TOXICOLOGICAL SCIENCES 83, 126–135 (2005) doi:10.1093/toxsci/kfi008 Advance Access publication October 13, 2004

Characterization of the N-methoxyindole-3-carbinol (NI3C)–Induced Cell Cycle Arrest in Human Colon Cancer Cell Lines Antje S. Neave,1 Sussi M. Sarup,1 Michel Seidelin,2 Fritz Duus, and Ole Vang3 Department of Life Sciences and Chemistry, Roskilde University, Roskilde, Denmark Received August 2, 2004; accepted September 28, 2004

Recent results have shown that indole-3-carbinol (I3C) inhibits the cellular growth of human cancer cell lines. In some cruciferous vegetables, another indole, N-methoxyindole-3-carbinol (NI3C), is found beside I3C. Knowledge about the biological effects of NI3C is limited. The aim of the present study was to show the effect of NI3C on cell growth of two human colon cancer cell lines, DLD-1 and HCT-116. For the first time it is shown that NI3C inhibits cellular growth of DLD-1 and HCT-116 and that NI3C is a more potent inhibitor of cell proliferation than I3C. In addition to the inhibition of cellular proliferation, NI3C caused an accumulation of HCT-116 cells in the G2/M phase, in contrast to I3C, which led to an accumulation of the colon cells in G0/G1 phase. Furthermore, NI3C delays the G1-S phase transition of synchronized HCT-116 cells. The indole-mediated cell-cycle arrest may be related to the increased levels of the CDK-inhibitors p21 and p27 (only induced by NI3C). Only an initial increase of cdc2 protein was observed, whereas prolonged treatment with NI3C or I3C downregulates the mRNA and proteins of cyclin-dependent kinases and cyclins. These results indicate that both NI3C and I3C inhibit the proliferation of human colon cells but via different mechanisms. Key Words: N-methoxyindole-3-carbinol; indole-3-carbinol; colon cells; cell cycle; cell proliferation; carcinogenesis.

INTRODUCTION

The preventive effect of a high intake of cruciferous vegetables against several cancer types has been shown in many epidemiological studies (Verhoeven et al., 1996), as well as in animal models showing reduction in colon cancer incidence (Barrett et al., 1998). The combined effect of several compounds found in cruciferous vegetables acting on different steps of carcinogenesis rather than a single compound, may explain the cancer-preventive effect of cruciferous vegetables. During digestion of a group of compounds known as glucosinolates, various indoles as well as isothiocyanates are formed, and both 1

Both authors have contributed equally to this article. Present address: Taconic M&B A/S, PO Box 1079, DK-8680 Ry, Denmark. 3 To whom correspondence should be addressed at Department of Life Sciences and Chemistry, Roskilde University, Universitetsvej 1, P.O. Box 260, DK-4000 Roskilde, Denmark. Fax: 145 4674 3011. E-mail: [email protected]. 2

Toxicological Sciences vol. 83 no. 1

#

types of compounds are proposed to have anticarcinogenic effects. In cruciferous vegetables, such as broccoli, two major indolylglucosinolates, glucobrassicin and neoglucobrassicin, are found in substantial amounts, together with 4-hydroxyglucobrassicin and 4-methoxyglucobrassicin, although the latter two compounds are present in lesser amounts (Vang et al., 2001). During degradation of glucobrassicin and neoglucobrassicin, indole-3-carbinol (I3C) and N-methoxyindole3-carbinol (NI3C) are formed, respectively. Indole-3-carbinol has been shown to have anticarcinogenic activity in various animal models (Stoner et al., 2002; Xu et al., 2001). The anticarcinogenic response of I3C in vivo and in vitro has been reviewed elsewhere (IARC, 2004; Vang and Dragsted, 1996). At the same time, several animal studies have shown a tumor-promoting activity, when the animals are exposed to I3C after exposure to carcinogens (Xu et al., 2001). This tumorpromoting effect of I3C has recently been supported by the observation that indolo[3,2-b]carbazole (ICZ), which is formed from I3C, inhibits gap junction intercellular communication (Herrmann et al., 2002). Besides ICZ, various dimeric (e.g., DIM), trimeric, and tetrameric indole compounds are formed from I3C in the acidic environment of the stomach, all with various different biological activities (Grose and Bjeldanes, 1992). Their biological effects are only sparsely evaluated. Until recently, the anticarcinogenic activity of I3C was suggested to be related to the ability to modulate xenobiotic metabolism. Indole-3-carbinol induces various cytochrome P-450 (CYP) isoforms in different organs (Manson et al., 1998), probably via the Ah receptor (Chen et al., 1996), and induces phase II enzymes (van Lieshout et al., 1998). The exposure to the carcinogens is suggested to be reduced by a shift in the metabolism of procarcinogens (Stresser et al., 1994). In 1998, new anticarcinogenic properties of I3C were identified by Cover et al., who showed that I3C inhibits DNA synthesis and cell division of human breast cancer cells, MCF-7 (Cover et al., 1998), by a G1 cell cycle arrest, via a specific downregulation of cyclindependent kinase 6 (CDK6) (Cover et al., 1999). A simultaneous increase of the CDK inhibitor p21WAF1/CIP1 was observed together with a dephosphorylation of retinoblastoma protein (Cover et al., 1999). An identical growth inhibition has been

Society of Toxicology 2005; all rights reserved.

