Preclinical evaluation of the nonsteroidal anti-inflammatory agent celecoxib on malignant mesothelioma chemoprevention

Abstract

Malignant mesothelioma (MM) remains the most lethal pleural, peritoneal and pericardial cancer. Here, we characterize the effects of nonsteroidal anti-inflammatory agents (NSAIDs) on in vitro and in vivo experimental MM models. Unlike primary normal mesothelial cells, the selective cyclooxygenase (COX)-2 inhibitor celecoxib reduced the in vitro proliferation of several MM cells derived from previously untreated MM patients. Moreover, celecoxib significantly inhibited MM cell colony formation in soft agarose (63–78% at 5 × 10⁻⁵ M; p ≤ 0.05) and it elicited remarkable antitumor activity, leading to long-term survival in >37% of nude mice bearing intraperitoneal MM. Celecoxib was more efficient in inhibiting MM cell growth than acetylsalicylic acid (10⁻⁶ M-10⁻² M), indometacin (10⁻⁶ M-10⁻² M) and the COX-2 inhibitor NS-398 (10⁻⁶ M-10⁻⁴ M). Efficacy of these different compounds was not related to the amount of COX-2 protein levels present on MM cells. Celecoxib, in a dose- and time-dependent manner, induced MM cell apoptosis, which involved decreased Akt phosphorylation, loss of Bcl-2 and Survivin protein expression and caspase-3 activation. Furthermore, vascular endothelial growth factor (VEGF), an MM autocrine growth factor and Akt inducer, rescued celecoxib-induced apoptosis and Akt dephosphorylation. When the VEGF receptor (KDR/Flk-1) inhibitor, SU-1498, was used in combination with celecoxib, IC₅₀ of celecoxib in vitro was reduced up to 65%. These data demonstrate that celecoxib may have antitumor properties in MM and provide a rationale for the therapeutic use of celecoxib in combination with a selective VEGF inhibitor. © 2004 Wiley-Liss, Inc.

Malignant mesothelioma (MM) is a fatal malignancy most often associated with asbestos exposure with a 3-year survival of <5%.1 Its incidence is still increasing with about 3,000 cases per year in the United States and 250,000 deaths expected in Western Europe in the next 30 years.2 Current approaches, ranging from aggressive surgical treatments to chemotherapy, fail to improve the prognosis of this disease.3 Most chemotherapeutic agents are not very effective against MM, with typical single-agent response rates of ≤20%.4 Therefore, multimodality treatments have been developed, and the most favorable outcomes are obtained with a combination of surgery, chemo- and radiation-therapy in highly selected groups of MM patients.3 However, mortality and morbidity are still high, and this strategy is not curative.5 Therefore, there is a need for innovative therapies that may offer hope for improved palliation, prolonged survival and even potential cure for MM.

Chemopreventive strategies have been investigated in a variety of other solid tumors, such as breast, colon, head and neck and liver cancers, suggesting the potential utility of this approach.6 Much attention has been given to the use of nonsteroidal anti-inflammatory compounds (NSAIDs), such as aspirin and indometacin, as agents able to block the sequence of carcinogenic events leading to an invasive malignancy.7 It is believed that these antitumor effects involve inhibition of both cyclooxygenase (COX) enzymes, namely COX-1 and -2, and modification of other non-COX targets.6 Although inhibition of COX-1 activity, which is constitutive in most tissues, can have adverse side effects, that of COX-2, which is inducible and overexpressed in many types of cancer,7, 8, 9 can be used as a new approach to adjuvant therapy in cancer treatment.7, 10 As a result, there was a rationale to develop selective COX-2 inhibitors, such as NS-398 and celecoxib, that block inducible COX-2 activity but that spare COX-1 activity and the normal physiologic functions of this enzyme.11, 12

Recent evidences indicate that COX-2 overexpression is associated with a shortened survival of patients affected by MM13 and that NS-398 suppresses in vitro MM cell proliferation.14 Therefore, our present study was undertaken to assess the feasibility of treating MM with NSAIDs. We first evaluated whether a series of COX-2 selective and nonselective NSAIDs can inhibit the in vitro growth of several MM cells derived from untreated patients and whether sensitivity to NSAIDs is related to COX-2 protein levels. We then developed a nude mouse model of MM to confirm the antitumor activity of NSAIDs. Our results demonstrate that the NSAID celecoxib has a marked chemopreventive effect on MM both in vitro and in vivo. We have also characterized the biochemical and cell cycle effects of celecoxib on MM cells, establishing its relationship with the antiapoptotic protein Akt.

