 3.4-fold increased expression of Bex1, Bex2, Bex3, Bex4 and Bex6 genes respectively in curcumin treated cells than controls, whereas expression of all these Bex genes in MPTQ treated N2a cells was comparable to control cells ( Fig. 5b-f). Furthermore, higher dose of MPTQ (25 μ M) also failed to induce any Bex genes as compared to curcumin ( Supplementary Fig. S1). In addition, effect of tunicamycin (TM), a well-known inducer of endoplasmic reticulum stress, on Bex gene induction in N2a cells was also studied. Results demonstrate that toxic dose of TM (0.1 μg/ml) failed to induce Bex genes as efficiently as curcumin (as shown in Supplementary Fig. S2). The expression of Bex3, 4 and 6 in TM treated N2a cells is comparable to controls. Therefore, induction of all Bex genes is a response, specific to curcumin in N2a neuroblastoma cells.

Association of Bex genes induction with curcumin-mediated N2a neuroblastoma cells apoptosis.
Involvement of Bex genes in cancer cell apoptosis warrants experiments that reduce Bex genes expression and correlates it with cancer cells death. Since curcumin-mediated induction of Bex genes was inhibited strongly by pifithrin-α , therefore, we wanted to study whether pifithrin-α -mediated inactivation of Bex genes also abrogates apoptosis in curcumin treated N2a cells. LIVE/DEAD assays demonstrated curcumin-mediated N2a cell deaths were reversed by pifithrin-α in a dose-dependent manner (Fig. 6c). Cell scoring analysis showed approximately 11 ± 1.4% cells death in curcumin treated N2a cells over control cells (2.3 ± 0.3%) at 24 hours, which was reduced significantly in a dose-dependent manner to 7.7 ± 1.4% and 4.2 ± 0.6% in 10 and 25 μ M of pifithrin-α pretreated cells respectively (Fig. 6d). In addition, pifithrin-α also inhibited curcumin-mediated activation of apoptotic factors, such as caspase-9, caspase-3 and proteolysis of PARP-1 (Fig. 6e top, middle and bottom panel respectively). In addition, N2a cells transfected with siRNA duplexes against Bex transcripts either individually or in combination showed inhibition towards curcumin-mediated N2a cell deaths (Fig. 6f). Analysis of ethidium homodimer positive (indicator of dead cell population) cells suggests siRNA against Bex6 and Bex1 transcripts have maximum inhibition followed by cocktail of all siRNAs, Bex4 and Bex2 in curcumin mediated N2a cell Supplementary Fig. S12. (b) Intensities of Bex1, Bex2, Bex3, Bex4 and Bex6 PCR products were measured and normalized values were displayed as histograms of mean ± standard error from three independent experiments. (c) LIVE/DEAD analysis was performed as described earlier except pifithrin-α was treated 30 minutes before curcumin treatment. Multiple images were captured and displayed with equal image settings. (d) Percentage of dead cells in each group was calculated from three independent experiments and plotted as histograms. (e) Cell lysates were prepared after 24 hours either from DMSO or curcumin (25 μ M) alone or along with pifithrin-α (25 μ M) treated N2a cells. Western blotting was performed using antibodies against caspase-9, caspase-3 or PARP1. GAPDH immunoblotting was used as loading and transfer control. The original immunoblot images are shown in Supplementary Fig. S13. (f) N2a cells were transfected with Bex1, 2, 4 or 6 siRNA duplexes either individually or in combination for 72 hours followed by curcumin (25 μ M) treatment for 24 hours. Scrambled siRNA transfected N2a cells were used as negative controls. Untransfected N2a cells treated with 25 μ M of curcumin or DMSO were used as positive and negative controls respectively. LIVE/DEAD assay was performed at 24 hours of curcumin treatment and fluorescent images from nine random fields were captured at 30X magnification indicating inhibition of Bex siRNAs against curcumin-mediated cell death. (g) The number of dead cells (red) were counted manually and plotted as histograms. P-values displayed as numbers or * were calculated using two tailed unpaired Student's t-test. P-value ≤ 0.05 is considered statistically significant. Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420 deaths (Fig. 6g). In order to authenticate our imaging strategy, images were also obtained from 4X objective lens to cover more area and from the center of the well where dead cells have a tendency to accumulate. Intensity from these images were obtained and plotted as histograms. Our results again suggest siRNA against Bex1 and Bex6 transcripts have maximum inhibitory effect followed by cocktail of all siRNAs and Bex4 in curcumin-mediated N2a cell deaths ( Supplementary Fig. S4). However, scrambled siRNA transfected N2a cells showed cell deaths comparable to curcumin treated N2a cells in both the experimental settings discussed. Collectively, all these results provide a strong evidence for the functional role of Bex genes in N2a neuroblastoma cell apoptosis and possible involvement of p53 in curcumin-mediated induction of Bex genes.

