Abstract

Background: The antitumor activity of cyclooxygenase-2 (COX-2) inhibitors is thought to involve COX-2 enzyme inhibition and apoptosis induction, but it is unclear whether COX-2 inhibition is required for apoptosis. Different COX-2 inhibitors have similar IC50 values (concentration for 50% inhibition) for COX-2 inhibition but differ considerably in their abilities to induce apoptosis, suggesting the involvement of a COX-2-independent pathway in apoptosis. To test this hypothesis, we investigated the effect of COX-2 depletion on apoptosis and performed a structure–activity analysis of the COX-2 inhibitor celecoxib in the androgen-independent prostate cancer cell line PC-3. Methods: Tetracycline-inducible (Tet-On) COX-2 antisense clones were isolated to assess the effect of COX-2 expression on cell viability and sensitivity to apoptosis induced by COX-2 inhibitors. Untreated Tet-On clones differentially expressed COX-2, and doxycycline-treated clones were depleted of COX-2. We synthesized and characterized various celecoxib derivatives with various COX-2 inhibitory activities and determined their apoptotic activity in PC-3 cells. Apoptosis was assessed with four tests. Results: In contrast to the effect of COX-2 inhibitors, which induced apoptosis, COX-2 depletion did not induce cell death. Susceptibility to COX-2 inhibitor-induced apoptosis was independent of the level of COX-2 expression. Structure–activity analysis found no correlation between apoptosis induction and COX-2 inhibition. Some celecoxib derivatives that lacked COX-2 inhibitory activity facilitated apoptosis and vice versa. Moreover, celecoxib and apoptosis-active celecoxib derivatives mediated cell death by inhibiting the same pathway. Conclusion: We have dissociated the apoptosis-inducing activity from the COX-2 inhibitory activity by structural modifications of the COX-2 inhibitor celecoxib. This separation of activities may provide a molecular basis for the development of new classes of apoptosis-inducing agents.

The signaling mechanism used by cyclooxygenase-2 (COX-2) inhibitors to mediate apoptotic death in cancer cells has been the focus of many investigations (110). A crucial issue yet to be resolved is whether COX-2 inhibition plays an obligatory role in the induction of apoptosis by COX-2 inhibitors (11). The premise that COX-2 inhibition is integral to this antitumor effect is based on the assumption that prostaglandins and other COX-2-generated downstream mediators promote tumor cell proliferation, survival, and angiogenesis in an autocrine and/or paracrine manner (1216). It has been reported that COX-2 overexpression leads to the inhibition of apoptosis or altered cell cycle kinetics in epithelial cells of the gastrointestinal system (17,18) and in PC-12 pheochromocytoma cells (19). In addition, knockout of the COX-2 gene can suppress tumorigenesis in mice that have a genetic predisposition to form polyps (20). Animal studies have demonstrated that efficient tumor growth requires the presence of COX-2 in the host (8) and that enhanced COX-2 expression in the host was sufficient to induce mammary gland tumorigenesis (21). In contrast, several lines of evidence indicate that a COX-2-independent mechanism may be involved in the antitumor effect of COX-2 inhibitors. For example, sulindac metabolites, which do not inhibit COX activity, were potent inducers of apoptosis in prostate cancer cells (22). In contrast to the situation in gastrointestinal cells and PC-12 cells, COX-2 overexpression in immortalized human umbilical vein endothelial, HEK-293, COS-7, and NIH 3T3 cells led to increased cell death and/or cell cycle arrest (23,24). The dichotomous effect of COX-2 overexpression on cell growth may, in part, be attributable to physiologic differences among different cell types. Moreover, malignant transformation in embryo fibroblasts was reportedly independent of the status of COX expression (9).

Previously, we demonstrated that celecoxib (Celebrex®) induced apoptosis in prostate cancer cells by interfering with multiple signaling targets, including Akt, ERK2, and endoplasmic reticulum Ca2+-ATPases (10,25). Disruption of these signaling pathways leads rapidly to apoptosis, an apoptotic mechanism distinctly different from that of conventional anticancer agents. It is noteworthy that the effect of celecoxib on apoptosis was independent of androgen responsiveness, the level of Bcl-2 expression, and the functional status of p53 in cancer cells (10,25). Nevertheless, this rapid induction of apoptosis was unique to celecoxib, because the abilities of other COX-2 inhibitors, including rofecoxib (Vioxx®), NS398, and DuP697, to induce apoptosis were much lower than that of celecoxib (25). This observation underscores differences in the mechanisms by which these COX-2 inhibitors mediate apoptosis in prostate cancer cells. To determine whether COX-2 inhibitor-induced apoptosis required the inhibition of COX-2 enzyme activity, we examined the effect of COX-2 depletion on apoptosis in PC-3 clones carrying tetracycline-on (Tet-On) antisense COX-2 complementary DNAs (cDNAs) and performed a structure–activity analysis of various celecoxib derivatives in androgen-independent PC-3 cells.