MATERIAL AND METHODS

Cell lines and culture conditions

Human MM cells, derived from previously untreated patients, and primary NM cells were established from patients in our laboratories at Ss. Antonio and Biagio Civil Hospital and identified morphologically and by extensive phenotypic analysis as previously described.15 After 2 weeks in culture, 100% of NM cells stained positive for calretinin. They were then expanded and utilized for the experiments. NM cells were used between the third and the seventh passage. H-Meso cell lines have been characterized previously.16 All cells were maintained in RPMI-1640 supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% L-glutamine, 1% penicillin-streptomycin (complete medium) (all from HyClone, Rome, Italy) at 37°C and 5% CO₂. Cells were screened periodically for Mycoplasma contamination (Gen-probe).

Drugs and reagents

Celecoxib was obtained from Searle Pharmacia (St. Louis, MO). A stock solution of 300 mM in DMSO was used. NS-398 and Indometacin was purchased from Cayman Chemical (Ann Arbor, MI), whereas ASA was from Sigma Chemical (St. Louis, MO). A stock solution of 500 mM in DMSO was used. Stock solutions were added to the test medium at dilutions of at least 1:1,000, and control samples were treated with vehicle (DMSO). The concentration of DMSO in the test medium never exceeded 0.1% (v/v). VEGF and FGF-2 were purchased from R&D Systems (Minneapolis, MN). SU-1498 was purchased from Calbiochem (Milan, Italy) and were added to test medium at described concentrations.

In vitro cytotoxicity assays

The in vitro drug sensitivity was assessed by MTS assay using the manufacturer's instructions (Promega, Madison, WI). The MTS assay is a colorimetric method in which the intensity of the dye is proportional to the number of the cells. Cells were seeded at 2,500–20,000 cells/well in 96-well flat-bottomed plates (Corning, Corning, NY) to allow exponential growth for the 3 days of the assay to give an absorbance of 1.0–2.2. The optimum number of cells required to reach an absorbance from 1.0 to 2.2 was determined for each cell line (data not shown). In a typical experiment, cells were trypsinized, seeded in 96-well plates and allowed to recover for 24 hr before the addition of NSAIDs in media containing reduced FBS (0.1%). Drug concentrations ranged from 10⁻⁷ to 5 × 10⁻⁴ M. All experimental points were set up in 4 replicate wells, performed in duplicate, and all experiments were repeated at least 3 times. The MTS assay was developed after a 24, 48 and 72 hr incubation. 20 μl of MTS in a final volume of 100 μl were added to each well 3 hr before termination of incubation. The absorbance was then measured by a spectrophotometer at 490 nm. For VEGF-dependent growth, exponentially growing cells were seeded into 96-well plates. Twenty-four hours later, NSAIDs were added as described above, and after 1 hr of incubation, VEGF (final concentration, 10 ng/ml) was added to all wells. Cell viability was assayed 48 hr later using the MTS assay.

Soft agarose assays

Cells (between 4 × 10³ and 1 × 10,4 depending on the cell types) were resuspended in 1.5 ml of 0.3% ultrapure agarose (Life Technologies, Milan, Italy) in complete culture media containing different concentrations of NSAIDs. This suspension was layered over 3 ml of 0.6% agarose in complete culture media in 6-well plates (Corning). Every 4 days, 250 μl of fresh complete culture media, alone or containing NSAIDs, were added to each of the wells. The cells were incubated for 2 weeks, and colonies measuring ≥50 μm in diameter were counted using an Omnicon FAS III image analyzer (Bausch & Lomb, Rochester, NY). All experiments were performed in duplicate and repeated 3 times. Colony numbers on 30 random fields were used to determine each treatment group and control group for each cell type. Colonies were photographed using an Olympus IX70 microscope (Melville, NY) with a DVC1310 digital video camera and a QED camera with Standalone 145 software (QED Imaging, Pittsburgh, PA).

