Chiral Resolution, Absolute Configuration Assignment, and Genotoxicity Evaluation of Racemic 3,4-Dihydroquinazoline as a Novel Anticancer Agent

Research Institute for Basic Sciences and Department of Chemistry, College of Sciences, Kyung Hee University, Seoul 02447, Republic of Korea ONCOZEN Co.,Ltd., ONCOZEN R&D Center, C-713, Beobwon-ro 11-gil, Songpa-gu, Seoul 05836, Republic of Korea Department of Pharmaceutical Biochemistry, College of Pharmacy, Kyung Hee University, Seoul 02447, Republic of Korea Department of Life and Nanopharmaceutical Sciences, Graduate School, Kyung Hee University, Seoul 02447, Republic of Korea KHU-KIST Department of Converging Science and Technology, Graduate School, Kyung Hee University, Seoul 02447, Republic of Korea


Introduction
e significance of chirality in drug design and development cannot be understated [1]. About more than half of the drugs currently in use are chiral compounds and near 90% of the last ones are marketed as racemates (or racemic mixtures) [2]. Although the enantiomers of chiral drugs have the same chemical connectivity of atoms, they would exhibit marked differences in their pharmacology, toxicology, pharmacokinetics, metabolism, etc.: while one enantiomer is beneficial to the body, the other enantiomer can be highly toxic to the body [3] . A well-known example of enantiomer-related toxicity is the R-and S-enantiomers of thalidomide [4]. e R-enantiomer is an effective sedative, but the S-enantiomer is known to cause teratogenic birth defects. During the late 1950s and early 1960s, thalidomide produced birth defects in children born to women who took the racemic mixture during pregnancy [5]. According to the United States Food and Drug Administration (US FDA) guidelines, it is mandatory to separate chiral drugs before they are marketed [6]. Rigorous justification is needed for the market approval of racemic drugs. FDA has approved the marketing of some racemic drugs such as flecainide, ketorolac, and prilocaine [7]. In this case, both enantiomers have equal therapeutic potency without any toxicity or they are known to racemize in the body.
In our previous work, a family of 3,4-dihydroquinazoline derivatives was discovered as a promising anticancer agent [8][9][10], among which KCP-10043F (OZ-001) represses the proliferation of human A549 lung cancer cells by caspasemediated apoptosis via STAT3 inactivation [11,12]. is compound was a racemate (±)-3,4-dihydroquinazoline derivative with a single chiral center (Figure 1). In this work, we separated the racemate (±)-KCP-10043F into two optically pure enantiomers (+)-KCP-10043F and (−)-KCP-10043F, which were assigned for their absolute configuration by the formation of chiral diastereomers and 1 H NMR anisotropy method. e bacterial reverse mutation test (Ames test) for two enantiomers and a racemate of KCP-10043F was performed to investigate their latent genotoxicity as well as their anticancer activities.

Chemistry.
Analytical thin-layer chromatography (TLC) was performed on silica gel precoated on glass-backed plates (Fluka Kieselgel 60F 254 , Merck). An UV light (λ � 254 nm) was used for the detection. Flash chromatography was performed on silica gel 60 (particle size 230-400 mesh, Merck). Commercially available reactants were supplied by Sigma-Aldrich and used without further purification.

Preparation of Compound (R)-6.
To a solution of (S, 4R)-8a (0.17 g, 0.28 mmol) in THF/H 2 O (1 : 1, v/v, 8 mL) LiOH (70 mg, 1.67 mmol) was added. e reaction mixture was stirred at room temperature for 72 h and concentrated under reduced pressure. e reaction mixture was dissolved with a mixed solvent of DCM (10 mL) and water (10 mL). e extracted aqueous layer was acidified with 0.5 N HCl until the pH of the aqueous layer was 2, extracted with DCM (15 mL). e extracted organic layer was dried with MgSO 4 and evaporated under reduced pressure to give desired product (R)-6 (140 mg, >99%): 1

