| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, The Ohio State University, Columbus, Ohio
Requests for reprints: Ching-Shih Chen, Division of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, Ohio State University, 336 Parks Hall, 500 West 12th Avenue, Columbus, OH 43210-1291. Phone: 614-688-4008; Fax: 614-688-8556. E-mail: chen.844{at}osu.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Key Words: Celecoxib • HUVEC • Cell cycle arrest • Phosphoinositide-dependent kinase-1 • Cyclin-dependent kinases • 07-02-00 Mechanisms of drug action/new molecular targets/therapeutics • 07-02-01 Cell cycle mechanisms of anticancer drugs • 07-02-02 Cellular responses to anticancer drugs • 07-05-09 Kinase and phosphatase inhibitors
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Nonetheless, information regarding the mechanism underlying celecoxib-mediated growth inhibition in angiogenic endothelial cells is fragmentary vis-à-vis cancer cells in the literature. At the cellular level, celecoxib inhibits COX-2, causes cell cycle arrest, and induces apoptosis at different effective concentrations. However, the effects on cell cycle progression and apoptosis in cancer cell systems in vitro require concentrations that are several orders of magnitude greater than that required to inhibit COX-2 activity (i.e., tens of micromolar concentrations versus 0.04 µmol/L). Such findings are among a growing body of evidence that celecoxib mediates the antitumor effects via both COX-2-dependent and COX-2-independent mechanisms (16, 18–24). Among several non-COX-2 targets proposed for celecoxib, suppression of Akt through the inhibition of phosphoinositide-dependent kinase-1 (PDK-1) is particularly noteworthy (24, 25). As dysregulated PDK-1/Akt signaling plays a key role in promoting cancer cell proliferation and survival (13, 26), its inhibition represents a major mechanism underlying celecoxib-mediated antitumor effects (19, 27). The unique ability of celecoxib to inhibit PDK-1, however, is not observed with other COX-2 inhibitors examined, such as rofecoxib, despite that they exhibit comparable COX-2 inhibitory potency (28, 29).
As Akt also plays a role in endothelial cell proliferation and survival (30), these findings raise a question of whether the growth inhibitory effects of celecoxib on the tumor vascular endothelium (15, 17) involve the inhibition of PDK-1/Akt signaling. To address this issue, we used human umbilical vein endothelial cells (HUVEC) as a model of the angiogenic endothelium to compare the growth inhibitory effects of celecoxib vis-à-vis a close, COX-2-inactive derivative of celecoxib. 4-[5-(2,5-Dimethylphenyl)-3(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide (DMC), although devoid of COX-2 inhibitory activity (31), retains the ability to induce apoptosis in PC-3 prostate cancer cells through the inhibition of PDK-1/Akt signaling (20, 24). Our present findings show that both celecoxib and DMC inhibited HUVEC growth with pharmacologic profiles reminiscent of those in PC-3 cells. The mechanism underlying celecoxib-mediated growth inhibition may differ between these two types of cells because of differences in signaling mechanisms governing cell cycle progression and apoptosis. For example, PC-3 cells are deficient in the expression of tumor suppressors such as PTEN, p53, and retinoblastoma. From a mechanistic perspective, the genomic integrity of the HUVEC system presents a vastly different intracellular context to examine how celecoxib acts to induce growth inhibition. Here, we obtain evidence that the antiproliferative effects of celecoxib in HUVECs and DMC at pharmacologically attainable concentrations (i.e., 
20 µmol/L) are attributable to G1 arrest through the concurrent inhibition of PDK-1/Akt signaling and cyclin-dependent kinases (CDKs).
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cell Culture
HUVECs were purchased from Glycotech, Inc. (Gaithersburg, MD) and were cultured in Medium 200 (Cascade Biologics, Portland, OR) supplemented with low-serum growth supplement (Cascade Biologics) at 37°C in a humidified CO2 incubator. Final concentrations of the following components in the medium were 2% fetal bovine serum, 1 µg/mL hydrocortisone, 10 ng/mL human epidermal growth factor, 3 ng/mL basic fibroblast growth factor, and 10 µg/mL heparin. All experiments were done with cells between passages 3 and 7. For the drug treatment of HUVECs, test compounds were added in the presence of the low-serum growth supplement (Cascade Biologics). Jurkat T cells, used as the source for preparing the CDK4 immune complex, were purchased from American Type Culture Collection (Manassas, VA). The cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum at 37°C in a humidified CO2 incubator.