In vivo study

BALB/c nude mice (6 weeks of age; weight approx. 18 g) were obtained from Harlan (Sprague Dawley, Madison, WI). All mice were maintained in a pathogen-free animal facility for at least 1 week before each experiment. H-Meso cells (1 × 10⁷) in 100 μl of serum-free RPMI 1640 medium were intraperitoneally (i.p.) injected into the animals according to our usual method.17 Beginning 1 day postimplant, 6–8 mice per group were randomly allocated to treatment with either 10, 30, 50, 75 or 100 mg/kg celecoxib in PBS or PBS alone (control) in a 150 μl volume, both given orally daily for 120 days. The doses of celecoxib used in our animal MM model is largely below that indicated as cytotoxic for normal gut.18 Treatment was initiated when approximately 3–4 mm tumor nodules were identified after tumor cell inoculation (data not shown). This was usually 7 days after injection of MM cells.17 Animals were sacrificed when they met predetermined criteria established for minimizing pain and suffering. After sacrifice, tumor and normal tissues were fixed in neutral-buffered formalin and processed for routine histology and immunohistochemistry analysis.

Apoptosis

Apoptosis was assessed by nucleosome formation and cell cycle analysis. To this end, 10⁶ cells were cultured in complete medium for 24 hr, then washed twice with fresh complete medium and exposed to different concentrations or times of NSAIDs in medium containing 0.1% FBS. Nucleosome formation was determined using a cell death detection ELISA kit (Roche Molecular Biochemicals, Milano, Italy) following the manufacturer's instructions. Rescue experiments were carried out by adding 10 ng/ml of VEGF (R&D System Europe, Abingdon, UK) 1 hr after celecoxib. Cell cycle was analyzed by flow cytometry of propidium iodide-stained nuclei using a FACScanTM and Lysis II software (Becton Dickinson, Milan, Italy). Cells grown in 25 cm² flasks were treated with various concentrations of NSAIDs for 48 hr. The cells were then digested by trypsin-EDTA, washed and resuspended in serum-free medium, counted and then fixed overnight in 75% ethanol at 4°C. The cells were then washed and resuspended in PBS (pH 7.4) containing 0.1% Triton X-100, 0.05 mg/ml DNase-free RNase A and 50 μg/ml propidium iodide at a concentration of 0.5 ml/10⁶ cells. The cells were incubated in the dark for 30 min at room temperature. One thousand nuclei were counted for each point.

Western blot analysis

Proteins (50–75 μg/lane) were separated on 8–10% SDS-polyacrylamide gels and transferred to nitrocellulose. Transfers were blocked for 2 hr at room temperature with 5% nonfat milk in TBS, 0.1% Tween 20 and then incubated overnight at 4°C in the primary antibody diluted (1/1,000) in 5% bovine serum albumin in TBS, 0.05% Tween 20. All of the primary antibodies used, excluding Akt and phosphorylated Akt (by Cell Signaling Technology, Beverly, MA), were obtained from Santa Cruz Biotechnology (Berkeley, CA). The transfers were rinsed with TBS, 0.05% Tween 20 and incubated for 1 hr at room temperature in horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse (Bio-Rad, Hercules, CA) diluted 1/5,000 in 5% nonfat milk in TBS, 0.1% Tween 20. The immunoblots were developed with the Super Signal reagent (Pierce, Milan, Italy). Protein concentration was determined by standard Bradford protein assay (Bio-Rad). All Western blots were reprobed with human β-actin antibody (sc-1616; 1/1,000) from Santa Cruz Biotechnology to confirm equivalent loading of protein samples.

Caspase 3 activity

Caspase 3 activity was measured using the Caspase 3 colorimetric assay kit from Clontech (Palo Alto, CA). Briefly, equal amounts of proteins from cellular lysates were added to 5 × 10⁻⁵ M substrate (Asp-Glu-Val-Asp-p-nitroanilide) in assay buffer. Samples were incubated at room temperature and analyzed using a spectrophotometer (ELISA microplate reader, Bio-Rad, Richmond, CA).

Statistical analyses

Results are expressed as mean ± SD (error bars). To assess the statistical significance of observed differences, we used Student's t-test, Wilcoxon 2-sample test and log-rank tests. The difference was considered significant when p ≤ 0.05. Kaplan-Meier survival curves were analyzed with the Mantel-Cox log-rank test.