Materials and Methods

Cells and Reagents

We used the following three different prostate cancer cell lines to assess the impact of androgen responsiveness and p53 functional status on the induction of apoptosis by celecoxib and its derivatives: androgen-responsive LNCaP (p53+/+), androgen-nonresponsive PC-3 (p53–/–), and DU-145 (p53–/–). The antisense COX-2 construct was a gift from Drs. Rebecca Chinery and Jason Morrow (Vanderbilt University Medical School, Nashville, TN). It contained an almost complete human COX-2 cDNA insert (1.93 kilobases) that was cloned into the XbaI/EcoRV sites in the tetracycline response element (TRE)-response plasmid pUHD.2neo (26). This Tet-On antisense COX-2 construct has been used in colorectal cancer cells to assess the role of prostaglandins in cell proliferation (26). Celecoxib and rofecoxib were from commercial Celebrex® and Vioxx® (Amerisource Health, Malvern, PA) capsules, respectively, by solvent extraction followed by recrystallization. DuP697 was a gift from Professor Hsin-Hsiung Tai (University of Kentucky, Lexington), and NS398 was from Calbiochem (La Jolla, CA). The following compounds were synthesized according to published procedures (27): 4-[5-(4-chlorophenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 1), 4-[5-phenyl-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 2), 4-[5-(4-aminophenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 3), 4-[5-(4-ethylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 4), 4-[5-(4-trifluoromethylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 5), 4-[5-(2,5-dichlorophenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 6), and 4-[5-(2,5-dimethylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (compound 7). Rabbit anti-COX-2 antibodies were from Cayman Chemicals (Ann Arbor, MI). Rabbit polyclonal antibodies against Akt, phospho-Ser473-Akt, ERK, and phospho-ERK were from New England Biolabs (Beverly, MA), and mouse anti-actin monoclonal antibody was from ICN Pharmaceuticals (Costa Mesa, CA). Goat anti-rabbit immunoglobulin G (IgG)-horseradish peroxidase conjugates were from Jackson ImmunoResearch Laboratories (West Grove, PA). Rabbit anti-poly(ADP-ribose) polymerase (PARP) antibodies were from BD PharMingen (San Diego, CA).

Development of PC-3 Tet-On Antisense COX-2 Clones

PC-3 cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) in T-25 flasks at 37 °C in a humidified CO2 incubator to 80% confluence. Each flask was washed with 6 mL of serum-free Opti-MEM (Invitrogen Life Technologies, Carlsbad, CA), and then 3 mL of serum-free Opti-MEM was added. Aliquots containing 0.12 μg of the Tet-On regulator plasmid pTet-On (Clontech, Palo Alto, CA) and 0.12 μg of the antisense COX-2 construct in 150 μL of serum-free Opti-MEM medium were preincubated with 3 μL of the Plus reagent from the LipofectAMINE Plus Reagent kit (Invitrogen Life Technologies) at 25 °C for 15 minutes, followed by 12 μL of the LipofectAMINE reagent in 150 μL of Opti-MEM medium. The resulting mixture was incubated at 25 °C for 15 minutes and then added to each flask with gentle mixing. After 5 hours at 37 °C, the transfection medium was replaced with 5 mL of RPMI-1640 medium containing 10% Tet-System-approved FBS (Clontech). After 48 hours, cells were cultured in fresh medium containing G418 (Calbiochem, La Jolla, CA) at 100 μg/mL to select for transfected clones. The G418-supplemented medium was changed every 4 days. After 3 weeks, G418-resistant cells were subcloned into 96-well plates by limiting dilution with a final cell density of approximately 0.5 cell per well. After 12 days with a change of G418-containing medium every 4 days, viable clones were further subcloned into 12-well plates. After 4 or 5 days, cells in each well were divided into three T-25 flasks. The level of COX-2 expression was determined 120 hours after cells were exposed to doxycycline (2 μg/mL) by western blot analysis. By this procedure, the following four independent clones (2F6, 1F2, 3D9, and 7D9) were selected for analyses; 2F6 was a COX-2-deficient clone, and 1F2, 3D9, and 7D9 expressed different levels of COX-2 in the absence of doxycycline.

Immunoblotting

For western blot analysis, cells were washed in phosphate-buffered saline (PBS), resuspended in sodium dodecyl sulfate (SDS) gel-loading buffer (50 mM Tris–HCl [pH 6.8], 100 mM dithiothreitol, 2% SDS, 0.1% bromophenol blue, and 10% glycerol), sonicated with an ultrasonic sonicator for 5 seconds (Virsonic 300, 4.5 output; VirTis, Gardiner, NY), and boiled for 5 minutes. After a brief centrifugation, equivalent protein amounts (60–100 μg) from the soluble fractions were resolved in 10% SDS–polyacrylamide gels on a Minigel (Amersham Pharmacia, Piscataway, NJ) apparatus and transferred to a nitrocellulose membrane in a semidry transfer cell. The transblotted membrane was washed twice with Tris-buffered saline (TBS) containing 0.05% Tween 20 (TBST). After blocking with TBS containing 5% nonfat milk for 60 minutes, the membrane was incubated with the appropriate primary antibody (anti-COX-2, anti-Akt, anti-P-473Ser Akt, anti-ERK, and anti-phospho-ERK antibodies diluted 1 : 1000; anti-actin monoclonal antibody, diluted 1 : 5000) in TBS–1% nonfat milk at 4 °C for 12 hours and washed twice with TBST. The membranes were probed with goat anti-rabbit IgG-horseradish peroxidase conjugates (diluted 1 : 5000) for 1 hour at room temperature and washed twice with TBST. Bands were visualized by enhanced chemiluminescence.

