SEARCH

SEARCH BY CITATION

Keywords:

  • rimonabant;
  • colon cancer;
  • mitotic catastrophe;
  • ACF formation

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The selective CB1 receptor antagonist rimonabant (SR141716) was shown to perform a number of biological effects in several pathological conditions. Emerging findings demonstrate that rimonabant exerts antitumor action in thyroid tumors and breast cancer cells. In our study, human colorectal cancer cells (DLD-1, CaCo-2 and SW620) were treated with rimonabant and analyzed for markers of cell proliferation, cell viability and cell cycle progression. Rimonabant significantly reduced cell growth and induced cell death. In addition, rimonabant was able to alter cell cycle distribution in all the cell lines tested. Particularly, rimonabant produced a G2/M cell cycle arrest in DLD-1 cells without inducing apoptosis or necrosis. The G2/M phase arrest was characterized by a parallel enhancement of the number of mitoses associated to elevated DNA double strand breaks and chromosome misjoining events, hallmarks of mitotic catastrophe. Protein expression analyses of Cyclin B1, PARP-1, Aurora B and phosphorylated p38/MAPK and Chk1 demonstrated that rimonabant-induced mitotic catastrophe is mediated by interfering with the spindle assembly checkpoint and the DNA damage checkpoint. Moreover, in the mouse model of azoxymethane-induced colon carcinogenesis, rimonabant significantly decreased aberrant crypt foci (ACF) formation, which precedes colorectal cancer. Our findings suggest that rimonabant is able to inhibit colorectal cancer cell growth at different stages of colon cancer pathogenesis inducing mitotic catastrophe in vitro. © 2009 UICC.

Rimonabant (SR141716) is the first described highly selective antagonist for the cannabinoid receptor type 1 (CB1).1 Because of its ability to block the CB1 receptor, which controls food intake at central and peripheral level, rimonabant has been in clinical development for the treatment of obesity and its metabolic complications, including dyslipidemia, type 2 diabetes and atherosclerosis.2 However, rimonabant shows a plethora of pharmacological effects in a number of physiopathological conditions including cancer.3

The first observation of a rimonabant potential antitumor action was provided by our group in rat thyroid cancer cells (KiMol) in vitro and in thyroid tumor xenografts induced by KiMol injection in athymic mice. In this model, rimonabant was able to partially prevent the antitumor effect of endocannabinoid degradation inhibitors as well as of a metabolically stable analogue of the endocannabinoid anandamide (2-methylarachidonyl-2′-fluoro-ethylamide, Met-F-AEA). However, rimonabant, when used alone in the same model and at the dose able to counteract the Met-F-AEA effect, did not enhance tumor growth but instead it exerted a small, although significant, antitumor effect on thyroid tumors both in vitro and in vivo.4 Other authors demonstrated a potential antiproliferative effect of rimonabant: it decreased viability of primary mantle lymphoma cells isolated from tumor biopsies,5 whilst a 48 hr incubation of C6 glioma cells with rimonabant produced an enhancement of anandamide-mediated apoptosis.6 Nonetheless, in C6 rat glioma cells, rimonabant failed to revert the antiproliferative effect of CB1 agonists, whereas a combination of CB and vanilloid (VR) receptor antagonists (rimonabant, SR144528 and capsazepine) blocked completely the antiproliferative effect of anandamide.7 Recently, we reported that rimonabant inhibits human breast cancer cell proliferation, being more effective in highly invasive metastatic MDA-MB-231 cells than in less-invasive T47D and MCF-7 cells, also depending on the presence and expression levels of the CB1 receptor. Rimonabant antiproliferative effect on MDA-MB-231 cells was not associated with apoptosis but characterized by a G1/S-phase cell cycle arrest, and required lipid rafts/caveolae integrity to occur.8 Interestingly, rimonabant used in combination with the CB1 agonist Met-F-AEA showed synergistic/additive effects in the blockade of human lymphocyte proliferation. Also in this case, the effect was not due to apoptosis but was characterized by a cell cycle arrest.9 Moreover, in a preliminary study by our group, micromolar concentrations of rimonabant significantly inhibited the viability of human adenocarcinoma DLD-1 cells after 24 and 48 hr of treatment in a concentration-dependent manner.10