RESULTS

Celecoxib and other NSAIDs inhibit MM cell proliferation

The effects of the NSAIDs on the growth of the MM cell lines, as well as of normal mesothelial cells, were examined with the tetrazolium conversion (MTS) assay. After 48 hr, indometacin, NS-398 and celecoxib dose-dependently inhibited cell proliferation of MPP89, H-Meso and Ist-Mes1 MM cells. In contrast, normal primary mesothelial cells, NM-1, were resistant to the growth inhibitory effect of each treatment (Fig. 1). The mean IC₅₀ for MM cell growth inhibition was 35 ± 12 μM for celecoxib and 100 ± 38 μM for indometacin and NS-398. After 24 hr, all NSAIDs had only minor effects on MM cell proliferation (data not shown). The acetylsalicylic acid (ASA) at concentrations from 10⁻⁷ to 5 × 10⁻⁴ M up to 72 hr, had no effect on cell proliferation in either MPP89, H-Meso, Ist-Mes1 or NM-1 cells (data not shown).

Figure 1. Cell-growth curves determined by MTS assay and expressed as percentage of control for 3 different MM cells (MPP89, H-Meso, Ist-Mes1) and 1 primary mesothelial cells (NM-1) after treatment with several concentrations of indometacin (open squares), NS-398 (shaded squares) and celecoxib (filled squares) for 48 hr. An amount of 5 × 10⁻⁴ M of celecoxib was not reported because it is a concentration that exceeded solubility limits. Data denote mean values ± SD repeated at least 3 times. *p ≤ 0.05 vs. control (Student's t-test).

Celecoxib reduces MM tumorigenicity in vitro and increases the survival of MM animal models

To further investigate the growth-inhibitory effects of NSAIDs, we examined cell growth in soft agarose. MPP89 cells were grown in presence of increasing concentrations of celecoxib, indometacin and NS-398. After 2 weeks, the total colony number measuring ≥50 μm in diameter was determined on 30 random fields. As Figure 2a shows, all NSAIDs inhibited colony formation in a dose-dependent manner but had differential inhibitory effects on MPP89 cells. In fact, MPP89 cells were less sensitive to the inhibitory effects of indometacin, have an intermediate sensitivity to NS-398 and are more inhibited by celecoxib treatment. The IC₅₀ in this assay with celecoxib was <50 μM, whereas for indometacin and NS-398, it was between 100 and 150 μM, consistent with the range of concentrations required to inhibit basal MM cell proliferation (see Fig. 1). A comparison of the mean values of colonies number between the control group and the treatment groups in each MM cell line was evaluated (Table I). For each cell line, a significant decrease (p ≤ 0.05; Wilcoxon 2-sample test) of the total colony number in the celecoxib group compared to controls was observed. In MPP89 cells, a similar decrease (p ≤ 0.05) of the total colony number was observed in the NS-398 group compared to controls. A representative experiment of MPP89 cells treated with each NSAIDs at 50 μM is shown (Fig. 2a, insert). These findings indicate that celecoxib was the most effective NSAIDs to MM cells, followed by NS-398 and indometacin, in that order, whereas our cells were rather resistant to ASA (data not shown).

Figure 2. Impact of celecoxib on MM tumorigenicity both in vitro and in vivo. (a) MPP89 cells (3 × 10⁴) were plated on 0.6% agarose, 10% FCS and 10 mM Hepes in the absence or presence of the indicated concentrations (25–150 μM) of indometacin (open squares), NS-398 (shaded squares) and Celecoxib (filled squares). After 14 days, colonies measuring ≥50 μm were counted as described in “Material and Methods.” Each data point represents the mean ± SD of triplicate dishes. A representative experiment (n = 3) of MPP89 cells grown in soft agarose in either absence (control) or presence of 50 μM indometacin, NS-398 and celecoxib is shown (insert). **p ≤ 0.05 vs. control (Wilcoxon 2-sample test). (b) Mouse model of MM either treated or not treated with 75 mg/Kg/day of celecoxib, as described in “Material and Methods.” The graph represents the percentage of survival done using the Kaplan-Meier method. **p ≤ 0.05 vs. control (log-rank test).

