Study of the antitumor mechanisms of apiole derivatives (AP-02) from Petroselinum crispum through induction of G0/G1 phase cell cycle arrest in human COLO 205 cancer cells

Figure 1a depicts the preparation of AP-02. Preparation began with base-mediated O-allylation of commercially available benzo[d][1,3]dioxol-5-ol (1) with allyl bromide (2) in acetone under reflux conditions to give 5-(allyloxy)benzo[d][1,3]dioxole (3) at a 75% yield. The subsequent microwave-induced Claisen rearrangement of 3 yielded 6-allylbenzo[d][1,3]dioxol-5-ol (4) at 68% yield. The final O-methylation of 4 with methyl iodide in the presence of potassium carbonate in dichloroethane under reflux conditions afforded the target compound, AP-02, at 88% yield.

Figure 1b shows preparation of AP-04, which was readily available via base-mediated O-methylation of commercially available benzo[d][1,3]dioxol-5-ol (1) with methyl iodide in dichloroethane under reflux conditions for 1 h to give 5-methoxybenzo[d][1,3]dioxole (AP-04) at 85% yield. AP-05 was purchased from Acros Co. (Geel, Belgium, Cn). The chemical structures of apiole derivatives (AP-02, AP-04, and AP-05) is summarized in Fig. 2.

Human breast (ZR75, MDA-MB-231), lung (A549, PE089), colon (COLO 205, HT 29) and hepatocellular (Hep G2, Hep 3B) cancer cells were used in this study. Normal cells were used as controls (human breast (MCF 10A), lung (HEL 299), liver (BNL CL.2, Clone 9), and colon (FHC) cells) and were treated with the same regimens. MDA-MB-231 and ZR-75 cells were derived from human mammary gland and from a metastatic site of pleural effusion, respectively (American Type Culture Collection, ATCC® HTB-26™ and CRL-1500™). A549 cells were derived from human alveolar basal epithelial cell adenocarcinoma (ATCC® CCL-185™ and CRL-1500™). PE089 cells were isolated from a female patient with lung adenocarcinoma with an EGFR exon 19 deletion (courtesy of K. J. Liu from the National Health Research Institute). COLO 205 and HT 29 cell lines were isolated from human colon adenocarcinoma (ATCC® HMIC-38™ and CCL-222™). Hep 3B and Hep G2 cell lines were derived from human hepatocellular carcinoma (ATCC® HB-8064™ and HB-8065™) (Knowles et al., 1980). MCF-10A cells were isolated from normal human epithelial cells of the mammary gland (ATCC® CDR-10317™). HEL-299 cells are human embryonic lung cells derived from embryonic lung tissue (ATCC® CCL-137™). BNL CL.2 is a normal murine liver cell line (ATCC® TIB-73™). Clone 9 is a normal rat liver epithelial cell line (ATCC® CRL-1439™). FHC cells are normal human colon epithelial cells (ATCC® CRL-1439™). The p53 gene in both Hep G2 and COLO 205 cells is wild type [14‐16], whereas p53 is partially deleted (7 kb) in the Hep 3B cells and mutated (codon 273) in the HT 29 cells [14, 17]. Cells were cultured in Eagle’s Minimal Essential Medium (for Hep 3BHep G2, and PE089 cells), Minimal Essential Medium (for HEL-299 and Clone 9 cells) or RPMI 1640 (for COLO 205, HT 29, FHC and A549 cells) supplemented with 50 μg/ml gentamycin, 0.3 mg/ml glutamine and 10% fetal calf serum (FCS). A 3:1 mixture of Ham’s F12 medium and DMEM (for MCF-10A cells) was supplemented with 10% FCS, 40 ng/ml hydrocortisone, 0.01 mg/ml cholera toxin, 0.005 mg/ml insulin, and 10 ng/ml epidermal growth factor. The BNL CL.2 cell line was selected to grow in medium containing ornithine and phenylalanine in place of arginine and tyrosine.

Cells were treated with compounds (Apiole, AP-02, AP-04, and AP-05) in a dose range from 4.5 to 600 μM for 24 h, and IC50 values were determined. Cell viability was determined at indicated apiole doses using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay. Briefly, cells were seeded in 96-well plates at a density of 1 × 10⁴ cells/well overnight. Next, medium in each well (200 μL) was replaced with 10 mmol/L HEPES (pH 7.4), and MTT dye (50 μL) was added to each well. Plates were incubated for 2~4 h at 37 °C in the dark. Media was removed and replaced with 200 μL phosphate-buffered saline (PBS) and 25 μL Sorensen’s glycine buffer. The absorbance ratio was measured at 570 nm using an ELISA plate reader.

