Celecoxib induces p53-PUMA pathway for apoptosis in human colorectal cancer cells

Contents lists available atScienceDirect

Chemico-Biological Interactions

Celecoxib induces p53-PUMA pathway for apoptosis in human

colorectal cancer cells

Huei-Fang Liu

_{, Po-Wen Hsiao}

_{, Jui-I Chao}

a_{Department of Biological Science and Technology, National Chiao Tung University, Hsinchu 300, Taiwan} b_{Institute of Pharmacology and Toxicology, Tzu Chi University, Hualien 970, Taiwan}

a r t i c l e i n f o

Received 1 April 2008

Received in revised form 17 July 2008 Accepted 28 July 2008

Available online 8 August 2008

Keywords: Apoptosis Colorectal cancer p53 PUMA Celecoxib

a b s t r a c t

Celecoxib, a clinical non-steroidal anti-inflammatory drug, displays anticarcinogenic and chemopreventive activities in human colorectal cancers, although the mechanisms of apop-tosis by celecoxib are poorly understood. The existence of functional p53 but not securin in colorectal cancer cells was higher on the induction of cytotoxicity than the p53-mutational colorectal cancer cells following celecoxib treatment. The p53-wild type HCT116 cells were more susceptible to increase∼25% cell death than the p53-null HCT116 cells after treatment with 100␮M celecoxib for 24 h. Transfection with a small interfering RNA of p53 reduced the celecoxib-induced cytotoxicity in the RKO (p53-wild type) colorectal cancer cells. Celecoxib (80–100␮M for 24 h) significantly increased total p53 proteins and the phosphorylated p53 proteins at serine15, 20, 46, and 392 in RKO cells. However, the phosphop53 (serine15, -20, and -392) proteins were presented on the nuclei of cells but the phospho-p53 (serine-46) protein was located on the cytoplasma of apoptotic cells following treatment with celecoxib. Interestingly, the p53 up-regulated modulator of apoptosis (PUMA) protein, which located on the mitochondria, was induced by celecoxib in the p53-functional colorectal cancer cells but not in the p53-mutational cells. Together, this study provides the first time that cele-coxib induces the various phosphorylated sites of p53 and activates p53-PUMA pathway, which potentiates the apoptosis induction in human colorectal cancer cells.

Crown Copyright © 2008 Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Colorectal cancer is one of most common cancers around the world[1]. Development of the novel anticancer drugs or strategies for colorectal cancer therapy is highly

Abbreviations: NSAID, non-steroidal anti-inflammatory drug; PUMA,

p53 up-regulated modulator of apoptosis; PI, propidium iodide; ERK, extracellular signal-regulated kinase; MTT, 3-(4,5-dimethyl-thiazol-2-yl) 2,5-diphenyl tetrazolium bromide; FITC, fluorescein isothiocyanate; FBS, fetal bovine serum; PBS, phosphate-buffered saline.

∗ Corresponding author at: Department of Biological Science and Tech-nology, National Chiao Tung University, 75 Bo-Ai Street, Hsinchu 300, Taiwan. Fax: +886 3 5556219.

E-mail addresses:[email protected],[email protected] (J.-I. Chao).

1 _{Equal contribution in this study.}

desired. Celecoxib, a non-steroidal anti-inflammatory drug (NSAID), which has been approved as adjuvant treatment in patients with familial adenomatous polyposis, exhibits anticarcinogenic and chemopreventive activities in human colorectal cancer[2–4]. It has been shown that celecoxib displayed anticancer ability by inducing apoptosis in tumor cells[5–8]. However, the mechanisms of celecoxib-induced apoptosis in human colorectal cancer cells are still unclear. Anticancer agents display the anticancer activities by inducing the various signal proteins for apoptosis in cancer cells [9]. p53 is a key tumor suppressor for controlling apoptosis in cancer cells[10–12]. The diverse phospho-rylation sites of p53 have been shown to play a variety of roles in the responses of cellular stresses[13–15]. The p53 up-regulated modulator of apoptosis (PUMA) was recently identified a p53 downstream protein for promoting apop-tosis[16]. PUMA encodes a BH3-only protein member of

0009-2797/$ – see front matter. Crown Copyright © 2008 Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cbi.2008.07.012

the BCL-2 family that is a key mediator to trigger apoptosis via a p53-dependent pathway[16]. Moreover, PUMA can sensitize cancer cells to chemotherapeutic agents[17,18]. Accordingly, the activation of p53-PUMA pathway by anticancer drugs may contribute to cancer therapy.

In this study, celecoxib induced the phosphorylation of p53 at various sites in human colorectal cancer cells. The existence of functional p53 in colorectal cancer cells was higher on the induction of apoptosis and PUMA protein expression following exposure to celecoxib. We provide that the phosphorylated sites of p53 and p53-PUMA path-way play an important role on the regulation of apoptosis in the celecoxib-treated colorectal cancer cells.