Table I. Comparison of the Mean Total Colony Numbers in the Control Group vs. Drug-Treatment Groups in Malignant Mesothelioma (MM) Cell Lines1

	Control group2	Indometacin (150 μM)	NS-398 (150 μM)	Celecoxib (50 μM)
1 Colonies measuring ≥50 μm in diameter were determined on 30 random fields. 2 The control group used contained the vehicle 0.1% absolute ethanol. 3 Mean ± SD of triplicate dishes. 4 Statistically significant (p ≤ 0.05; Wilcoxon 2-sample test) compared to control values.
MPP89	250.6 ± 10.93	190.8 ± 17.6	140.3 ± 12.24	80.8 ± 114
H-Meso	140.5 ± 13	120.4 ± 16.1	115.4 ± 8.6	84.8 ± 6.74
Ist-Mes1	60.9 ± 11.5	60.7 ± 9.7	70.9 ± 9.3	30.7 ± 104

To study the inhibitory effect of celecoxib in vivo, we used a mouse model of MM based on injection of MM cells into the peritoneal cavity of nude mice.17 H-Meso cells can be injected intraperitoneally (i.p.), where they form diffuse tumors throughout the peritoneal cavity, similar to the presentation of peritoneal human MM. Two groups (n = 6/8) of tumor-bearing mice were treated with saline or with 75 mg/Kg celecoxib, both given orally daily for as long as the mice survived. As shown in Figure 2b, treatment with celecoxib led to significant (p = 0.02 by log-rank test) increases in median survival from 45 days (control) to 62 days, with 3 long-term (>120 days) survivors.

Celecoxib induces apoptosis in MM cells

We further studied whether the antitumor effect of celecoxib could be attributed to its ability to induce MM cell death. Celecoxib induced a time-dependent nucleosome formation in MM cells that was significant after 48 hr and maximal after 72 hr of treatment (Fig. 3a). MM cells were more sensitive to apoptosis than NM cells, with an increase of 183 ± 24% of nucleosomes over control after 72 hr of celecoxib (50 μM). In contrast, NM-1 and NM-2 cells did not exhibit significant nucleosome formation even at 72 hr of exposure (Fig. 3a). Indomethacin (up to 100 μM), NS-398 (up to 50 μM) or ASA (up to 150 mM) did not significantly increase nucleosome formation after 48 or 72 hr of exposure (results not shown). The apoptotic effect of celecoxib was also dose-dependent at concentrations from 1 to 100 μM in MPP89 cells, with a mean IC₅₀ of 38 ± 7.6 μM (Fig. 3b). Similar results were also obtained for H-Meso or Ist-Mes1 cells (data not shown). Cell cycle analysis confirmed that celecoxib induced apoptosis in MPP89 cells (51.4 %) compared to control (4.2%) (Fig. 3c).

Figure 3. Detection of apoptosis induced by incubation with celecoxib. (a) An amount of 10⁶ cells/ml of MM cells (MPP89, H-Meso and Ist-Mes1) and normal mesothelial cells (NM1 and NM2) were exposed to 50 μM of celecoxib in RPMI medium with 0.1% FBS for the indicated times. Nucleosome formation was assessed using a cell death detection ELISA kit. (b) Dose-response effect of celecoxib on MM cell apoptosis was measured at 48 hr of incubation. Results are reported as the absorbance ratio (percentage of control) between treated and untreated cells and depicted as mean ± SD from n = 3 with duplicate measurements. *p ≤ 0.05 vs. control (Student's t-test). (c) Flow cytometry profiles of MPP89 cells treated without (control) or with 50 μM celecoxib for 48 hr. Bars represent the percentage of apoptotic cells. These experiments were repeated at least 3 times with similar results.

Effect of celecoxib on proteins regulating apoptosis in MM cells

We analyzed whether sensitivity to celecoxib may be associated with COX-2 protein expression. Western blot analysis showed that COX-2 protein levels was highest in Ist-Mes1 cells and relatively high in H-Meso cells, whereas MPP89 and 2 normal mesothelial cells (NM-1 and NM-2) showed modest to weak expression of COX-2. In contrast, COX-1 protein is expressed at the same degree in all cell types (Fig. 4a). Therefore, it appears that the effects of celecoxib were not associated with COX-2 protein levels.