NI3C-INDUCED CELL CYCLE ARREST

observed in another breast cancer cell line (T47D) (Ge et al., 1999), in prostate cells (Chinni et al., 2001), and in colon cells (Hudson et al., 2003; Zheng et al., 2002). Both I3C and NI3C are formed in dominating levels in broccoli, for example, during ingestion, but in contrast to I3C, very little is known about the biological effects of NI3C. As found for I3C, NI3C induces CYP1A1 in vitro (Stephensen et al., 2000) and in vivo in rats (Bradfield and Bjeldanes, 1987), probably via activation of the Ah-receptor (Stephensen et al., 2000). A similar induction of hepatic and colonic CYP1A proteins and activities was found when rats were exposed to glucobrassicin, neoglucobrassicin, or a mixture of these two glucosinolates and 4-methoxyglucobrassicin (Bonnesen et al., 1999). While the major part of the literature considers the indoles as anti-carcinogenic agents, this study investigates their effects as cancer chemotherapeutic agents against human colon cancer cell lines. N-methoxy-indole-3-carbinol is not commercially available; therefore, to investigate the cellular effects of NI3C, it has to be produced synthetically. Using the synthesized NI3C, the experiments show that both NI3C and I3C inhibit the proliferation of human colon cancer cells, but NI3C is a more potent inhibitor. NI3C causes an accumulation of the cells in the G2/M phase, whereas I3C accumulates the cells in the G0/G1 phase of the cell cycle. MATERIALS AND METHODS Materials. McCoys 5A medium containing glutamax I (a stabilized form of L-glutamine), calcium-free and magnesium-free phosphate-buffered saline (PBS), trypsin, Versene, gentamicin, TBE-buffer, ethidium bromide (EtBr) and Trizol Reagent were obtained from Gibco BRL (InVitrogen, Taastrup, Denmark). Fetal bovine serum (FBS) was obtained from Seromed (Biochrom KB, Berlin, Germany), and RNAse was obtained from Amersham BioScience (Hørsholm, Denmark). NP40 was from Roche Diagnostic Scandinavia (Copenhagen, Denmark). Propidium iodide, dimethylsulfoxide (DMSO), sodium deoxycholate, PMSF, aprotinin, sodium orthovanadate, and indole3-carbinol (I3C) were purchased from Sigma-Aldrich (Copenhagen, Denmark). Indole-3-carbinol was recrystallized in hot toluene, prior to use. N-methoxyindole-3-carbinol was synthesized and purified in this laboratory, 1H- and 13 C-NMR data were consistent with previously reported values (Hanley et al., 1990). High vacuum destillation afforded a 100% pure NI3C according to high performance liquid chromatography (HPLC) analysis. Cell culture. The human colon cancer cell lines HCT-116 and DLD-1 were obtained from American Type Culture Collection (Manassas, VA) and cultured in McCoys 5A medium supplemented with 10% FBS, 0.1% gentamicin in a 5% CO2 atmosphere at 37
C. The cells were grown as monolayers, subcultured twice a week. DMSO was used as solvent in all experiments, and all the cells were exposed to the same concentration in each experiment (50.1%, which is a noncytotoxic concentration). Cell proliferation. The DLD-1 and HCT-116 cells were seeded at a density of 2.6 3 104 cells/cm2 in Nunc 24-well culture dishes and allowed to adhere overnight. The cells were then treated with NI3C (0–100 mM), I3C (0–450 mM) or DMSO (vehicle control) and incubated for 48 h. The cells were washed with PBS (1 ml/well), trypsinized (0.2 ml/well), and counted using a Coulter counter (Coulter Z2, Beckman Coulter). Particles with diameters greater than 10.5 mm were included.