Growth Inhibition Assay
The assay was carried out according to a slight modification of the procedure suggested by the National Cancer Institute Angiogenesis Resources Center.1HUVECs were plated into 96-well plates at 1.5 x 103 cells per well in 100 µL of the aforementioned medium. After 24 hours (day 0), the test compound in 100 µL of the same medium was added to each well at twice the desired concentration. On day 0, one plate was stained with 0.5% crystal violet in 20% methanol for 10 minutes, rinsed with water, and air-dried. The remaining plates were incubated at 37°C for 72 hours, stained with 0.5% crystal violet in 20% methanol, rinsed with water, and air-dried. The crystal violet stain was eluted with a mixture of ethanol-0.1 mol/L sodium citrate (1:1), and absorbance at 540 nm was measured with an ELISA reader (Dynatech Laboratories, Chantilly, VA). Day 0 absorbance was subtracted from the absorbance values of the 72-hour samples, and data were plotted as percentage of control proliferation (vehicle-treated cells) to calculate IC50 values (drug concentrations that cause 50% inhibition).
Apoptosis Analysis
Two methods were used to assess drug-induced apoptotic cell death: detection of DNA fragmentation by the Cell Death Detection ELISA kit (Roche Diagnostics, Mannheim, Germany) and Western blot analysis of PARP cleavage. The ELISA was based on the quantitative determination of cytoplasmic histone-associated DNA fragments in the form of mononucleosomes or oligonucleosomes generated in the course of apoptotic death. In brief, 1.5 x 106 HUVECs were cultured in a T-75 flask for 24 hours before treatment. Cells were treated with the DMSO vehicle or the test agent at the indicated concentrations for indicated intervals and were collected, and cell lysates equivalent to 104 cells were used in the ELISA. For the PARP cleavage assay, drug-treated cells were collected, washed with ice-cold PBS, and resuspended in lysis buffer containing 20 mmol/L Tris-HCl (pH 8), 137 mmol/L NaCl, 1 mmol/L CaCl2, 10% glycerol, 1% NP40, 0.5% deoxycholate, 0.1% SDS, 100 µmol/L 4-(2-aminoethyl)benzenesulfonyl fluoride, 10 µg/mL leupeptin, and 10 µg/mL aprotinin. Soluble cell lysates were collected after centrifugation at 10,000 x g for 5 minutes. Equivalent amounts of proteins (60–100 µg) from each lysate were resolved in 10% SDS-PAGE. Bands were transferred to nitrocellulose membranes. Western blotting with an anti-PARP antibody was carried out, and apoptosis was detected by monitoring proteolysis of the 116-kDa native PARP enzyme to the apoptosis-specific 85-kDa fragment.
Flow Cytometry for Cell Cycle Analysis
A detergent-trypsin method was used for the preparation of nuclei for flow cytometric DNA analysis (33). In brief, HUVECs were treated with DMSO or the test agent at the indicated concentration for 24 hours. The harvested cells (1x 106) were suspended in 400 µL of 40 mmol/L citrate buffer (pH 7.6), containing 250 mmol/L sucrose and 10% DMSO, and stored at –80°C until analysis. The cells were centrifuged and resuspended in 500 µL solution A (3.4 mmol/L trisodium citrate, 0.5 mmol/L Tris, 0.1% NP40, and 1.5 mmol/L spermine tetrahydrochloride with final pH 7.4) containing 15 µg/mL trypsin and 10 µg/mL EDTA. After incubating at 37°C for 30 minutes, 500 µL solution B, containing 0.5 mg/mL trypsin inhibitor and 0.1 mg/mL DNase-free RNase A, was added. After another incubation at 37°C for 30 minutes, 500 µL solution C, containing 0.05 mg/mL propidium iodide and 1.2 mg/mL spermine tetrahydrochloride, was added and incubated on ice for 1 hour. Cell cycle phase distributions were determined on a FACScan flow cytometer (Beckman-Coulter, Mountain View, CA).