Time-Dependent Density Functional eory (TDDFT)
Calculation. e simulated ECD spectrum of (R)-(−)-KCP-10043F was calculated with TDDFT approach at B3LYP/ TZVP level based on the structure optimized at the same level in Gaussian09 program package [13].
To assign the absolute configuration at C-4 position of the quinazoline scaffold, we decided to investigate the 1 H NMR  data of the separated esters 8a and 8b, respectively. Before the analysis of the chemical shift data of 1 H NMR data, molecular mechanics (MM2) calculation and 3D-conformation analysis showed the considerable energy difference (ΔE � 1.02 kcal/ mol) between (S, 4R)-8a and (S, 4S)-8b as shown in Figure 3. N-methyl group inside the white-colored dotted circle in (S, 4R)-8a was completely located above the diamagnetic (shielding) zone of benzene ring (4Å: yellow-colored dotted circle) as shown in Figure 3(a), while the corresponding methyl group of (S, 4S)-8b was located within the paramagnetic (deshielding) zone (Figure 3(b)). According to the calculated anisotropic effect zone of benzene ring (Supplemental data: Figure S5) [21], the protons of the methyl groups of (S, 4R)-8a were anticipated to be largely upfield-shifted by the diamagnetic anisotropy effect of the phenyl moiety compared to those of (S, 4S)-8b. In experimental 1 H NMR data, the chemical shifts of N-methyl groups in (S, 4R)-8a and (S, 4S)-8b exhibited 2.07 and 2.57 ppm, respectively (Figures 4(b) and 4(c)). Based on the consistency between the theoretical data and experimental data, the chiral center of (−)-KCP-10043F can be designated R configuration and thus (+)-KCP-10043F can be readily assigned as S configuration. Additionally, the ECD spectrum of (−)-KCP-10043F was further measured to afford a positive band at 268 nm and a negative band at 303 nm, which was consistent with Gaussian TDDFT-calculated R configuration of KCP-10043F as shown in Figure 5 [22].
Each compound was studied at concentrations of 20∼5,000 μg/mL in the presence or absence of metabolic activation (S9 fraction). As shown in Table 1, a revertant analysis showed no significant differences between the treatment doses (20∼5,000 μg/mL) and the negative control (DMSO) regardless of +S9 and −S9, when compared to the positive control and included all strains of S. typhimurium and E. coli EP2 (Supplemental data: Figures S6-S10). All of these Ames test results showed that none of all stereoisomers of KCP10043F were genotoxic regarding frameshift mutations (TA98 and TA1537) and base pair substitutions {TA100, TA1535 and E. coli WP2 uvrA[pKM101]} regardless of the treatment dose [24].  Journal of Chemistry

Conclusions
(R)-(−)-KCP10043F and (S)-(+)-KCP10043F were successfully separated from racemate (±)-KCP-10043F (OZ-001) by the chiral technique of supercritical fluid chromatography and assigned to their absolute configuration by the preparation of chiral diastereomers and 1 H NMR anisotropy method as well as the experimental ECD. e bacterial reverse mutation test (Ames test) for racemate KCP-10043F and its two enantiomers was performed to investigate their latent genotoxicities. All stereoisomers were fortunately found to be non-genotoxic against five bacterial strains with/without metabolic activation. In addition, (R)-(−)-KCP-10043F displayed almost equal cytotoxic activity to (S)-(+)-KCP-10043F against three cancer cell lines. Based on these overall results, racemate (±)-KCP-10043F (OZ-001) could be used for our ongoing preclinical and clinical studies without the expensive asymmetric process and/or chiral separation.  Strain e results were obtained on the baselines by using the calculation of Xenometrix (Xenometric AG, Swiss); After calculating the baseline of DMSO (negative control), the compound was regarded as a mutagen when the induction is more than 2-fold over the baseline and binomial P value ≥0.99; x means nongenotoxicity (negative response). Table 2: Cytotoxicity of racemate (±)-KCP-10043F and its two enantiomers against human cancer cell lines.