Synthesis, antitumour and antioxidant activities of novel α,β-unsaturated ketones and related heterocyclic analogues: EGFR inhibition and molecular modelling study

Abstract New α,β-unsaturated ketones 4a,b; 5a–c; and 6a,b; as well as 4-H pyran 7; pyrazoline 8a,b; isoxazoline 9; pyridine 10–11; and quinoline-4-carboxylic acid 12a,b derivatives were synthesized and evaluated for in vitro antitumour activity against HepG2, MCF-7, HeLa, and PC-3 cancer cell lines. Antioxidant activity was investigated by the ability of these compounds to scavenge the 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) radical cation (ABTS•+). Compounds 6a, 6b, 7, and 8b exhibited potent antitumour activities against all tested cell lines with [IC50] ≅5.5–18.1 µΜ), in addition to significantly high ABTS•+ scavenging activities. In vitro EGFR kinase assay for 6a, 6b, 7, and 8b as the most potent antitumour compounds showed that; compounds 6b, and 7 exhibited worthy EGFR inhibition activity with IC50 values of 0.56 and 1.6 µM, respectively, while compounds 6a and 8b showed good inhibition activity with IC50 values of 4.66 and 2.16 µM, respectively, compared with sorafenib reference drug (IC50 = 1.28 µM). Molecular modelling studies for compounds 6b, 7, and 8b were conducted to exhibit the binding mode towards EGFR kinase, which showed similar interaction with erlotinib.

Taking all the aforementioned facts into account in our continuous efforts to develop new structures to serve as antitumour and antioxidant agents, we synthesized new a,b-unsaturated ketones, 4-H pyran, pyrazoline, isoxazoline, pyridine, and quinoline derivatives (G; Figure 1). The rationale for evaluating the antitumour, antioxidant, and EGFR kinase inhibition activities of the designed molecules (G; Figure 1) was as follows: (i) design the structure-activity relationship for compounds incorporating a,b-unsaturated ketones with diverse substituent groups; (ii) recognise the effectiveness of the cyclic a,b-unsaturated ketones versus the acyclic derivatives; (iii) thus, compare the cycloalkanones and their piperidinone analogues; (iv) heterocyclic compounds resulting from the addition reaction of a,b-unsaturated ketones such as pyrane, pyrazoline, oxazoline, and pyridine derivatives were also included in the study in order to cover the most relative analogues.
Furthermore, the most active antitumour compounds were subjected to EGFR kinase inhibition test and docked into the binding sites of EGFR kinase enzyme to explore their complementarity with the specified binding pockets.

Chemistry
Melting points ( C, uncorrected) were measured using a Fisher-Johns apparatus. Elemental analyses were carried out at the microanalytical unit, Cairo University. IR spectra (potassium bromide [KBr]) were acquired using a Mattson 5000 FT-IR spectrometer ( in cm À1 ). 1 H NMR and 13 C NMR spectra were obtained in deuterated dimethyl sulphoxide (DMSO-d 6 ) or deuterated chloroform (CDCl 3 ) on Bruker 400 and 100 MHz instruments, respectively, using tetramethyl silane (TMS) as an internal standard. Chemical shifts were reported downfield from TMS in ppm, d units. Mass spectrometry (MS) measurements were performed on a JEOL JMS-600H spectrometer. The purities of the compounds were evaluated by thin layer chromatography (TLC), which was performed on silica gel G (Merck), and spots were visualised by irradiation with ultraviolet light (UV; 254 nm). Compound 3, 4-(cyclopentyloxy)benzaldehyde, was synthesized in accordance with the method described in the literature 61 .
General method for the synthesis of a,b-unsaturated ketone derivatives (4a,b; 5a-c; and 6a,b) A solution of 4-(cyclopentyloxy)benzaldehyde 3 (1.9 g, 0.01 mol) in ethanol (20 ml) was added to a stirred solution of the appropriate ketone (0.03 mol) in ethanol (20 ml) containing NaOH (0.8 g, 0.02 mol). The reaction mixture was refluxed for 8 h, cooled and the solvent was evaporated under reduced pressure. The resulting solid was triturated with diethyl ether, filtered, dried, and crystallised from the appropriate solvent.