Prostaglandin E2 (PGE2) Immunoassay

Parental and transfected cells, with or without a doxycycline (2 μg/mL) pretreatment, were grown to 10 × 106 cells in T-75 flasks in RPMI-1640 medium containing 10% FBS. Culture medium was changed, and 24 hours later, conditioned medium was collected to assay PGE2. Conditioned medium was centrifuged to remove particulate material, and then attached cells were collected by scraping to determine the protein concentration. PGE2 was assayed in 100 μL of medium in triplicate, according to the manufacturer's instruction (R&D Systems, Minneapolis, MN). PGE2 data were normalized to protein content.

Cell Viability

Parental or transfected PC-3 cells with or without doxycycline (2 μg/mL) pretreatment were plated in 12-well plates and cultured in RPMI-1640 medium supplemented with 10% FBS in the absence or presence of doxycycline (2 μg/mL) for 48 hours. COX-2 inhibitors (at various concentrations) were dissolved in dimethyl sulfoxide (DMSO; final concentration = 0.1% after addition to medium) and were then added to the cells in serum-free RPMI-1640 medium. Control cells received DMSO at the same concentration. During treatment, the percentage of floating cells increased over time. At the end of the treatment, adherent cells were harvested by trypsinization, and floating cells were recovered by centrifugation at 3200g for 5 minutes. Cell morphology was assessed with a light microscope at ×400. Both adherent and floating cells were combined, and cell viability was assessed by trypan blue dye exclusion.

Analysis of Apoptosis

In addition to the morphologic changes of intact cells observed by phase-contrast microscopy, four methods were used to assess drug-induced apoptotic cell death.

Phosphatidylserine externalization.

Approximately 2.5 × 105 cells were grown on glass coverslips for 24 hours. At various times after drug treatment, cells were washed gently with PBS and then exposed to 0.5 mL of binding buffer (10 mM HEPES [pH 7.4], containing 150 mM NaCl, 2.5 mM CaCl2, 1 mM MgCl2, and 4% bovine serum albumin), followed by 0.6 mL of annexin V-fluorescein isothiocynate (FITC) (200 μg/mL) for 30 minutes. After washing with binding buffer, apoptotic cells were identified directly as cells with annexin V-FITC on their outer membrane by fluorescence microscopy. In a set of controls, cells received medium containing DMSO vehicle in lieu of the test agent.

4`,6-Diamidino-2-phenylindole (DAPI) staining of nuclei.

At various times after treatment with various test agents, morphologic changes were detected in the nuclei of apoptotic cells by staining with the DNA-binding fluorochrome DAPI. For adherent PC-3 cells, cells were grown on glass coverslips until approximately 70% confluent and exposed to the test agent at 50 μM for various times. Supernatants then were carefully removed, adherent cells were washed with PBS, DAPI (0.5 μg/mL) was added in a fixation solution (4% paraformaldehyde, 2 mM EGTA [ethylene glycol bis(β-aminoethyl ether)-N,N,N`,N`-tetraacetic acid], and 13.7% sucrose in PBS), and the mixture was incubated at room temperature for 20 minutes in the dark. Cells were then washed for two 20-minute periods with PBS. Floating PC-3 cells were examined by a modification of the above method as follows. PC-3 cells were cultured in T-25 flasks and treated with the test agent. Floating cells then were collected, washed, and stained with DAPI as described above. Cells were allowed to attach to poly-l-lysine-coated coverslips and viewed by microscopy at a magnification of ×400.

Apoptosis detection by an enzyme-linked immunosorbent assay (ELISA).

Induction of apoptosis was also assessed by using a Cell Death Detection ELISA (Roche Diagnostics, Mannheim, Germany) by following the manufacturer's instructions. This test is based on the quantitative determination of cytoplasmic histone-associated DNA fragments in the form of mononucleosomes and oligonucleosomes after induced apoptotic death. In brief, 2.5 × 106 PC-3 cells were cultured in a T-75 flask 24 hours before the experiment. Cells were washed twice in 5 mL of serum-free RPMI-1640 medium and then treated with a test agent or the DMSO vehicle, as indicated. Both floating and adherent cells were collected, and cell lysates equivalent to 104 cells were used in the ELISA.

Western blot analysis of PARP cleavage.

Drug-treated cells were collected, washed with ice-cold PBS, and resuspended in lysis buffer [20 mM Tris–HCl (pH 8), 137 mM NaCl, 1 mM CaCl2, 10% glycerol, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 100 μM 4-(2-aminoethyl)benzenesulfonyl fluoride, leupeptin at 10 μg/mL, and aprotinin at 10 μg/mL]. Soluble cell lysates were collected after centrifugation at 1500g for 5 minutes. Equivalent amounts of protein (60–100 μg) from each lysate were resolved in 10% SDS–polyacrylamide gels. Bands were transferred to nitrocellulose membranes and analyzed by immunoblotting with anti-PARP antibodies, as described above.

Statistical Analysis

Each experiment was performed in triplicate. All experiments were carried out at least two times on different occasions. Where appropriate, the data are presented as the mean ± 95% confidence interval.