Activation of p53 through hyperphosphorylation at ser15 in curcumin treated N2a cells prior
to Bex genes induction. Inhibition of curcumin-mediated induction of Bex genes by pifithrin-α , an established inhibitor of p53 23 and increased TUNEL positive cells in curcumin treated N2a cells hint the involvement of p53. Therefore, activation of p53 by hyperphosphorylation at ser15 was investigated. Western blot analysis demonstrates increased p53-ser15 phosphorylation after 15 and 30 minutes of curcumin treatment, which decreased thereafter (Fig. 7a, top panel). Unlike increased p53-ser15 phosphorylation, the level of p53 decreased in a time-dependent manner in curcumin treated N2a cells than controls (Fig. 7a, middle panel). Densitometric analysis of phospho-p53-ser15 intensity demonstrated significant increased p53 phosphorylation by 2.08 ± 0.4 and 2.33 ± 0.18-fold after 15 and 30 minutes of curcumin treatment respectively. By 2 hours, disappearance of phospho-p53-ser15 level reached statistically significant (Fig. 7b, right panel). Densitometric analysis also indicated that p53 level decreased significantly by 37.8 ± 15.3% and 80.6 ± 15.4% after 1 and 2 hours respectively but marginally after 15 and 30 minutes of curcumin treatment (Fig. 7b, left panel). Images from immunocytochemistry exhibit increased nuclear phospho-p53-ser15 intensity in curcumin treated cells than control cells (Fig. 7c). Cell count analysis show approximately 42 ± 4.25% and 6.4 ± 3.8% phospho-p53-ser15 positive cells in curcumin treated and control N2a cells respectively and the difference between the treatments is highly significant, p = 6E-07 (Fig. 7e). However, the number of p53 positive cells shows a marginal decrease in curcumin treated N2a cells than control cells at 30 minutes of treatment ( Fig. 7d and e), which paralleled the results from western blot analysis at similar time point.  (Fig. 7f-j respectively). Densitometric analysis of PCR amplicons was performed and intensities of ChIP DNA amplicons was normalized with input DNA amplicons. The fold change between two groups was calculated and histograms were plotted indicating 1.61 ± 0.37 (p = 0.01), 1.59 ± 0.36 (p = 0.01), 3.57 ± 3.94 (p = 0.24), 2.55 ± 1.33 (p = 0.05) and 1.57 ± 0.31 (p = 0.01) fold more phospho-p53-ser15 binding to Bex1, Bex2, Bex3, Bex4 and Bex6 promoters respectively in curcumin treated cells than controls (Fig. 7f-j respectively). Thus, our results clearly indicate, induction of all endogenous Bex genes in curcumin treated N2a neuroblastoma cells are regulated by p53.