2. Materials and methods

2.1. Reagents and antibodies

Celecoxib was purchased from Toronto Research Chemi-cal Inc. (North York, Canada). BODIPY FL phallacidin (B-607) and Mitotracker orange (M-7510) were purchased from Invitrogen (Carlsbad, CA). 3,3-dihexiloxadicarbocyanine [DiOC6(3)] was purchased from Calbiochem (San Diego, CA). Hoechst 33258, propidium iodide (PI), and 3-(4,5-dimethyl-thiazol-2-yl) 2,5-diphenyl tetrazolium bromide (MTT) were purchased from Sigma Chemical Co. (St. Louis, MO). Anti-extracellular signal-regulated kinase-2 (ERK-2) (C-14), anti-PUMA (N-19), anti-COX-2 (sc-19999), and the FITC (fluorescein isothiocyanate)-labeled goat anti-mouse IgG (sc-2010) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The Cy5-labeled goat anti-rabbit IgG was purchased from Amersham Pharma-cia Biotech (Little Chalfont Buckinghamshire, UK). Anti-p53 (AHO0132) was purchased from BioSource International, Inc. (Camarillo, CA). Anti-phospho-p53 (serine-15, -20, -46, and -392) antibodies, control siRNA, and p53 siRNA were purchased from Cell Signaling Technology, Inc. (Beverly, MA). The monoclonal anti-securin (ab-3305) was pur-chased from Abcam (Cambridgeshire, UK).

2.2. Cell lines and cell culture

RKO colorectal carcinoma cells can express the func-tional p53 proteins[19]. SW480 was a colorectal carcinoma cell line that contained double mutations at codon 273 (G to A mutation) and codon 309 (C to T mutation) in the p53 gene[20]. The wild type p53 and securin, securin-null, and p53-null HCT116 colorectal carcinoma cell lines were kindly provided by Dr. B. Vogelstein of Johns Hop-kins University (Baltimore, MD). RKO and SW480 cells were maintained in DMEM medium (Invitrogen). The dif-ferent genotypes of HCT116 cells were cultured in McCoy’s 5A medium (Sigma Chemical). The complete medium was supplemented with 10% fetal bovine serum (FBS). These cells were maintained at 37◦C and 5% CO2 in a

humidified incubator (310/Thermo, Forma Scientific, Inc., Marietta, OH).

2.3. Cytotoxicity assay

Briefly, the cells were plated in 96-well plates at a den-sity of 1× 104_{cells/well for 12–16 h. Following celecoxib}

treatment, the cells were washed with phosphate-buffered saline (PBS) and were replaced fresh medium for cul-tured 2 days. Thereafter, the medium was replaced and the cells were incubated with 0.5 mg/ml of MTT in complete medium for 4 h. The surviving cells converted MTT to for-mazan that generates a blue-purple color when dissolved in dimethyl sulfoxide. The intensity of formazan was mea-sured at 565 nm using a plate reader (Molecular Dynamics, OPTImax). The relative percentage of cell viability was cal-culated by dividing the absorbance of treated cells by that of the control in each experiment.

2.4. Apoptotic morphology analysis

To examine the effect of celecoxib on apoptosis, RKO or SW480 cells were cultured on coverslip in a 60-mm Petri dish for 12–16 h before treatment. At the end of treat-ment, the cells were carefully and gently washed with isotonic PBS (pH 7.4), and then fixed with 4% paraformalde-hyde solution in PBS for 1 h at 37◦C. The actin filament (F-actin) was stained with 20 U/ml BODIPY FL phallacidin for 30 min at 37◦C. The nuclei were stained with 2.5␮g/ml Hoechst 33258 for 30 min at room temperature. The cell morphology of apoptosis was confirmed by the observa-tion of nuclear fragmentaobserva-tion, cell membrane blebbing, and F-actin cytoskeleton disruption under a fluorescence microscope.

2.5. The analysis of sub-G1fraction by flow cytometry

The sub-G1fractions were determined by flow

cytome-ter. Briefly, RKO or SW480 cells were plated at a density of 1× 106_{cells per 60-mm Petri dish in complete medium for}

16–20 h. At the end of treatment, the cells were collected and fixed with ice-cold 70% ethanol overnight at−20◦_C.

The cell pellets were treated with 4␮g/ml PI solution con-taining 1% Triton X-100 and 100␮g/ml RNase for 30 min. To avoid cell aggregation, the cell solutions were filtrated through nylon membrane (Becton-Dickinson, San Jose, CA). The percentage of sub-G1fractions was quantified by using

CellQuest software (Becton-Dickinson). 2.6. Immunofluorescence staining and confocal microscopy

To view the protein localization and expression of p53 and PUMA after celecoxib treatment, the cells were subjected to immunofluorescence staining and confocal microscopy as described [21]. After fixation with 4% paraformaldehyde solution, the cells were washed three times with PBS, and non-specific binding sites were blocked in PBS containing 10% FBS and 0.3% Triton X-100 for 1 h. Thereafter, the cells were separately incubated with rabbit anti-phospho-p53 (1:100), mouse anti-p53 (1:50) and goat anti-PUMA (1:250) antibodies in PBS containing 10% FBS overnight at 4◦C, and washed three times with 0.3% Triton X-100 in PBS. Then the cells were incubated with goat anti-rabbit Cy5 (1:100), goat anti-mouse FITC (1:50), or donkey anti-goat FITC (1:50) in PBS for 2.5 h at 37◦C. F-actin and nuclei were stained with BODIPY FL phallacidin and Hoechst 33258, respectively. The samples

were examined under a laser scanning confocal microscope (Leica, Wetzlar, Germany).