Results

Isolation and Characterization of Tet-On Antisense COX-2 Clones

To assess the effect of COX-2 enzyme activity on cell growth, we prepared Tet-On antisense COX-2 clones by transfecting parental PC-3 cells with an antisense COX-2 cDNA construct under the control of a tetracycline-inducible promoter. Four Tet-On antisense COX-2 clones, 2F6, 1F2, 3D9, and 7D9, were selected after subcloning the G418-resistant cells by limiting dilution. 2F6 is a COX-2-deficient clone in which the COX-2 protein was virtually undetectable, and 1F2, 3D9, and 7D9 are antisense COX-2 clones that express COX-2 protein at different levels. In 1F2, 3D9, and 7D9 in the absence of doxycycline, COX-2 is expressed, but in the presence of doxycycline, COX-2 is depleted by antisense inhibition (Fig. 1, B). Fig. 1, A, shows a western blot analysis of the effect of doxycycline (2 μg/mL) on COX-2 expression in the Tet-On clone 3D9 over a 10-day period. By 4 days after the addition of doxycycline, COX-2 protein had been depleted in 3D9 cells, and as long as doxycycline was present, COX-2 was not detected. Because COX-1 expression was negligible in these clones (data not shown), COX-2 produced most of the prostaglandins. Accordingly, the level of COX-2 expression reflected the level of PGE2 production. Although the levels of PGE2 in 1F2 cells and parental PC-3 cells were comparable, those in 3D9 were twofold higher and those in 7D9 cells were 10-fold higher than levels in parental cells (Fig. 1, C). However, when treated with rofecoxib, NS398, or DuP697 at 50 μM, PGE2 production was reduced to the same extent in 1F2, 3D9, 7D9, and parental PC-3 cells (data not shown; it could not be determined in celecoxib-treated cells whether PGE2 was produced by rapid apoptosis). Because the decreased expression of the antiapoptotic protein Bcl-2 has been implicated in the apoptotic mechanism of the COX-2 inhibitors SC-58125 and NS398 (1,2), it is also noteworthy that COX-2 ablation had essentially no impact on the expression of Bcl-2 (Fig. 1, D).

Effect of COX-2 Ablation on Apoptosis in Prostate Cancer Cells

Using these COX-2 antisense clones, we obtained two lines of evidence that the effect of COX-2 inhibitors on apoptosis was independent of their COX-2-inhibitory activity. First, although both antisense COX-2 cDNA and COX-2 inhibitors completely blocked prostaglandin production, their effects on cell viability were markedly different. Treatment of PC-3 cells or any of the four clones with individual COX-2 inhibitors led to apoptotic death, whereas depletion of COX-2, and thus inhibition of PGE2 production with the antisense cDNA, did not adversely affect the viability of these antisense clones, i.e., this treatment did not induce cell death. Second, although the basal levels of COX-2 in these four clones varied, all clones and parental PC-3 cells were equally susceptible to apoptosis induced by COX-2 inhibitors, and this susceptibility did not change after doxycycline-induced COX-2 depletion. In other words, susceptibility to COX-2-inhibitor-induced apoptosis was independent of the level of COX-2 expression. Fig. 2, A, shows the time course of cell death in the presence of 50 μM celecoxib in PC-3 cells, COX-2-deficient 2F6 cells, and COX-2-overexpressing 7D9 cells with and without COX-2 depletion, i.e., grown in the presence and absence of doxycycline, respectively. For all three of these clones incubated with 50 μM celecoxib, the time required for 50% cell death (T1/2) was approximately 2 hours. Similar times were noted with 1F2 and 3D9 cells.

Previously, we demonstrated that celecoxib induced rapid apoptotic death in both androgen-responsive LNCaP (p53+/+) and androgen-nonresponsive PC-3 (p53–/–) prostate cancer cells by inhibiting the Akt and ERK signaling pathways (10,25). In this study, the apoptotic death induced in all four clones was also associated with decreased phosphorylation of Akt and ERK2, as observed in parental PC-3 cells. In addition, the time course for the dephosphorylation of Akt and ERK2 in 7D9 cells (Fig. 2, B) was consistent with that for cell death. Similar results were obtained with the three other clones.

The effect of COX-2 depletion on apoptosis induced by three other COX-2 inhibitors—rofecoxib, NS398, and DuP697—was also examined in the COX-2 antisense clones. These compounds triggered apoptosis in parental PC-3 cells by a mechanism distinctly different from that induced by celecoxib (25). The onset of apoptosis was substantially delayed compared with that induced by celecoxib, with T1/2 values for rofecoxib, NS398, or DuP697 (each at 100 μM) ranging from 96 hours to 120 hours. Fig. 3 shows the cell viability after exposure of 7D9 cells to 100 μM rofecoxib with or without pretreatment with doxycycline. In line with the celecoxib data, loss of COX-2 expression did not alter the susceptibility of these clones to the induction of apoptosis by rofecoxib. Similar results were obtained with NS398 and DuP697 (data not shown).

Structure–Activity Analysis

Structural modifications of celecoxib were carried out to dissociate COX-2 inhibition and the induction of apoptosis. We synthesized a series of celecoxib derivatives with different substituents at the terminal phenyl ring and examined the apoptosis-inducing potency of each. Fig. 4 summarizes the structures, the COX-2 inhibitory activity (IC50 = concentration of drug-inhibiting COX-2 activity by 50%), and the apoptosis-inducing activity (T1/2) of celecoxib and seven representative analogues.