Chromatin immunoprecipitation demonstrates p53-mediated regulation of
Wortmannin, SP600125 and pifithrin-α inhibited p53-ser15 phosphorylation and it's binding to Bex genes promoter in curcumin treated N2a neuroblastoma cells. Since, pretreatment of N2a (c) and p53 (d) level at 30 minutes of curcumin treated or untreated N2a cells. (e) Percent phospho-p53-ser15 and p53 positive cells were calculated in both the groups and plotted as histograms. (f) Interaction of activated p53 to the promoters of Bex genes by chromatin immunoprecipitation (ChIP) assay. N2a cells were treated with curcumin (25 μ M) for 30 minutes followed by ChIP using anti-phospho-p53-ser15 antibody. PCR amplification and densitometric analysis (see material and methods) of p53 binding elements of Bex1 (f), Bex2 (g), Bex3 (h), Bex4 (i) and Bex6 (j) genes to phospho-p53-ser15 show at least one p53 binding element of each Bex gene interact more with phospho-p53-ser15 in curcumin treated N2a cells than corresponding controls. (k) Immunofluorocytochemical analysis demonstrates wortmannin, SP600125 and pifithrin-α reduces p53-ser15 phosphorylation at 30 minutes of curcumin treated N2a cells. (l) Average nuclear phospho-p53-ser15 intensities with or without inhibitors were measured and plotted as histograms from three independent experiments. The original gels images used for f-j are shown in Supplementary Fig. S15. (m) Effect of wortmannin (4 nM), SP600125 (300 nM) and pifithrin-α (25 μ M) on phospho-p53-ser15 DNA binding activity to Bex genes at 30 minutes of curcumin (25 μ M) treated N2a cells. Chromatin immunoprecipitation was performed as described above. PCR amplification was performed corresponding to p53 binding element/s of Bex1, Bex2, Bex3, Bex4 and Bex6 gene promoters. Input DNA PCR was also used as loading control. The original gel images are shown in Supplementary Fig. S16. (n) Band intensities of ChIP PCR products were normalized with corresponding input DNA PCR products and fold change was calculated. Values were plotted as histograms. P-values displayed were calculated using two-tailed, unpaired Student's t-test and * = p ≤ 0.05 is considered statistically significant. cells with wortmannin (4 nM), SP600125 (300 nM) or pifithrin-α (25 μ M) inhibited curcumin-mediated induction of Bex genes, we therefore investigated their effect on p53-ser15 phosphorylation prior to the induction of Bex genes. Immunofluorescent images demonstrate curcumin-mediated p53-ser15 phosphorylation is reduced notably in wortmannin and SP600125 but to lesser level in pifithrin-α pretreated N2a cells than cells treated with curcumin only (Fig. 7k). Average nuclear phospho-p53-ser15 intensity shows approximately 2.57-fold increased p53-ser15 phosphorylation in curcumin treated cells than controls and the difference between untreated and curcumin treated samples is highly significant (p = 4E−08). Wortmannin, SP600125 and pifithrin-α inhibited curcumin-mediated p53-ser15 phosphorylation significantly by 92.2%, 91.4% and 52.6% respectively than N2a cells treated with curcumin only (Fig 7l). Furthermore, we investigated the effect of these inhibitors on curcumin-mediated binding of phospho-p53-ser15 onto Bex gene promoters prior to the induction of Bex genes. Results from chromatin immunoprecipitation showed that curcumin-mediated binding of phospho-p53-ser15 to Bex1 (− 2000 to − 2023 bp), Bex2 (− 2462 to − 2493 bp), Bex3 (− 1883 to − 1914 bp), Bex4 (− 912 to − 943 bp) and Bex6 (− 458 to − 482 bp) promoters is reduced in wortmannin and pifithrin-α but not much in SP600125 pretreated N2a cells than cells treated with curcumin only (Fig. 7m). However, binding of phospho-p53-ser15 to  (Fig. 7n). Collectively, these results clearly indicate the direct binding of phospho-p53-ser15 onto Bex genes promoters in curcumin treated N2a neuroblastoma cells.

Discussion
Re-expression/induction of all endogenous Bex genes by a safe and small molecule was our main interest. In addition, we also wanted to investigate their function in neuroblastoma, which is completely unknown. Our results clearly demonstrated that curcumin, the active principle of turmeric has the potential to induce Bex1 and Bex2 and re-express Bex3, Bex4 and Bex6 genes. Induction of these genes functions as tumor suppressors in N2a neuroblastoma cells.
At first, we established a robust curcumin-mediated apoptosis in N2a neuroblastoma cells. In this study, we observed the cytotoxic effect of curcumin on N2a cells similar to earlier studies 24 . Curcumin activated caspase-9 (an intrinsic apoptotic factor) 25 and caspase-3 in N2a cells, which are consistent with earlier studies 18 but not caspases-8, an extrinsic apoptotic factor 26 . Proteolytic inactivation of PARP-1 (115 kDa) to 89 kDa fragment by activated caspase-3 is another hallmark of apoptosis 27 . Here we show for the first time that, curcumin significantly inhibited PARP-1 in N2a cells, which resembles earlier studies on breast cancers 28 and glioma cells 29 . Therefore, curcumin-mediated N2a cells death confirmed intrinsic apoptotic pathway activation and making it a suitable neuroblastoma model to test the induction of endogenous Bex genes by a single agent, curcumin and to study their functions in neuroblastoma cell apoptosis.