Cells were washed with PBS three times 24 h after plating and incubated with media containing 0.04% fetal calf serum (FCS) for an additional 24 h. Under these conditions, cells are arrested in G0/G1 phase based on flow cytometry analysis as reported in our previous study [18]. Next, synchronized cells (cultured in 0.04% FCS) were challenged with the addition of media containing 10% FCS. Apiole- and PBS-treated groups were assessed via flow cytometry analysis to determine cell cycle distribution. Cells were stained with propidium iodide (50 μg/ml) (Sigma Chemical Co., St. Louis, MO, USA), and DNA content was assessed using a FACScan laser flow cytometry analysis system (Becton-Dickinson, San Jose, CA, USA), with 15,000 events being analyzed for each sample.

Cell lysates were prepared, electrotransferred, immunoblotted with antibodies, and then visualized via incubation with colorigenic substrates (nitroblue tetrazolium, NBT, and 5-bromo-4-chloro-3-indolyl phosphate, BCIP) (Sigma Chemical Co., St. Louis, MO, USA). Expression of GAPDH was used as a control to ensure equal protein loading [19].

Immunodetection was performed by probing with appropriate dilutions of specific antibodies at room temperature for 2 h. Anti-p53, anti-p27/Kip1, anti-p21/Cip1, anti-GAPDH monoclonal antibodies (Santa Cruz, Inc. CA, USA), anti-E cadherin [SP64] (ab227639) (Abcam Inc. Shanghai, China), and anti-cyclin D1 monoclonal antibodies (Transduction Laboratories, Lexington, KY) were used. Membranes were incubated with secondary alkaline phosphatase-coupled anti-mouse and anti-rabbit antibodies (Jackson, Westgrove, PA, USA) at room temperature for 1 h at dilutions of 1:5,000 and 1:1,000, respectively. Immunoreactive proteins were visualized with a chemiluminescent detection system (PerkinElmer Life Science, Inc., Boston, MA, USA) and BioMax LightFilm (Eastman Kodak Co., New Haven, CT, USA) according to the manufacturer’s instructions. Results were analyzed by densitometry analysis.

Four-week-old athymic nude mice were purchased from the National Science Council Animal Center, Taipei, Taiwan. Five animals in each cage were fed and acclimatized in our experimental animal center for 2 weeks in rectangular cages, and a small wood strip served as environmental enrichment. Animals were fed with the lab diet 5 k52 formulation (6% fat), and water was accessible at all times. Our animal facility was maintained under standard specific-pathogen-free conditions and a 12 h/12 h light/dark cycle. The total number of animals used in this experiment was 21, and the animals were equally divided among the 3 groups based on tumor volume. Sample size calculation was considered as follows: effect size = total number of animals − total number of groups, and the number of animals in each experimental group was n = 7. If the tumor size was greater than 4 cm³ or the mouse weight was 15% below the original weight, euthanasia was performed by placing the mice in a chamber and piping in carbon dioxide (CO₂) at increasing concentrations until the animals became unconscious and died.

COLO 205 cells were cultured in media as described above. Cells (5 × 10⁶) were suspended in 0.2 ml medium and injected subcutaneously between the scapulae of nude mice (purchased from National Science Council Animal Center, Taipei, Taiwan). For determination of tumor growth, the tumor volume was measured according to the following formula: tumor volume (mm³) = L x W²/2, where L is the length and W is the width [19]. Once tumors reached a mean size of 200 mm³, experimental animals were treated with either 25 μl PBS or 1 to 5 mg/kg AP-02 via intraperitoneal injection three times per week for 6 weeks. Control animals were treated with PBS at the same volume.

For each analysis, data are represented as the mean ± SEM of at least three independent experiments. For comparison, statistical significance was tested using t-tests. All p-values were based on two-sided statistical analyses, and p < 0.05 was considered statistically significant.