Commonly used anticancer drugs eliminate tumor cells predominantly by triggering apoptosis. However, many cancer cell types especially those lacking functional p53, escape apoptosis and enhance their survival potential thus resulting in drug resistance. Nevertheless, these cancer cells can be still eliminated through nonapoptotic mechanisms by causing a cell cycle arrest, which ultimately leads to senescence, autophagy or mitotic catastrophe.11 Mitotic catastrophe is defined as a type of mammalian cell death characterized by entering mitosis with DNA damage. Hallmarks of this process are mitotic arrest, mitotic spindle disorganization, abnormal chromosome segregation and formation of aneuploid cells. The mechanisms at the basis of mitotic catastrophe are still poorly understood. It has been proved that the main events leading to mitotic catastrophe result from both a combination of deficient cell cycle checkpoints and DNA damage12, 13 and by the inactivation or dysfunction of the chromosomal passenger complex (CPC), which is necessary to coordinate all chromosomal and cytoskeletal events during mitosis.14 Starting from these observations, we investigated the effects of rimonabant on the proliferation of DLD-1, Caco-2 and SW620 cells and analyzed in detail the molecular mechanisms underlying this effect on DLD-1 cells. We also explored the antitumoral activity of rimonabant in a murine model of colon carcinogenesis, to assess whether rimonabant could act as an antitumor agent also in vivo in the early stage of colon cancer pathogenesis.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Drugs

Rimonabant (SR141716, N-(piperidino-1-yl)-5-(4-chlorophenyl)-1-(2,4dichlorophenyl)-4-methyl-pyrazole-3-carboxamide) was kindly provided by Sanofi-aventis (Montpellier, France) and dissolved in dimethylsulfoxide (DMSO). DMSO (0.1% final concentration in culture) used as solvent control did not induce any positive result in all in vitro assays. For in vivo experiments, azoxymethane (AOM) was purchased from Sigma (Milan, Italy) and dissolved in saline; rimonabant was first dissolved in DMSO and then suspended in a lipophilic solution (Peceol/Gelucire 44/14, gift from Indena, Milan, Italy) (0.4% DMSO, 96.6% lipophilic solution). The drug vehicle (50 μl/mouse) had no effect on the response under study.

Animals

Experiments were performed on female 5-week-old C57BL/6N mice (Harlan Italy, Corezzana, MI; 20–22 g). Mice were housed in polyethylene cages and given rodent chow and water ad libitum, except for the 12 hr immediately preceding the sacrifice of the animals. Principles of laboratory animal care (NIH publication No. 86–23, revised 1985) were followed. Moreover, all animal experiments complied with the Italian D.L. n 116 of January 27, 1992 and associated guidelines in the European Communities Council Directive of November 24, 1986 (86/609/ECC).

Cell cultures

Human colorectal cancer cells DLD-1, undifferentiated Caco-2 and SW620 were obtained from the ICLC (IST, Genoa, Italy). DLD-1 cells having a pseudo-normal karyotype 2n = 46, were routinely grown in monolayers in RPMI supplemented with 10% heat inactivated fetal bovine serum (FBS) and 2 mM L-glutamine. Caco-2 and SW620 cells were grown in Dulbecco's Modified Eagle medium (DMEM) plus 10% FBS and 2 mM L-glutamine. One percent of nonessential aminoacids was added to the Caco-2-culture medium. All cell lines were incubated at 37°C in a humidified atmosphere containing 5% CO2. All the cell culture reagents were purchased from Sigma.

Cell growth and colony formation assays

Exponentially growing cells (2 × 106) were seeded in 6-well plates. Cells were exposed to increasing concentrations of rimonabant for 24 and 48 hr prior harvesting (trypsinization), stained with trypan blue (0.5% solution, Sigma) and counted by using a hemocytometer.

Cells (200 cells/well) were seeded in 6-well plates and soon after rimonabant was added to the culture medium at indicated concentrations. Cells were incubated for 10 days. Colonies were stained with crystal violet (Sigma) and counted manually.

Cell cycle and apoptosis assays

Cells (1 × 106) were plated in 100 mm culture dishes. Cells were left to adhere to the wells, then culture medium was removed and cells were starved overnight to enrich cells in G1 phase. After the starvation period, the medium at 0.5% FBS was removed and replaced with complete cell culture medium. Rimonabant was added at indicated concentrations, and the cells were incubated for another 24 hr. To analyze cell cycle progression, both floating and adherent cells were collected, washed twice with PBS and resuspended in 300 μl of PBS; fixed in 700 μl of ethanol and kept at −20°C for at least 6 hr. Propidium iodide (PI 10 μg/ml) (Sigma) dissolved in PBS containing 100 U/ml DNase-free Rnase was added to the cells. Cells were subjected to flow cytometric analysis using ModFit LT v3.0 from Verity Software House (Topsham, ME) program. To analyze apoptosis, annexin V staining was performed. Briefly, collected cells were washed twice with PBS, resuspended in annexin V binding buffer and stained with AnnexinV-FITC for 15 min at room temperature. Samples were protected from light. Subsequently, cells were collected, washed and resuspended in Annexin buffer. PI was added to the cells before flow cytometric analysis. Data were analyzed using Cell Quest software (BD Biosciences). For flow cytometric evaluation of both cell cycle and apoptosis, 10,000 events corrected for debris and aggregate were analyzed for each sample.