Our results from curcumin treated N2a neuroblastoma cells indicated the induction of Bex1, Bex2 and Bex3 genes prior to activation of apoptotic pathways and continued to last for 24 hours post treatment. Adenoviral or trichostatin-A-mediated induction of Bex1 and Bex2 genes decreased glioma cells proliferation and tumor formation indicating their function as tumor suppressors in malignant glioma 12 . Re-expression of Bex1 gene also functions as tumor suppressor in acute myeloid leukemia 30 and pediatric intracranial ependymoma 31 . Similarly, Bex3 has been shown to interact with p75NTR and mediates apoptosis 32 . Very recently, Bex4 gene has been reported to be epigenetically silenced in oral squamous cell carcinoma (OSCC) and overexpression of Bex4 reduced OSCC proliferation in vitro and subcutaneous tumor volume in Bex4 overexpressing CAL27 mouse xenograft models in vivo 33 . However, there is no report on Bex4 expression or function in the field of neuroblastoma. Furthermore, there is no report on the association of Bex6 gene with any type of cancer cells death. In this study, we report for the first time that, Bex4 and Bex6 genes are re-expressed by curcumin and are associated with apoptotic N2a neuroblastoma cell deaths. Since Bex proteins are intrinsically disordered proteins, we also believe that these genes may not act in a similar manner in all cancer cell types and in normal cell processes 10 . In this study, we observed that all these genes were induced almost in a similar manner and remain overexpressed for a long period of time in response to curcumin-mediated N2a cells apoptotic death. Therefore, these results indicate us to interpret Bex genes as pro-apoptotic genes at least in N2a neuroblastoma cells. Furthermore, unlike curcumin, induction of Bex genes was not observed in MPTQ (a known anti-cancer agent)-mediated N2a cells death even at 25 μ M dosage, known to induce DNA fragmentation and apoptosis approximately 3-fold more than curcumin 22 . In addition, tunicamycin, which is a well-known inducer of ER stress-mediated cytotoxicity in N2a cells 34 , also failed to induce Bex genes at cytotoxic dosage (0.1 μ g/ml) in N2a cells. This observation not only eliminates the role of curcumin-induced ER stress on Bex genes induction in N2a cells but also strengthens the specific role of curcumin in Bex genes induction. Thus, curcumin is a specific chemical inducer of all the Bex genes and upregulation of these genes might be involved in N2a neuroblastoma apoptosis.
Next, we wanted to establish the mechanism behind curcumin-mediated induction of Bex genes. We utilized inhibitors of cell signaling, which are commonly used in experiments on cell physiology. In parallel with our decreased ERK1/2 phosphorylation prior to Bex genes induction, pretreatment with PD98059 and U0126 failed to reduce curcumin-mediated induction of Bex genes in N2a cells suggesting ERK1/2 activation is not required in Bex genes induction. In addition, reappearance of phosphorylation level of ERK1/2 at later stage suggests dephosphorylation of ERK1/2 might be an immediate early response to curcumin treatment in N2a cells. Curcumin induces reactive oxygen species (ROS) in human neuroblastoma cells 35 and ROS generation can inhibit ERK1/2 phosphorylation in cancer cells 36 . It has also been reported that curcumin dephosphorylate ERK1/2 transiently up to 2 hours followed by reappearance of ERK1/2 phosphorylation by 6 and 12 hour of treatment in olfactory ensheathing cells 37 . In addition, curcumin induces ROS in mouse L929 cells within 5 minutes of treatment, which decreases to normal level by 60 minutes 38 . Therefore, inhibition of pERK1/2 phosphorylation at 15 min post curcumin treatment might be due to rapid generation of ROS, which might have decreased over time resulting in reappearance of phospho-ERK1/2 in N2a cells at later period of treatment. Furthermore, reduction in curcumin-mediated Bex gene induction was not seen in SB202190 pretreated N2a cells indicating p38 MAPK pathway is also not involved. However, wortmannin, SP600125 and pifithrin-α significantly inhibited curcumin-mediated induction of all the Bex genes in N2a cells. Wortmannin is a PI3-kinases inhibitor, which also inhibits PI3-kinase like kinases, such as ATM, ATR and DNA-PK 39,40 . ATM and DNA-PK