We have found that the synthetic compound CC-5079 potently inhibits cancer cell growth in vitro and in vivo by a novel combination of molecular mechanisms. CC-5079 inhibits proliferation of cancer cell lines from various organs and tissues at nanomolar concentrations. Its IC50 value ranges from 4.1 to 50 nmol/L. The effect of CC-5079 on cell growth is associated with cell cycle arrest in G2-M phase, increased phosphorylation of G2-M checkpoint proteins, and apoptosis. CC-5079 prevents polymerization of purified tubulin in a concentration-dependent manner in vitro and depolymerizes microtubules in cultured cancer cells. In competitive binding assays, CC-5079 competes with [3H]colchicine for binding to tubulin; however, it does not compete with [3H]paclitaxel (Taxol) or [3H]vinblastine. Our data indicate that CC-5079 inhibits cancer cell growth with a mechanism of action similar to that of other tubulin inhibitors. However, CC-5079 remains active against multidrug-resistant cancer cells unlike other tubulin-interacting drugs, such as Taxol and colchicine. Interestingly, CC-5079 also inhibits tumor necrosis factor-α (TNF-α) secretion from lipopolysaccharide-stimulated human peripheral blood mononuclear cells (IC50, 270 nmol/L). This inhibitory effect on TNF-α production is related to its inhibition of phosphodiesterase type 4 enzymatic activity. Moreover, in a mouse xenograft model using HCT-116 human colorectal tumor cells, CC-5079 significantly inhibits tumor growth in vivo. In conclusion, our data indicate that CC-5079 represents a new chemotype with novel mechanisms of action and that it has the potential to be developed for neoplastic and inflammatory disease therapy. (Cancer Res 2006; 66(2): 951-9)
- multidrug resistance
- novel small molecules
Microtubules, a major type of cytoskeletal filament in cells, are formed from tubulin subunits, including α-tubulin and β-tubulin. Microtubules are the target of a large and diverse group of anticancer drugs derived from natural products. Given the success of this class of drugs in cancer treatment, microtubules can been considered one of the best cancer target identified to date ( 1). Increasingly, anticancer drugs targeting microtubules are being discovered from small molecules derived from either terrestrial plants or marine organisms. These small molecules can alter the dynamics of microtubule polymerization, leading to mitosis arrest, and consequently resulting in apoptosis of cancer cells ( 2). They can be separated into two groups according to their effect on microtubules: inhibitors of microtubule polymerization, such as vincristine, vinblastine, colchicines, cryptophycins, and combretastatins, and microtubule stabilizers, such as paclitaxel (Taxol), docetaxel (Taxotere), discodermolide, epothilones, laulimalide, and eleutherobins ( 3). These small molecules bind to a variety of sites on tubulin and have diverse effects on microtubule dynamics. In spite of the initial clinical success of some of these agents, the majority of cancer types are either inherently resistant to these agents or eventually develop resistance. In addition, there are problems of high toxicity (especially neurotoxicity), marginal oral bioavailability, poor solubility, difficulty of synthesis or isolation from natural sources, and, most importantly, development of drug resistance ( 4). Therefore, new synthetic compounds with oral bioavailability and an improved therapeutic index for first-line and second-line therapy are highly attractive therapeutics.
Tumor necrosis factor-α (TNF-α) is a cytokine produced in vivo mainly by activated macrophages/monocytes, with multiple effects on normal and malignant cells. Although initial TNF-α expression in response to infection or injury is considered beneficial, excessive production typically by activated monocytes and macrophages can produce significant pathologic changes. Recent research has indicated that elevated levels of TNF-α may contribute to the pathogenesis of cancers, possibly promoting cell growth and inhibiting apoptosis ( 5, 6). In addition, elevated levels of TNF-α may allow malignant cells to escape immune surveillance ( 7). Therefore, TNF-α may be a promising drug target for cancer. Recently, small-molecule TNF-α inhibitors, such as thalidomide and its analogues, have shown significant activity in the treatment of cancer ( 8, 9).