127

To determine whether NI3C-induced growth inhibition is also timedependent and reversible, DLD-1 and HCT-116 cells were seeded at a density of 1.8 3 104 cells/cm2 in 24-well culture dishes and incubated overnight. The cells were treated with 0, 30, and 60 mM NI3C (4 wells/treatment) and incubated. At the denoted time points for each condition, the cells were washed, trypsinized, and counted. After 48 h of NI3C treatment, the NI3C-containing medium, in a subset of the cell cultures, was replaced with fresh growth medium containing NI3C or DMSO, and the time course was continued for an additional 48 h. 5-Bromo-20 -deoxyuridine (BrdU) incorporation. Using the Cell Proliferation ELISA Biotrak system (Amersham BioScience), the effects of NI3C and I3C on the DNA synthesis in HCT-116 cells were examined. HCT-116 cells were seeded at a density of 3000 cells/well in 96-well culture dishes and allowed to adhere overnight. The cells were then treated with NI3C (0–120 mM) and I3C (150 and 300 mM). The dishes were incubated for 24 and 48 h, the last 4 h with BrdU (10 mM). Subsequently, the assay was performed as described by the manufacturer. To adjust for the variation in the cell numbers after the treatment, HCT-116 cells were simultaneously seeded in 24-well culture dishes (1.8 3 104 cells/cm2), treated with the indoles as described above, and counted. Flow cytometric analyses of DNA content. To investigate the effects of NI3C and I3C on the cell cycle phase distribution, the cells were analyzed by flow cytometry. HCT-116 cells were plated onto NUNC dishes (5.7 3 104 cells/cm2) and incubated overnight before addition of the indoles. The cells were then treated with 0–30 mM NI3C or 0–250 mM I3C (three dishes/treatment) and incubated for 48 h. The treated cells were removed from culture dishes by trypsinization, collected by centrifugation, and washed with PBS. 5 3 105 cells from each sample were fixed in ice-cold 70% ethanol and incubated on ice for at least 30 min. Cells were then washed in PBS, resuspended in 400 ml PBS, and 50 ml RNAse (1 mg/ml), and 50 ml propidium iodide (0.4 mg/ml) were added. After incubation (1 h at RT) the stained nuclei were analyzed with a flow cytometer (FACSCalibur, BD Biosciences, Denmark). Cell cycle distribution was based on 2N and 4N DNA content. Cells with less than 2N DNA content were indicative of apoptotic cells. The percentages of cells within the G0/G1, S, and G2/M phases of the cell cycle were analyzed in ModFit LT (Verity Software House, Inc.). Furthermore, the effects of the indoles on the cell cycle phase distribution in synchronized HCT-116 cells were analyzed. When HCT-116 cells are grown in a low serum concentration for 72 h, the cells accumulate in the G0/G1 phase. HCT-116 cells were plated onto Nunc dishes (2.85 3 104 cells/cm2) and incubated overnight. The growth medium was then removed, and after cells were washed with PBS, a medium containing 0.2% FBS was added, and the cells were further incubated for 72 h. This starvation of the cells resulted in 60–70% of the cells accumulating in the G0/G1 phase of the cell cycle. The now synchronized cells were then washed with PBS, and medium containing 10% FBS and the final concentrations of the indoles 0 and 30 mM NI3C and 250 mM I3C (three dishes/ treatment) were added. The cells were incubated at 37
C for 0, 4, 8, 12, 16, 24, and 48 h, and the DNA content was analyzed as described above. Northern blot analysis. Genes related to cell cycle control were analyzed in asynchronous HCT-116 cells treated with 0–30 mM NI3C and 0–250 mM I3C for 48 h, and in synchronized HCT-116 cells treated with 0, 30 mM NI3C and 250 mM I3C for 0, 4, 8, 12, 16, and 24 h (three dishes/treatment) by Northern blotting. RNA was extracted from the cells using 1 ml Trizol Reagent per dish. Isolation of total RNA was performed as described by the manufacturer (Gibco BRL). The northern blotting was performed as described previously (Herrmann et al., 2002). The cDNA templates used for probe synthesis were constructed using reverse transcribed (RT) DLD-1 cell total RNA and subsequent amplification by polymerase chain reaction (PCR) (Advantage RT-for-PCR kit, Clontech Laboratories Inc., BD Biosciences, Denmark) (Herrmann et al., 2002). Specific primers were designed from the known cDNA sequences published in GeneBank of the National Center of Biotechnology: cyclin A2 (acc. no. NM_001237, nt: 344–775), cyclin D1 (acc. no. NM_001758, nt: 901–1200), cyclin E1 (acc. no. NM_001238, nt: 414–1174), CDC2 (acc. no. NM_001786, nt: 255–692), CDK2 (acc. no. NM_001798, nt: 77–399), CDK4 (acc. no. NM_000075, nt: 1047–1366), CDK6 (acc. no. NM_001259, nt: 792-1196), and CDKN1B (p27),

128

NEAVE ET AL.

(acc. no. NM_004064, nt: 180–495). CDKN1A (p21) (nt: 194–585), and cyclin B1 (nt: 1244–1476) cDNA were a kind gift from A. Lu¨tzen, Roskilde University. Protein extraction and Western blot analysis. Asynchronous HCT-116 cells were plated onto Nunc dishes, at a density of 1 3 105/ml and allowed to adhere for 24 h. The cells were then treated with 75.6 mM NI3C, 330.0 mM I3C or DMSO (vehicle control) and incubated for 6, 24, and 48 h. After incubation and addition of RIPA buffer [PBS, 1% NP40, 0,5% sodium deoxycholate, 0,1% SDS, PSMF (10 ml/ml), aprotinin (10 mg/ml), sodium orthovanadate (10 ml/ml)], the cells were harvested by scraping the cells from the culture dishes and collected by centrifugation. Protein concentration was then measured using DC Protein Assay reagents (Bio-Rad, Copenhagen Denmark) as described by the manufacturer. Cell extracts containing equal amounts of protein were mixed with loading buffer and fractionated eletrophoretically on a 10% Tris-HCl Ready gel (Bio-Rad). Kaleidoscope marker (Bio-Rad) was used as the molecular weight standard. Proteins were electrically transferred to a PVDF (polyvinylidene difluoride) membrane (Amersham LIFE SCIENCE) and blocked overnight, at 4
C, with blocking buffer containing 5% non-fat dry milk. The membranes were subsequently incubated with monoclonal antibodies against p21 (1:500), p27 (1:2500), Cyclin B (1:1000), Cdk2 (1:2500), or Cdc2 (1:2500) (BD Biosciences, Denmark), for 2 h at room temperature, washed with TBSa containing 0.05% Tween 20, and incubated with, a secondary antibody, horseradish peroxidase-conjugated antimouse IgG (DAKO, Glostrup, Denmark) for 1 h at room temperature (1:2000). The signal was then detected using the chemiluminescent detection system Supersignal West Femto, Maximum sensitivity substrate (Pierce Biotechnology, Bie & Berntsen, Denmark). Statistics. The results are expressed as means, and differences were tested statistically by analysis of variance (ANOVA) in Systat version 6.0.1 (SPSS Inc. Chicago, IL). Dunnets or Fischers least significant difference (LSD) test were used to determine differences at p 5 0.05.