2.7. Mitochondrial hyperpolarization

Mitochondrial function was evaluated by the cells stained with the mitochondrial sensitive probe DiOC6(3). Lipophilic cation DiOC6(3) accumulated in the mitochon-drial matrix driven by the electrochemical gradient[22]. RKO cells were cultured in 60-mm Petri dish at a density of 1_{× 10}6_{cells for 12–16 h. After treatment with or}

with-out celecoxib, the cells were washed with ice-cold PBS. The cells were trypsinized and collected by centrifugation. The cell pellets were fixed in 70% ethanol at−20◦_{C for 2 h.}

Then cell pellets were incubated with 500 nM DiOC6(3) at 37◦C for 30 min (in the dark). Finally, the collected cell pel-lets were resuspended in 0.5 ml ice-cold PBS and analyzed by flow cytometer (FACScan, Becton Dickinson, San Jose, CA).

2.8. Western blot

The total cellular protein extracts were prepared as described [21]. Briefly, proteins were separated on 10–12% sodium dodecyl sulfate-polyacrylamide gels, and electrophoretic transfer of proteins onto polyvinylidene difluoride membranes. The membranes were sequentially hybridized with primary antibody and followed with a horseradish peroxidase-conjugated secondary antibody. The protein bands were visualized on the X-ray film using the enhanced chemiluminescence detection system (PerkinElmer Life and Analytical Sciences, Boston, MA). Western analyses of various phospho-p53, total p53, PUMA, and ERK-2 were performed using specific antibodies. To verify equal protein loading and transfer, ERK-2 was used as the protein loading control. A gel digitizing software, Un-Scan-It gel (Ver. 5.1, Silk Scientific, Inc.), was used to analyze the intensity of bands on X-ray film by semi-quantification. 2.9. Transfection

The cells (5× 105_{cells/35-mm Petri dish) were}

trans-fected with control siRNA or p53 siRNA by using LipofectamineTM _{2000 (Invitrogen) in serum-free DMEM}

medium according to the manufacturer’s recommenda-tions.

2.10. Statistical analysis

Each experiment was repeated at least three times. Data were analyzed by one-way or two-way analysis of variance (ANOVA), and further post hoc tests using the statistic soft-ware of GraphPad Prism 4 (GraphPad softsoft-ware, Inc. San Diego, CA). A p value of <0.05 was considered as statisti-cally significant in each experiment. Differences between control and celecoxib-treated samples were compared by one-way ANOVA with post hoc Tukey’s tests. Differences between two different genotype cell lines from control and celecoxib treatments were compared by two-way ANOVA with Bonferroni post-tests.

3. Results

3.1. The existence of functional p53 but not securin in colorectal cancer cells increases cytotoxicity after treatment with celecoxib

A variety of human colorectal cancer cell lines includ-ing RKO, SW480, HCT116 (securin-wild type), and HCT116 (securin-null) expressed the high levels of COX-2 proteins (Fig. 1A). The protein levels of p53 presented in these cell lines; however, SW480 cells expressed the mutant p53 (contained a G to A mutation in codon 273 and a C to T mutation in codon 309). The securin proteins existed in RKO, SW480, and HCT116 (securin-wild type) cells but did not express in the securin-null HCT116 cells (Fig. 1A). ERK-2

Fig. 1. The effect of celecoxib on the cell viability in a variety of human

colorectal cancer cell lines. (A) The total protein extracts from various col-orectal cancer cell lines were subjected to Western blot analysis using specific anti-COX-2, anti-p53, anti-securin, and anti-ERK-2 antibodies. Representative Western blot data were shown from one of three sepa-rate experiments with similar findings. (B) The cells were treated with 0–80␮M celecoxib for 24 h. After drug treatment, the cells were washed with PBS, and incubated for 2 days before they were measured by MTT assay. Results were obtained from 6 to 12 experiments and the bar repre-sents the mean± S.E. *p < 0.05 and **p < 0.01, indicate significant difference between control and celecoxib treated samples in various colorectal can-cer cell lines.