In the course of our drug screening efforts to discover novel anticancer and immunomodulatory agents from synthetic compounds, a novel class of cytokine inhibitory drugs has been developed with distinct and multiple activities (i.e., inhibition of both mitosis and TNF-α synthesis). We report here that CC-5079, as a representative of this class of compound, binds tubulin and destabilizes microtubules, leading to mitotic arrest and cell death. CC-5079 inhibits TNF-α production from lipopolysaccharide (LPS)–stimulated human peripheral blood mononuclear cells (PBMC) by inhibiting the enzymatic activity of phosphodiesterase type 4 (PDE4), an essential cyclic AMP (cAMP)–metabolizing enzyme in immune and inflammatory cells for endotoxin-activated TNF-α responses ( 10). CC-5079 displayed broad anticancer activity against a variety of cancer cell lines independent of multidrug resistance (MDR) phenotype status. Finally, CC-5079 exhibited significant in vivo antitumor activity in a severe combined immunodeficient (SCID) mouse xenograft model of human colorectal cancer.
Materials and Methods
Chemicals and biologicals. CC-5079 ( Fig. 1A ) was synthesized at Celgene Corp. (Summit, NJ). General chemicals as well as tubulin inhibitors Taxol, vinblastine, and colchicine were purchased from Sigma (St. Louis, MO). All compounds were dissolved in 100% DMSO before further dilution in cell culture medium. Final DMSO concentrations were kept at a constant 0.1% for all samples, including controls, unless otherwise stated. Streptavidin-coated yttrium scintillation proximity assay (SPA) beads were obtained from Amersham Pharmacia Biotech (Piscataway, NJ). [3H]Colchicine was from New England Nuclear (Boston, MA); [3H]Taxol and [3H]vinblastine were from Morevek Biochemicals (Brea, CA). Purified tubulin and biotinylated microtubule-associated protein-free bovine brain tubulin were from Cytoskeleton, Inc. (Denver, CO).
Cancer cell lines and primary cells. The human tumor cell lines HT29 (colon adenocarcinoma, HTB-38), HT-144 (melanoma, HTB-63), HCT-116 (colorectal carcinoma, CCL-247), A549 (non–small cell lung cancer, CCL-185), NIH:OVCAR-3 (ovary adenocarcinoma, HTB-161), PC-3 (prostate adenocarcinoma, CRL-1435), HCT-15 (colorectal adenocarcinoma, CCL-225), MCF-7 (breast adenocarcinoma, HTB-22), MES-SA (uterine sarcoma, CRL-1976), MES-SA/MX2 (CRL-2274), and MES-SA/Dx5 (CRL-1977) were purchased from American Type Culture Collection (Manassas, VA). MCF-7/Adr (also known as NCI/Adr-Res) was kindly provided by the Signal Research Division of Celgene. All cell lines were cultivated at 37°C, 5% CO2 in medium as published or as stated by American Type Culture Collection. The detailed characteristics of human parental MCF-7, MES-SA cell lines, and multidrug-resistant, P-glycoprotein (P-gp) 170–overexpressing MCF-7/Adr, MES-SA/MX2, MES-SA/Dx5, and HCT-15 cell lines have been reported ( 11, 12). Human umbilical vein endothelial cells (HUVEC) were kindly provided by Celgene Cellular Therapeutics (Cedar Knolls, NJ). PBMCs from normal donors were obtained by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) density centrifugation.
Cell proliferation assay. Cell proliferation was assessed in cancer cell lines, HUVECs, and human PBMCs by [3H]thymidine incorporation assay. Briefly, cells were seeded on 96-well microtiter plates 24 hours before addition of compound to allow them to adhere to plates. Each compound was tested at serial dilutions in triplicate. Following compound treatment, the cells were incubated at 37°C for additional 72 hours. [3H]Thymidine (1 μCi in 20 μL medium) was added to each well for the last 6 hours of incubation time. The cells were then harvested for detection of tritium incorporation with a TopCount Microplate Scintillation Counter (Packard Instrument Co., Meriden, CT). IC50s were calculated using a nonlinear regression analysis using GraphPad Prism program (San Diego, CA).
Flow cytometric analysis. For cell cycle analysis, cells were harvested following treatment with test agents for 24 hours and stained with propidium iodide (PI) according to the instruction of Cycle Test Plus DNA Reagent kits from Becton Dickinson (San Jose, CA). Samples were examined using FACSCalibur instrument (Becton Dickinson). Cell cycle distribution was analyzed with CellQuest version 3.1 acquisition software and the ModFit version 2.0 program.
For apoptosis analysis, cells were treated with test agents for 48 hours and then harvested. Double staining for FITC-Annexin V binding and for DNA using PI was done as described previously ( 13).