Mitotic index, chromosome aberrations and polyploidy

DLD-1 colon cancer cells (1 × 106) were plated in 100 mm culture dishes 24 hr before performing each experiment, starved and treated as in the cell cycle analysis. To maintain a stable karyotype, cells were stored in liquid nitrogen and used at no more than 2 passages. Cells were treated at the indicated concentrations of rimonabant. To visualize metaphase cells, Colcemid (Sigma, 0.2 μg/ml) was added 2 hr before harvesting. At the end of the incubation period, cells were removed from each culture dish by trypsinization and centrifuged at 1,200 rpm for 10 min. Harvested cells were treated with a prewarmed hypotonic solution (0.075 M KCl) for 20 min and fixed in cold methanol/acetic acid (3:1) for 20 min. Fixed cells were dropped onto slides and stained with a 5% Giemsa solution and analyzed under a light microscope at 1,000× magnification.

Mitotic index was determined as the percentage of metaphases over a total of 1,000 nuclei analyzed at random. Polyploid cells and structural chromosome aberrations were determined scoring 50 metaphases for each treatment on blindly coded slides.8 The aberration frequency was calculated as total number of aberrations/total number of cells analyzed, × 100 and the percentage of aberrant cells was estimated as the total number of cells with at least one chromosome aberration/total number of cells scored, × 100.15

Immunoprecipitation and western analysis

DLD-1 cells, starved and treated with rimonabant (2.5 μM) were collected by centrifugation, washed twice with PBS and resuspended in RIPA buffer (NaCl 150 mM, 1% triton X-100 pH 8.0, 0.5% sodium deoxycholate, 0.1% SDS, 50 Mm Tris, pH 8.0) at 4°C and centrifuged at 15,000 rpm for 30 min. Supernatants were collected and protein concentration determined by Bio-Rad protein assay.

Equal amounts of protein extracts (40 μg) were boiled in Laemmli's buffer, fractionated on 12% SDS-PAGE and then transferred to nitrocellulose membranes (Amersham, GE Healthcare). Membranes were blocked in TBS-T (50 mM Tris, 135 mM NaCl, and 5 mM KCl, 0.1% Tween-20) containing 5% nonfat dry milk, washed in TBS-T, then incubated overnight at 4°C with total anti-Aurora B, Chk1 and p38/MAPK, anti phosphoChk1 (Ser296), anti phosphop38/MAPK (Thr180/Tyr182) (all from Cell Signaling Technology) and anti PARP-1 (Santa Cruz Biotechnology). Anti α-tubulin was used as loading control (Santa Cruz Biotechnology). Blots were probed with mouse or rabbit secondary antibodies for 1 hr and then developed using enhanced chemiluminescence (ECL) system (Amersham, GE Healthcare).

The level of cyclin B1 was assessed by immunoprecipitation and Western blotting. Briefly, collected cells were lysed. The lysates (total cell extract 500 μg) were incubated for 12 hr at 4°C with 3–4 μg of mouse monoclonal antibodies directed against cdk1 and protein A/G agarose (Santa Cruz Biotechnology). The eluates from beads were analyzed by immunoblotting using a mouse monoclonal primary antibody against human cyclin B1 (Santa Cruz Biotechnology). Western analysis was performed at least 3 times and more representative results are shown.

Aberrant crypt foci determination

Aberrant crypt foci (ACF) were induced by the genotoxic chemical AOM.16 Mice were randomly divided into 3 groups (n = 9 per group) as follows:

Group 1 was treated with the vehicle used to dissolve AOM;

Group 2 was treated with AOM plus the vehicle used to dissolve rimonabant;

Group 3 was treated with AOM plus rimonabant (3 mg/kg).

AOM (16 mg/kg in total, i.p.) was administered during the first (3 m/kg at day 1 and 5), third (3 mg/kg at day 1 and 5) and at the 17th week (2 mg/kg at day 1 and 5) of treatment.