RESULTS

Synthesis of NI3C N-methoxyindole-3-carbinol was synthesized in three steps: First, N-methoxyindole was synthesized, and then the compound was converted, via Vilsmeier-Haak reaction, to N-methoxyindole-3-carbaldehyde, which was finally reduced to N-methoxyindole-3-carbinol. The synthesis of NI3C was carried out several times, and approximately the same yields were obtained. In total, we synthesized 4.2 g NI3C with an overall yield of 25%. 1 H-NMR and 13C-NMR data were consistent with previously reported values (Hanley et al., 1990). High performance liquid chromatography analysis of NI3C showed a purity of 98%, but subsequent high vacuum distillation afforded 100% pure NI3C, which was used in the following experiments. The Effects of NI3C and I3C on the Growth of HCT-116 and DLD-1 Cells The initial test of the growth regulation by I3C and NI3C was performed in two human colon epithelial cell lines, DLD-1 and HCT-116, and the cells were treated with 0–100 mM NI3C or 0–450 mM I3C for 48 h. Both NI3C and I3C caused a dosedependent growth inhibition in both cell lines (Fig. 1). In HCT-116 cells, the mean values (6STD) of the 50% inhibitory

FIG. 1. Dose-dependent inhibition of cell proliferation. Human colon cells, DLD-1 (-*-) and HCT-116 (-&-), were exposed to increasing doses of NI3C (A) or I3C (B). The numbers of cells after 48 h of exposure were determined by counting and are shown as means 6 S.E.M. of four determinations. The numbers are relative to the level in the control groups. The experiment was performed three times with similar results.

concentration (IC50) for NI3C and I3C were 35.3 mM (64.8) and 250 mM (667), respectively. In the DLD-1 cells, the mean IC50 values for NI3C and I3C were 25.9 mM (63.8) and 270 mM (638), respectively. These sets of data show that NI3C is a sevenfold to tenfold more potent inhibitor than I3C. Furthermore, the inhibitory effect of NI3C on cell proliferation was time-dependent and reversible (Fig. 2). Time course studies of NI3C addition and withdrawal demonstrated that withdrawal of both 30 mM and 60 mM NI3C after 48 h of exposure caused a complete restoration of the growth of HCT-116 cells, whereas the growth-inhibitory effect of DLD-1 cells was not reversible for the highest concentration of NI3C. A dose-dependent inhibition of DNA synthesis was observed in HCT-116 cells

NI3C-INDUCED CELL CYCLE ARREST

129

FIG. 2. Time-dependent inhibition of cell number. Human colon cells, DLD-1 (A) or HCT-116 (B), were exposed to 0, 30, or 60 mM NI3C (*, &, ~). The number of cells per well were determined (means 6 S.E.M. of four wells) at 24, 48, 72, and 96 h. At 48 h, the medium was replaced with fresh growth medium without NI3C (open symbols, dotted lines) or with the same concentration of NI3C (filled symbols, solid lines) for an additional 24–48 h.

for both NI3C and I3C after 24 and 48 h of treatment. On the other hand, the relative DNA synthesis per cell was only slightly affected by the indoles (data not shown). The Effects of NI3C and I3C on the Cell Cycle Phase Distribution of HCT-116 Cells The effects of the indoles on the cell cycle were tested in both asynchronous and synchronous HCT-116 cells. When asynchronously growing HCT-116 cells were treated with NI3C or I3C for 48 h, 30 mM NI3C significantly increased the number of cells in the G2/M phase (27% of the treated cells compared to 22% of the control cells). In comparison, 250 mM I3C significantly accumulated the cells in the G0/G1 phase (55% of the treated compared to 32% of the control cells; Fig. 3).

FIG. 3. The effects of NI3C and I3C on the cell cycle progression in asynchronous HCT-116 cells. Cell fractions in the G0/G1 phase (A), in the S phase (B), and the G2/M phase (C) are given. HCT-116 cells were exposed to 0, 15, or 30 mM NI3C or 150 or 250 mM I3C for 48 h. The numbers are means 6 S.E.M. of three dishes in one experiment. Two additional experiments were done with similar results. Statistical differences were analyzed by ANOVA and significantly different values according to Dunnett’s test ( p 5 0.05) are indicated by different letters.