Tubulin polymerization assay. The polymerization of purified tubulin was monitored using the CytoDYNAMIX Screen (Cytoskeleton). Tubulin polymerization was monitored spectrophotometrically by the change in absorbance at 340 nm. The absorbance was measured at 1-minute intervals for 60 minutes using a PowerWave HT microplate reader (Bio-Tek Instruments, Highland Park, VT).
Immunofluorescence microscopy. Detection of α-tubulin in A549 cells by immunofluorescence was done as described previously ( 14). Briefly, cells were treated with test compounds for 24 hours and washed with PBS. Cells were then fixed and permeabilized with warm PBS buffer containing 3.7% formaldehyde and 1% Triton X-100 for 30 minutes. After washing cells twice with PBS and saturation with 1% mouse blocking serum in PBS for 30 minutes, staining was done with an anti-β-tubulin-FITC antibody (Sigma) alone or in the presence of 100 μg/mL PI. Cells were observed under an epifluorescence microscope (Nikon Instruments, Melville, NY) and imaged with a CCD camera using Image-Pro (Media Cybernetics, Silver Spring, MD).
Tubulin competitive binding SPA assay. The tubulin-binding assay was done as reported previously ( 15) using biotin-labeled tubulin, streptavidin-coated yttrium SPA beads, and 3H-labeled ligands ([3H]colchicine, [3H]Taxol, or [3H]vinblastine). Briefly, the binding mixture includes 0.08 μmol/L 3H-labeled ligand, 1 mmol/L GTP, and 0.5 μg biotinylated tubulin in 100 μL assay buffer containing 80 mmol/L PIPES (pH 6.9), 1 mmol/L MgCl2, 1 mmol/L EGTA, and 5% glycerol. The test compound and 3H-labeled ligand were added before tubulin. After incubation at 37°C for 2 hours, 20 μL SPA beads (80 μg in the assay buffer) were added. After further incubation for 30 minutes under agitation at room temperature, the SPA beads were allowed to settle down for 45 minutes, and scintillation counting was done on the TopCount Microplate Scintillation Counter.
Caspase assay. Caspase activity was determined according to the instructions from the assay kit supplier (R&D Systems, Minneapolis, MN). The results were expressed as fold change in caspase activity of drug-treated cells over the vehicle control cells.
Immunoblot analysis of cell cycle regulatory proteins. Cancer cells were treated with test compounds or 0.1% DMSO for 24 hours. Cells were trypsinized, spun down for 6 seconds in a microfuge, and immediately lysed in 0.1 mL lysis buffer containing 10 mmol/L Tris-HCl (pH 8.0), 10 mmol/L EDTA, 150 mmol/L NaCl, 1% NP40, 0.5% SDS, 1 mmol/L DTT, 1 mmol/L Na3VO4 plus Complete protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN) and then spun through a Qiashredder (Qiagen, Valencia, CA) for 1 minute and frozen on dry ice. Samples were diluted with 3× SDS sample buffer (New England Biolabs, Beverly, MA) and boiled 5 minutes. The mixture (∼30 μL) was loaded per lane on Tris-glycine polyacrylamide gels (Invitrogen, Carlsbad, CA), electrophoresed, and transferred to polyvinylidene difluoride (PVDF) membranes (Invitrogen). PVDF membranes were blocked for 1 hour at room temperature in PBS containing 0.05% Tween 20 and 5% nonfat milk powder and then blotted overnight at 4°C with antibodies against either MPM-2 (Upstate Biotechnology, Lake Placid, NY), bcl-2, cyclin B1, p53, p21, Cdc25C, or α-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA). Membranes were washed and incubated with horseradish peroxidase–conjugated anti-rabbit or anti-mouse IgG (Santa Cruz Biotechnology; 1:1,000 dilution) for 60 minutes at room temperature, washed thrice, and then developed using the Enhanced Chemiluminescence Plus Detection System (Amersham Biosciences, Piscataway, NJ).
PBMC culture and ELISA for TNF-α. PBMCs were prepared by density centrifugation on Ficoll-Hypaque. PBMCs, resuspended at 1 × 106/mL in complete RPMI 1640/10% FCS, were stimulated with LPS (1 μg/mL; Escherichia coli serotype 0127:B8; Sigma) in 24-well plates by incubation at 37°C in 5% CO2 for 24 hours with or without compounds (0.1-100 μmol/L). Cell-free supernatants were collected and stored in aliquots at −70°C until assayed by ELISA. Supernatants were assayed for TNF-α using an assay procedure and reagents provided by R&D Systems.