The structure–activity analysis found no association between the COX-2 inhibitory and apoptosis-inducing activities. Increased polarity (i.e., 4-aminoin compound 3) or bulkiness (i.e., 4-ethyl-, 4-trifluoromethyl-, 2,5-dichloro-, and 2,5-dimethylin compounds 47, respectively) of the terminal phenyl ring reduced the ability of these compounds to inhibit COX-2 activity. In contrast, a certain degree of bulkiness and hydrophobicity in the substituted phenyl ring was essential to the apoptosis-inducing activity. For example, compound 2, in which the 4-methyl moiety of celecoxib was replaced by a hydrogen atom, was a highly potent COX-2 inhibitor (IC50 = 32 nM) but lacked apoptosis-inducing activity (T1/2 >100 h). Conversely, compounds 6 and 7 had no COX-2 inhibitory activity (IC50>100 μμ for each) but were highly potent mediators of apoptotic death in PC-3 cells. The time- and dose-dependent effect of compound 7 on cell viability is shown in Fig. 5, A. Exposure of cells to 50 μM compound 7 resulted in a 50% decrease in cell viability within 1 hour compared with 2 hours for celecoxib. Characteristics of apoptotic death in PC-3 cells are shown in Fig. 5, B–D. Fig. 5, B, shows pronounced changes in morphology and membrane compositions after treatment with compound 7. As shown by phase-contrast microscopy, treated cells shrank, rounded, and detached from the dish, and bleb formation was evident 1 hour after compound 7 was added (left panel). Morphologic evidence of apoptosis was assessed as nuclear fragmentation detected by staining cells with DAPI, and cells treated with compound 7 were found to have condensed and fragmented nuclei (center panel). Another test for apoptosis induced by compound 7 was the externalization of phosphatidylserine as detected by FITC-conjugated annexin V and fluorescence microscopy (right panel). Among other parameters, degradation of DNA to nucleosomal fragments and cleavage of PARP to the apoptosis-specific 85-kDa fragment are also well characterized events of apoptotic cell death. Results of these tests are shown in Fig. 5, C and D, respectively.

Moreover, it is noteworthy that the mechanism used by compounds 1 and 47 to facilitate apoptosis was the same mechanism used by the parent compound celecoxib, i.e., concurrent dephosphorylation of Akt and ERK2 independent of Bcl-2. Fig. 6, A–C, illustrates the effect of compound 7 on the expression level of Bcl-2 and the phosphorylation status of Akt and ERK2, respectively, in PC-3 cells, which was reminiscent of that observed with celecoxib.

In addition to PC-3 cells, we also tested the effect of compounds 1 and 47 on apoptosis in the androgen-responsive LNCaP (p53+/+) and the androgen-independent DU-145 (p53–/–) prostate cancer cell lines. Apoptosis was induced by celecoxib and compounds 1 and 47 in LNCaP and DU-145 cells with susceptibility identical to that of PC-3 cells (data not shown). Thus, the mechanism by which these active compounds mediate apoptosis appears to be independent of androgen status, p53 functional status, and the level of Bcl-2 expression.

Discussion

Although the clinical relevance of COX-2 inhibitors in chemoprevention has been demonstrated, the antitumor mechanism used by these compounds is not well defined (11). Mechanisms involving different signaling targets have been proposed to account for nonsteroidal anti-inflammatory drug-induced apoptosis in cancer cells. These putative mechanisms include inhibition of Bcl-2 expression (1,2), accumulation of arachidonic acid (28), stimulation of ceramide production (3), dephosphorylation of Akt and ERK2 proteins (10,25), inhibition of PPARδ (peroxisome proliferator-activated receptor) (29), and interference with the nuclear factor NF-κB signaling pathway (30). It is plausible that different COX-2 inhibitors mediate apoptosis via distinct mechanisms. However, whether COX-2 inhibition plays an obligatory role in the apoptotic effect of COX-2 inhibitors is yet to be resolved. In this study, we used COX-2 depletion via inducible expression of an antisense COX-2 cDNA and structure– activity analysis of celecoxib derivatives to address this issue.

Use of the Tet-On antisense COX-2 clones was advantageous because these clones represented syngeneic cell lines, thereby abating concerns about the effects of genetic variations among different cell lines, and they displayed differential COX-2 expression that could be shut off by doxycycline treatment. We determined that the effect of COX-2 inhibitors on apoptosis was independent of the COX-2 inhibitory activity. Although both COX-2 depletion and COX-2 inhibitors inhibited PGE2 production, these two treatments had very different effects on cell viability. Treatment with COX-2 inhibitors led to cell death, but COX-2 depletion did not. The sensitivity to COX-2-inhibitor-induced apoptosis was independent of the level of COX-2 expression in the antisense clones and was, in fact, unaltered from that of parental PC-3 cells.

We were able to dissociate the apoptosis-inducing activity of celecoxib from the COX-2 inhibitory activity via structural modification of this drug. This finding is reminiscent of the previous report that sulindac metabolites—sulindac sulfide and sulindac sulfone—could induce apoptosis in prostate cancer cells via a COX-independent mechanism (22). Several celecoxib derivatives, although lacking COX-2 inhibitory activity, were as potent in eliciting apoptosis in PC-3 cells as the parent compound. These compounds induce apoptosis in both hormone-responsive and hormone-nonresponsive prostate cancer cells. Furthermore, the mechanism by which these compounds and the parental compound celecoxib induce apoptosis remained the same, i.e., facilitating the dephosphorylation of Akt and ERK2.

From a clinical perspective, the separation of the COX-2 inhibitory activity of celecoxib from the apoptosis-inducing effect provides a molecular basis for the design of new classes of apoptosis-inducing agents. It is noteworthy that these molecules mediate apoptosis independent of the cell's androgen sensitivity, p53 functional status, and level of Bcl-2 expression. These features make these apoptosis-inducing agents potential candidates for the prevention and treatment of human prostate cancer.

Fig. 1.