To get a more accurate picture of the effects of the indoles on the cell cycle phase distribution, the HCT-116 cells were synchronized. In synchronized cells it is possible to follow the indole-induced retardation of the cells in the various cell

130

NEAVE ET AL.

cycle phases. Furthermore, experiments with synchronized cells show the effects on the dynamics of the cell cycle. HCT-116 cells were synchronized by cultivating the cells with 0.2% serum for 72 h, and they were released from the G0/G1 phase by replacement with medium containing 10% serum. At the same time the treatment with 0, 30 mM NI3C or 250 mM I3C was initiated. The phase distribution was analyzed for up to 48 h, by flow cytometry. Figure 4 shows that 30 mM NI3C delays the cells transition from the G0/G1 phase to the S phase compared to cells treated with the vehicle control or 250 mM I3C. In accordance with the experiments with the asynchronous cells, 250 mM I3C results in an accumulation of the cells in the G0/G1 phase after 48 h of exposure (57% compared to 36% of the control cells), whereas 30 mM NI3C cause an accumulation of the cells in the G2/M phase (37% compared to 23% of the control cells).

The Effects of NI3C and I3C on the Expression of Cell Cycle Regulatory Genes To identify the effect of the indoles on the gene expression in relation to the cell growth inhibition and the specific cell cycle stop in the G2/M and G0/G1 phases for NI3C and I3C, respectively, the expression of a number of genes related to cell cycle was analyzed by Northern blotting. In synchronous HCT116 cells, exposure to 30 mM NI3C caused a significant increase of p21 mRNA content at 8 h and a significant downregulation of the CDK4 mRNAs beyond 16 h of treatment (Fig. 5). Exposure to I3C increases p21 mRNA levels beyond 24 h and lowers the level of CDK4 mRNAs, but neither of these effects was significant. Furthermore, I3C downregulates CDC2 expression at 24 h, an effect also observed after 48 h of treatment of asynchronous HCT-116 cells (Fig. 6). A general down-regulation of many of the cell-cycle-related genes was observed after 48 h of treatment of asynchronous HCT-116 cells with either NI3C or I3C. Only cyclin E1 mRNA levels were significantly upregulated by NI3C in asynchronous HCT-116 cells (Fig. 6).

The Effects of NI3C and I3C on Levels of Cell Cycle Regulatory Proteins The effects of the indoles on proteins involved in the cell cycle were then analyzed. The protein levels of the two CDK inhibitors, p21 and p27, were increased with time in control cells. The level of p21 protein was increased further (twofold) by both indoles compared with the control cells, whereas the p27 protein level was increased at all time points by NI3C only (Fig. 7). After 6 h of exposure, cdc2 was clearly enhanced by both NI3C and I3C, but I3C reduced the cdc2 level at 24 and 48 h. CDK2 protein is increased during the exponential growth in control cells, but I3C, and to a lesser degree NI3C, reduces the protein level of CDK2. Last, I3C clearly reduces the cyclin B level at both 24 and 48 h.

FIG. 4. The effects of NI3C and I3C on the cell cycle progression in synchronous HCT-116 cells. HCT-116 cells were synchronized, and at the time-release of the growth brake (by addition of serum) DMSO (-*-), 30 mM NI3C (-!-), or 250 mM I3C (-&-) was added. After the incubation the cell cycle phase distribution was analyzed 0, 4, 8, 12, 16, 24, and 48 h after initiation of the cell cycle. Cell fractions in the G0/G1 phase (A), in the S phase (B), and the G2/M phase (C) are given. The numbers are means 6 S.E.M. of three dishes from one experiment. The experiment was been performed twice more with similar results. Statistical differences were analyzed by ANOVA and significantly different values according to Fischers least significant difference (LSD) test ( p 5 0.05) are indicated with different symbols as follows: *NI3C different from the control; ¤ NI3C different from I3C; # I3C different from the control.

131

NI3C-INDUCED CELL CYCLE ARREST

FIG. 5. mRNA levels of cell cycle-related genes in synchronous HCT-116 cells. HCT-116 cells were synchronized, and at the time-release of the growth brake (by addition of 10% serum) DMSO (-*-), 30 mM NI3C (-!-), or 250 mM I3C (-&-) was added. The RNA was isolated after 0, 4, 8, 12, 16, and 24 h and analyzed by northern blotting. A typical northern blot is shown (A): lane 1: 4 h, DMSO; lane 2: 4 h, 30 mM NI3C; lane 3: 4 h, 250 mM I3C; lane 4: 8 h, DMSO; lane 5: 8 h, 30 mM NI3C; lane 6: 8 h, 250 mM I3C; lane 7: 16 h, DMSO; lane 8: 16 h, 30 mM NI3C; lane 9: 16 h, 250 mM I3C; lane 10: 24 h, DMSO; lane 11: 0 h; lane 12: 24 h, 250 mM I3C. mRNA levels of p21 (B), cyclin E1 (C), CDC2 (D), and CDK4 (E) are expressed relative to 18 S rRNA. The numbers are means 6 S.E.M. of three dishes from one experiment. Statistical differences were analyzed by ANOVA and significantly different values according Fischer’s LSD test (p  0.05) are indicated with different symbols as follows: *NI3C different from the control; ¤ NI3C different from I3C.