PDE4 assay. PDE4 purification from human histiocytic lymphoma U937 cells was done as described previously ( 16). Briefly, cells (1 × 109) were washed in PBS and lysed in cold homogenization buffer [20 mmol/L Tris-HCl (pH 7.1), 3 mmol/L 2-mercaptoethanol, 1 mmol/L MgCl2, 0.1 mmol/L EGTA, 1 μmol/L phenylmethylsulfonyl fluoride, and 1 μg/mL leupeptin]. Following homogenization, the supernatant was collected by centrifugation and loaded onto a Sephacryl S-200 column equilibrated in homogenization buffer. Phosphodiesterase was eluted in homogenization buffer and rolipram-sensitive fractions were pooled and stored in aliquots. Phosphodiesterase activity was assayed by a procedure described by DiSanto and Heaslip ( 17) and in the presence of varying concentrations of compounds. IC50 was determined from dose-response curves derived from three independent experiments.
Human tumor xenograft model. Female CB17 SCID mice (6-8 weeks old) were obtained from the Charles River Laboratory (Wilmington, MA) and maintained in microisolator cages under sterile conditions. The right hind legs of the mice were inoculated s.c. with HCT-116 cells suspended in sterile PBS (2 × 106 cells per mouse). On day 6, tumors of all mice were measured with a digital caliper and volumes were calculated with a formula of W2 × L / 2, where W is width (short axis) and L is length (long axis). Mice bearing tumor size ranging between 75 and 125 mm3 were pooled together and randomly distributed into cages. The mice were then ear tagged and cages were randomly assigned to treatment groups. On day 7, the tumors were measured and considered as starting volumes, and the mice were then given i.p. with either vehicle (N-methyl-2-pyrrolidone/polyethylene glycol 400/saline, 1:9:10), CC-5079 (5 and 25 mg/kg), or positive control Camptosar (10 mg/kg; Pfizer, Inc., New York, NY). Both doses of CC-5079 were initiated as once daily administrations for the first 5 days (days 7-11); the dosing of CC-5079 at 25 mg/kg was then switched to q3d × 4 (days 14, 17, 20, and 23) due to potential toxicity of the compound. Camptosar was given as q4d × 5 (days 7, 11, 15, 19, and 23). Mice were monitored daily for health status as well as tumor growth. Tumors were measured twice weekly.
Effects of CC-5079 on various cancer cells, HUVECs, and PBMCs. To explore the effect of CC-5079 on cancer cell proliferation, we treated human cancer cell lines from colon, prostate, breast, lung, ovary, uterus, and skin with different concentrations of compounds using Taxol and colchicine as positive controls. Cell proliferation was measured by [3H]thymidine incorporation assay. Proliferation of all cell lines was inhibited by CC-5079 in a concentration-dependent manner. The growth inhibition constants (IC50) of the different tumor cell lines ranged from 4.1 to 50 nmol/L for CC-5079.
One major mechanism of MDR of anticancer drugs is mediated by the overexpression of the multidrug transporter P-gp ( 18, 19). The antitumoral efficacy of CC-5079 was compared with Taxol and colchicine in a cytotoxicity assay using four MDR cell lines with high levels of P-gp: HCT-15, MCF-7/Adr, MES-SA/MX2, and MES-SA/Dx5. These cell lines were markedly resistant to Taxol and colchicine, whereas no significant resistance to CC-5079 was observed. As judged from the resistance factors (the ratio of the IC50 of the resistant cell line relative to the IC50 of its parental cell line), MDR cells were >2,000- and >300-fold resistant to Taxol and colchicine, respectively. In contrast, the cytotoxic efficacy of CC-5079 against tumor cells was unaltered by the MDR phenotype ( Table 1 ).
Because CC-5079 shows significant antiproliferative effects on various cancer cells, including MDR cell lines, it was tested for its toxicity on normal primary cells, including HUVEC and PBMC. As shown in Fig. 1B, HUVECs were found to be more sensitive to the cytotoxic effects of CC-5079 than cancer cells. The IC50 for CC-5079 on HUVEC was 0.17 nmol/L, which was 24- to 294-fold lower than that of cancer cells. This selectivity on HUVEC indicates that CC-5079 may be a potential antiangiogenic agent. The cytotoxicity of CC-5079 was also tested on PBMC at concentrations ranging from 1 to 100 μmol/L. Results ( Fig. 1C) showed that CC-5079 did not affect unstimulated PBMC DNA synthesis but showed inhibitory effects on phorbol 12-myristate 13-acetate (PMA) and ionomycin-stimulated PBMC. These results suggest that CC-5079 has significant cytotoxic effects on fast-proliferating cells but not on quiescent cells.