Rimonabant (3 mg/kg) was given i.p. every other day for the whole duration of the experiment (starting from a week before the first injection of AOM). This dose was selected on the basis of previous studies.17, 18 All animals were euthanized by asphyxiation with CO2 6 months after the first injection of AOM. Based on our laboratory experience, this time (at the dose of AOM used) was associated with the occurrence of a significant number of ACF, with no appearance of polyps or tumors.19 The lack of polyps and tumors formation could be due to the use of C57BL/6N mice, which are probably more resistant compared to other strains such as C57B1/6.20

For ACF determination, the colons were rapidly removed after sacrifice, washed with saline, opened longitudinally, laid flat on a polystyrene board and fixed with 10% buffered formaldehyde solution before staining with 0.2% methylene blue in saline. The colons were examined using a light microscope at 40× magnification. Aberrant crypts were identified as previously described.16, 19 Briefly, in comparison to normal crypts, aberrant crypts have greater size, larger and often elongated openings, thicker lining of epithelial cells, compression of adjacent crypts and are more darkly stained with methylene blue. ACF containing large ACF with 4 or more crypts per focus were analyzed. To determine crypt multiplicity, the number of aberrant crypts in each focus was recorded.

Statistical analysis

Data were expressed as mean ± standard deviation (SD). Statistical differences between the treatments and the control were evaluated by one-way analysis of variance (ANOVA). In the case of a significant result in the ANOVA, Student's t test was performed for all in vitro experiments.7 For the in vivo experiments, ANOVA was followed by the Dunnett's test. The Chi-Squared test was used to evaluate the significance between animals with or without ACF.15 A p value less than 0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Rimonabant reduces colon cancer cell growth

DLD-1, Caco-2 and SW620 cells were cultured in the presence and absence of rimonabant at increasing concentrations ranging from 0.1 to 20 μM, and cell recovery was determined at 24 and 48 hr after treatments. Results show that rimonabant caused a concentration- and time-dependent decrease of cell growth in all the cell lines tested (Fig. 1). The reduction of cell proliferation was associated with an increase of cell death at both 24 and 48 hr of treatment (Fig. 1, low panels). The effects on cell proliferation and cell death were higher in SW620 and DLD-1 cells than in Caco-2 cells.

thumbnail image

Figure 1. Rimonabant decreases cell recovery and induces cell death in colon cancer cells. DLD-1, Caco-2 and SW620 cells were treated with rimonabant (0–20 μM) for 24 and 48 hr. Cell recovery was analyzed using a hemocytometer. Results were significant starting from 2.5 μM concentration of rimonabant (p < 0.05). Cell death was assessed by trypan blue staining, and the percentage of cell killing is reported (*p < 0.05 and **p < 0.01 vs. control). All data are shown as mean ± SD of three independent experiments each done in triplicate.

Download figure to PowerPoint

Rimonabant inhibits cell cycle progression

To better characterize the inhibition of cell growth induced by rimonabant and to correlate this effect with cell cycle progression, we analyzed cell cycle distribution at concentrations able to induce the highest delay of cell recovery without inducing a severe cytotoxicity. After 24 hr treatment, rimonabant (from 2.5 to 10 μM) induced a significant decrease of the percentage of cells in the G1 phase. It also induced a G2/M accumulation particularly at 2.5 μM concentration in all the cell lines (Fig. 2). Moreover, an increase in S phase was detectable in Caco-2 and SW620 cells. Notably, DLD-1 cells exhibited the highest blockade of cell cycle progression in G2/M phase, resulting in a 2- and 1.5-fold increase in the proportion of cells within G2/M at 2.5 and 5 μM of rimonabant, respectively. No significant changes in S phase were observed in DLD-1 cells (Fig. 2).

thumbnail image

Figure 2. Rimonabant affects cell cycle phase distribution in colon cancer cells. DLD-1, Caco-2 and SW620 were exposed to rimonabant for 24 hr (0–10 μM). More representative flow cytometric determinations are shown. The histograms report the percentage of cells in each cell cycle phase (mean ± SD of three separate experiments). *p < 0.05 and **p < 0.01 vs. control.

