Design and synthesis of novel quinazolinone-based derivatives as EGFR inhibitors with antitumor activity

Abstract Nineteen new quinazolin-4(3H)-one derivatives 3a–g and 6a–l were designed and synthesised to inhibit EGFR. The antiproliferative activity of the synthesised compounds was tested in vitro against 60 different human cell lines. The most potent compound 6d displayed superior sub-micromolar antiproliferative activity towards NSC lung cancer cell line NCI-H460 with GI50 = 0.789 µM. It also showed potent cytostatic activity against 40 different cancer cell lines (TGI range: 2.59–9.55 µM). Compound 6d potently inhibited EGFR with IC50 = 0.069 ± 0.004 µM in comparison to erlotinib with IC50 value of 0.045 ± 0.003 µM. Compound 6d showed 16.74-fold increase in total apoptosis and caused cell cycle arrest at G1/S phase in breast cancer HS 578T cell line. Moreover, the most potent derivatives were docked into the EGFR active site to determine their binding mode and confirm their ability to satisfy the pharmacophoric features required for EGFR inhibition.


Introduction
Cancer is a widespread and lethal noncommunicable disease with rapidly growing incidence and mortality worldwide 1 . Breast cancer is the leading cause of death among women with a reported incidence of approximately 2.3 million and 685,000 deaths worldwide in 2020 2 .
Molecular targeting therapy approach, that targets a crucial cancer-related enzyme or receptor, increases the tumour specificity and decreases the side effects [3][4][5] . The epidermal growth factor receptor (EGFR) is mainly involved in the growth factor signalling. Abnormal signalling and overexpression of this receptor enhance downstream effects such as cell survival, cell proliferation and angiogenesis which lead to uncontrolled cell proliferation and metastasis, which ultimately promote tumour growth (Figure 1) 6 . The EGFR is one of the members of ErbB tyrosine kinase receptors family 7 , and consists of two domains; an extracellular receptor domain connected via a transmembrane region to an intracellular domain with tyrosine kinase function 8 . Inhibition of EGFR by tyrosine kinase inhibitors (TKIs) delays these downstream effects and lead to inhibition of tumour growth. Many breast cancers express 2 Â 10 6 EGFR molecules per cell which is more than 20-fold the expression of EGFR in normal cells 9,10 . This overexpression of EGFR in breast cancer as well as other cancer cells including colon and lung cancers made this a potential molecular target for inhibition [11][12][13][14] .
Three generations of small-molecule EGFR TKIs have been developed so far 15,16 . The first-generation, such as gefitinib and erlotinib usually with 4-anilinoquinazoline motifs, achieves initial good responses in the treatment of cancer patients with overexpression of EGFR but unfortunately, resistance to this class was acquired by most patients within 1 year 17 . The second-generation of EGFR TKIs, such as neratinib and afatinib, were developed to overcome the acquired resistance 18,19 , however their associated side effects 20,21 lead to the design of a novel class of third-generation EGFR TKIs, including osimertinib, rociletinib and olmutinib. This class has been designed and developed to overcome resistance and reduce side effects 22 . The aim of this study is to design, synthesise and evaluate compounds with the same pharmacophoric features as the third generation EGFR TKIs.
In previous studies in our lab, we synthesised different TKIs with high therapeutic index [23][24][25][26][27] . In the present study, the design of the new compounds started by finding common pharmacophoric structural features of the newer generations of EGFR TKI without including the acrylamide moiety. The acrylamide moiety in the second and the third generations gives strong covalent binding to the SH group in Cys797 residue of the receptor. However, these compounds are prone to drug resistance when this irreversible covalent interaction is lost due to C797S mutation 28,29 . The common structural features were found to be (A) an upper aryl ring attached to (B) 2-aminopyrimidines or 2-amino arylfusedpyrimidines and the 2-amino is attached to either C) phenoxy group or piperazinyl aryl group ( Figure 2). The design of the novel compounds was done to mimic these structural features of the third generation EGFR TKIs to have 2-aminoarylquinazoline linked to an upper aryl ring and the 2-amino is attached to either substituted phenoxy group (3a-d) or piperazinyl aryl (or its isostere) (3e-g). We also designed congener compounds with a 2-thioacetyl spacer between the pyrimidine ring and either phenoxy group (6a-f) or piperazinyl aryl (6g-l) groups to investigate the effect of this spacer on the activity of the compounds ( Figure  2). The designed compounds were synthesised and subjected to in vitro screening of their cytotoxic activity against a panel of 60 different cancer cell lines. The compounds were investigated for their ability to inhibit EGFR, their effect on cancer cell cycle and their ability to induce apoptosis. Molecular modelling studies were done as well to rationalise the biological activity of the synthesised compounds.