phosphorylate p53 at ser15 in response to DNA damage, which is crucial for p53-dependent transactivation of downstream genes 41,42 . Therefore, inhibition of curcumin-mediated induction of Bex genes by wortmannin suggests the involvement of ATM or DNA-PK leading to p53 activation. JNK has been known to phosphorylate p53 at ser33 and ser15 in response to reactive oxygen species in breast cancer and colon carcinoma cell lines 43 . SP600125, an established JNK inhibitor has been shown to inhibit glutamate-induced p53-ser15 phosphorylation in murine neuronal Wortmannin, SP600125 and pifithrin-α abrogate curcumin-mediated induction of Bex genes suggesting the involvement of PI3-kinases, JNKs and p53respectively. Curcumin activated p53 phosphorylation at ser15 and inactivated MEKs at a similar time but well before the induction of Bex genes. Phosphorylation of p53-ser15 is inhibited by wortmannin, SP600125 and pifithrin-α indicating a molecular association of p53 activation with curcumin-mediated Bex genes induction. Furthermore, inhibition of Bex genes by pifithrin-α also inhibited curcumin-mediated N2a cell deaths suggesting a strong correlation of Bex genes induction with apoptotic neuroblastoma cells death. hippocampal cell line 44 suggesting SP600125-mediated inhibition of Bex genes induction in curcumin treated N2a cells might involves JNK-mediated p53 activation. Pifithrin-α has been shown to inhibit DNA binding activity of p53 without affecting its phosphorylation status 23 . Results from cell signaling studies show pifithrin-α is the most potent inhibitor of curcumin-mediated induction of all the Bex genes, suggesting the involvement of p53 activation in curcumin-mediated induction of Bex genes. Results from p53 activation indicated a significant reduction in p53 level by 60 minutes in curcumin treated N2a cells than untreated controls. Increased p53-ser15 phosphorylation but decreased p53 level have been reported in acetaminophen treated glioma cells, which has been linked with ubiquitin-dependent degradation of p53 45 . Similarly, Kuenzi et al., have reported increased p53-ser15 phosphorylation but decreased expression of p53 in fungal alkaloid, militarinone A-treated N2a cells 46 , indicating the activation of p53 by increased phosphorylation at ser15 might reduce p53 level in curcumin treated N2a cells. In addition, our results also indicated p53-ser15 hyperphosphorylation immediately (15-30 minutes) after curcumin treatment but before Bex genes induction in neuroblastoma cells.
Furthermore, to establish p53 as a direct transcriptional activator of Bex genes, the promoters of Bex genes were screened manually for p53 binding elements as reported by Deiry El et al. 47 (5′ -RRRCWWGYYY (0-13 bases) RRRCWWGYYY-3′ ) and Veprintsev et al. 48 (5′ -NDRCATGYYY (0-13 bases) NDRCATGYYY-3′ or 5′ -NNDCWWGYHN (0-13 bases) NNDCWWGYHN-3′ ) 47,48 . DNA binding elements of p53 were found in the promoter regions of Bex1, Bex2, Bex3, Bex4 and Bex6 genes as shown in Supplementary Fig. S5. Our results from chromatin immunoprecipitation utilizing anti-phospho-p53-ser15 antibody for the first time demonstrate that all the Bex genes have at least one functional p53 binding elements in their promoters and p53 can transactivate all the Bex genes in response to curcumin treatment in N2a cells. As suggested earlier, inhibitors of curcumin-mediated induction of Bex genes, wortmannin, SP600125 and pifithrin-α also reduced p53-ser15 phosphorylation significantly indicating the involvement of PI3-kinase like kinases and JNKs in activating p53, which might be associated with curcumin-mediated induction of Bex genes in N2a neuroblastoma cells. Interestingly, the time of p53 activation paralleled with the time of ERK1/2 inactivation. Inhibition of ERK1/2 stimulated p53 activation in MCF7 cells that led to the upregulation of downstream targets, like p21 and Bax 49 and activation of p53 stimulate intrinsic apoptotic pathway 50,51 . In addition, wortmannin, SP600125 and pifithrin-α reduced curcumin-mediated binding of phospho-p53-ser15 onto p53 DNA binding elements in Bex gene promoters. Collectively, our results for the first time provide a strong evidence for presence of functional p53 DNA binding elements in all Bex genes, which can be epigenetically activated by curcumin.