Effect of CC-5079 on molecular events of cell cycle progression of cancer cells. The strong antiproliferative activity of CC-5079 on cancer cells prompted us to test its effects on the cell cycle progression. As shown in Table 2 , there was an accumulation of MCF-7 cells with 4N DNA content (G2-M phase) and a concomitant decrease in cells in G0-G1 or S phase after treatment with either 20 nmol/L Taxol, 20 nmol/L colchicine, or 20 to 100 nmol/L CC-5079. In contrast, the cell cycle distribution of Taxol- and colchicine-treated MCF-7/Adr cells looked similar to the DMSO controls, whereas CC-5079 was equally effective in both cell lines. The accumulation of cells with 4N DNA content increased with time followed by a decrease in cells with 4N DNA content and an increase in subdiploid cells at later time points (48-72 hours), indicative of apoptotic cells (data not shown). As expected, the effect of CC-5079 on the cell cycle distribution was found to be independent of MDR status of cells.
We also assessed the effect of CC-5079 on the phosphorylation of mitotic regulatory proteins and other molecular events involved in cell cycle progression. As shown in Fig. 2A and B , treatment of MCF-7 cells with CC-5079, Taxol, or colchicine for 24 hours resulted in phosphorylation of proteins involved in mitosis as judged from slower migrating forms of the Cdc25C phosphatase or the antiapoptotic protein bcl-2. The changes in Cdc25C and bcl-2 coincided with the appearance of phosphoepitopes recognized by MPM-2, an antibody that recognizes phosphorylated polypeptides found only in mitotic cells ( 20). In contrast, the MDR derivative MCF-7/Adr cells seemed to be resistant to Taxol treatment, as there was a significant decrease in phosphorylation of Cdc25C and proteins detected by MPM-2. Similar to Taxol and colchicine, CC-5079 also markedly induced the accumulation of cyclin B. These results indicate that CC-5079 arrests the cell cycle and is capable of inducing mitotic arrest in cells, which express P-gp and are resistant to known antimitotic drugs.
Effect of CC-5079 on p53 and p21 expression, caspase activity, and apoptosis of cancer cells. p53 protects mammals from neoplasia by inducing apoptosis, DNA repair, and cell cycle arrest in response to a variety of stresses ( 21). When we examined the effect of CC-5079 on the expression of p53 in several cancer cell lines, we found that CC-5079 treatment of MCF-7 cells led to a marked increase in the level of p53 protein ( Fig. 2A). Similar effects were observed in cells treated with drugs, such as colchicine or Taxol. Unlike the MCF-7 cells, the MDR derivative MCF-7/Adr cells were found to constitutively express high levels of p53. CC-5079, Taxol, and colchicine were unable to increase the level of p53 protein in this MDR cell line. When p21 was examined in these treated MCF-7 cells, it was also appreciably induced by CC-5079 as well as by Taxol and colchicine. In contrast, p21 was undetectable by Western blot analysis in MCF-7/Adr cells treated with or without these test compounds ( Fig. 2A). These results indicate that constitutively expressed p53 in MCF-7/Adr failed to transactivate p21 expression whether cells were incubated in the presence and absence of test drugs and that the antiproliferative effects of CC-5079 in MCF-7/Adr cells are p53 independent.
To determine whether the antimitotic effects of CC-5079 eventually leads to apoptosis of cancer cells, we examined the ability of CC-5079 to activate caspase-3, caspase-8, and caspase-9 in A549 cells. Caspase-3, considered an “executioner” caspase, is implicated in the last and irreversible phase of the apoptotic caspase pathway and is activated by upstream “initiator” caspases, such as caspase-8 and caspase-9 ( 22). Indeed, CC-5079 (25-400 nmol/L), similar to Taxol (100 nmol/L), was able to induce a concentration-dependent increase in activity of these caspases in A549 cells ( Fig. 2C). A549 cells incubated with 100 nmol/L CC-5079 showed a time-dependent increase in caspase-3 activity compared with untreated cells, reaching a maximum at 30 hours (data not shown). Fluorescein-labeled Annexin V binding in conjunction with a PI dye exclusion test was also used to discriminate intact cells (FITC−/PI−), apoptotic cells (FITC+/PI−), and necrotic cells (FITC+/PI+) following the treatment of A549 cells with CC-5079 or Taxol. Exposure of A549 cells to 100 nmol/L Taxol or 100 to 1,000 nmol/L CC-5079 for 48 hours generally resulted in an increase in the percentages of apoptotic and necrotic cells and a reduction in the percentage of intact cells when compared with the DMSO control ( Fig. 2D).
