Clonogenic long-term survival assay of HCT 116 colorectal cancer cells after treatment with the synthesized diphenyl imidazoline derivatives

Fourteen diphenyl imidazoline derivatives were designed, synthesized, and identified using NMR spectroscopy and high-resolution mass spectrometry. Their cytotoxicities in HCT 116 colorectal cancer cell lines were measured using a clonogenic long-term survival assay and the half-maximal cell growth inhibitory concentration (GI50) values were in the range 3.1–58.4 μM. As the anticancer effects of diphenyl imidazolines were reported to be caused by the inhibition of mouse double minute 2 homolog (MDM2), the inhibitory effects of the most potent derivative on MDM2 were assessed through Western blotting analysis. In silico docking experiments revealed the binding mode between this derivative and MDM2.

Colorectal cancer refers to the cancer of the large intestine, colon, and the last part of the colon, the rectum. It is the third most common cancer in the USA, with more than 100,000 patients diagnosed in 2016, and it is the thirdleading cause of cancer-related deaths [12]. The possible treatments include surgery, radiation therapy, and chemotherapy, but research into new chemotherapeutic agents is in progress. There are many methods to measure the anticancer effects of small compounds. One method is the measurement of their cytotoxicity in cancer cell lines [13][14][15]. Although the clonogenic long-term survival assay requires a time period of several days to obtain results for cytotoxic activities, it can distinguish the cytotoxicities between compounds containing a common skeleton [16]. To find diphenyl imidazolines that show cytotoxicity against colorectal cancer cell lines, 14 diphenyl-imidazoline derivatives were designed and synthesized. Their cytotoxicities in the HCT 116 colorectal cancer cell line were measured using a clonogenic long-term survival assay and their half-maximal cell growth inhibitory concentration (GI 50 ) values were observed at the micromolar level. Recently, it was reported that the anticancer effects of  [17][18][19][20]. Therefore, the most potent compound, (4R,5R)-2-(naphthalen-2-yl)-4,5-diphenyl-imidazoline, synthesized in this study was assessed for its inhibitory effect on MDM2 by Western blotting analysis. The binding mode between this compound and MDM2 was elucidated using in silico docking. The goal of this research was to study the structure-activity relationships (SARs) to suggest the structural features that contribute to the compound's strong cytotoxicity.

Materials and methods
Preparation of diphenyl imidazoline derivatives (1R, 2R)-(?)-1,2-diphenylethylenediamine (DPEN, 11 mmol, 2.33 g) and p-methoxy benzaldehyde (10 mmol, 1.36 g) were dissolved in 100 mL t-BuOH. The mixture was stirred at room temperature for 30 min after which potassium carbonate (30 mmol, 4.14 g) and iodine (12.5 mmol, 3.15 g) were added to the mixture. The reaction mixture was stirred at 70°C for 3 h. When the completion of the reaction was confirmed by thin layer chromatography, the reaction mixture was quenched with saturated Na 2 S 2 O 5 solution until the iodine color disappeared. The precipitate was filtered under vacuum and dried. The crude solid was recrystallized from ethanol. Yield 60%, m.p. 216-218 (white solid). The synthetic process is summarized in Scheme 1. Of 14 derivatives synthesized here, three derivatives 1, 4, and 9 have been reported previously.

Identification of diphenyl imidazoline derivatives
The structures and names of the diphenyl imidazolines synthesized in this work are listed in Table 1. The compounds were dissolved in deuterated dimethyl sulfoxide (DMSO-d 6 ), chloroform (CHCl 3 -d), and acetone (CH 3-COCH 3 -d 6 ) for nuclear magnetic resonance (NMR) spectroscopy and adjusted to a concentration of approximately 50 mM. These solutions were transferred to 2.5-mm NMR tubes. All NMR spectra were collected using an Avance 400 spectrometer system (9.4 T; Bruker, Karlsruhe, Germany) at 25°C. The NMR experimental conditions used are similar to those employed in a previous study [24]. To confirm the structures of the diphenyl-imidazoline derivatives, ultraperformance liquid chromatography-hybrid quadrupole-time-of-flight mass spectrometry was performed using a Waters Acquity UPLC system (Waters Corp., Milford, MA) with the help of Prof. Choong Hwan Lee at Konkuk University, Korea [25]. All high-resolution mass spectrometric (HRMS) data consisted of M ? H ions.  Melting points and IR spectra were measured using a Mel-Temp II (LabX, Midland, ON, Canada) and an FT-IR 4200 spectrometer (JASCO, Easton, MD) equipped with an ATR (Attenuated Total Reflection, ATR PR0450-S), respectively [26]. The melting points and spectroscopic data for the 14 diphenyl imidazoline derivatives obtained from NMR spectroscopy, HRMS, and IR are summarized in the references and notes.