Characterization of antisense cyclooxygenase-2 (COX-2) PC-3 transfectants. A) Western blot analysis showing the time course of COX-2 depletion in the representative antisense COX-2 clone 3D9 in response to doxycycline. (Results from other clones were similar.) Cells were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum in the presence of doxycycline at 2 μg/mL, as indicated. B) Western blot analysis showing the COX-2 protein levels in parental PC-3 cells and four independent antisense COX-2 clones in the presence (+) or absence (–) of doxycycline (Dox; 2 μg/mL) for 10 days. C) Prostaglandin E2 (PGE2) production in parental PC-3 cells and four antisense COX-2 clones without (–) or with (+) doxycycline pretreatment, as indicated.D) Level of Bcl-2 expression in parental PC-3 cells, 2F6 cells (no doxycycline treatment), and 7D9 cells in the presence (+) or absence (–) of doxycycline (2 μg/mL) for 10 days. Data are expressed as means ± 95% confidence intervals (error bars) (n = 3). Western blots are representative of three experiments, all with similar results.

Fig. 1.

Characterization of antisense cyclooxygenase-2 (COX-2) PC-3 transfectants. A) Western blot analysis showing the time course of COX-2 depletion in the representative antisense COX-2 clone 3D9 in response to doxycycline. (Results from other clones were similar.) Cells were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum in the presence of doxycycline at 2 μg/mL, as indicated. B) Western blot analysis showing the COX-2 protein levels in parental PC-3 cells and four independent antisense COX-2 clones in the presence (+) or absence (–) of doxycycline (Dox; 2 μg/mL) for 10 days. C) Prostaglandin E2 (PGE2) production in parental PC-3 cells and four antisense COX-2 clones without (–) or with (+) doxycycline pretreatment, as indicated.D) Level of Bcl-2 expression in parental PC-3 cells, 2F6 cells (no doxycycline treatment), and 7D9 cells in the presence (+) or absence (–) of doxycycline (2 μg/mL) for 10 days. Data are expressed as means ± 95% confidence intervals (error bars) (n = 3). Western blots are representative of three experiments, all with similar results.

Fig. 2.

Susceptibility of prostate cancer cells to celecoxib-induced apoptosis is independent of cyclooxygenase-2 (COX-2) expression levels. A) Left panel. Effect of 50 μM celecoxib on the viability of parental PC-3 cells and the COX-2-deficient clone 2F6. Right panel. Effect of 50 μM celecoxib on the viability of the COX-2 antisense clone 7D9 with (+) or without (–) a doxycycline (Dox) pretreatment (2 mg/mL). Data represent means ± 95% confidence intervals (error bars) (n = 3). B) Effect of celecoxib on the phosphorylation status of Akt and ERK2 in the COX-2 antisense clone 7D9 with or without doxycycline treatment (2 mg/mL), as indicated. Western blots are representative of three independent experiments. These data indicate that the mechanism underlying celecoxib-induced apoptotic death in the 7D9 cells remained unaltered after COX-2 depletion.

Fig. 2.

Susceptibility of prostate cancer cells to celecoxib-induced apoptosis is independent of cyclooxygenase-2 (COX-2) expression levels. A) Left panel. Effect of 50 μM celecoxib on the viability of parental PC-3 cells and the COX-2-deficient clone 2F6. Right panel. Effect of 50 μM celecoxib on the viability of the COX-2 antisense clone 7D9 with (+) or without (–) a doxycycline (Dox) pretreatment (2 mg/mL). Data represent means ± 95% confidence intervals (error bars) (n = 3). B) Effect of celecoxib on the phosphorylation status of Akt and ERK2 in the COX-2 antisense clone 7D9 with or without doxycycline treatment (2 mg/mL), as indicated. Western blots are representative of three independent experiments. These data indicate that the mechanism underlying celecoxib-induced apoptotic death in the 7D9 cells remained unaltered after COX-2 depletion.

Fig. 3.

Effect of rofecoxib and dimethyl sulfoxide (DMSO) vehicles on the cell viability of the cyclooxygenase-2 (COX-2) antisense clone 7D9 with or without doxycycline (Dox) treatment (2 mg/mL). Data are the means ± 95% confidence intervals (error bars) (n = 3).

Fig. 3.

Effect of rofecoxib and dimethyl sulfoxide (DMSO) vehicles on the cell viability of the cyclooxygenase-2 (COX-2) antisense clone 7D9 with or without doxycycline (Dox) treatment (2 mg/mL). Data are the means ± 95% confidence intervals (error bars) (n = 3).

Fig. 4.

Structures and characteristics of celecoxib and compounds 17. The general structure of these molecules is shown at the top. COX-2 = cyclooxygenase-2; IC50 = concentration inhibiting COX-2 activity by 50%; T1/2 = time required for 50% cell death. IC50 values were from (27).

Fig. 4.

Structures and characteristics of celecoxib and compounds 17. The general structure of these molecules is shown at the top. COX-2 = cyclooxygenase-2; IC50 = concentration inhibiting COX-2 activity by 50%; T1/2 = time required for 50% cell death. IC50 values were from (27).

Fig. 5.