Apoptotic Effects of NI3C and I3C From the FACS data, the level of sub G1/G0 cells indicates apoptosis. Table 1 indicates that 30 mM NI3C induce apoptosis, whereas 250 mM I3C does not induce apoptosis in HCT-116 cells.

DISCUSSION

In contrast to the numerous articles showing cellular effects of indole-3-carbinol, very few have investigated the cellular effects of N-methoxyindole-3-carbinol. Previous reports

132

NEAVE ET AL.

FIG. 6. mRNA levels of cell cycle–related genes in asynchronous HCT-116 cells. Asynchronous HCT-116 cells were exposed to 0, 15, or 30 mM NI3C (A) or 0, 150, or 250 mM I3C (B) for 48 h. mRNA levels of cyclin A2, cyclin B1, cyclin D1, cyclin E1, CDC2, CDK2, CDK4, CDK6, p21, p27 are expressed relative to 18 S rRNA. The numbers are means 6 S.E.M. for three dishes relative to the controls. Statistically different levels were identified by using Dunnett’s test (p  0.05) and are marked with an asterisk ().

showed that NI3C induces CYP1-related activities in rat liver and intestine (Bradfield and Bjeldanes, 1987) and in mouse hepatoma cells in vitro (Stephensen et al., 2000). This is the first report, showing that NI3C also inhibits cellular growth, and that NI3C is a more potent inhibitor than I3C of human colon cancer cell proliferation. The mean IC50 for I3C was 250–270 mM in the human colon cancer cell line analyzed here (HCT-116 and DLD-1 cells). These levels are somewhat higher than observed for other cell types. In two tumor-derived human colon cell lines,

SW480 and HT-29, and the ‘‘normal’’ derived cell line HCEC, the IC50 ranges between 120 and 165 mM (Hudson et al., 2003). The reduced IC50 may be explained in part by the prolonged exposure time used (168 h compared to 48 h in the present experiment). In Colo 320 growth inhibition is observed at 400 mM I3C after 48 h (Mori et al., 2001). The observed IC50 was 600 mM for I3C in two human colon cell lines (Caco-2 and LS-174) following a 24 h exposure (Bonnesen et al., 2001). The present experiments do not reveal whether the biological responses are caused by I3C or

NI3C-INDUCED CELL CYCLE ARREST

133

FIG. 7. The effects of I3C and NI3C on expression of cell cycle proteins in HCT-116 cells. (A) HCT-116 cells were treated with 75.6 mM NI3C, 330.0 mM I3C or DMSO for the indicated times. The protein levels were determined by Western analysis using specific antibodies against p21, p27, cdc2, CDK2, and cyclin B. Densitometry estimation of p21 (B), p27 (C), cdc2 (D), CDK2 (E), and cyclin B (F).

by the products of I3C formed in the medium during the incubation. The effect of I3C on DNA synthesis has been analyzed in parallel with the cell number. In our study, neither NI3C nor

I3C seems to inhibit the relative DNA synthesis per cell in the colon cancer cells. This is in contrast to the inhibition of DNA proliferation observed in colo-320 cells, where the DNAproliferation was reduced to 45% of the control values after

134

NEAVE ET AL.

48 h of exposure to 40 mM I3C (Zheng et al., 2002). In the present experiment the growth arrest by NI3C was interestingly accompanied by an accumulation of HCT-116 cells in the G2/M phase. This is in contrast to the observed accumulation in the G0/G1 phase when HCT-116 cells were exposed to I3C. In the present experiment, NI3C also caused a transient retardation of the G1-S phase transition in synchronized HCT-116 cells, indicating that molecular events related to G1-S phase transition were slightly affected, as shown by the upregulation of p21 mRNA (8 h) and p21 protein. The different cellular effects of I3C and NI3C that modulates different phases of the cell cycle indicate different molecular points of action for the two indoles. Of the analyzed proteins, only p21 responds equally to the treatment of both indoles. The p27 protein level was upregulated by NI3C alone at all time points, and cdc2 was downregulated by I3C. Exposure to both NI3C and I3C caused a small but general downregulation of most of the genes analyzed in the HCT-116 cells after 48 h. We therefore assume that the general downregulation of the transcription of cell cycle–related genes is a consequence of the cell proliferation stop, without relation to a specific point of action in the cell cycle. The NI3C-mediated delayed transition from G1 to S phase, and the I3C-induced accumulation in the G0/G1-phase, are possibly linked to the increased expression of p21 mRNA TABLE 1 Apoptotic Effects of NI3C and I3C Control

250 mM I3C

30 mM NI3C

3.1 6 0.3%

8.0 6 1.0%

34.6 6 2.2%

The relative number of cells in the sub-G0/G1-fraction observed in the FACS analysis, when the HCT-116 cells were exposed to 30 mM NI3C or 250 mM I3C for 48 h. The numbers are means of three dishes 6 S.T.D. The experiment was done three times with similar results.