Figure 4. (a) Protein levels of COX-2 (top) and COX-1 (bottom) in 3 MM cells (MPP89, H-Meso, Ist-Mes1) and 2 NM cells (NM-1 and NM-2). (b) MPP89 cells were treated with 50 μM celecoxib for different times (3, 6, 18 and 24 hr). Total protein (50 μg) was probed by Western blot with different antibodies for p53, Bcl-2, Bax, Survivin, phosphorylated-Akt (p-Akt) and total Akt. α-tubulin was used as an internal control. (c) MPP89 cells were incubated with 50 μM of celecoxib either in presence or absence of 100 μM of zVAD-fmk. Caspase 3 activity was determined as indicated in “Material and Methods.” The p-value is also indicated (Student's t-test).

To examine the mechanisms involved in propagating and executing apoptosis in celecoxib-treated MM cells, time-course experiments were carried out to determine the levels of protein expressed by apoptosis-related genes. In MPP89 cells, the tumor suppressor p53 endogenous levels were unchanged (Fig. 4b), suggesting that antitumor activity of celecoxib can also elicit in the absence of wild-type p53. In contrast, celecoxib exposure resulted in a decrease of the antiapoptotic protein Bcl-2 by 24 hr and in an increase of the pro-apoptotic protein Bax as early as 6 hr (Fig. 4b). Expression of the pro-apoptotic proteins Bcl-X_S or Bad and the antiapoptotic protein Bcl-X_L, were unaltered by celecoxib treatment (data not shown).

We also examined the effects of celecoxib treatment on the phosphorylation of known antiapoptotic proteins involved in MM cell growth. MPP89 cells expressed the same Akt protein levels over time, but the phosphorylated-Akt was reduced after celecoxib exposure (Fig. 4b). Moreover, celecoxib strongly decreased survivin protein levels (Fig. 4b). Survivin has been shown to inhibit apoptosis by binding to active caspase-3 and caspase-7.19 Consistently, celecoxib exposure significantly enhanced caspase-3 activity in MPP89 cells, and this effect was blunted by the caspases inhibitor, zVAD-fmk (Fig. 4c). Therefore, celecoxib treatment in MM cells results in a marked shift from survival to apoptotic proteins, which activate a caspase-dependent programmed cell death.

VEGF antagonizes celecoxib-induced apoptosis and Akt dephosphorylation

VEGF is an important autocrine growth factor for MM cells.20 To determine whether modulation of VEGF concentration in culture medium influences the antiproliferative actions of celecoxib on MM cells, we treated MPP89 cells with VEGF or FGF-2, known as a general cell mitogen and overexpressed in MM cells.21, 22 Unlike FGF-2, VEGF diminished the efficacy of celecoxib to induce MM cell apoptosis, reducing nucleosome formation to 126.7 ± 8.1% over control (Fig. 5a). Furthermore, MPP89 cells were treated with 50 μM celecoxib in presence or absence of VEGF, and phosphorylated-Akt protein was examined. The Akt protein was selected for this experiment because its phosphorylated form is reduced most markedly in response to celecoxib (Fig. 4b), and Akt is a downstream target of survival effects induced by VEGF.23 VEGF treatment (10 ng/ml) increased the amount of phosphorylated-Akt protein in MPP89 cells treated with celecoxib, restoring the control levels (Fig. 5b). Thus, the autocrine/paracrine loop of VEGF protects MM cells from celecoxib-induced apoptotic effects.

Figure 5. VEGF antagonizes celecoxib-induced apoptosis and Akt dephosphorylation. (a) MPP89 cells were either untreated (open bars) or treated (filled bars) for 48 hr with 50 μM celecoxib in absence (control) or presence of 10 ng/ml VEGF and FGF-2. Nucleosomes were determined by the cell death detection ELISA kit. Data are expressed as a percentage of the control and represent the mean ± SD from 3 independent experiments done in duplicate. Statistically significant changes in nucleosomes are shown (Student's t-test). (b) MPP89 cells were treated without or with 50 μM celecoxib in presence of 10 ng/ml VEGF for either 6 or 24 hr. Cells lysates (50 μg) were electrophoresed and probed by Western blot with anti-P-⁴⁷³Ser-Akt and anti-Akt antibodies. A representative experiment (n = 3) is shown.