Fig. 2. The comparison of cytotoxicity between the p53-wild type and

p53-null HCT116 colorectal cancer cells following celecoxib treatment. The cells were treated with or without 100␮M celecoxib for 24 h. After treatment, the cell viability was measured by MTT assay. Results were obtained from 6 to 7 experiments and the bar represents the mean± S.E. **p < 0.01, indicates between the control and celecoxib treated samples in the p53-wild type or p53-null HCT116 cells.#_{p < 0.05, indicates between}

the p53-wild type and p53-null HCT116 cells by the control and celecoxib treated samples.

was used as a loading control protein. The roles of p53 and securin on the regulation of cytotoxicity in human colorec-tal cancer cells were investigated by exposure to celecoxib. Celecoxib (40–80␮M, 24 h) significantly induced cancer cell death in the p53-wild type cell lines including RKO, securin-wild type and securin-null HCT116 but not SW480 cells (Fig. 1B). Moreover, the securin-wild type HCT116 cells were similar to the securin-null HCT116 cells on the cyto-toxicity following treatment with celecoxib (Fig. 1B). To further determine the role of p53 on the celecoxib-induced colorectal cancer cell death. We have further compared the p53-wild type and p53-null HCT116 cells on the celecoxib-induced cytotoxicity. As shown inFig. 2, both the p53-wild type and p53-null cells were reduced the cell viability fol-lowing celecoxib treatment. However, the p53-wild type cells were more susceptible to increase ∼25% cell death than the p53-null cells after treatment with 100␮M cele-coxib for 24 h (Fig. 2).

3.2. Celecoxib increases the total p53 protein expression and p53 phosphorylated proteins

Treatment with 40–100␮M celecoxib for 24 h increased total p53 proteins and the phosphorylated proteins of

Fig. 3. The effect of celecoxib on the protein levels of total p53 and various

phospho-p53 in RKO cells. The cells were treated with 0–100␮M celecoxib for 24 h. The total protein extracts were subjected to Western blot anal-ysis using specific anti-p53 and anti-phopho-p53 (serine-15, -20, -46, or -392) antibodies. The relative protein intensities under each blot were the average of three-independent experiments.

p53 at serine-20, -46, and -392 (Fig. 3). The protein phosphorylation of p53 (serine-15) was elevated at the relative higher concentrations of 80–100␮M celecoxib (p < 0.05) (Fig. 3). Celecoxib slightly reduced the phospho-p53 (serine-15) proteins at 40–60␮M treatment but not significant difference after statistic analysis (p > 0.05). ERK-2 protein was used as an internal control, which was not altered by celecoxib. To further vision the local-ization of phosphorylated p53 proteins after treatment with celecoxib, the cells were subjected to immunoflu-orescence staining and confocal microscopy. In these experiments, we therefore selected the concentrations of celecoxib at range 80–100␮M by inducing the high lev-els of phospho-p53 proteins in colorectal cancer cells. The red fluorescence (Cy5) of phospho-p53 (serine-15) pro-teins was increased by celecoxib in RKO cells (Fig. 4). The phosphorylated p53 (serine-15) proteins were concen-trated and located on the nuclei in the celecoxib-treated cells (Fig. 4, arrows). The phosphorylated p53 proteins at serine-20 and -392 were also located on the nuclei by celecoxib treatment (data not shown). However, the pro-tein phosphorylation of p53 (serine-46) was located in the cytoplasma of the celecoxib-induced apoptotic cells (Fig. 5, red color). Celecoxib induced nuclear fragments (Figs. 5 and 6A, arrows) and the disruption of F-actin cytoskeleton and membrane blebbing (Figs. 5 and 6A, stars).

3.3. The existence of p53 in colorectal cancer cells increases apoptosis by celecoxib treatment

To investigate the role of p53 on apoptotic induction in the celecoxib-exposed cells, the p53-wild type and

Fig. 4. The expression and localization of phospho-p53 (serine-15) proteins following treatment with celecoxib in RKO cells. The cells were treated with

or without 80␮M celecoxib for 24 h. After celecoxib treatment, the cells were incubated with rabbit anti-phospho-p53 (serine-15) antibodies and then incubated with goat anti-rabbit Cy5. F-actin and nuclei were stained with BODIPY FL phallacidin and Hoechst 33258, respectively. The arrows indicate phospho-p53 (serine-15) proteins located in the nuclei.

p53-mutant colorectal cancer cells were compared with apoptosis by treatment with celecoxib. Celecoxib induced the cytoskeleton disruption and apoptotic nuclei in RKO cells (Fig. 6A). Counting the apoptotic cell number, cele-coxib (60␮M for 24 h) significantly increased apoptosis in RKO cells; however, the apoptotic cell number was not altered by celecoxib in SW480 cells (Fig. 6B). Moreover, treatment with 40–80␮M celecoxib for 24 h increased the sub-G1fractions via a concentration-dependent manner in

RKO cells but not in SW480 cells by flow cytometry analysis (Fig. 6C). In addition, transfection with a p53 siRNA (50 nM, 24 h) partially restored the celecoxib-induced cell death in RKO cells (Fig. 6D).