and protein and to the down-regulation of CDK4 mRNA, as both proteins are known to be involved in the regulation of G1-S transition (Fig. 8). Similarly the NI3C induced upregulation of p27 protein levels may also play a part in this delayed G1-S transition as both CDK inhibitors regulate the G1-S phase transition by inhibiting the CDK2/cyclin E complex and p21 inhibits the activity of CDK4. Inhibition of the CDK2/cylin E complex and downregulation of CDK4 may in turn reduce the activation of retinoblastoma protein, which regulates the transcription of S-phase genes including cdc2. The CDK inhibitor p21 is also involved in the G2-M phase transition by inhibiting the cdc2/cyclin B complex. During the normal cell cycle, the cyclin B starts to be synthesized from the end of the S phase, and cyclin B mRNA is thought to be more stable in the G2 phase than in the G1 phase (Smits and Medema, 2001). It is interesting that cyclin B1 mRNA (and to a degree the cyclin B protein) is observed to be downregulated in both NI3C-treated and I3C-treated cells. One explanation could be that the downregulation of cyclin B1 in combination with enhanced p21 protein level in NI3C-treated cells is the cause of the G2/M arrest, whereas the I3C-induced G0/G1 arrest would cause a lower level of cyclin B1 mRNA. It is not possible to identify the specific mechanism for NI3C-induced G2/M arrest. The G2/M phase accumulation is often related to induction of apoptosis in p53 functioning cell lines (Russo et al., 2002). Indeed, NI3C increases the number of apoptotic cells dramatically in contrast to I3C, which in our experiments led to only a small increase in apoptotic HCT-116 cells. Previously, apoptosis has been observed in human colon cells (Colo 320) exposed to 400 mM I3C (Mori et al., 2001; Zheng et al., 2002). The discrepancy from our I3C data is not clear, but the Colo 320 cell line may be more sensitive to apoptosis than the HCT-116 cell line used in our experiments, although both cell lines have functional p53 (Russo et al., 2002).

FIG. 8. A molecular model for the NI3C-mediated and I3C-mediated cell-cycle regulation in HCT-116 cells.

NI3C-INDUCED CELL CYCLE ARREST

The anticarcinogenic effect of cruciferous vegetables, such as broccoli, is likely caused by the combined effect of various compounds found in cruciferous vegetables including I3C and NI3C and their parental substances glucobrassicin and neoglucobrassicin. The indication that both I3C and NI3C inhibit the growth of human colon cancer cells by different mechanisms is very promising, as the indoles may then have a synergistic effect on colon cancer cell growth when present together in the same food item. Furthermore, based on these in vitro data alone one may assume that NI3C may be a more potent anticarcinogenic compound than I3C. This needs to be proved in an animal study.

ACKNOWLEDGMENTS We thank Ms. Mia Skaaning and Anne Lise Maarup for technical assistance in performing the DNA proliferation analysis. The work was supported in part by the Danish Cancer Society with a grant to A.S.N.

REFERENCES Barrett, J. E., Klopfenstein, C. F., and Leipold, H. W. (1998). Protective effects of cruciferous seed meals and hulls against colon cancer in mice. Cancer Lett. 127, 83–88.

135

in the enzymatic hydrolysis of indole glucosinolates. J. Chem. Soc. Perkin Trans. 1, 2273–2276. Herrmann, S., Seidelin, M., Bisgaard, H. C., and Vang, O. (2002). Indolo[3,2-b]carbazole inhibits gap junctional intercellular communication in rat primary hepatocytes and acts as a potential tumor promoter. Carcinogenesis 23, 1861–1868. Hudson, E. A., Howells, L. M., Gallacher-Hofrley, B., Fox, L. H., Gescher, A., and Manson, M. M. (2003). Growth-inhibitory effects of the chemopreventive agent indole-3-carbinol are increased in combination with the polyamine putrescine in the SW480 colon tumour cell line. BMC Cancer 3, 2. IARC (International Agency for Research in Cancer). (2004). IARC Handbooks of Cancer Prevention. Cruciferous Vegetables, Isothiocyanates and Indoles, IARC Press, Lyon. Manson, M. M., Hudson, E. A., Ball, H. W., Barrett, M. C., Clark, H. L., Judah, D. J., Verschoyle, R. D., and Neal, G. E. (1998). Chemoprevention of aflatoxin B1-induced carcinogenesis by indole-3-carbinol in rat liver— Predicting the outcome using early biomarkers. Carcinogenesis 19, 1829– 1836. Mori, H., Niwa, K., Zheng, Q., Yamada, Y., Sakata, K., and Yoshimi, N. (2001). Cell proliferation in cancer prevention; effects of preventive agents on estrogen-related endometrial carcinogenesis model and on an in vitro model in human colorectal cells. Mutat. Res. 480, 201–207. Russo, P., Malacarne, D., Falugi, C., Trombino, S., and O’Connor, P. M. (2002). RPR-115135, a farnesyltransferase inhibitor, increases 5-FU-cytotoxicity in ten human colon cancer cell lines: Role of p53. Int. J. Cancer 100, 266–275. Smits, V. A., and Medema, R. H. (2001). Checking out the G(2)/M transition. Biochim. Biophys. Acta 1519, 1–12.

Bonnesen, C., Eggleston, I. M., and Hayes, J. D. (2001). Dietary indoles and isothiocyanates that are generated from cruciferous vegetables can both stimulate apoptosis and confer protection against DNA damage in human colon cell lines. Cancer Res. 61, 6120–6130.