Akt Kinase Assay
Akt kinase assay was carried out according to a modified published procedure (34). Briefly, HUVECs were treated with DMSO vehicle or the test agents at the indicated concentrations for 12 hours, lysed, and homogenized in lysis buffer [50 mmol/L Tris-HCl (pH 7.5), 120 mmol/L NaCl, 1% (v/v) NP40, 1 mmol/L EDTA, 50 mmol/L NaF, 40 mmol/L ß-glycerophosphate, and 1 µg/mL each of aprotinin, pepstatin, and leupeptin]. Cell lysates were centrifuged at 13,000 x g for 10 minutes, and the supernatants were collected for the kinase assay. Protein concentrations in the supernatants were determined by the Bradford method (Bio-Rad, Hercules, CA). The kinase assay was done by adding 10 µL substrate peptide (RPRAATF; 
80 µmol/L final assay concentration) and 10 µL [
-32P]ATP (1 mCi/mL) to equivalent amounts of supernatant (20 µg of total protein). After incubation for 30 minutes at 30°C, 25 µL of each reaction mixture were slowly spotted onto P81 phosphocellulose paper. After three washes with 0.75% phosphoric acid, the papers were transferred to scintillation vials containing 5 mL liquid scintillation cocktail. Radioactivity was measured in a scintillation counter.
Transient Transfection by Calcium Phosphate Coprecipitation Method
The constitutively active Akt construct HA-PKB-T308D/S473D was kindly provided by Dr. Brian Hemmings (Friedrich Miescher Institute, Basel, Switzerland). HUVECs were seeded into T-75 flasks (1 x 106 per flask). Various amounts of the plasmid were added to 450 µL TE buffer [10 mmol/L Tris-HCl, 1 mmol/L EDTA (pH 7.2)] followed by the addition of 50 µL of 2.5 mol/L CaCl2 solution [10 mmol/L HEPES (pH 7.2)]. After briefly mixing the DNA-CaCl2 solution, it was added dropwise to 500 µL of 2x HBS (280 mmol/L NaCl, 10 mmol/L KCl, 12 mmol/L dextrose, 50 mmol/L HEPES, and 1.5 mmol/L Na2HPO4). After 30 minutes of incubation, the mixture was applied to HUVECs directly. The calcium phosphate–containing medium was replaced by normal medium after 6 hours of incubation, and the transfected cells were harvested after additional 3 days for experiments.
CDK1/Cyclin B Kinase Assay
The reagents used in this assay were prepared according to the manufacturer's instructions (Upstate, 14-450). Briefly, for each assay, 10 µL of assay dilution buffer [20 mmol/L MOPS (pH 7.2), 25 mmol/L ß-glycerophosphate, 5 mmol/L EGTA, 1 mmol/L sodium orthovanadate, and 1 mmol/L DTT] containing the test agent at varying concentrations, 10 µL inhibitor cocktail (20 µmol/L protein kinase C inhibitor peptide, 2 µmol/L protein kinase A inhibitor peptide, and 20 µmol/L compound R24571), and 10 µL active CDK1/cyclin B (20 ng) were mixed and incubated for 15 minutes. Then, 10 µL of histone H1 (0.5 mg/mL) and 10 µL of diluted [-32P]ATP solution were added to the mixture. After reacting at 30°C for 30 minutes, 25 µL of mixture were transferred to P81 paper. The paper was washed with 0.75% phosphoric acid thrice and acetone once and put into 5 mL scintillation cocktail. The radioactivity was measured by the scintillation counter.
CDK2/Cyclin E Kinase Assay
The reagents used in this assay were prepared according to the manufacturer's instructions (Upstate, 14-475). Briefly, for each assay, 10 µL of reaction buffer [50 mmol/L MOPS (pH 7.0) and 2.5 mmol/L EDTA] containing the test agents at varying concentrations, and 2.5 µL of active CDK2/cyclin E (20 ng) were mixed and incubated for 15 minutes. Then, 2.5 µL of histone H1 (1 mg/mL) and 10 µL of diluted [-32P]ATP solution were added into the mixture. After reacting at 30°C for 30 minutes, 20 µL of mixture were transferred to P81 paper. The paper was washed with 0.75% phosphoric acid thrice and acetone once and transferred to 5 mL scintillation cocktail. The radioactivity was measured by the scintillation counter.