3.4. The existence of wild type p53 in colorectal cancer cells elevates the PUMA protein expression following celecoxib treatment

We have examined the effect of celecoxib on the PUMA protein expression in human colorectal cancer cells. Cele-coxib increased PUMA protein expression in RKO cells but not SW480 cells (Fig. 7A). The semi-quantified data from Western blot showed that celecoxib (40–100␮M for 24 h) significantly increased the protein levels of PUMA via a concentration-dependent manner in RKO cells (Fig. 7B). To investigate the expression and location of PUMA proteins relative to mitochondrial damage, the cells

Fig. 5. The expression and localization of phospho-p53 (serine-46) proteins following treatment with celecoxib. RKO cells were treated with or without

100␮M celecoxib for 24 h. After celecoxib treatment, the cells were incubated with rabbit anti-phospho-p53 (serine-46) antibodies and then incubated with goat anti-rabbit Cy5. F-actin and nuclei were stained with BODIPY FL phallacidin and Hoechst 33258, respectively. The arrow indicates the location of nuclear fragments. The stars indicate the location of the disrupted cytoskeleton of F-actin and the membrane blebbing. (For interpretation of the references to color in the citation of figure, the reader is referred to the web version of the article.)

were subjected to immunofluorescence and mitochon-drial staining. Celecoxib increased the green fluorescence intensities of FITC exhibited by PUMA proteins in RKO cells (Fig. 8A). The fluorescence intensities of Mitotracker orange (a mitochondria-sensitive probe) were elevated in the celecoxib-treated cells. The yellow color of the merged pictures indicated the co-localization of PUMA proteins and mitochondria (Fig. 8A, arrows). Similar to Mitotracker orange, celecoxib also increased the fluorescence inten-sity of DiOC6(3) in RKO cells (Fig. 8B). The quantified data showed that treatment with celecoxib (40–100␮M for

24 h) significantly increased the levels of DiOC6(3) intensi-ties in a concentration-dependent manner (Fig. 8B). 4. Discussion

This is the first report providing that celecoxib induces the various phosphorylated sites of p53 proteins, which located on both of cytoplasma and nuclei in human colorectal cancer cells. Celecoxib induced the phosphory-lated p53 proteins (serine-15, -20, and -392) presenting on the nuclei but the phosphorylated p53 (serine-46)

Fig. 6. The effect of celecoxib on apoptosis in colorectal cancer cells. (A) RKO cells were treated with or without 60␮M celecoxib for 24 h. The F-actin and nuclei were stained with BODIPY FL phallacidin and Hoechst 33258, respectively. The arrows indicate the apoptotic nuclei and nuclear fragments. The stars indicate the cytoskeleton disruption of F-actin. (B) The cells were treated with or without 60␮M celecoxib for 24 h. The percentage of apoptosis was scored by the apoptotic cells under a fluorescence microscope. Results were obtained from three experiments and the bar represents the mean± S.E. *p < 0.05 indicates significant difference between control and celecoxib treated samples in RKO cells. (C) After celecoxib treatment, the cells were trypsinized and then subjected to flow cytometry analyses. The populations of sub-G1phases were quantified. *p < 0.05 and **p < 0.01, indicate significant difference

between RKO and SW480 from the same concentrations of celecoxib. D, RKO cells were transfected with 50 nM control siRNA or p53 siRNA for 24 h. After transfection, the cells were treated with or without 60␮M celecoxib for 24 h. The cell viability was measured by MTT assay. *p < 0.05 indicates between control siRNA and p53 siRNA transfection in the celecoxib-treated cells. **p < 0.01 indicates significant difference between untreated and celecoxib treated samples by control siRNA transfection. The bar represents the mean± S.E.

proteins were located on the cytoplasma of apoptotic cells. The diverse phosphorylation sites of p53 have been indicated to play different roles by the cellular stresses [13–15]. The phosphorylation of p53 at serine15 and -20 sites have been shown to cause the activation and stabilization of p53 proteins[13,23,24]; however, the phos-phorylation of p53 at serine-46 mediates the induction of apoptosis[14,25]. The tumor suppressor p53 has been pro-posed as an important target for cancer treatment[26,27]. Development of anticancer strategies by transcription-dependent and transcription-intranscription-dependent p53 functions is useful for cancer therapy [27]. p53 is a transcription

factor, which presents on nuclei to regulate its down-stream gene expression, but the cytoplasmic p53 may be through the transcription-independent proapoptotic effects of p53 on mitochondrial membrane damage for apoptosis[26,27]. We suggest that p53 proteins presenting in the nuclei may regulate its downstream gene expres-sion after treatment with celecoxib; however, p53 locating in the cytoplasma may interact with signal proteins for executing apoptosis. Nevertheless, the precise mechanisms of varying phosphorylated p53 sites induced by cele-coxib on the regulation of apoptosis need to be further determined.

Fig. 7. The effect of celecoxib on PUMA protein expression in the p53-functional and p53-mutational colorectal cancer cells. (A) RKO and SW480 cells were

treated with or without celecoxib for 24 h. The total protein extracts were subjected to Western blot analysis using anti-PUMA and anti-ERK-2 antibodies. (B) The relative protein intensities of PUMA were from Western blot by semi-quantification. *p < 0.05 and **p < 0.01, indicate significant difference between RKO and SW480 from the same concentrations of celecoxib.