Stephensen, P. U., Bonnesen, C., Schaldach, C., Andersen, O., Bjeldanes, L. F., and Vang, O. (2000). N-methoxyindole-3-carbinol is a more efficient inducer of cytochrome P-450 1A1 in cultured cells than indol-3-carbinol. Nutr. Cancer 36, 112–121.

Bonnesen, C., Stephensen, P. U., Andersen, O., Sørensen, H., and Vang, O. (1999). Modulation of cytochrome P-450 and glutathione S-transferase isoform expression in vivo by intact and degraded indolyl glucosinolates. Nutr. Cancer 33, 178–187.

Stoner, G., Casto, B., Ralston, S., Roebuck, B., Pereira, C., and Bailey, G. (2002). Development of a multi-organ rat model for evaluating chemopreventive agents: Efficacy of indole-3-carbinol. Carcinogenesis 23, 265–272.

Bradfield, C. A., and Bjeldanes, L. F. (1987). Structure-activity relationships of dietary indoles: A proposed mechanism of action as modifiers of xenobiotic metabolism. J. Toxicol. Environ. Health 21, 311–323. Chen, I., Safe, S., and Bjeldanes, L. (1996). Indole-3-carbinol and diindolylmethane as aryl hydrocarbon (Ah) receptor agonists and antagonists in T47D human breast cancer cells. Biochem. Pharmacol. 51, 1069–1076.

Stresser, D. M., Bailey, G. S., and Williams, D. E. (1994). Indole-3-carbinol and b-naphthoflavone induction of aflatoxin B-1 metabolism and cytochromes P-450 associated with bioactivation and detoxification of aflatoxin B-1 in the rat. Drug Metab. Dispos. 22, 383–391. van Lieshout, E. M., Posner, G. H., Woodard, B. T., and Peters, W. H. M. (1998). Effects of the sulforaphane analog compound 30, indole-3-carbinol, D-limonene or relafen on glutathione S-transferases and glutathione peroxidase of the rat digestive tract. Biochim. Biophys. Acta 1379, 325–336.

Chinni, S. R., Li, Y. W., Upadhyay, S., Koppolu, P. K., and Sarkar, F. H. (2001). Indole-3-carbinol (I3C) induced cell growth inhibition, G1 cell cycle arrest and apoptosis in prostate cancer cells. Oncogene 20, 2927–2936.

Vang, O., and Dragsted, L. (1996). Naturally occurring antitumourigens—III. Indoles. Nordic Council of Ministers, Copenhagen.

Cover, C. M., Hsieh, S. J., Cram, E. J., Hong, C., Riby, J. E., Bjeldanes, L. F., and Firestone, G. L. (1999). Indole-3-carbinol and tamoxifen cooperate to arrest the cell cycle of MCF-7 human breast cancer cells. Cancer Res. 59, 1244–1251.

Vang, O., Frandsen, H., Hansen, K. T., Sorensen, J. N., Sørensen, H., and Andersen, O. (2001). Biochemical effects of dietary intake of different broccoli samples. I. Differential modulation of cytochrome P-450 activities in rat liver, kidney, and colon. Metabolism 50, 1123–1129.

Cover, C. M., Hsieh, S. J., Tran, S. H., Hallden, G., Kim, G. S., Bjeldanes, L. F., and Firestone, G. L. (1998). Indole-3-carbinol inhibits the expression of cyclin-dependent kinase-6 and induces a G1 cell cycle arrest of human breast cancer cells independent of estrogen receptor signaling. J. Biol. Chem. 273, 3838–3847.

Verhoeven, D. T., Goldbohm, R. A., van Poppel, G., Verhagen, H., and Van den Brandt, P. A. (1996). Epidemiological studies on brassica vegetables and cancer risk. Cancer Epidemiol. Biomark. Prev. 5, 733–748.

Ge, X. K., Fares, F. A., and Yannai, S. (1999). Induction of apoptosis in MCF-7 cells by indole-3-carbinol is independent of p53 and bax. Anticancer Res. 19, 3199–3203. Grose, K. R., and Bjeldanes, L. F. (1992). Oligomerization of indole-3-carbinol in aqueous acid. Chem. Res. Toxicol. 5, 188–193. Hanley, A. B., Parsley, K. R., Lewis, J. A., and Fenwick, G. R. (1990). Chemistry of indole glucosinolates: Intermediacy of indole-3-ylmethyl isothiocyanates

Xu, M. R., Orner, G. A., Bailey, G. S., Stoner, G. D., Horio, D. T., and Dashwood, R. H. (2001). Post-initiation effects of chlorophyllin and indole-3-carbinol in rats given 1,2-dimethylhydrazine or 2-amino-3-methylimidazo[4,5-f]quinoline. Carcinogenesis 22, 309–314. Zheng, Q., Hirose, Y., Yoshimi, N., Murakami, A., Koshimizu, K., Ohigashi, H., Sakata, K., Matsumoto, Y., Sayama, Y., and Mori, H. (2002). Further investigation of the modifying effect of various chemopreventive agents on apoptosis and cell proliferation in human colon cancer cells. J. Cancer Res. Clin. Oncol. 128, 539–546.