In this study, apiole and its derivatives (AP-02, 04, 05) were evaluated for their antitumor activities. To determine whether these compounds induce cell death in human cancer cells derived from different organs, we selected human breast (ZR75, MDA-MB-231), lung (A549, PE089), colon (COLO 205, HT 29) and liver (Hep G2, Hep 3B) cancer cells and treated cells with each compound (apiole, Ap-02, Ap-04, and Ap-05; 4.5–600 μM) for 48 h. IC50 values were then determined (Fig. 3 and Table 1). Normal human breast (MCF 10A), lung (HEL 299), liver (BNL CL.2, Clone 9), and colon (FHC) cells were treated in the same manner and served as controls. Among tested compounds, AP-02 was the most effective drug for inhibiting the proliferation of cancer cell lines compared to AP-04 and AP-05 (Table 1). We further observed that the cytotoxic effects of AP-02 were effective against human colon cancer cells (COLO 205 and HT 29, IC50 = 16.57 and 38.45 μΜ, respectively) compared to normal colon (FHC, IC50 = 263.76 μΜ) cells (Table 1, **p < 0.01). Interestingly, the cytotoxicity of AP-02 in normal colon (FHC) epithelial cells was reduced compared to other normal cells derived from the breast, lung and liver (MCF-10A, HEL299, BNL CL.2, and Clone 9, respectively). These results imply that compared to AP-04 and AP-05, AP-02 is valuable for targeting colon cancer cells (Table 1 and Fig. 3, **p < 0.01). Colon cancer cell lines (HT 29 and COLO 205) and normal (FHC) cell line were then treated with AP-02 across a range of doses (5–150 μM, for 24 h). Results indicated that AP-02 was the most effective compound with respect to induction of cytotoxicity, specifically in COLO 205 cells (Fig. 4, red bars, **p < 0.01).

Based on the results described above, we next investigated the specific mechanisms of AP-02 against cancer cells and selected COLO 205 cells as an in vitro cell model to test for cell cycle inhibition effects using flow cytometry (Fig. 5). We synchronized COLO 205 cells by culturing them in 0.04% FCS medium for 24 h according to our previous studies [18, 20]. Synchronized cells were than treated with 10% FCS to reactivate cell proliferation (Fig. 5, red bars). Our previous studies [18, 20] showed that the major differences in G0/G1 cell populations between serum starved and reactivated (10% FCS-treated) COLO 205 cells were observed 15 h after the media was replenished with complete media (Fig. 5a). COLO 205 cells in the PBS-treated group were in S phase at this time point (15 h) [20, 21]. According to these data, we selected this time point (15 h) to test AP02-induced G0/G1 arrest effects across a range of doses. The minimal dose of AP-02 that induced G0/G1 arrest was evaluated via flow cytometry analysis (Fig. 5b). Results indicated that the minimal dose of AP-02 needed to induce significant G0/G1 arrest was 5 μΜ (Fig. 5b, green vs. red bars, *p < 0.05). We further demonstrated that higher doses of AP-02 (> 5 μM for 24 h) induced appearance of a significant subG1 phase cell population (Fig. 5b, yellow and blue bars, *p < 0.05). These results indicate that AP-02-induced cancer cell death was occurring in high-dose groups.

To evaluate AP-02-induced antitumor effects, we used athymic nude mice bearing COLO 205 tumor xenografts as an in vivo animal model. Palpable tumors were established (mean tumor volume, 200 mm³). Next, COLO-205 tumor xenograft animals were treated with AP-02 (1 and 5 mg/kg, I.P. injection) or treated with the same volume of PBS as a control. Results indicated that AP-02 (> 1 mg/kg) treatment significantly reduced tumor volume compared to the control group (Fig. 6). No gross signs of AP-02-induced drug toxicity were detected in body weight changes (Fig. 6b). The general and microscopic appearance of individual organ tissues was also assessed (data not shown).

In this study, we found that high doses (> 25 μΜ) of AP-02 induced the appearance of a significant subG1 phase cell population in COLO 205 cells compared to the 10% FCS-treated group (Fig. 5b, *p < 0.05). However, in the lower dose Apo2-treated group (< 5 μΜ), only G0/G1 cell cycle arrest was detected (Fig. 5b, green bar, *p < 0.05). Based on these results, we suggest that G0/G1 cell cycle regulatory proteins are likely involved in AP-02-induced effects. We next confirmed these in vivo observations in AP-02-treated (1 and 5 mg/mL) COLO 205-xenograft tumor tissues. We found that expression of p53 and p21/Cip1 proteins was significantly induced, while cyclin D1 was downregulated in AP-02-treated tumor tissues (> 1 mg/kg) compared to controls (Fig. 7). To test whether AP-02 effectively inhibited metastasis-related signals in COLO 205-xenograft tumor cells, we selected E-cadherin, which functions as an invasion suppressor, as an indicator and determined whether it was upregulated in AP-02-treated tumors. We demonstrated that AP-02 treatment upregulates expression of E-cadherin protein (> 1 mg/kg) (Fig. 7). These results suggest that induction of E-cadherin by AP-02 inhibits COLO 205 cancer cell migration. The E-cadherin protein has also been used as a target for therapeutic purposes to prevent colorectal cancer cell metastasis [22]. Therefore, AP-02 may have clinical utility as a novel small molecule to prevent colorectal cancer cell metastasis.