Characteristics of compound 7. A) Dose- and time-dependent effect of compound 7 on the cell viability of parental PC-3 cells. Data are the means ± 95% confidence intervals (error bars) (n = 3–6). B) Effect of compound 7 on the morphology and membrane composition of PC-3 cells. PC-3 cells were grown on coverslips and treated with dimethyl sulfoxide (DMSO) vehicles (upper panels) or 50 μM compound 7 (lower panels) for 1–2 hours, as indicated below. Left panel. Phase-contrast micrograph of PC-3 cells 2 hours after treatment. Center panel. Nuclear fragmentation viewed after DNA staining with 4`,6-diamidino-2-phenylindole (DAPI) with the use of fluorescence microscopy 2 hours after treatment. Condensed and fragmented nuclei (arrow) were observed in drug-treated cells. Right panel. Detection of annexin V binding to the surface of apoptotic cells by fluorescence microscopy 1 hour after treatment. Note that some blebbing cells show strong annexin V labeling of their surface. C) Time course of the formation of nucleosomal DNA in PC-3 cells treated with DMSO vehicles or compound 7 (50 μM). The formation of nucleosomes was quantitatively measured by Cell Death Detection ELISA (enzyme-linked immunosorbent assay) with lysates equivalent to 104 cells for each assay. Data are the means ± 95% confidence intervals (error bars) (n = 3). D) Induction of poly(ADP-ribose) polymerase (PARP) cleavage by compound 7 in PC-3 cells. PC-3 cells were treated with 50 μM compound 7, as indicated. PARP proteolysis of the 116-kDa native enzyme to the apoptosis-specific 85-kDa fragment was monitored by western blotting.

Fig. 5.

Characteristics of compound 7. A) Dose- and time-dependent effect of compound 7 on the cell viability of parental PC-3 cells. Data are the means ± 95% confidence intervals (error bars) (n = 3–6). B) Effect of compound 7 on the morphology and membrane composition of PC-3 cells. PC-3 cells were grown on coverslips and treated with dimethyl sulfoxide (DMSO) vehicles (upper panels) or 50 μM compound 7 (lower panels) for 1–2 hours, as indicated below. Left panel. Phase-contrast micrograph of PC-3 cells 2 hours after treatment. Center panel. Nuclear fragmentation viewed after DNA staining with 4`,6-diamidino-2-phenylindole (DAPI) with the use of fluorescence microscopy 2 hours after treatment. Condensed and fragmented nuclei (arrow) were observed in drug-treated cells. Right panel. Detection of annexin V binding to the surface of apoptotic cells by fluorescence microscopy 1 hour after treatment. Note that some blebbing cells show strong annexin V labeling of their surface. C) Time course of the formation of nucleosomal DNA in PC-3 cells treated with DMSO vehicles or compound 7 (50 μM). The formation of nucleosomes was quantitatively measured by Cell Death Detection ELISA (enzyme-linked immunosorbent assay) with lysates equivalent to 104 cells for each assay. Data are the means ± 95% confidence intervals (error bars) (n = 3). D) Induction of poly(ADP-ribose) polymerase (PARP) cleavage by compound 7 in PC-3 cells. PC-3 cells were treated with 50 μM compound 7, as indicated. PARP proteolysis of the 116-kDa native enzyme to the apoptosis-specific 85-kDa fragment was monitored by western blotting.

Fig. 6.

Time-dependent effect of compound 7 on Bcl-2 expression levels (A), the phosphorylation status of Akt (B), and the phosphorylation status of ERK2 (C) in PC-3 cells. The western blots are representative of three independent experiments, all with similar results.

Fig. 6.

Time-dependent effect of compound 7 on Bcl-2 expression levels (A), the phosphorylation status of Akt (B), and the phosphorylation status of ERK2 (C) in PC-3 cells. The western blots are representative of three independent experiments, all with similar results.

Supported by Public Health Service grants CA92307 and CA94829 to C.-S. Chen, National Cancer Institute, National Institutes of Health, Department of Health and Human Services.
X. Song and H.-P. Lin contributed equally to this paper.

We thank Dr. Rebecca Chinery and Dr. Jason Morrow at Vanderbilt University Medical School for providing the antisense COX-2 cDNA construct and Dr. Hsin-Hsiung Tai at the University of Kentucky for providing DuP697.

References

1
Sheng H, Shao J, Morrow JD, Beauchamp RD, DuBois RN. Modulation of apoptosis and Bcl-2 expression by prostaglandin E2 in human colon cancer cells.
Cancer Res
 
1998
;
58
:
362
–6.
2
Liu XH, Yao S, Kirschenbaum A, Levine AC. NS398, a selective cyclooxygenase-2 inhibitor, induces apoptosis and down-regulates bcl-2 expression in LNCaP cells.
Cancer Res
 
1998
;
58
:
4245
–9.
3
Chan TA, Morin PJ, Vogelstein B, Kinzler KW. Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis.
Proc Natl Acad Sci U S A
 
1998
;
95
:
681
–6.
4
Piazza GA, Rahm AL, Krutzsch M, Sperl G, Paranka NS, Gross PH, et al. Antineoplastic drugs sulindac sulfide and sulfone inhibit cell growth by inducing apoptosis.
Cancer Res
 
1995
;
55
:
3110
–6.
5
Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Staiano-Coico L, et al. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway.
Biochem Pharmacol
 
1996
;
52
:
237
–45.
6
Thompson HJ, Jiang C, Lu J, Mehta RG, Piazza GA, Paranka NS, et al. Sulfone metabolite of sulindac inhibits mammary carcinogenesis.
Cancer Res
 
1997
;
57
:
267
–71.
7
Jones MK, Wang H, Peskar BM, Levin E, Itani RM, Sarfeh IJ, et al. Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implications for cancer growth and ulcer healing.
Nat Med
 