Biological evaluation of diphenyl imidazoline derivatives
HCT 116 human colorectal cancer cells were used for the cytotoxicity tests of the diphenyl imidazoline derivatives. The long-term clonogenic survival data were collected in the absence or presence of different concentrations (0, 1, 5, 10, and 20 lM) of the 14 derivatives. The procedures adhered to previously reported methods [27]. The results were analyzed using densitometry (MultiGauge, Fujifilm, Japan), and the GI 50 values were determined using Sig-maPlot software (SYSTAT, Chicago, IL, USA) ( Table 1) [28].

Statistical analysis
Statistical analyses were conducted using a one-way analysis of variance (ANOVA), followed by Kruskal-Wallis test, by using GraphPad Instat software (La Jolla, CA). All experiments were performed in triplicate [30].

In silico docking
For in silico docking experiments, a three-dimensional (3D) structure of derivative 13 was obtained from the crystal structure of a ligand for 1ttv.pdb and modified [31]. From the 3D structures of MDM2 deposited in the PDB, 4ode.pdb was selected for this experiment because it has the longest chain of the N-terminal p53 interaction domain (6-110 residues) at a high resolution (1.8 Å ) [32]. In silico docking experiments were performed on an Intel Core 2 Quad Q6600 (2.4 GHz) Linux PC with Sybyl 7.3 software (Tripos) [33], and LigPlot software was used to identify the binding sites [34]. All 3D images were generated using the PyMOL software (The PyMOL Molecular Graphics System, Version 1.0r1, Schrödinger, LLC).