Effect of CC-5079 on tubulin polymerization and tubulin binding of colchicine, vinblastine, and Taxol. Microtubules are a major component of the mitotic spindle assembly, which pulls the chromosomes apart at mitosis and then splits the dividing cell into two. The majority of antimitotic agents induce mitotic arrest by interacting with tubulin. We tested CC-5079 for its effect on tubulin. The polymerization of purified tubulin at 37°C in the presence of test compounds or DMSO control was monitored spectrophotometrically. Without treatment of drugs (DMSO control), tubulin subunits heterodimerize and self-assemble to form cylindrical microtubules in a time-dependent manner. Taxol treatment enhanced tubulin polymerization, whereas microtubule-depolymerizing agents colchicine and vinblastine, as well as CC-5079, prevented tubulin polymerization ( Fig. 3A ). Results in Fig. 3A also indicate that CC-5079 inhibited tubulin polymerization in a concentration-dependent manner.
To test whether CC-5079 also affects microtubule polymerization in cells, we treated A549 cells with or without test compounds for 24 hours and subsequently fixed and stained with FITC-conjugated anti-β-tubulin antibody and PI for in situ observation of microtubule network and nuclei with fluorescence microscopy. A549 cells without treatment ( Fig. 3B, a, DMSO control) exhibited the characteristic staining of individual microtubules with fluorescent anti-tubulin as indicated by a fine network “mesh” of microtubular materials. Nuclei of these control cells were intact and appeared normal as visualized by staining with PI. Taxol (40 nmol/L)–treated cells ( Fig. 3B, c) exhibited a significant increase of microtubule bundles in cytoplasm due to enhanced tubulin polymerization and highly condensed chromatin. CC-5079-treated cells exhibited condensed chromatin and a decreased amount of microtubular materials, which almost completely disappeared in the presence of 40 to 200 nmol/L CC-5079 presumably due to depolymerization of tubulin ( Fig. 3B, e and f). However, colchicine (40 nmol/L) and lower concentrations of CC-5079 (8 nmol/L) had only a slight effect on microtubules and nuclei ( Fig. 3B, b and d).
The binding sites for three tubulin ligands, Taxol, vinblastine, and colchicine, form the pharmacologic map of the surface of tubulin; they each bind at separate noncompeting sites ( 15). Using a competitive binding SPA, we determined if CC-5079 interacted directly with tubulin by binding to these ligand-binding sites. We found that CC-5079 competitively inhibited [3H]colchicine binding to biotinylated tubulin (IC50, 0.8 μmol/L), whereas it did not compete with [3H]vinblastine and [3H]Taxol ( Fig. 3C). Thus, we conclude that the antiproliferative activity of CC-5079 is based on binding to tubulin at the colchicine-binding site.
Effect of CC-5079 on TNF-α production from human PBMC and PDE4 enzymatic activity. Besides the antimitotic effect of CC-5079, this compound was also tested for an effect on cytokine production from PBMCs. CC-5079 at nontoxic concentrations was found to inhibit TNF-α production from LPS-stimulated PBMC in a dose-dependent manner. In contrast, Taxol and colchicine had no marked effect on TNF-α production at low concentrations. The IC50 values for CC-5079, Taxol, and colchicine are 270 nmol/L, >100 μmol/L, and 2.0 μmol/L, respectively. Because PDE4 is an essential cAMP-metabolizing enzyme involved in immune and inflammatory cells for LPS-activated TNF-α responses, these compounds were tested in a PDE4 enzymatic assay. CC-5079 was found to inhibit enzymatic activity of PDE4 purified from U937 cells in a dose-dependent manner (IC50, 350 nmol/L). However, neither Taxol nor colchicine had any significant effect on PDE4 activity at concentrations up to 100 μmol/L, although colchicine weakly inhibited TNF-α production from PBMC. These additional biological activities differentiate CC-5079 from other known tubulin interactive agents. Statistical analysis of results from structure-activity relationships (SAR) screening of 50 analogues of CC-5079 indicated that the inhibition of TNF-α from LPS-stimulated PBMC by CC-5079 and its analogues is correlated with their inhibitory effect on PDE4 activity ( Fig. 4A ), indicating that CC-5079 impairs TNF-α production by inhibiting PDE4 activity in monocytes. However, the inhibition of the growth of cancer cells and HUVEC by CC-5079 (and its analogues) was found to be unrelated with their inhibitory effect on PDE4 activity ( Fig. 4B).