The physiologic events of p53 are mediated through its downstream genes[10–12]. PUMA belongs BCL-2 family and is a key mediator to trigger apoptosis via a p53-dependent pathway[16]. PUMA can sensitize cancer cells to anticancer agents[17,18]. We found that celecoxib induced the PUMA protein expression in the p53-functional cells but not in the p53-mutational cells. Active cytochrome c release is associated with early mitochondrial hyperpo-larization that relates to caspase activation and apoptosis [28,29]. Celecoxib induced the activation of casapse-3 and caspase-9 that mediated by mitochondrial pathway in cancer cells[6,30]. We found that celecoxib increased the PUMA proteins, which localized on the mitochon-dria. The fluorescence intensities of Mitotracker orange and DiOC6(3) in colorectal cancer cells were elevated by celecoxib indicating celecoxib-induced mitochondrial hyperpolarization. However, we cannot exclude that cele-coxib may induce the loss of mitochondrial membrane potential during late apoptotic stages. The up-regulation of PUMA by celecoxib has been shown to induce apoptosis in human gastric carcinoma cells[31]. The gene knockdown of PUMA by siRNA inhibited the celecoxib-induced apopto-sis via a mitochondrial dysfunction pathway[31]. We also found that celecoxib induced the capase-3 activation for apoptosis in colorectal cancer cells [8]. Accordingly, our findings suggest that the activation of p53-PUMA may be through the mitochondrial pathway for inducing apoptosis in the celecoxib-treated colorectal cancer cells. In addition

to PUMA, for example, Bax is another important down-stream of p53 by inducing mitochondrial dysfunction for apoptosis[32]. Thus, the effects of Bax and other p53 down-stream proteins on the regulation of celecoxib-induced apoptosis in human colorectal cancer cells need further investigation.

The p53-functional colorectal cancer cells are more vul-nerable than p53-mutational cells to the celecoxib-induced cytotoxicity and apoptosis. Blockade of p53 by transfec-tion with a p53 siRNA reduced the celecoxib-induced cell death. Although statistically significant between control siRNA and p53 siRNA transfection in the celecoxib-treated cells, it only restored ∼8% survival by treatment with p53 siRNA (Fig. 6D). We have compared the p53-wild type and p53-null colorectal cancer cells on the celecoxib-induced cytotoxicity. Celecoxib celecoxib-induced the cytotoxicity in both p53-wild type and p53-null HCT116 cells. The p53-wild type cells were more susceptible to increase ∼25% cell death than the p53-null cells after treatment with celecoxib (Fig. 2). These results indicate that p53 pathway is partially involved in the celecoxib-induced apoptosis in human colorectal cancer cells. Indeed, we have reported that celecoxib induced p38 MAP kinase pathway to induce apoptosis in colorectal cancer cells[8]. Block-age of survivin protein expression by celecoxib caused the capase-3 activation and apoptosis [8]. Additionally, the 3-phosphoinositide-dependent protein kinase-1/Akt sig-naling pathway mediates celecoxib-induced apoptosis in

Fig. 8. The effect of celecoxib on the location of PUMA and the hyperpolarization of mitochondria in colorectal cancer cells. (A) RKO cells were treated

with or without 60␮M celecoxib for 24 h. The cells were incubated with goat anti-PUMA antibody and subsequently incubated with donkey anti-goat FITC. Mitochondria and nuclei were stained with Mitotracker orange and Hoechst 33258, respectively. The yellow color indicated the localization of PUMA on mitochondria (arrows). (B) After treatment with or without celecoxib, the cells were incubated with 500 nM DiOC6(3) and then analyzed by a flow cytometer. The fluorescence intensities of DiOC6(3) were quantified by a CellQuest software. Results were obtained from 4 experiments and the bar represents the mean± S.E. *p < 0.05 and **p < 0.01, indicate significant difference between control and celecoxib treated samples. (For interpretation of the references to color in the citation of figure, the reader is referred to the web version of the article.)

prostate cancer cells[33]. We suggest that celecoxib can induce different signal transduction pathways for inducing apoptosis. Nonetheless, celecoxib elevated the cell death at similar degree in both of the wild type and securin-null colorectal cancer cells. Therefore, celecoxib induces apoptosis via a securin-independent pathway in colorectal cancer cells.

Celecoxib exhibits high selectivity for the COX-2 enzyme and exerts anticarcinogenic and chemopreventive activi-ties[2–4]. However, Pyrko et al. recently reported that the induction of apoptosis by celecoxib via a COX-independent pathway[34]. In addition, celecoxib mediates cell cycle arrest and apoptosis via COX-2-independent pathways in human cholangiocarcinoma and prostate cancer cells

[33,35]. We found that various genotype of human colorec-tal cancer cell lines expressed high levels of COX-2 proteins (Fig. 1A). Moreover, the phosphorylated p53 and apopto-sis were induced at the relative higher concentrations of celecoxib (80–100␮M). Therefore, we suggest that the acti-vation of p53 and apoptosis by celecoxib may result from a COX-2-independent pathway in colorectal cancer cells.