Download figure to PowerPoint

Rimonabant does not induce apoptosis and inhibits colony formation of DLD-1 cells

To assess in detail the mechanisms underlying the observed effects, we focused on DLD-1 cells determining whether cell death was ascribable to apoptosis. Flow cytometric analysis of cell death at 24 hr treatment, revealed an increase in PI positive cells at 5 and 10 μM showing a cytotoxic activity of the compound not accompanied by apoptosis (Table I) or by a reduction in DNA synthesis, as assessed by 3H-thymidine-incorporation assay (data not shown). Similar results were obtained in Caco-2 and SW620 cells (data not shown). Moreover, to test rimonabant's sensitivity on DLD-1 cell proliferation after a prolonged in vitro exposure, we also performed human tumor clonogenic assay. Results demonstrated that rimonabant was able to inhibit colony formation in a dose-related manner, reducing the number and dimension of colonies. The effectiveness of rimonabant was marked at all concentrations tested (Fig. 3).

thumbnail image

Figure 3. Long-term effect of rimonabant on DLD-1 cells. Colony formation was analyzed after 10 days treatment with rimonabant (0–10 μM); the figure is one representative of six.

Download figure to PowerPoint

Table I. FLOW CYTOMETRIC DETERMINATION OF THE PERCENTAGE OF CELL DEATH IN 24-HR RIMONABANT-TREATED CELLS
Treatment PI+Annexin V+/PI+Annexin V+
  1. Results are given as mean ± SD of three independent experiments. *p < 0.05 vs. control.

Control 29.3 ± 9.02.5 ± 1.40.30 ± 0.08
Rimonabant2.5 μM37.8 ± 7.91.6 ± 0.30.17 ± 0.10
5 μM51.1 ± 2.2*2.7 ± 0.50.20 ± 0.10
10 μM58.4 ± 4.0*2.5 ± 1.70.10 ± 0.03

Rimonabant induces mitotic catastrophe

On the basis of previous findings, we obtained evidence that rimonabant could reduce cell growth and induce cell death, but apoptosis had no role in this phenomenon. Other forms of cell death are known to occur in the presence of an early arrest of cells in G2/M phase of the cell cycle, especially in cancer cells that are null or mutant for p53, as in the case of DLD-1 cells. To characterize this nonapoptotic cell death, the role of mitotic catastrophe was assessed in rimonabant-treated cells. Hallmarks of mitotic catastrophe are the entry of cells into mitosis, despite the presence of damaged DNA, and the increase of polyploid cells as a consequence of the mitotic spindle damage.11, 12 To determine whether rimonabant could induce this type of cell death, the mitotic index, chromosome aberrations and polyploidy were analyzed. Results showed (Fig. 4a) a marked and statistically significant enhancement of the number of cells in mitotic division in rimonabant-treated cells compared to the controls. The finding that rimonabant, after 24 hr treatment, blocked the cells in the G2/M phase of the cell cycle enhancing cells in mitosis, suggest that the drug specifically caused mitotic arrest. Moreover, we observed a dose-related increase in the percentage of polyploid cells particularly from 2.5 to 5.0 μM of rimonabant (Fig. 4a). Chromosome aberration analysis (Fig. 4b) showed that rimonabant-treatment produced a marked increase of both the percentage of aberrant cells (about 35%) and structural aberrations (about 50%). Cytogenetic assay also indicated that most aberrant cells contained multiple structural aberrations (Fig. 4b), mainly chromatid breaks and end-to-end fusions, (Fig. 4c) thus providing evidence that rimonabant induced an extensive DNA damage and chromosome misjoining events in these cells.

thumbnail image

Figure 4. Rimonabant enhances mitotic index, polyploidy and chromosome aberrations. Cells were treated with rimonabant (0–10 μM) for 24 hr. Mitotic index and polyploidy (a) and the percentage of aberrant cells and structural chromosome aberrations were determined (b). Results are shown as mean ± SD of three separate experiments (*p < 0.05 and **p < 0.01 vs. control). (c), representative normal metaphase having a model chromosome number of 46, in absence of rimonabant treatment (i) and metaphases treated with rimonabant 2.5 μM (ii, iii) undergoing chromatid/chromosome breaks (arrows) and homologous as well as nonhomologous, complete or incomplete, end-toend fusion (arrows with enlarged head).