Finally, to link the function of Bex genes induction with neuroblastoma cell apoptosis, experiments involving reduction of Bex mRNAs either by siRNA or pifithrin-α (most potent inhibitor of Bex gene induction) treatment along with reduction of curcumin-mediated N2a cell deaths were performed. Our results clearly demonstrated, both pifithrin-α and Bex1, 2, 4 and 6 gene siRNAs either individually or in combination demonstrated significant reduction in curcumin-mediated cell death in N2a cells. Although the use of siRNA against Bex genes are limited but a recent report suggests siRNA-mediated inhibition of only Bex4 gene promotes proliferation of OSCC cells 33 suggesting the tumor suppressor activity of Bex4 in OSCC cells. Pifithrin-α also significantly inhibited proteolytic activation of apoptotic factors such as caspase-9, caspase-3 and inactivation of PARP-1, providing a strong evidence for a functional association of Bex genes in neuroblastoma cell apoptosis mediated by curcumin.
In conclusion, the present study indicates for the first time the induction of all Bex genes by ectopic application of curcumin, a commonly used phytochemical. Moreover, curcumin is a specific inducer of Bex genes unlike MPTQ and tunicamycin. Cell signaling studies for the first time demonstrate the association of p53, PI3-Kinase and JNK in curcumin-mediated induction of endogenous Bex genes and the activation of intrinsic apoptotic pathway (Fig. 8). This study for the first time also reports, Bex genes are novel downstream targets of p53 and possess active p53 DNA binding elements in the promoter region of all Bex genes. In addition, based on the experimental evidence, induction of Bex genes are associated with curcumin-mediated apoptosis in N2a neuroblastoma cells. Therefore, induction of Bex genes (especially Bex6, Bex1 and Bex4 genes) functions as tumor suppressors at least in N2a neuroblastoma cells and may serve as an alternative for the treatment of neuroblastomas and other cancers with silenced Bex genes. Antibodies. Anti-caspase-3, anti-caspase-9, anti-phospho-p53-ser15 and anti-p53 antibodies were purchased from Cell Signaling Technology Inc., USA. Anti-PARP-1, anti-phospho ERK1/2, anti-ERK1/2 and anti-GAPDH antibodies were purchased from Santa Cruz Biotechnology, USA. Anti-cleaved caspase-3 antibody was purchased from Imgenex Corp., USA. Anti-caspase-8 antibody was purchased from R&D systems, Inc., USA. Horseradish peroxidase (HRP)-conjugated secondary antibodies and Alexa-Fluor 647 conjugated F(ab') 2 fragment antibodies were purchased from Thermo-Pierce and Molecular probes respectively. Bright Field Imaging. N2a cells were seeded at a density of 20000 cells/well in 24 well plates. After 2 days, cells were serum starved for 2 hours followed by treatment with 50 μ M of curcumin or equivalent amount of DMSO as controls. After 24 hours of treatment, bright field images of cells were captured using 40X objective lens with Hoffman modulation and thickness correction rings in Nikon TS100 inverted microscope supported by NIS elements BR (2.3 version) software. LIVE/DEAD Assay. N2a cells were seeded at a density of 5000 cells/well in 96 well plates. Curcumin-mediated N2a cell death was studied using LIVE/DEAD assay reagents (Invitrogen) as described earlier 22 . Briefly, cells were serum starved for 2 hours and treated with 10, 25, 50 μ M of curcumin or equal volume of DMSO as control for 24 hours. Cells were then incubated in media containing calcein-AM and ethidium homodimer for 30 minutes at room temperature in dark. Random images (3)(4) were captured using 40X objective lens in Nikon TS100 inverted microscope supported by NIS elements BR (2.3 version) software with same exposure settings. Number of live (green) and dead (red) cells were counted using ImageJ software.