In summary, we provide the p53-PUMA pathway that plays an important role for celecoxib-induced apopto-sis in human colorectal cancer cells. Understanding the mechanisms by which p53-PUMA and other signal trans-duction pathways regulate apoptosis following treatment with celecoxib may contribute to the novel therapeutic strategies in colorectal cancers.

Conflict of interest None declared. Acknowledgements

We thank Dr. B. Vogelstein of Johns Hopkins University for providing the wild type p53 and securin, securin-null, and p53-null HCT116 colorectal cancer cell lines. This work was supported by the grant from NSC 96-2311-B-320-006-MY3.

References

[1] A. Jemal, T. Murray, E. Ward, A. Samuels, R.C. Tiwari, A. Ghafoor, E.J. Feuer, M.J. Thun, Cancer statistics, 2005, CA Cancer J. Clin. 55 (2005) 10–30.

[2] J.J. Keller, F.M. Giardiello, Chemoprevention strategies using NSAIDs and COX-2 inhibitors, Cancer Biol. Ther. 2 (2003) S140–S149. [3] K. Kismet, M.T. Akay, O. Abbasoglu, A. Ercan, Celecoxib: a potent

cyclooxygenase-2 inhibitor in cancer prevention, Cancer Detect. Prev. 28 (2004) 127–142.

[4] G. Steinbach, P.M. Lynch, R.K. Phillips, M.H. Wallace, E. Hawk, G.B. Gordon, N. Wakabayashi, B. Saunders, Y. Shen, T. Fujimura, L.K. Su, B. Levin, The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis, N. Engl. J. Med. 342 (2000) 1946–1952. [5] X. Liu, P. Yue, Z. Zhou, F.R. Khuri, S.Y. Sun, Death receptor regulation

and celecoxib-induced apoptosis in human lung cancer cells, J. Natl. Cancer Inst. 96 (2004) 1769–1780.

[6] H. Ding, C. Han, J. Zhu, C.S. Chen, S.M. D’Ambrosio, Celecoxib deriva-tives induce apoptosis via the disruption of mitochondrial membrane potential and activation of caspase 9, Int. J. Cancer 113 (2005) 803–810.

[7] D.S. Dandekar, M. Lopez, R.I. Carey, B.L. Lokeshwar, Cyclooxygenase-2 inhibitor celecoxib augments chemotherapeutic drug-induced apop-tosis by enhancing activation of caspase-3 and -9 in prostate cancer cells, Int. J. Cancer 115 (2005) 484–492.

[8] P.W. Hsiao, C.C. Chang, H.F. Liu, C.M. Tsai, T.H. Chiu, J.I. Chao, Activation of p38 mitogen-activated protein kinase by celecoxib oppositely reg-ulates survivin and gamma-H2AX in human colorectal cancer cells, Toxicol. Appl. Pharmacol. 222 (2007) 97–104.

[9] C. Brantley-Finley, C.S. Lyle, L. Du, M.E. Goodwin, T. Hall, D. Szwedo, G.P. Kaushal, T.C. Chambers, The JNK, ERK and p53 pathways play distinct roles in apoptosis mediated by the antitumor agents vin-blastine, doxorubicin, and etoposide, Biochem. Pharmacol. 66 (2003) 459–469.

[10] S. Bates, K.H. Vousden, p53 in signaling checkpoint arrest or apopto-sis, Curr. Opin. Genet. Dev. 6 (1996) 12–18.

[11] A.J. Levine, p53, the cellular gatekeeper for growth and division, Cell 88 (1997) 323–331.

[12] L.J. Hofseth, S.P. Hussain, C.C. Harris, p53: 25 years after its discovery, Trends Pharmacol. Sci. 25 (2004) 177–181.

[13] D.W. Meek, Mechanisms of switching on p53: a role for covalent modification? Oncogene 18 (1999) 7666–7675.

[14] K. Oda, H. Arakawa, T. Tanaka, K. Matsuda, C. Tanikawa, T. Mori, H. Nishimori, K. Tamai, T. Tokino, Y. Nakamura, Y. Taya, p53AIP1, a potential mediator of p53-dependent apoptosis, and its regulation by Ser-46-phosphorylated p53, Cell 102 (2000) 849–862. [15] S. Saito, H. Yamaguchi, Y. Higashimoto, C. Chao, Y. Xu, A.J. Fornace Jr.,

E. Appella, C.W. Anderson, Phosphorylation site interdependence of human p53 post-translational modifications in response to stress, J. Biol. Chem. 278 (2003) 37536–37544.

[16] K. Nakano, K.H. Vousden, PUMA, a novel proapoptotic gene, is induced by p53, Mol. Cell 7 (2001) 683–694.

[17] J. Yu, W. Yue, B. Wu, L. Zhang, PUMA sensitizes lung cancer cells to chemotherapeutic agents and irradiation, Clin. Cancer Res. 12 (2006) 2928–2936.

[18] F. Yu, R.N. Watts, X.D. Zhang, J.M. Borrow, P. Hersey, Involve-ment of BH3-only proapoptotic proteins in mitochondrial-dependent phenoxodiol-induced apoptosis of human melanoma cells, Anti-cancer Drugs 17 (2006) 1151–1161.