In vivo antitumor activity of CC-5079 in mouse xenograft model. We have shown that CC-5079 has dual biological activities (i.e., inhibition of the in vitro growth of different human cancer cell lines by tubulin binding and impairment of the TNF-α production from activated PBMC by blocking PDE4 activity). To test whether CC-5079 also affects growth of solid tumors in vivo, we evaluated the activity in the HCT-116 colorectal cancer model in CB17 SCID mice. Exponentially growing HCT-116 cells were injected s.c. into SCID mice. When the tumors reached ∼100 mm3, the mice were given i.p. with vehicle, CC-5079, or the positive control Camptosar, an anticancer drug specifically targeting DNA topoisomerase I ( 23). CC-5079 at 5 mg/kg inhibited tumor growth by 22%, whereas CC-5079 at 25 mg/kg showed a significant inhibition of HCT-116 tumor growth by 46% (P < 0.001; Fig. 5 ). The body weights of these mice were not significantly affected, when CC-5079 at these doses was administrated.
In our systematic effort to discover novel anticancer and immunomodulatory agents from synthetic small molecules, a novel class of compound with a diarylalkene structure has been synthesized. These compounds show properties similar to our previously reported selective cytokine inhibitory drugs but have greater potency in inhibiting tumor cell growth. Distinct from the previous class of cytokine inhibitory compounds, these novel compounds are found to interact directly with tubulin, arrest the cell cycle, and induce apoptosis of tumor cells ( 24). In the present study, we have characterized the activities of CC-5079, one of the most potent among these compounds. We first tested CC-5079 in a cancer cell proliferation assay and found that CC-5079 inhibited the growth of various cancer cells with IC50 values at the nanomolar level ( Table 1). We also tested if CC-5079 has any antiproliferative effect on normal cells and found that CC-5079 was 10- to 100-fold more effective in inhibiting proliferation of endothelial cells than that of cancer cells, suggesting that CC-5079 may be useful in antiangiogenic therapy of cancer and other diseases ( 25). We found that CC-5079 at nontoxic concentrations was able to inhibit the capillary tube formation in vitro and adhesion and migration of endothelial cells (data not shown). Using the chick chorioallantoic membrane–layered expression scanning, CC-5079 was found to inhibit the expression of αvβ3 significantly ( 26). Our data further confirm the antiangiogenic potential of CC-5079. On the other hand, CC-5079 was found not to impair the viability of PBMC at concentrations as high as 100 μmol/L, although it did inhibit the proliferation of PMA plus ionomycin–stimulated human PBMC.
Our data clearly show that CC-5079 is efficacious in inhibiting the proliferation of a broad range of cancer cells as well as endothelial cells. CC-5079 was found to significantly induce cell cycle arrest in G2-M phase. The antimitotic effect of CC-5079 was further confirmed by the fact that treatment of cells with this compound triggered several molecular events involved in the mitotic signaling cascade ( 20, 27). These included phosphorylation of some important mitotic regulatory proteins, such as Cdc25C, bcl-2, and MPM-2 epitopes, and accumulation of cyclin B.
As a cellular gatekeeper for growth and division, the tumor suppressor gene p53 plays an essential role in sensing various stress signals and serves as a focal point of signal integration to decide whether cells will undergo growth arrest or apoptosis ( 28). Several antimitotic agents have been reported to induce p53 and inhibit cyclin-dependent kinases, p21WAF1/CIP1 (p21), and activate/inactivate several protein kinases, including Ras/Raf, mitogen-activated protein kinases, and p34Cdc2 ( 29). These kinases are associated directly or indirectly with phosphorylation of bcl-2 ( 2). Similar to the microtubule inhibitors Taxol and colchicine, CC-5079 induces p53 and p21 expression in cancer cells. It has been established that phosphorylation of bcl-2 and the elevations of p53 and p21 may lead to apoptosis ( 30). Treatment of cancer cells with CC-5079 induces activation of caspase-3, caspase-8, and caspase-9 and eventually results in cell undergoing apoptosis as shown by bivariate FITC-Annexin V/PI flow cytometric analysis ( Fig. 2D).