Download figure to PowerPoint

Rimonabant affects DNA damage and spindle assembly checkpoints

Aimed to characterize the mitotic catastrophe observed cytologically and to identify the molecular targets of rimonabant, we investigated the expression level of proteins essential for the regulation of M-phase progression, chromosome segregation and DNA repair. Because the increase of cyclin B1 expression and the formation of cyclin B1/cdk1 complex are the rate-limiting steps during mitosis and the disruption of cyclin B1 is normally prevented until mitosis has been completed,12, 21 we analyzed cyclin B1 protein expression levels after the treatment with 2.5 μM rimonabant, that is, the concentration at which we evidenced the best induction of mitotic catastrophe. Results showed a persistent increase of cyclin B1 starting from 1 hr of treatment and until 4 hr (Fig. 5). To assess whether rimonabant could modulate the spindle assembly checkpoint, we examined the expression of the main subunit of the CPC, Aurora B.14, 22 We observed a general reduction of total protein expression levels of Aurora B (up to 1 hr and from 4 to 24 hr). At 2 hr, the level of Aurora B is slightly higher in the rimonabant-treated cells than in the control cells. Furthermore, we determined the expression of phosphorylated p38MAPK and Poly(ADP-ribose) polymerase (PARP-1) that are major proteins involved in maintaining genomic integrity and activated as a consequence of genotoxic stress.23–25 Rimonabant-treatment inhibited total PARP-1 expression and phosphorylation of p38/MAPK in a time-dependent manner. The inhibition of p38/MAPK begins at 30 min and persists up to 1 hr. In contrast, PARP-1 expression level decreases after 1 hr, and this reduction is observed up to 4 hr of treatment. Notably, no cleaved form of PARP-1 was observed (Fig. 5).

thumbnail image

Figure 5. The spindle assembly checkpoint and the DNA damage response checkpoint are perturbed by rimonabant. DLD-1 cells were treated with 2.5 μM rimonabant for the indicated time. Cell lysates were immunoprecipitated with anti-cdk1 antibody. The amounts of cdk1 and cyclin B1 into immunocomplexes were analyzed by Western blotting with specific antibodies (upper panel). Cell lysates were analyzed by Western blotting with antibodies against Aurora B, PARP-1 and phosphorylated p38MAPK and Chk1 (lower panels). Same filters were stripped and reprobed with both total Chk1 and p38/MAPK and anti α-tubulin as loading control. Immunoreactive bands were quantified using quantity one program. Each experiment was performed three times, and more representative results are shown (raw data are given as Supporting Information material Fig. 1).

Download figure to PowerPoint

Finally, we investigated the expression of phosphorylated Chk1, which has been considered one of the major regulators of cell cycle progression.13 Chk1 also regulates the DNA repair machine, and its activation is correlated with the activation of the ATM/ATR pathway.26 We found that rimonabant reduced Chk1 phosphorylation from 30 min to 1 hr thus suggesting that the rimonabant-induced mitotic catastrophe is also mediated by downregulation of the Chk1 enzyme. No differences in protein expression levels were observed by analyzing total p38/MAPK and Chk1 (Fig. 5). Similar results were obtained by treating cells with 5.0 μM rimonabant (data not shown).

Rimonabant prevents ACF formation in mice

Because we studied the anticancer effects of rimonabant on human colon cancer cells mutated for p53, and this circumstance occurs as a late event in colorectal cancer pathogenesis,27 we investigated whether rimonabant could exert also a chemopreventive action on earlier neoplastic lesions. We used the murine model of AOM-induced colon carcinogenesis and evaluated the efficacy of rimonabant to inhibit the formation of ACF. The dose and administration of rimonabant were selected on the basis of previous published data dealing with the effects of the drug in subchronic or chronic experiments.17

The percentage of mice with ACF, the percentage of mice with ACF containing 4 or more crypts, the total number of ACF/mouse and the number of ACF/mouse with 4 or more crypts, observed after the treatment with rimonabant either in the presence or in the absence of AOM- are shown in Table II. AOM given alone induced the appearance of ACF in all the animals treated, with 57% of mice presenting in the colon ACF with 4 or more crypts. The average number of ACF for each mouse was 12.8 ± 2.4 (0.57 ± 0.20 ACF with 4 or more crypts). Notably, rimonabant significantly reduced the percentage, as well as the number of mice with ACF containing 4 or more crypts, without reducing the total number of ACF/mouse.

Table II. REDUCTION IN THE FORMATION OF ABERRANT CRYPT FOCI (ACF) BY RIMONABANT
TreatmentMice with ACF (%)Mice with ACF containing ≥4 crypts (%)No of ACF/mouseNo of ACF/mouse containing ≥4 crypts
  • *

    Azoxymethane (AOM) (16 mg/kg in total, IP) was administered during the first (3m/kg at day 1 and 5), third (3 mg/kg at day 1 and 5) and at the 17th week (2 mg/kg at day 1 and 5). Rimonabant (3 mg/kg, IP) was administered every other day for the 6 month-period of the experiment; then, colons were collected and ACF were determined. Results are given as mean ± SD. Representative images of the data are given as Supporting Information material, Figure 2. *p < 0.05 vs. AOM (n = 6–9 mice for each experimental group).