[19] M.K. Bhat, C. Yu, N. Yap, Q. Zhan, Y. Hayashi, P. Seth, S. Cheng, Tumor suppressor p53 is a negative regulator in thyroid hormone receptor signaling pathways, J. Biol. Chem. 272 (1997) 28989–28993. [20] N.R. Rodrigues, A. Rowan, M.E. Smith, I.B. Kerr, W.F. Bodmer, J.V.

Gan-non, D.P. Lane, p53 mutations in colorectal cancer, Proc. Natl. Acad. Sci. U.S.A. 87 (1990) 7555–7559.

[21] J.I. Chao, H.F. Liu, The blockage of survivin and securin expression increases the cytochalasin B-induced cell death and growth inhibi-tion in human cancer cells, Mol. Pharmacol. 69 (2006) 154–164. [22] E. Bedner, X. Li, W. Gorczyca, M.R. Melamed, Z. Darzynkiewicz,

Anal-ysis of apoptosis by laser scanning cytometry, Cytometry 35 (1999) 181–195.

[23] N. Dumaz, D.W. Meek, Serine15 phosphorylation stimulates p53 transactivation but does not directly influence interaction with HDM2, EMBO J. 18 (1999) 7002–7010.

[24] P.F. Lambert, F. Kashanchi, M.F. Radonovich, R. Shiekhattar, J.N. Brady, Phosphorylation of p53 serine 15 increases interaction with CBP, J. Biol. Chem. 273 (1998) 33048–33053.

[25] L.D. Mayo, Y.R. Seo, M.W. Jackson, M.L. Smith, J. Rivera Guzman, C.K. Korgaonkar, D.B. Donner, Phosphorylation of human p53 at serine 46 determines promoter selection and whether apoptosis is attenuated or amplified, J. Biol. Chem. 280 (2005) 25953–25959.

[26] A.M. Bode, Z. Dong, Post-translational modification of p53 in tumori-genesis, Nat. Rev. Cancer 4 (2004) 793–805.

[27] J.J. Fuster, S.M. Sanz-Gonzalez, U.M. Moll, V. Andres, Classic and novel roles of p53: prospects for anticancer therapy, Trends Mol. Med. 13 (2007) 192–199.

[28] M. Poppe, C. Reimertz, H. Dussmann, A.J. Krohn, C.M. Luetjens, D. Bockelmann, A.L. Nieminen, D. Kogel, J.H. Prehn, Dissipation of potas-sium and proton gradients inhibits mitochondrial hyperpolarization and cytochrome c release during neural apoptosis, J. Neurosci. 21 (2001) 4551–4563.

[29] A. Gross, J.M. McDonnell, S.J. Korsmeyer, BCL-2 family members and the mitochondria in apoptosis, Genes Dev. 13 (1999) 1899–1911. [30] U. Sinha-Datta, J.M. Taylor, M. Brown, C. Nicot, Celecoxib disrupts the

canonical apoptotic network in HTLV-I cells through activation of Bax and inhibition of PKB/Akt, Apoptosis 13 (2008) 33–40.

[31] T. Ishihara, T. Hoshino, T. Namba, K. Tanaka, T. Mizushima, Involve-ment of up-regulation of PUMA in non-steroidal anti-inflammatory drug-induced apoptosis, Biochem. Biophys. Res. Commun. 356 (2007) 711–717.

[32] J.E. Chipuk, T. Kuwana, L. Bouchier-Hayes, N.M. Droin, D.D. Newmeyer, M. Schuler, D.R. Green, Direct activation of Bax by p53 mediates mito-chondrial membrane permeabilization and apoptosis, Science 303 (2004) 1010–1014.

[33] S.K. Kulp, Y.T. Yang, C.C. Hung, K.F. Chen, J.P. Lai, P.H. Tseng, J.W. Fowble, P.J. Ward, C.S. Chen, 3-Phosphoinositide-dependent protein kinase-1/Akt signaling represents a major cyclooxygenase-2-independent target for celecoxib in prostate cancer cells, Cancer Res. 64 (2004) 1444–1451.

[34] P. Pyrko, N. Soriano, A. Kardosh, Y.T. Liu, J. Uddin, N.A. Petasis, F.M. Hofman, C.S. Chen, T.C. Chen, A.H. Schonthal, Downregulation of sur-vivin expression and concomitant induction of apoptosis by celecoxib and its non-cyclooxygenase-2-inhibitory analog, dimethyl-celecoxib (DMC), in tumor cells in vitro and in vivo, Mol. Cancer 5 (2006) 19. [35] C. Han, J. Leng, A.J. Demetris, T. Wu, Cyclooxygenase-2 promotes

human cholangiocarcinoma growth: evidence for cyclooxygenase-2-independent mechanism in celecoxib-mediated induction of p21waf1/cip1 and p27kip1 and cell cycle arrest, Cancer Res. 64 (2004) 1369–1376.