Material and Methods
MTT Assay. MTT assay was performed as described earlier 52 . Briefly, 5000 cells/well were cultured in a 96-well plate for 72 hours. Cells were treated with indicated concentration of curcumin after 2 hours of incubation in serum free DMEM. After 24 hours, the media was completely replaced with serum free DMEM containing 1 mg/ml of MTT reagent (Sigma) and cultured for another 4 hours. The media containing MTT was completely removed and intracellular formazan products were dissolved in MTT solvent (4 mM HCl and 0.1% NP-40 in isopropanol) for 15 minutes with shaking. The absorbance of the samples was measured at 570 nm using a multi-well plate reader. DNA Fragmentation. N2a cells were seeded at a density of 24,000 cells/well in 24-well plates for 72 hours.
Cells were then treated with varying amount of curcumin for 24 hours or treated with equal volume of DMSO as controls after 2 hours of incubation in serum free DMEM. Cells were then lysed using 100 μ l of digestion buffer (100 mM NaCl, 10 mM Tris-Cl pH 8.0, 1 mM EDTA (pH 8.0), 0.1% SDS and 0.1 mg/ml proteinase-K) at 50 °C for 3 hours. Samples were mixed with 5 μ l of SDS-OUT reagent and incubated on ice for 20 minutes followed by centrifugation at 10,000 × g for 10 minutes at 4 °C. Supernatants were collected and treated with RNase for 20 minutes on ice. After DNA quantification, 2.5 μ g of DNA from each sample was resolved in ethidium bromide stained 1.5% agarose gels and imaged in a gel-doc system. Integrated density of fragmented DNA normalized with total DNA intensity was measured and fold change was calculated. TUNEL Assay. N2a cells were seeded at a density of 0.5 × 10 6 cells per 25 cm 2 flasks and cultured for 2 days.
Cells were then incubated in serum free DMEM for 2 hours and treated with indicated amounts of curcumin or equal volume of DMSO as control. After 24 hours, TUNEL assay was performed in suspension as described earlier 22 . Labeled cells were spotted on slides coated with tissue adhering solution (Cancer diagnostics) and mounted using anti-fade gold mounting media containing DAPI (Invitrogen). Fluorescent images were captured using 40X lens using similar exposure settings in a Nikon Ti eclipse microscope supported by Metamorph software (version 7.7.0.0). TUNEL positive cells were counted and percent positive cells were plotted as histograms.
RNA Isolation. N2a cells treated with varying amount of curcumin for 2 hours or with 25 μ M of curcumin for indicated time points after 2 hours of serum starvation. Equivalent amount of DMSO treated cells were used as controls. Similar approach was adopted for N2a cells treated with different dosages of MPTQ for 4 hours. For inhibitor assay, N2a cells were pretreated with 4 nM of wortmannin, 300 nM of SP600125 and 25 μ M of pifithrin-α for 30 minutes followed by 25 μ M of curcumin treatment for 4 hours. DMSO treated and only curcumin treated N2a was used as negative and positive controls of the experiment respectively. Total RNA was isolated using Trizol (Invitrogen) as per manufacturer's instructions and quantified by utilizing Nanovue spectrophotometer (GE healthcare). Total RNA was treated with 2 units of amplification grade DNase I (Life technologies, USA) for 15 minutes at room temperature. The DNase activity was inhibited using EDTA to a final concentration of 2 mM and incubated at 65 °C for 10 minutes. DNA free RNA was further quantified by Nanovue spectrophotometer.
Scientific RepoRts | 7:41420 | DOI: 10.1038/srep41420 Western blotting. Western blot analysis was performed using earlier protocol 53 . Briefly, cells were collected after the treatments and centrifuged at 2000 rpm for 5 minutes to obtain cell pellet. Supernatant was removed completely and cell lysates were prepared in lysis buffer. Lysates were sonicated, cleared by centrifugation at 500 × g for 5 minutes at 4 °C, aliquoted and stored at − 70 °C deep freezer. Protein content of each lysate was estimated by bicinchoninic acid assay using micro BCA protein estimation kit (Thermo-Pierce). Equal amount of proteins (~60 μ g) from each sample were separated by SDS-PAGE and transferred onto nitrocellulose or PVDF membranes. After blocking, membranes were blotted with primary antibodies at 4 °C for overnight. Horseradish peroxidase conjugated secondary antibodies were used to develop blots. Bands were detected using West Pico supersignal chemiluminiscent substrate (Pierce). GAPDH immunoblotting was used to normalize protein loading and transfer between the samples.