Control
AOM1005712.8 ± 2.40.57 ± 0.20
AOM +rimonabant10014*10.0 ± 1.20.29 ± 0.29*

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Colorectal cancer is a leading cause of morbidity and mortality in Western countries.28 In this article, we demonstrated that rimonabant induced anticancer effects in human adenocarcinoma cells in vitro and prevented the formation of aberrant crypt foci in a mouse model of chemically-induced colon carcinogenesis. The effect of rimonabant on cancer cell growth was characterized by the inhibition of cell proliferation associated to increased cell death. These results, together with the observation that rimonabant-induced S-G2/M arrest in both Caco-2 and SW620 cells and G2/M blockade in DLD-1 cells without inducing apoptosis, support our data previously obtained in MDA-MB-231 breast cancer cells.8 However, it appears that in colon cancer cells tested in our study, higher concentrations of rimonabant, likely much higher than those required to antagonize the CB1 receptor, are required to achieve a reduction of cell proliferation comparable to that observed in breast cancer cells.4, 8 Furthermore, our findings that rimonabant increases the number of DLD-1 mitotic cells suggest that the cell cycle moves beyond G2 and arrests early within mitosis. The mitotic index is inversely correlated with the inhibition of cell proliferation and survival rate, indicating that the mitotic arrest could be responsible for the rimonabant-induced cell death. These observations and the findings that rimonabant produced a marked increase of structural chromosome aberrations, which are the clear sign of the entry in mitosis of cells with unrepaired DNA strand breaks,29 unequivocally show that the efficacy of rimonabant to inhibit DLD-1 cell proliferation is ascribable to the induction of mitotic catastrophe. Indeed, mitotic catastrophe (also known as mitotic/proliferative cell death) is a form of death, not yet completely understood, resulting from a prolonged cell growth arrest during mitosis and extensive DNA damage or disturbed spindle formation, as the consequence of deficient cell cycle checkpoints (the DNA structure checkpoint and the spindle assembly checkpoint).13, 30 In line with the hypothesis that rimonabant could affect mitotic spindle assembly and chromosome segregation, we observed a slight but significant increase of the percentage of polyploid cells, indicating that a number of cells endoduplicated DNA but were incapable to achieve cytokinesis. These results also agree with the S-G2/M arrest observed in Caco-2 and SW620 cells. Molecular data on DLD-1 cells also corroborate results obtained by cytogenetic and cell cycle progression analyses, as we evidenced that rimonabant-induced mitotic catastrophe was mediated by a sustained increase of cyclin B1/cdk1 complex associated with a decrease of both Aurora B and Chk1, which are proteins needed for G2/M transition, CPC formation and correct chromosome segregation during mitosis.12–14, 22, 26 Our results are in agreement with the hypothesis that rimonabant-induced inhibition of Chk1 could facilitate M-phase entry by activation of cdk1 kinase and accumulation of cyclin B1. In contrast, Chk1 downregulation might be also responsible for the inhibition of anaphase as already proposed for other Chk1 inhibitors.31 Moreover, rimonabant decreased Aurora B expression at all time points but 2 hr. This fluctuation could be due to the involvement of Aurora B in different events of mitosis (e.g. spindle assembly checkpoint, DNA damage and anaphase checkpoints).14 The rimonabant-mediated inhibition of Aurora B and phosphorylated Chk1 kinases could be of particular concern because recent data pointed out that selective suppression of these proteins could be a promising therapeutic strategy for cancer treatment.31–33 To study the molecular mechanisms underlying the induction of chromosome aberrations, we also studied p38MAPK and PARP-1, which are two key proteins required in maintaining genomic integrity.23–25 Recent evidence showed that in PARP-1−/− cells p38MAPK activation is defective, suggesting a common activation pathway.34 Our results demonstrated that rimonabant inhibited Chk1 and p38MAPK activation and reduced PARP-1 expression, thus providing evidence that the DNA damage checkpoint response was negatively regulated by rimonabant. Interestingly, rimonabant did not cause the cleavage of PARP-1 protein and this corroborates the lack of caspase-dependent apoptosis. In summary, in vitro data suggest that rimonabant induces mitotic catastrophe in colon cancer cells. In DLD-1 cells this occurs via a dual mechanism: the downregulation of DNA damage response checkpoint and the abrogation of the spindle assembly checkpoint, both mechanisms leading to cell growth arrest and finally to cell death.