Immunoprecipitated Cyclin E Kinase Assay
Cyclin E immunoprecipitates were prepared from the lysates of HUVECs as follows. HUVECs were collected by scraping from T-75 flasks at 80% confluence and resuspended in lysis buffer [50 mmol/L Tris-HCl (pH 7.5), 120 mmol/L NaCl, 1% (v/v) NP40, 1 mmol/L EDTA, 50 mmol/L NaF, 40 mmol/L ß-glycerophosphate, and 1 µg/mL each of aprotinin, pepstatin, and leupeptin]. After homogenization, the lysates were centrifuged at 13,000 x g for 10 minutes. The supernatants were collected and diluted with the same buffer at a 1:500 ratio. The diluted lysates were then subjected to immunoprecipitation at 4°C for 2 hours with antibodies (1:100) against cyclin E (Santa Cruz Biotechnology). Protein A-agarose beads were then added followed by an additional 1 hour of incubation at the same temperature. The beads with captured immune complexes were washed thrice with lysis buffer and used for the CDK2 kinase assay. The assay was done as described above for the CDK2/cyclin E kinase assay, except for the use of the cyclin E immunoprecipitates instead of the recombinant enzymes and the inclusion of a short centrifugation step to pellet the beads before transfer of supernatants to P81 paper.
CDK4 Kinase Assay
CDK4 immunoprecipitates were prepared from the lysates of Jurkat T cells as follows. Jurkat T cells (8 x 108) were lysed with lysis buffer [50 mmol/L Tris-HCl (pH 7.5), 120 mmol/L NaCl, 1% (v/v) NP40, 1 mmol/L EDTA, 50 mmol/L NaF, 40 mmol/L ß-glycerophosphate, 0.1 mmol/L phenylmethylsufonyl fluoride, and 1 µg/mL each of aprotinin, pepstatin, and leupeptin] followed by homogenization and centrifugation at 13,000 x g for 10 minutes. The supernatant was incubated with anti-CDK4 antibodies (Santa Cruz Biotechnology) at 4°C for 2 hours followed by addition of protein A-agarose beads for 1 hour. The beads with captured immune complexes were washed thrice by lysis buffer and twice by CDK4 assay buffer [50 mmol/L MOPS (pH 7.0) and 2.5 mmol/L EDTA] and subjected to CDK4 kinase assay. In brief, 5 µL of assay buffer with test agent at the desired concentration and 25 µL of the same buffer containing CDK4 immunoprecipitates were incubated together for 15 minutes. Then, 10 µL of assay buffer containing retinoblastoma substrate peptide (2.5 ng) and 10µL of diluted [-32P]ATP solution were added into the mixture. The reaction mixture was incubated at 30°C for 1.5 hours followed by brief centrifugation. Thirty microliters of the supernatant were carefully transferred to P81 paper. The paper was washed thrice with 0.75% phosphoric acid followed by one wash with acetone and transferred to 5 mL scintillation cocktail. Radioactivity was measured by the scintillation counter.
Quantitative Determination of pRbThr821 and Total Retinoblastoma in HUVECs
The levels of phosphorylated retinoblastoma and total retinoblastoma in cell lysates were determined using Human pRbThr821 ELISA kit and Human Total Rb ELISA kit, respectively (both from Biosource International, Camarillo, CA). HUVEC cells were seeded on 0.1% gelatin-coated T-75 flasks 1 day before treatments. The cells were treated with 20 µmol/L celecoxib or DMC or with DMSO vehicle for 24 or 72 hours. For the 24-hour treatment, cells were plated at a density of 1.5 x 106 cells per flask, with 12 flasks per treatment group. For the 72-hour treatment, cells were plated at a density of 8 x 105 cells per flask to avoid overconfluence, with 20 flasks per treatment group. Cells were harvested by scraping and treated with lysis buffer [50 mmol/L Tris-HCl (pH 7.5), 120 mmol/L NaCl, 1% NP40 (v/v), 1 mmol/L EDTA, 50 mmol/L NaF, 40 mmol/L ß-glycerophosphate, 0.1 mmol/L phenylmethylsufonyl fluoride, and 1 mg/mL each of aprotinin, pepstatin, and leupeptin], briefly sonicated, and centrifuged at 13,000 x g for 25 minutes. The supernatants were collected for immediate analysis of pRbThr821 and total retinoblastoma contents by the immunoassays. The ELISAs were done according to the manufacturer's instructions.