Synthesis of compounds 8a and 8b
A mixture of compound 4b (0.91 g, 0.003 mol) and hydrazine hydrate (0.15 g, 0.003 mol) in absolute ethanol (30 ml) or phenyl hydrazine (0.32 g, 0.003 mol) in glacial acetic acid (5 ml) was heated under reflux for 9-10 h. After cooling, the separated products were filtered, dried, and crystallised from ethanol to yield the title compounds. Biological testing

Antioxidant assay
The absorbance (A control ) of a green-blue solution (ABTS þ radical solution) resulted from a mixture of ABTS and manganese dioxide (MnO 2 ) and was recorded at k max 734 nm, according to the reported procedure 64,65 . The absorbance (A test ) was measured upon the addition of 20 ml of 1 mg/ml solution of the test sample in spectroscopic grade methanol/phosphate buffer (1:1 v/v) to the ABTS solution. The decrease in absorbance is expressed as % inhibition, which can be calculated from the following equation: L-Ascorbic acid 20 ml (2 mM) solution was used as standard antioxidant (positive control). A blank sample was run using only methanol/phosphate buffer (1:1), while the negative control was run with ABTS and the methanol/phosphate buffer.
EGFR kinase inhibition assay EGFR kinase activity was determined via EGFR Human In-Cell ELISA Kit in 96-well plates according to the manufacturer's instructions (EGFR Kinase Assay Kit Catalog # ab126419 of ABCAM, Cambridge, MA), as supplemental information 66 . The EGFR kinase activities for each compound were expressed as IC 50 values using seven concentrations (10.0, 5.0, 2.5, 1.25, 0.625, 0.31, and 0.15 mM).

Docking methodology
All modelling experiments were conducted with MOE programs running on PC computer (MOE 2008.10 of Chemical Computing Group. Inc, Montreal, QC, Canada) 67 . Starting coordinates of the Xray crystal structure of EGFR enzyme in complex with eroltinib (PDB code 1M17) is obtained from the RCSB Protein Data Bank. All the hydrogen was added and enzyme structure was subjected to a refinement. The docking methodology was similar to that described in our previous reports 5, [68][69][70] . aromatic signals at 7.85-7.32 ppm in addition to a singlet signal at 2.36 ppm due to the presence of a 4-methyl group. The presence of a new peak at 198.2 ppm due to a carbonyl (CO) group was demonstrated in 13 C NMR spectrum. 1 H NMR spectra of compounds 5a-c were characterised by the presence of cycloalkane protons at 4.90-1.00 ppm. The 1 H NMR spectrum of compound 6a is characterised by the presence of a singlet peak at 2.3 ppm corresponding to the methyl protons of the N-CH 3 group, while a triplet-quartet pattern characteristic of an ethyl group (N-CH 2 CH 3 ) was identified in the 1 H NMR spectrum of compound 6b at 2.65 and 3.70 ppm, respectively. Synthesis of 4-H pyran derivative (7) was achieved by stirring 4-(cyclopentyloxy)benzaldehyde (3), malononitrile, and ethyl acetoacetate in ethanol in the presence of a catalytic amount of sodium benzoate at room temperature. The infra-red (IR) spectrum of compound 7 exhibited bands at 3401, 3326 (NH 2 ), 2221 (CN), and 1697

Results and discussion
. Meanwhile, the 1 H NMR spectrum showed a triplet and quartet at 1.20 and 4.10 ppm integrating for the COOCH 2 CH 3 group, respectively. In addition, presence of two singlet peaks at 5.70 and 8.30 ppm for the methyl (CH 3 ) and amine (NH 2 ) groups, respectively.