The fact that CC-5079 induced G2-M arrest suggests that the observed antiproliferative and apoptosis-inducing effect may be related to its effect on mitotic spindles in cells. Indeed, we found that CC-5079 disrupts microtubule assembly. The in vitro tubulin polymerization study shows that CC-5079 completely blocked microtubule assembly. The microtubule network in cytoplasm was drastically disrupted, when tumor cells were treated with CC-5079. Results from competitive binding studies clearly show that CC-5079 binds to tubulin at site distinct from vinblastine and Taxol and blocks colchicine binding to tubulin, indicating that CC-5079 affects tubulin dynamics by binding the colchicine-binding site.
The dramatic antimicrotubule effect of CC-5079 suggests that it has potential for cancer chemotherapy as an antimitotic agent. This is further supported by the in vivo data showing that CC-5079 has activity in the HCT-116 colorectal cancer xenograft model (P < 0.001). Although antimitotic agents have been widely used in the clinic to treat patients with neoplastic disease, a major drawback is the loss of efficacy over time because of the development of resistance. Drug resistance often develops through the expression of efflux pumps, such as P-gp and other MDR proteins ( 31). However, unlike colchicine, vinblastine, and Taxol, CC-5079 was equipotent in parental cancer cells versus MDR variants as we observed in cytotoxicity tests, cell cycle analysis, phosphorylation of mitotic regulatory proteins, and induction of p53 and p21 expression. Therefore, CC-5079 may circumvent P-gp-mediated drug resistance seen with most chemotherapeutic agents. CC-5079 may represent a new class of anti-tubulin agents with properties that may provide advantages over the other tubulin inhibitors.
CC-5079 also potently inhibits TNF-α production from monocytes by inhibiting cAMP-specific phosphodiesterase PDE4 with submicromolar activity. The second messenger cAMP is involved in a multitude of cellular processes, including growth and differentiation. PDE4s are critical components of cAMP signaling ( 32). Inhibition of PDE4 activity can induce growth suppression, apoptosis, and p53 and p21 proteins in human acute lymphoblastic leukemia cells ( 33, 34). After statistical analysis of results from testing of CC-5079 and 50 analogues, we can conclude that the TNF-α inhibitory effect of CC-5079 is due to its suppressive effect on PDE4. However, its antiproliferative activity was found not to correlate with PDE4 inhibitory activity. There is convincing evidence that an increase in the cellular levels of cAMP and inhibition of PDE4 lead to the induction of apoptosis in cancer cells ( 34, 35). TNF-α is a prominent member of the multifunctional TNF superfamily and has important roles in immunity as well as influencing apoptosis and cell survival. CC-5079 inhibits TNF-α production from monocytes but induces apoptosis of cancer cells, indicating that the apoptosis-inducing effect of the compound is mediated by pathways other than TNF/death receptor pathway. Accumulating evidence implicates TNF-α in inflammatory pathways that increase tumorigenesis ( 36, 37). TNF-α under specific conditions is a tumor promoter and helps to produce the toxic effects associated with conventional cancer therapy ( 5, 37). The PDE4 and TNF-α inhibitory effects of CC-5079, along with its antimitotic effect, may contribute to its anticancer effects and give this compound potential for therapy of some immunoinflammatory disorders.
In summary, our studies suggest that CC-5079 is a unique antimitotic and anti-TNF-α agent because of its two distinct biological functions. The first is the ability to bind directly to tubulin and to thereby perturb microtubule polymerization and the function of the spindle apparatus, which causes cells to arrest in mitosis and undergo apoptosis in cancer cells, whether cells overexpress P-gp or not. The second is the ability to inhibit PDE4 and thus TNF-α production from monocytes. In view of these dual effects, CC-5079 is a novel drug lead that might be particularly effective for the treatment of neoplastic and inflammatory diseases. CC-5079 and its analogues are currently under development for oncology by EntreMed, Inc. (Rockville, MD).
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.
We thank Dr. William E. Fogler (EntreMed) and Bernd Stein (Celgene) for critical review of the article.
- Received June 15, 2005.
- Revision received September 19, 2005.
- Accepted November 9, 2005.
- ©2006 American Association for Cancer Research.