Western Blot Analysis of Signaling Components
Drug-treated cells were collected as described above, washed with PBS, resuspended in SDS-PAGE loading buffer [100 mmol/L Tris-HCl (pH 6.8), 4% (w/v) SDS, 0.2% (w/v) bromophenol blue, 20% (v/v) glycerol, and 200 mmol/L DTT], sonicated for 5 seconds, and boiled for 5 minutes. After brief centrifugation, equivalent amounts of soluble proteins, as determined by the Bradford method, were resolved in 8% to 15% SDS-polyacrylamide minigels, depending on the size of desired protein, and transferred to nitrocellulose membranes with the use of a semidry transfer cell (Bio-Rad). The membranes were washed twice with TBS [0.3% (w/v) Tris, 0.8% (w/v) NaCl, and 0.02% (w/v) KCl] containing 0.05% Tween 20 (TBST) and then incubated with TBS containing 5% nonfat dry milk for 20 minutes to block nonspecific antibody binding. Each membrane was then incubated at 4°C for 12 hours with a primary antibody specific for cyclins A, B1, D1, and E, CDKs 1, 2, 4, and 6, p21Cip1, p27Kip1, ß-actin, ERK, or phospho-ERK, which was diluted 1:1,000 in TBS containing 1% nonfat dry milk. The membranes were washed twice with TBST and then incubated at room temperature for 1 hour with a horseradish peroxidase–conjugated goat anti-rabbit or anti-mouse immunoglobulin G diluted 1:5,000 in TBS containing 1% nonfat dry milk. The membranes were washed twice with TBST, and bound antibody was visualized by enhanced chemiluminescence Western blotting detection reagents (Amersham Pharmacia Biotech, Little Chalfont, United Kingdom). Unphosphorylated ERK2, as immunostained with anti-ERK2 antibodies, were used as internal standards for the comparison of phospho-ERK2 levels among samples of different exposure intervals.
Chicken Chorioallantoic Membrane Assay
The chorioallantoic membrane (CAM) assay was done in accordance with procedures described by Marks et al. (35). Eight-day-old fertile White Leghorn chicken eggs were obtained from the Ohio State University Department of Animal Sciences and CBT Farms (Chestertown, MD). Eggs were candled to ensure fertility and viability of the embryos before inclusion in the experiment. A small hole (1 cm diameter) was made in the shell over the air sac through which treatment solutions were directly pipetted onto the surface of the CAM. Treatments included celecoxib, DMC, rofecoxib, indomethacin, and roscovitine, which were prepared as suspensions in PBS and applied to the CAM in a total volume of 15 µL containing 0.15, 1.5, or 15 nmol of drug. Controls received 15 µL PBS only. Each treatment group contained eight eggs. The holes were then sealed with Micropore surgical tape (3M Health Care, St. Paul, MN), and the eggs were incubated in a humidified egg incubator (Murray McMurray Hatcheries, Webster City, IA) at 37°C for 72 hours. After treatment, the area of each CAM to which the treatment agent had been applied was quickly excised and fixed in 4% paraformaldehyde (w/v in PBS). Images of the CAMs were acquired with a digital camera (Nikon Coolpix 990, Melville, NY) and visualized for quantitation of vascularity in Adobe Photoshop 6.0 software. Each digital CAM image was overlaid with a 1 x 1-cm grid, and the number of blood vessel branch points within the grid was counted. Vascular densities in the CAMs were expressed as a percentage of the blood vessel branch points in the vehicle-treated control CAMs. For determination of phospo-Akt status, CAMs treated with 15 nmol of DMC were excised after 24, 48 and 72 hours of treatment, quickly frozen in liquid nitrogen, and stored at –80°C until analysis. CAMs were homogenized in the previously described lysis buffer, and Western blot analysis was performed as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
30 µmol/L for both agents. In contrast, rofecoxib could only cause partial inhibition of HUVEC growth at concentrations exceeding 50 µmol/L. These data suggest that the in vitro growth inhibitory effects of celecoxib on HUVECs involved COX-2-independent mechanisms.