In vitro antitumor effects of celecoxib in combination with anti-VEGF agents

We further investigated whether adjunct use of SU-1498, the VEGF receptor (KDR/Flk-1) inhibitor,24 can enhance the sensitivity to celecoxib of MM cells in vitro. The inhibitory effects of SU-1498 at 0, 1, 10 or 50 μM were evaluated in combination with the IC₅₀ value of celecoxib (35 μM) in MPP89 cells. The use of 50 μM SU-1498 in combination with celecoxib resulted in the reduction of IC₅₀ by 63 ± 5%, whereas the use of 1 μM SU-1498 resulted in the reduction of IC₅₀ by 36 ± 2%. In addition, the use of 10 μM SU-1498 has an adjunct that yielded 53 ± 2% reductions in IC₅₀ (Fig. 6a). An isobologram was constructed based on the dose-response curves for SU-1498 to examine its synergistic effects with celecoxib in MPP89 cells.25, 26 It was clearly shown that supra-additive effects could be obtained using celecoxib in combination with 10–50 μM SU-1498, whereas a marginal supra-additive effect was observed in combinatorial use with 1 μM SU-1498 (Fig. 6b). These findings indicated that the chemosensitivity to celecoxib could be synergistically enhanced in MM cells by the adjunct use of anti-VEGF agents, such as SU-1498.

Figure 6. Chemosensitivity of MPP89 cells to celecoxib in combination with the VEGF inhibitor SU-1498. (a) Reduction of IC₅₀ of celecoxib by the adjunct use of SU-1498 as determined by MTS assay. (b) Supra-additive enhancement of chemosensitivity to celecoxib is seen as a result of the adjunct use of SU-1498 at 1 (lightly shaded squares), 5 (darker shaded squares) and 10 (filled squares) μM. Solid and broken lines represent Mode I and Mode II in the IC₅₀ isobologram, respectively, which were constructed from the dose-response curves of SU-1498 and celecoxib, respectively. The doses of SU-1498 used in our study were 0.5, 1, 5, 10, 25 and 50 μM. To produce 50% inhibition in combination with a chosen dose of SU-1498, the dose of celecoxib required was 35 μM (Fig. 1). The area delineated by Mode I and Mode II lines represents the envelope of additivity, whereas the area below this envelope represents the supra-additive area and the area above it the sub-additive area.

DISCUSSION

The objective of our study was to examine the chemopreventive efficacy of celecoxib in experimental models of MM tumorigenesis. The rationale for examining the potential efficacy of celecoxib as an MM chemopreventive agent was at least 3-fold. First, recent evidences suggest that NSAIDs, which inhibit arachidonic acid metabolism, are effective inhibitors of MM cells.14, 27 Second, celecoxib is one of a new class of inhibitors that specifically target COX-2 but not COX-1. This drug does not appear to have certain side effects associated with NSAIDs.28 Third, the finding that COX-2 is overexpressed in human MM predicts that compounds targeted against this enzyme specifically are likely to be effective against this tumor.13

To distinguish prevention from regression effects, MM cells were resuspended in soft agarose in the presence of celecoxib to study in vitro MM tumorigenicity. Moreover, we i.p. injected MM cells into nude mice and treated these early (from 7 days after injection of MM cells until sacrifice). At this time, there are very few tumor nodules in control mice, and these are quite small (3–4 mm). In both cases, treatment with celecoxib significantly suppressed the tumorigenicity potential of MM cells (Fig. 2). These findings suggest that celecoxib may provide a means for safe and effective chemopreventive MM therapy. However, clinical trials with celecoxib in patients with predisposing diseases (i.e., asbestos-exposed people) will be required to prove clinical benefit.

The degree of MM growth inhibition induced by celecoxib and other typical NSAIDs was different. In fact, celecoxib was the most effective, indometacin was the least efficient, and NS-398 had an intermediate effect (Figs. 1 and 2a). This variability may be not associated with COX-2 protein levels (Fig. 4a), suggesting that the higher antineoplastic effects of celecoxib in respect to other NSAIDs are not only due to prostaglandin synthesis inhibition but also to other non-COX targets. Similar findings have also been observed in prostate and pancreatic cancer cell lines in which downregulation of Akt and extracellular signal-regulated kinase 1/2 (ERK1/2) activity have been hypothesized as one of the mechanisms involved in celecoxib-induced apoptosis.6, 29

Akt activation has been shown to induce survival and to suppress apoptosis through increased phosphorylation of Bad and subsequent activation of antiapoptotic mechanisms involving Bcl-2 family members.30 We observed decreased levels of phosphorylated-Akt in MPP89 cells after treatment with celecoxib. Moreover, Bcl-2 protein levels decreased after celecoxib exposure, and this reduction was coupled to an increased expression of the pro-apoptotic protein Bax in the MPP89 cells (Fig. 4b). These data suggest that decreased expression of Bcl-2 in MM cells could be dependent on Akt dephosphorylation.