1999
;
5
:
1418
–23.
8
Williams CS, Tsujii M, Reese J, Dey SK, DuBois RN. Host cyclooxygenase-2 modulates carcinoma growth.
J Clin Invest
 
2000
;
105
:
1589
–94.
9
Zhang X, Morham SG, Langenbach R, Young DA. Malignant transformation and antineoplastic actions of nonsteroidal antiinflammatory drugs (NSAIDs) on cyclooxygenase-null embryo fibroblasts.
J Exp Med
 
1999
;
190
:
451
–9.
10
Hsu AL, Ching TT, Wang DS, Song X, Rangnekar VM, Chen CS. The cyclooxygenase-2 inhibitor celecoxib induces apoptosis by blocking Akt activation in human prostate cancer cells independently of Bcl-2.
J Biol Chem
 
2000
;
275
:
11397
–403.
11
Marx J. Cancer research. Anti-inflammatories inhibit cancer growth—but how?
Science
 
2001
;
291
:
581
–2.
12
Hla T, Ristimaki A, Appleby S, Barriocanal JG. Cyclooxygenase gene expression in inflammation and angiogenesis.
Ann N Y Acad Sci
 
1993
;
696
:
197
–204.
13
Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, et al. Cyclooxygenase in biology and disease.
FASEB J
 
1998
;
12
:
1063
–73.
14
Taketo MM. Cyclooxygenase-2 inhibitors in tumorigenesis (part I).
J Natl Cancer Inst
 
1998
;
90
:
1529
–36.
15
Hla T, Bishop-Bailey D, Liu CH, Schaefers HJ, Trifan OC. Cyclooxygenase-1 and -2 isoenzymes.
Int J Biochem Cell Biol
 
1999
;
31
:
551
–7.
16
Prescott SM, Fitzpatrick FA. Cyclooxygenase-2 and carcinogenesis.
Biochim Biophys Acta
 
2000
;
1470
:
M69
–78.
17
Tsujii M, DuBois RN. Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2.
Cell
 
1995
;
83
:
493
–501.
18
DuBois RN, Shao J, Tsujii M, Sheng H, Beauchamp RD. G1 delay in cells overexpressing prostaglandin endoperoxide synthase-2.
Cancer Res
 
1996
;
56
:
733
–7.
19
McGinty A, Chang YW, Sorokin A, Bokemeyer D, Dunn MJ. Cyclooxygenase-2 expression inhibits trophic withdrawal apoptosis in nerve growth factor-differentiated PC12 cells.
J Biol Chem
 
2000
;
275
:
12095
–101.
20
Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, et al. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2).
Cell
 
1996
;
87
:
803
–9.
21
Liu CH, Chang SH, Narko K, Trifan OC, Wu MT, Smith E, et al. Overexpression of cyclooxygenase-2 is sufficient to induce tumorigenesis in transgenic mice.
J Biol Chem
 
2001
;
276
:
18563
–9.
22
Lim JT, Piazza GA, Han EK, Delohery TM, Li H, Finn TS, et al. Sulindac derivatives inhibit growth and induce apoptosis in human prostate cancer cell lines.
Biochem Pharmacol
 
1999
;
58
:
1097
–107.
23
Narko K, Ristimaki A, MacPhee M, Smith E, Haudenschild CC, Hla T. Tumorigenic transformation of immortalized ECV endothelial cells by cyclooxygenase-1 overexpression.
J Biol Chem
 
1997
;
272
:
21455
–60.
24
Trifan OC, Smith RM, Thompson BD, Hla T. Overexpression of cyclooxygenase-2 induces cell cycle arrest. Evidence for a prostaglandin-independent mechanism.
J Biol Chem
 
1999
;
274
:
34141
–7.
25
Johnson AJ, Song X, Hsu A, Chen C. Apoptosis signaling pathways mediated by cyclooxygenase-2 inhibitors in prostate cancer cells.
Adv Enzyme Regul
 
2001
;
41
:
221
–35.
26
Chinery R, Coffey RJ, Graves-Deal R, Kirkland SC, Sanchez SC, Zackert WE, et al. Prostaglandin J2 and 15-deoxy-delta12,14-prostaglandin J2 induce proliferation of cyclooxygenase-depleted colorectal cancer cells.
Cancer Res
 
1999
;
59
:
2739
–46.
27
Penning TD, Talley JJ, Bertenshaw SR, Carter JS, Collins PW, Docter S, et al. Synthesis and biological evaluation of the 1,5-diarylpyrazole class of cyclooxygenase-2 inhibitors: identification of 4-[5-(4-methylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (SC-58635, celecoxib).
J Med Chem
 
1997
;
40
:
1347
–65.
28
Cao Y, Pearman AT, Zimmerman GA, McIntyre TM, Prescott SM. Intracellular unesterified arachidonic acid signals apoptosis.
Proc Natl Acad Sci U S A
 
2000
;
97
:
11280
–5.
29
He TC, Chan TA, Vogelstein B, Kinzler KW. PPARdelta is an APC-regulated target of nonsteroidal anti-inflammatory drugs.
Cell
 
1999
;
99
:
335
–45.
30
Yin MJ, Yamamoto Y, Gaynor RB. The anti-inflammatory agents aspirin and salicylate inhibit the activity of I(kappa)B kinase-beta.
Nature
 
1998
;
396
:
77
–80.