Results and discussion
The cytotoxicities of the diphenyl imidazolines were measured using a clonogenic long-term survival assay, as shown in Fig. 1. The GI 50 values were in the range 3.1-58.4 lM, as listed in Table 1, and the bar graph (including errors) is shown in Fig. 2.
Derivatives 12 and 13 both contained a 2-naphthalenyl group, but the former was 2-naphthalen-1-yl and the latter was 2-naphthalen-2-yl. The GI 50 values of 8.1 lM and 3.1 lM, respectively, indicated that the 2-naphthalen-1-yl group was a better pharmacophore than the 2-naphthalen-1-yl group was. Derivatives 5, 7, and 8 contained the halogens bromine, chlorine, and fluorine at the para position of the 2-phenyl group, respectively, with GI 50 values of 13.4 lM, 18.1 lM, and 23.0 lM, respectively. Consequently, it was shown that the electronegativity affected the cytotoxicity. Derivative 1, with a p-methoxy phenyl group, showed a higher GI 50 value than did derivative 4 that contained p-toluene. A comparison of the derivatives 1, 2, Fig. 2 The half-maximal cell growth inhibitory concentration (GI 50 ) values in the HCT 116 human colorectal cancer cell lines of 14 diphenyl imidazolines with error bars Fig. 3 Effect of derivative 13 on the phosphorylation of MDM2 at S166. HCT 116 cells were treated with 40 lM derivative 13 for different time periods. Whole cell lysates were separated by 10% SDS-PAGE, and the phosphorylation status of MDM2 on S166 was determined by immunoblotting. GAPDH was used as an internal control and 11 revealed that the cytotoxicity was improved by an increase in the number of the methoxy groups of diphenyl imidazolines. This phenomenon agreed with the behavior of derivatives 12 and 14. Derivative 14, with a 4-methoxynaphthalenyl group, showed a lower GI 50 than did derivative 12 without the 4-methoxy group. The GI 50 value of derivative 9, 2-((4R,5R)-4,5-diphenyl-imidazolin-2yl)pyridine, was better than that of derivative 10, 3-((4R,5R)-4,5-diphenyl-imidazolin-2-yl)pyridine, which indicated that the 2-pyridinyl group was a better pharmacophore than the 3-pyridinyl group was. This phenomenon might be caused by the different distances between the nitrogen of pyridine and the nitrogens of imidazoline. Derivative 6, with a thiopen-3-yl group, and derivative 10, with a pyridin-3-yl group, showed less cytotoxicity than did derivative 9, with a pyridin-2-yl group. The distance between the two nitrogens of the former compound was 4.5 Å and that of the latter was 3.0 Å .
The p53 protein functions as a tumor-suppressor gene, which plays a critical role in the induction of cell cycle arrest, DNA repair, and apoptosis after diverse DNA damage responses [35][36][37]. MDM2 is an oncoprotein that prevents apoptosis by inducing the degradation of p53 [38]. MDM2 is activated by the phosphorylation of S166, which is mediated by the MEK/ERK or PI3 K/AKT pathways [39,40]. As mentioned above, the anticancer effects of diphenyl imidazolines were attributable to their inhibitory effects against MDM2 [17][18][19][20]. To investigate whether the phosphorylation of MDM2 at S166 was affected by diphenyl imidazoline derivative 13 showing the most potent cell growth inhibitory effect, (4R,5R)-2-(naphthalene-2-yl)-4,5-diphenyl-imidazoline, HCT 116 cells were subjected to treatment with 40 lM derivative 13 for different periods of time. The results indicated that the phosphorylation of MDM2 at S166 was decreased in a time-dependent manner (Fig. 3). Thus, derivative 13 efficiently inhibited the phosphorylation of MDM2 on S166.
The 12 residues in the binding pocket were used for the docking experiment of derivative 13. After docking the derivative 13 into MDM2, 30 complexes were generated from 30 docking iteration processes. The binding energy of the 30 complexes was in the range -17.80 to -10.69 kcal/mol. Among the 30 complexes, the seventh complex, with -12.14 kcal/mol of binding energy, was selected to clarify the binding mode based on both the binding pose and the binding energy. This complex was analyzed using the LigPlot program. Twelve residues were involved in the binding of derivative 13 with MDM2: M6,  Q18, L54, G58, I61, K70, Y67, H73, V93, K94, H96, and I99 (Fig. 6). Two residues, V93 and H96, formed H-bonds with the nitrogen of derivative 13 and the other 10 residues showed hydrophobic interactions with the derivative. Nine residues (M6, L54, G58, I61, Y67, H73, V93, K94, and H96) were observed in both binding pockets of the original ligand of 4ode.pdb, 2U0, and derivative 13. The inside of the binding pocket was split into three small pockets. In the crystal structure of 4ode.pdb, the ligand was a good fit into these pockets (Fig. 7A) as the cyclopropyl, 4-chloro-3fluorophenyl, and 3-chlorophenyl groups of the ligand were positioned into each of the smaller pockets. The naphthalenyl ring of derivative 13 bound to one of the pockets (Fig. 7B), which provided an explanation of the increased inhibitory activity of derivative 13 (3.1 lM) compared with derivative 12 (8.1 lM). The LigPlot analysis of the derivative 12-MDM2 complex is shown in Fig. 8A. Unlike derivative 13, derivative 12 interacts with seven residues of the protein. The binding pocket for the naphthalenyl ring, which is narrow and deep, is suitable for derivative 13, which contains 2-naphthalen-2-yl. Four residues (L54, G58, I61, and I99) bound the 2-naphthalen-2-yl group of derivative 13 (blue) via hydrophobic interactions, as shown in Fig. 8B. In contrast, it is too narrow to bind with derivative 12 (orange), which contains 2-naphthalen-1-yl. A hydrophobic interaction was observed between MDM2 and the 2-naphthalen-1-yl group of derivative 12.
As previously mentioned, derivative 2, with tri-methoxy groups at the 2-phenyl ring, showed better inhibition compared with the activities of derivatives 1-11. This suggested that the bulky group at C-2 of imidazoline was necessary for docking into the binding pocket of MDM2.
In conclusion, we synthesized 14 diphenyl imidazoline derivatives and examined their cytotoxicities. Although the diphenyl imidazoline derivatives used in this study have only small structural differences, the GI 50 values of the 14 derivatives in HCT 116 cells varied from 3.1 to 58.4 lM. Based on the results of SARs and in silico docking studies, we therefore proposed that bulky groups, such as naphthalene or a highly methoxylated phenyl ring, were favored at the C-2 position of imidazoline. Additionally, branched structures were suitable for tight binding because of the presence of three small pockets.