|
|
|
|
As celecoxib inhibits PDK-1 activity through ATP competition, a crucial issue that could shed light on alternative targets is the specificity of its kinase inhibition, particularly with respect to CDKs in light of their direct involvement in cell cycle progression. As CDKs phosphorylate retinoblastoma to enable cells to progress from G1 to S phase (39, 40), we analyzed the phosphorylation status of retinoblastoma as an indicator of CDK activity in drug-treated HUVECs. Specifically, we evaluated the phosphorylation level of Thr821, a preferential phosphorylation site for CDK2 (41), by ELISA following exposure of HUVECs to 20 µmol/L celecoxib or DMC for 24 and 72 hours. As shown in Fig. 4A, a time-dependent decrease in the ratio of pRbThr821 to total retinoblastoma was observed in both treatment groups. As intracellular CDK2 forms complexes with cyclin E and several cell cycle regulators such as p21Cip1 and p27Kip1 (42), we examined the effect of celecoxib and DMC on the CDK2 kinase activity of immunoprecipitated cyclin E from HUVECs (Fig. 4B). Both agents exhibited significant inhibitory activity on the cyclin E immune complexes with IC50 of 10 µmol/L.
|
In addition to CDK2, celecoxib and DMC also inhibited recombinant CDK1/cyclin B1 and CDK4 immune complexes. The estimated IC50 values for celecoxib and DMC were 34 and 22 µmol/L for recombinant CDK1 and 16 and 14 µmol/L for the CDK4 immune complex, respectively.
Earlier reports indicate that celecoxib mediated growth arrest by altering the expression levels of various cyclins or CDK inhibitors in different cancer cell systems (18, 21, 22, 44). In light of this transcriptional regulation, we examined the effect of celecoxib at 10, 20, and 30 µmol/L on the expression levels of various cell cycle regulatory proteins in HUVECs after 72 hours of exposure, which included cyclins A, B1, D1, and E, p21, p27, and CDKs 1, 2, 4, and 6 (Fig. 5A). However, celecoxib treatment did not alter the expression of any of these cell cycle regulatory proteins, which again suggests differences in signaling mechanisms governing celecoxib-mediated cell growth inhibition between cancer cells and HUVECs.
|
B kinase ß, protein kinase A, p70S6K, protein kinase C , mitogen-activated protein kinase kinase 1, mitogen-activated protein kinase 2, platelet-derived growth factor receptor 
, c-RAF, and c-Src. However, none of these kinases was inhibited by 30 µmol/L celecoxib (data not shown). We reported previously that treatment of PC-3 cells with 50 µmol/L celecoxib or DMC under serum-free conditions resulted in rapid dephosphorylation of ERK (45). However, within the growth inhibitory range of 10 to 30 µmol/L in medium containing 2% serum (as recommended by the HUVEC supplier, Cascade Biologics), neither agent caused an appreciable change in the level of phospho-ERKs in HUVECs (Fig. 5B). This finding suggests that ERK inhibition does not play a role in celecoxib-induced G1 arrest under these conditions.
Chicken Chorioallantoic Membrane Assay
The above in vitro data provided a rationale to examine the in vivo efficacy of celecoxib and DMC in the inhibition of neovascularization by using a CAM assay (Fig. 6). The CAMs of fertile 8-day-old White Leghorn chicken eggs were treated in vivo with 0.15, 1.5, or 15 nmol celecoxib or DMC suspended in 15 µL PBS. In addition, the COX-2 inhibitor rofecoxib, the nonselective COX inhibitor indomethacin, and the CDK inhibitor roscovitine were also examined for their in vivo antiangiogenic activities for comparison (n = 8 for all treatment groups). Figure 6A shows that, in comparison with the untreated group, DMC reduced vascular densities in the CAMs at all levels tested. Meanwhile, celecoxib, rofecoxib, indomethacin, and roscovitine also caused a significant reduction in neovascularization but only at the highest dose (P < 0.05). This finding is in line with the reported effects of a nonsteroidal anti-inflammatory drug (NSAID) on angiogenesis (46) and suggests the involvement of CDKs in the antiangiogenic action of celecoxib and DMC.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Both celecoxib and DMC induced G1 arrest, in the absence of significant apoptosis, in HUVECs at concentrations up to 30 and 20 µmol/L, respectively. In contrast, rofecoxib required at least 50 µmol/L to exhibit appreciable inhibition. These findings suggest the involvement of COX-2-independent mechanisms in the antiproliferative action of celecoxib in HUVECs. These in vitro data were reflected in the results of the CAM assay in which both celecoxib and DMC reduced neovascularization. The in vivo efficacy of DMC seemed to be higher than that of celecoxib, which might be due to its slightly higher potency in inhibiting non-COX-2 targets or improved bioavailability due to increased hydrophobicity.