We have previously reported that the Bcl protein family members are involved in the MM cells apoptosis induced by antiangiogenic agents.31 Treatment with the angiostatic compound fumagillin resulted in a downregulation of the antiapoptotic Bcl-2 protein and a marked decrease in the Bcl-2:Bax ratio, which could be blocked by Bcl-2 overexpression. Therefore, celecoxib as well as fumagillin altered the expression of the same Bcl-family protein members, resulting in a shift in their ratios favoring apoptosis. Moreover, treatment with celecoxib resulted in a strong decrease in survivin expression over time, resulting in undetectable levels in MM cells after 24 hr. Survivin is overexpressed in the majority of human cancers, including MM,32 and it has been shown to inhibit apoptosis by binding to active caspase-3 and caspase-7.19 In our study, we observed that treatment with celecoxib after 24 hr activated caspase-3/7. This correlated with decreased levels of the survivin protein (Figs. 4b,c). Thus, the caspases-mediated apoptosis, in response to decreased in survivin expression, could be responsible for celecoxib-induced apoptosis. Recent reports associate survivin and Bcl-2 expression in cervical carcinoma tissues.33, 34 Our results support this association, as we observed decreased survivin and Bcl-2 proteins expression in MM cells treated with celecoxib.

Since we have shown that VEGF plays a key role on MM biology20 and Akt is an important molecular target of VEGF,23 we further investigated whether the VEGF pathway inhibits the antiproliferative effects of celecoxib protecting MM cells from celecoxib-induced apoptosis. We show that VEGF rescued celecoxib-induced Akt dephosphorylation and apoptosis in MM cells. A similar mechanism is involved in pancreatic cancer cells, in which IGF-I, a component of serum, by antagonizing the downregulation of Akt activity has been hypothesized as one of the factors involved in the resistance to celecoxib-induced antineoplastic actions.35 Thus, our data also clarify that the cytotoxic effect of celecoxib on cancer cells may be influenced by the extracellular environment and, in particular, by the autocrine/paracrine loop of VEGF.

Any risk of toxicity is an important consideration if a drug is to be considered for prevention in relatively healthy individuals at risk for MM, particularly because exposure to the drug may need to be continuous and of a potentially long duration. In contrast to typical NSAIDs, celecoxib retained chemopreventive efficacy within a dose range that did not cause any evident side effects on normal gut.18 Moreover, our physiologic data indicated a complete absence of any GI ulcerations or bleeding. Furthermore, no weight loss or impairment of weight gain was observed at 75 mg/kg/day of celecoxib used, implying no untoward side effects (data not shown). However, early studies suggest that celecoxib could be nephrotoxic.36 In this regard, an important preclinical consequence of our present study is that celecoxib and SU-1498, a VEGF antagonist,24 when administered together, were more effective in inhibiting MM growth in vitro than the treatment with these agents alone. Moreover, SU-1498 reduced IC₅₀ of celecoxib in vitro up to 65% (Fig. 6). Thus, it is anticipated that the administration of a combination of chemopreventive agents, such as celecoxib and angiostatic drugs, would avoid the kidney necrosis associated with celecoxib.

In conclusion, our present data demonstrate that celecoxib is effective for inhibiting the in vitro and in vivo tumorigenicity potential of MM cells when administered early. Furthermore, celecoxib specifically induces MM cell apoptosis by affecting various proteins regulating these processes. These results provide the first evidence that celecoxib are effective for the prevention and regression of MM cells in experimental models of MM and strongly support ongoing clinical trials in MM patients.

Acknowledgements

This work was supported by grants from the Associazione Italiana per la Ricerca sul Cancro (AIRC) and from the Italian Ministero dell'Università e della Ricerca Scientifica (ex 60%) to A.P. A.C. is a fellow of Fondazione Italiana per la Ricerca sul Cancro (FIRC).