Rimonabant is a CB1 receptor antagonist/inverse agonist. Our previous finding showed that the antiproliferative effects of micromolar concentration of rimonabant can be potentiated rather than reverted by CB1 agonists.9 In addition, it has been reported that agonism as opposed to antagonism of CB1 receptors can reduce the proliferation of colorectal cancer cells, and that nanomolar concentrations of rimonabant are sufficient to antagonize endocannabinoid-mediated inhibition of cell proliferation in human colorectal carcinoma cell lines.35, 36 Because the rimonabant concentrations used in our study are at least 10-fold higher than those required to antagonize the CB1 receptors and we evidenced that rimonabant, at concentrations of 2.5 and 5.0 μM, failed to inhibit cell proliferation or to increase mitotic index in CHO cells that lacks this receptor (data not shown), it is likely that CB1 receptor inverse agonism rather than antagonism, could be involved in our observed effects.

We studied the antiproliferative effects of rimonabant on human colon cancer cells mutated for p53, a late event in colorectal cancer pathogenesis.27 Therefore, we aimed to investigate whether rimonabant could exert also a chemopreventive action on preneoplastic lesions by using the in vivo model of AOM-induced colon carcinogenesis. ACF are the earliest identifiable neoplastic lesions, they are believed the precursors of colon cancer in humans and ACF containing 4 or more aberrant crypts correlate with colon tumor outcome.37 Our results indicated that rimonabant is able to protect colon cancer cells from chemically-induced carcinogenesis, even though the dose of rimonabant used in this article (i.e. 3 mg/kg) is higher than that used in clinical trials. However, this dose has been used in other experimental studies demonstrating that rimonabant could prevent acquisition of drinking behavior in alcohol-preferring rats17 or to exacerbate intestinal inflammation in mice.18

Our previous results demonstrated that increased endocannabinoid levels are able to reduce the development of precancerous lesions in the mouse colon19 and the data presented here indicate that rimonabant decreases the number of mice with ACF containing 4 or more crypts. This implies that CB1 agonists and rimonabant have similar effects. A possible explanation for such a paradoxical in vivo effect could be that after CB1 receptor blockade by rimonabant, the released endocannabinoids could reduce ACF formation. Furthermore, the finding that genetic blockade of this receptor in CB1 null mice does not alter the susceptibility to develop ACF19 suggests that rimonabant acts as an inverse agonist rather than an antagonist in reducing ACF formation.

Recent data have pointed out that the CB1 antagonist/inverse agonist AM251 is able to inhibit NCM460 cell proliferation.38 On the basis of this observation, obtained in cells exhibiting a nontransformed phenotype, the authors proposed that rimonabant might affect negatively the integrity and maintenance of the intestinal mucosa as a consequence of its continuous intake during the pharmacotherapy of obesity. They also suggest that rimonabant-induced mucosal damage might also trigger persistent regenerative processes leading to colon cancer. Indeed, these results have been obtained by using another drug and in our study, we clearly demonstrate that rimonabant inhibits human colon cancer cell growth and reduces the formation of precancerous lesions in the mouse colon.

In conclusion, this is the first report providing evidence that rimonabant exhibits antitumor actions in colon cancer cells through induction of mitotic catastrophe and prevents colon carcinogenesis in vivo. The efficacy of rimonanabant to inhibit cell growth in vitro involves the suppression of key cell signaling pathways, the spindle assembly checkpoint and the DNA damage checkpoint, both regulating the M-phase transition. Nonapoptotic forms of cell death are crucial for effective anticancer therapy and, at present, few conventional anticancer treatments aim to produce cell death by mitotic catastrophe. In this view, rimonabant might represent an useful tool for colon cancer management, notably in obese patients, who have an increased risk of developing colorectal cancer.39, 40

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

This work was supported by Sanofi-aventis (M.B.) and the Associazione Educazione e Ricerca Medica Salernitana (ERMES). A.S was supported by a fellowship from Associazione Italiana Ricerca sul Cancro (AIRC), and S.P. was supported by a fellowship from Fondazione Italiana per la Ricerca sul Cancro (FIRC).

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_24483_sm_suppfig1.pdf196KSupplementary Information.
IJC_24483_sm_suppfig2.pdf64KSupplementary Information.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.