Beyond their utility as a model of the angiogenic endothelium, HUVECs provide an interesting contrast to the PC-3 prostate cancer cells in that they differ in many aspects of intracellular signaling. In contrast to HUVECs, PC-3 cells lack functional PTEN, p53, and retinoblastoma. Of particular interest is the loss of PTEN function in PC-3 cells, which results in elevated Akt signaling. Inhibition of PDK-1 and the consequent suppression of Akt activation by celecoxib cause growth inhibition and apoptosis in PC-3 cells. Consistent with our identification of PDK-1 as a major non-COX-2 target in PC-3 prostate cancer cells (24), Akt kinase activity was also suppressed in HUVECs treated with either celecoxib or DMC. However, enforced expression of constitutively active Akt kinase alone failed to abrogate celecoxib-mediated G1 arrest, suggesting the involvement other targets in the HUVEC system.
Celecoxib has been shown to inhibit PDK-1 through the competition for ATP binding, a mechanism reminiscent to that of many other kinase inhibitors like flavopiridol and UCN-01. This mode of kinase inhibition, however, often gives rise to the cross-inhibition of related kinases. For example, flavopiridol, an inhibitor of several CDKs including CDK1, CDK2, CDK4, CDK6, and CDK7, was originally reported to inhibit epidermal growth factor receptor and protein kinase A at higher concentrations (47), and UCN-01 has been shown to inhibit Ca2+-dependent protein kinase C, CDK1, CDK2, Chk1, and PDK-1 (47, 48). Our data indicate that celecoxib inhibited different CDKs with varying potency over a concentration range that was associated with G1 arrest in HUVECs. Celecoxib at 30 µmol/L, however, did not affect the activity of a wide range of kinases examined, indicating the existence of a certain degree of specificity.
The ability of NSAIDs, including celecoxib, to compete for ATP binding sites has also been reported for several other non-COX-2 targets. Aspirin, sodium salicylate, sulindac, and sulindac sulfide have been shown to down-regulate nuclear factor-B signaling by inhibiting IB kinase ß activity (49). For aspirin and sodium salicylate, this inhibition was attributed to the occupation of the ATP binding site. Our own work showed that celecoxib inhibits endoplasmic reticulum Ca2+-ATPases (19) and PDK-1 through competition for ATP binding (24). A similar mechanism was reported to mediate the inhibition of adenylyl cyclase by celecoxib with an IC50 of 375 µmol/L (50). These findings underscore the pharmacologic complexity of the antiproliferative activities of celecoxib.
Whereas our findings show direct inhibition of CDK activity by celecoxib, others have reported indirect mechanisms. The intracellular elevation of the CDK inhibitors, p21Cip1 and p27Kip1, by celecoxib in several cancer types has been reported (18, 21, 22, 43). In the present study, similar effects on these proteins were not observed in celecoxib-treated HUVECs over the concentration range (10–30 µmol/L), which induced growth arrest without substantial apoptosis, and were lower than those used in the cited studies (
50 µmol/L). A notable exception is the work of Narayanan et al. (21), which showed elevated expression levels of p21Cip1 and p27Kip1 after treatment with 10 and 20 µmol/L celecoxib in carcinogen-induced rat prostate cancer cells. In several glioblastoma cell lines, antiproliferative effects of celecoxib were attributed to the transcriptional down-regulation of cyclins A and B and the subsequent reduction in activities of corresponding CDKs in the absence of significant elevations in p21Cip1 and p27Kip1 (23). In our study, reductions in cyclin A or B1 were not observed with the concentration range we examined (Fig. 5). These discrepancies may result from differences in the concentrations of celecoxib used or differences in signaling pathways governing cell proliferation between cancer cells and HUVECs.