Synthesis of compounds 8-12 (Scheme 2)
The compound 3-(4-(cyclopentyloxy)phenyl)-1-(4-methylphenyl)prop-2-en-1-one (4b) was heated under reflux with hydrazine hydrate or phenylhydrazine in ethanol or glacial acetic acid, resulting in the corresponding pyrazoline derivatives 8a and 8b. 1 H NMR spectra of compounds 8a and 8b were characterised by the disappearance of the olefinic protons with the appearance of pyrazoline protons at 6.85-6.75, 3.90-3.75, and 3.50-3.30 ppm. Moreover, facile cyclocondensation of compound 4b with hydroxylamine hydrochloride in ethanolic potassium hydroxide gave the corresponding isoxazoline (9). The 1 H NMR spectrum of compound 9 was characterised by the disappearance of the olefinic protons with the appearance of isoxazoline protons at 6.80-6.70 and 3.90-3.80 ppm. Reaction of the a,b-unsaturated ketone 4b with ethylcyanoacetate or malononitrile in ethanol in the presence of ammonium acetate yielded the cyanopyridine derivatives 10 and 11, respectively. IR spectra of compounds 10 and 11 were used to verify their structures through the appearance of characteristic absorption bands due to nitrile groups at 2215 and 2212 cm À1 , respectively. In addition, a singlet peak at 8.07 ppm corresponding to the NH proton appeared in the 1 H NMR spectrum of compound 10, while a singlet peak at 8.10 ppm was assignable to the NH 2 group in compound 11, and both were deuterium oxide (D 2 O) exchangeable. Quinoline-4-carboxylic acid derivatives 12a,b were prepared by condensation of 4-(4-(cyclopentyloxy)phenyl)but-3-en-2-one (4a) and isatin derivatives in ethanolic potassium hydroxide 71 . The IR spectrum of compound 12b was characterised by the presence of absorption bands at 3421 cm À1 and 1690 cm À1 , representing hydroxy (OH) and carbonyl (C ¼ O) groups, respectively. Moreover, a broad singlet at 11.40 ppm assignable to the exchangeable OH group was seen in the 1 H NMR spectrum, and the 13 C NMR spectrum showed the presence of a signal for the carbonyl group at 183.20 ppm.

Biological evaluation
Antitumour evaluation using MTT assay The designed compounds were evaluated for their in vitro antitumour effects via the standard 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) method against a panel of four human tumour cell lines; namely, hepatocellular carcinoma Scheme 2. Synthesis of the designed pyrazoline, isoxazoline, cyanopyridine, and quinoline-4-carboxylic acid derivatives. Antioxidant activity using ABTS þ radical-scavenging assay The assay is based on measuring the ability of the tested compounds to scavenge the long-life radical cation of ABTS 22,43,64,65 . In this study, all the newly synthesized compounds 4-12 and Lascorbic acid, as a positive control, were evaluated and showed considerable free radical-scavenging activities. The reduction in colour intensity was expressed as inhibition percentage of the ABTS þ as shown in Table 2. From the listed results, we concluded that all tested compounds exhibited more than 50% inhibition of the ABTS radical cation except derivatives 5a, 10, 11, and 12a. It is clear that the conversion of a,b-unsaturated ketones 4a,b (% inhibition ¼52%) to the corresponding heterocyclic molecules generally led to sharp increases in antioxidant effects. Among them, the N-phenylpyrazoline derivative 8b displayed the highest free radical-trapping properties, with 88.5% inhibition, which was comparable to L-ascorbic acid at 90.0%. Moreover, 4-H pyran 7 and isoxazoline 9 derivatives showed inhibition of 75.8% and 60.0%, respectively. Conversion of the acyclic a,b-unsaturated ketones 4a,b (52% inhibition) into their corresponding cyclic a,b-unsaturated ketones 5a-c (51% inhibition) and 6a,b (54-55% inhibition) showed no change in activity. However, we concluded that compounds characterised by having pyrane 7, pyrazoline 8b, and isoxazoline 9 ring systems were among the most active compounds (60-88.5% inhibition), indicating that these core structures may play a role in trapping ABTS free radicals.

Correlations between antioxidant and antitumour activities
The correlation between the antioxidant and the antitumour activities was investigated using SigmaPlot software (London, UK) 73 . The overall correlation between the antioxidant and antitumour activities of the synthesized compounds against individual cancer cell lines is shown in Figure 2. Most of the synthesized compounds showed moderate correlation (a moderate uphill relationship) between antioxidant and antitumour activities, as indicated by their coefficients of determination (R 2 ). These   activities, which lead to the conclusion that antioxidant activity is not the only mechanism responsible for antitumour activity.

EGFR inhibitory activity
The antitumour activity results of compounds 6a, 6b, 7, and 8b encourage us to study the mechanism of antitumour activity using ELISA-based EGFR-TK assay with sorafenib as the reference drug 66 . The % inhibition and IC 50 values of the tested compounds were calculated and are listed in Table 3 . We concluded, based on these results, that the designed compounds such as 6a, 6b, 7, and 8b are EGFR inhibitors which could be a new scaffold for the design of future analogues.