Immunocytochemistry. N2a cells were plated at a density of 10,000 cells/well onto Poly-D-Lysine (0.05 mg/ ml) coated coverslips placed in 24 well plates. After 72 hours, cells were serum starved for 2 hours and treated with 25 μ M of curcumin or with equal amount of DMSO (controls) for 30 minutes. Cells were treated with pharmacological inhibitors 30 minutes prior to curcumin treatment. Cells were quickly washed with 1X PBS and fixed with 4% paraformaldehyde for 30 minutes at RT. Cells were then permeabilized with 0.3% triton X-100 followed by blocking (PBS with 0.1% BSA and 10% normal goat serum) for 1 hour at RT. Cells were incubated with primary antibodies in blocking buffer at 4 °C overnight in a humidified chamber. After several washes, cells were incubated with appropriate secondary antibody conjugated with Alexa Fluor 647 for 1 hr at RT. Cells were washed several times and mounted with anti-fade mounting medium containing DAPI. Fluorescent images were captured using 40X plan Fluor lens in Nikon Ti eclipse microscope and at same exposure settings 54 . Fluorescence intensity was measured by multi-wavelength cell scoring module of Metamorph software (version 7.7.0.0) with identical settings. Images were displayed with same image scale using Metamorph software.

Transfection of siRNA duplex into N2a neuroblastoma cells. N2a cells were transfected with 200
pico moles of Bex siRNAs either individually or all four (Bex1, Bex2, Bex4 and Bex6) in combination at 30-40% confluency using lipofectamine 2000 (Invitrogen, USA) reagent. The media was replaced after 24 hours of transfection. At 72 hours of transfection, N2a cells were serum starved for 2 hours followed by curcumin (25 μ M) treatment. Control cells were treated with equivalent amount of DMSO. After 24 hours, LIVE/DEAD assay was performed to establish the relation between Bex gene induction and cell death. Fluorescent images were captured from nine random fields per sample using 20X objective lens in Nikon Ti eclipse microscope and at same exposure settings. Number of ethidium homodimer positive (dead and red) cells were counted per field, analyzed and plotted as histograms. Fluorescent images from 4X objective lens were also obtained to cover more area and from the center of the well where dead cells have a tendency to accumulate. The intensity of ethidium homodimer (indicator of dead cell) was measured from 4X magnified image by utilizing Metamorph software and analyzed for graphical presentation.
Chromatin Immunoprecipitation. Approximately 70-80% confluent N2a cells were serum starved for 2 hours followed by treatment with 25 μ M of curcumin or with equal amount of DMSO (controls) for 30 minutes. For inhibitor assay, cells were pretreated with 4 nM of wortmannin, 300 nM of SP600125 and 25 μ M of pifithrin-α followed by 25 μ M of curcumin treatment for 30 minutes. Chromatin immunoprecipitation was performed as reported earlier 57 . Briefly, cells were cross-linked with 1% paraformaldehyde for 10 min at room temperature. Cells were washed with 1X PBS twice and cell lysates were prepared in lysis buffer (50 mM Tris-Cl pH 8.0, 1% SDS, 10 mM EDTA, 1X protease inhibitor cocktail and 1X phosphatase inhibitor cocktail) for 30 minutes on ice. Lysates were sonicated and supernatants were recovered by centrifugation at 12000 rpm for 10 minutes at 4 °C. Concentration of protein in each lysate was measured as described earlier and 250 μ g of protein equivalent lysate was diluted to 1 ml using dilution buffer (150 mM NaCl, 20 mM Tris-Cl pH 8.0, 2 mM EDTA and 1% Triton X-100). The diluted lysates were precleared and immunoprecipitated with 5 μ l of anti-phospho-p53-ser15 antibody, 2 μ g of sheared salmon sperm DNA and 20 μ l of Protein A-sepharose beads at 4 °C overnight. After several washings in dilution buffer at 4 °C, the antibody, antigen and DNA complex was eluted in elution buffer