This study used HUVECs as a model to investigate the direct inhibitory effects of celecoxib on the angiogenic endothelium. Our findings show that celecoxib can induce growth arrest in these cells by interfering with multiple signaling molecules involved in the regulation of cell proliferation and suggest that its antiangiogenic activity in the tumor vasculature includes a role for COX-2-independent mechanisms. The results of the CAM assay are consistent with the report that the in vivo antiangiogenic actions of NSAIDs may involve both COX-2-dependent and COX-2-independent mechanisms (46). The relative contribution of COX-2 versus non-COX-2 mechanisms in the context of suppressing tumor-associated angiogenesis, however, is unclear and may vary with cell types or in vivo models. For instance, Masferrer et al. have elegantly shown the involvement of COX-2 inhibition in the antiangiogenic effect of celecoxib in a rat cornea model (15, 17). More recently, rofecoxib has also been shown to suppress microvessel density in human colorectal cancer liver metastases (51). In our CAM study, rofecoxib and indomethacin inhibited neovascularization but only at the highest dose used. A recent report indicates that the ability of indomethacin to inhibit angiogenesis also involved inhibition of ERK2 activity and interference with ERK nuclear translocation (46). These data further underscore the involvement of COX-2-independent signaling mechanisms in the antiproliferative activities of NSAIDs.
In summary, this study represents part of an emerging body of work that underscores the complexity of the pharmacologic actions of celecoxib. This complexity may confer broad anticancer efficacy and may underlie the established clinical value and potential of celecoxib. Meanwhile, the role of CDK inhibition in celecoxib-mediated antitumor effects in cancer cell systems (i.e., PC-3 cells) that lack functional retinoblastoma is currently under investigation.
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 http://dtp.nci.nih.gov/aa-resources/aa_index.html. 
Received 5/ 5/04; revised 9/ 3/04; accepted 10/18/04.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Kardosh, E. B. Golden, P. Pyrko, J. Uddin, F. M. Hofman, T. C. Chen, S. G. Louie, N. A. Petasis, and A. H. Schonthal Aggravated Endoplasmic Reticulum Stress as a Basis for Enhanced Glioblastoma Cell Killing by Bortezomib in Combination with Celecoxib or Its Non-Coxib Analogue, 2,5-Dimethyl-Celecoxib Cancer Res., February 1, 2008; 68(3): 843 - 851. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Amrite and U. B. Kompella Celecoxib Inhibits Proliferation of Retinal Pigment Epithelial and Choroid-Retinal Endothelial Cells by a Cyclooxygenase-2-Independent Mechanism J. Pharmacol. Exp. Ther., February 1, 2008; 324(2): 749 - 758. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-C. A. Marshall, B. F. Fernandes, S. Di Cesare, S. C. Maloney, P. T. Logan, E. Antecka, and M. N. Burnier Jr The use of a cyclooxygenase-2 inhibitor (Nepafenac) in an ocular and metastatic animal model of uveal melanoma Carcinogenesis, September 1, 2007; 28(9): 2053 - 2058. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Pyrko, A. Kardosh, Y.-T. Liu, N. Soriano, W. Xiong, R. H. Chow, J. Uddin, N. A. Petasis, A. K. Mircheff, R. A. Farley, et al. Calcium-activated endoplasmic reticulum stress as a major component of tumor cell death induced by 2,5-dimethyl-celecoxib, a non-coxib analogue of celecoxib Mol. Cancer Ther., April 1, 2007; 6(4): 1262 - 1275. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Grosch, T. J. Maier, S. Schiffmann, and G. Geisslinger Cyclooxygenase-2 (COX-2)-independent anticarcinogenic effects of selective COX-2 inhibitors. J Natl Cancer Inst, June 7, 2006; 98(11): 736 - 747. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kardosh, N. Soriano, Y.-T. Liu, J. Uddin, N. A. Petasis, F. M. Hofman, T. C. Chen, and A. H. Schonthal Multitarget inhibition of drug-resistant multiple myeloma cell lines by dimethyl-celecoxib (DMC), a non-COX-2 inhibitory analog of celecoxib Blood, December 15, 2005; 106(13): 4330 - 4338. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||