Molecular docking results
The preceding results encouraged us to study the molecular docking of the most active compounds 6b, 7, and 8b using EGFR, which are overexpressed in numerous tumours such as prostate (PC-3), breast (MCF-7), hepatocellular carcinoma (HepG2), and human cervical (HeLa) cancer cell lines [24][25][26][27][28][29][30][31][32] . All docking calculations were performed using MOE 2008.10 software 67 . The docked compounds 6b, 7, 8b, and the reference inhibitor erlotinib (Protein Data Bank [PDB] code 1M17) 33 into the putative active site of EGFR are shown in Figure 3. The molecular modelling results of the compound, 6b, demonstrated an approximate orientation of the molecule in comparison with erlotinib inside the putative binding site of receptor pocket with some additional hydrogen bond interactions with surrounding amino acids. These docking results showed three classical and five non-classical hydrogen bonds, where the distinctive residue Thr 766 formed bifurcated hydrogen bonds with oxygen and carbon atoms of the piperidin-4-one ring system (Figure 3, middle left panel). In addition, the amino acid residue Thr 830 formed bifurcated hydrogen    Similarly, compound 7 binds into the putative active site of EGFR with three classical and one non-classical hydrogen bond. It was found that the amino acid Thr 766 formed bifurcated classical hydrogen bonds with the 2-amino moiety and the oxygen atom of the 4-H pyran ring system (Figure 3, middle right panel). Moreover, the distinctive amino acid residue Met 769 was involved in two hydrogen bonds: with the oxygen atom and with alkyl moieties of the ester group.
Moreover, compound, 8b, demonstrated similar results as compounds 6b and 7 inside the putative binding site of receptor pocket. These docking results showed two classical hydrogen bonds, where the distinctive residue Thr 766 formed bifurcated classical hydrogen bonds with nitrogen atoms of the pyrazoline ring system (Figure 3, lower panel). In addition, three non-classical hydrogen bonds formed with surrounding amino acids, as shown in Figure 3 (lower panel). The amino acid residue Leu 768 formed bifurcated hydrogen bonds through NH-Ar-CH interaction and one with the methyl group of the 4-tolyl moiety (NH-aliphatic-CH), while the third non-classical hydrogen bond was observed between the amino acid Thr 830 and an aromatic ring through the OH-Ar-CH interaction. Additionally, the surrounding amino acids Leu 768 , Leu 820 , and Thr 766 showed hydrophobic interactions with aromatic rings through CH-p and OH-p ( Figure 3, lower panel).

Conclusions
Novel a,b-unsaturated ketone 4-6a,b, 4-H pyran 7, pyrazoline 8a,b, isoxazoline 9, pyridine 10-11, and quinoline-4-carboxylic acid 12a,b derivatives have been synthesized, and the antitumour, antioxidant, and EGFR kinase inhibition activities have been evaluated. It is clear that most of the synthesized compounds exert significant antitumour activities. Among the tested derivatives, 6a, 6b, 7, and 8b showed potent IC 50 values ffi 5.5-18.1 mM, which were comparable to that of 5-FU (IC 50 ffi 4.8-8.3 mM) and afatinib (IC 50 ffi 5.4-7.6 mM). Moreover, compound 8b has been shown promising, broad spectrum antitumour activity against the tested cell lines with an IC 50 range of 5.5-7.8 mM. Additionally, compounds 6a, 6b, 7, 8b, and 9 exhibited the highest antioxidant effects using the ABTS radical-scavenging assay. Moreover, we observed a moderate relationship between the antitumour activity and the antioxidant effects of the tested compounds, which suggested that antioxidant effect is not the major role in the antitumour activity. Additionally, compounds 6b, and 7 exhibited excellent inhibition towards EGFR kinase enzyme with IC 50 values range of 0.56-1.6 mM, respectively, while compounds 6a and 8b have good activity with IC 50 ¼ 4.66 and 2.16 mM, respectively, compared with the reference drug sorafenib (IC 50 ¼ 1.28 mM). Molecular docking studies were conducted for compounds 6b, 7, and 8b into putative binding sites of EGFR kinase enzyme, which showed similar binding modes to erlotinib (EGFR kinase inhibitor).

Disclosure statement
No potential conflict of interest was reported by the authors.