General
Melting points were obtained on a Griffin apparatus and were uncorrected. Microanalyses for C, H and N were carried out at the Regional Centre for Mycology and Biotechnology, Faculty of Pharmacy, Al-Azhar University. IR spectra were recorded on Shimadzu IR 435 spectrophotometer (Shimadzu Corp., Kyoto, Japan) Faculty of Pharmacy, Cairo University, Cairo, Egypt, and values were represented in cm À1 . 1 H NMR spectra were carried out on Bruker 400 MHz (Bruker Corp., Billerica, MA, USA) spectrophotometer, Faculty of Pharmacy, Cairo University, Cairo, Egypt. The chemical shifts were recorded in ppm on d scale, coupling constants (J) were given in Hz and peak multiplicities are designed as follows: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; m, multiplet 13 C NMR spectra were carried out on Bruker 100 MHz spectrophotometer, Faculty of Pharmacy, Cairo University, Cairo, Egypt. Mass spectra were recorded with Advion expression V R CMS, Nawah Scientific, Cairo, Egypt. Progress of the reactions were monitored by TLC using precoated aluminium sheet silica gel MERCK 60 F 254 and was visualised by UV lamp. The original NMR spectra of the investigated compounds are provided as supporting information. Compounds 1a,b 30,31 , 2a,b 31 , 4a,b 32,33 and 5a,b 32,33 were prepared as reported.

Biological assays
The biological assays were carried out according to the previously reported procedures and have been provided in the Supplementary Materials; antiproliferative activity screening and five dose concentrations assay by NCI [34][35][36][37] , in vitro EGFR inhibitory assay 38 , cell cycle analysis 39 , apoptosis assay 40 , caspase-3 enzyme assay 41 .

Molecular modelling studies
The crystallographic structure of EGFR protein (PDB: 1M17) was obtained from the protein data bank website, (http://www.pdb.org) with resolution of 2.60 Å. All the molecular modelling studies were carried out using Molecular Operating Environment (MOE 2020.09; Chemical Computing Group, Canada) as the computational software. The hydrogen atoms were added, the protonation states of the amino acid residues were assigned, and the partial charges of atoms Scheme 1. Synthesis of quinazolin-4(3H)-one derivatives 3a-g and 6a-l. Reagents and conditions: (i) NH 2  were added using Protonate 3D algorithm. Compounds were modelled using MOE builder, and the structure was energy minimised using the MMFF94x force field. Using the MOE induced-fit Dock tool, docking studies of the synthesised compound into the active site was done and the final docked complexes of ligand-enzyme was selected according to the criteria of binding energy score combined with geometrical matching quality.

Statistical analysis
Data are represented as mean ± SD. Significant differences between groups were analysed by using Graphpad Prism 9.1.0. Differences were considered significant at p < 0.05.
On the other hand, alkylation of 2-thioxo-2,3-dihydroquinazolin-4(1H)-one derivatives 1a,b with ethyl bromoacetate at room temperature in dry acetone containing anhydrous potassium carbonate afforded ethyl 2-((4-oxo-3-aryl-3,4-dihydroquinazolin-2yl)thio)acetate 4a,b 32,33 . The ester derivatives 4a,b were further reacted with hydrazine hydrate 99% to afford the corresponding acetohydrazide derivatives 2-((4-oxo-3-aryl-3,4-dihydroquinazolin-2-yl)thio)acetohydrazide 5a,b 32,33 . Compounds 5a,b were finally reacted with 4-substituted benzaldehyde derivatives to yield the desired hydrazone derivatives 6a-l. The 1 H NMR spectra of derivatives 6a-l revealed that the hydrazones existed in a keto amide/ enol amide mixture (Figure 3) 43 . 1 H NMR spectra of series 6a-l showed two singlet signals with an integration of 2H due to SCH 2 C¼O protons at the ranges of d 4.45-4.47 ppm and d 3.97-4.01 ppm corresponding to keto amide and enol amide tautomers, respectively. Moreover, two singlet signals with an integration of 1H in 1 H NMR spectra of 6a-l appeared due to the azomethine protons of both tautomers at the ranges of d 8.07-8.24 and d 7.90-8.04 ppm. Additionally, two D 2 O exchangeable peaks due to OH proton in enol and NH in keto tautomers with 1H total integration appeared at the ranges of d 11.49-11.73 ppm and d 11.39-11.64 ppm, respectively. On the other hand, the appearance of the OCH 3 group in derivatives 6c and 6f as two singlet signals with total integration of 3H at d 3.77-3.78 ppm and d 3.76-3.77 ppm is in concordance with the presence of tautomerism. 13 C NMR spectra of derivatives 6a-l showed two peaks at the ranges d 35.8-35.9 ppm and d 34.6-34.9 ppm corresponding to SCH 2 C¼O in enol and keto tautomers, respectively.

Evaluation of antiproliferative activity against a panel of 60 human cancer cell lines
Two series of substituted quinazolinone derivatives 3a-g and 6a-l were evaluated for their antiproliferative activity at a single dose (10 lM) using 60 human cancer cell lines, by the National Cancer Institute (NCI), USA. The screening was achieved under the Developmental Therapeutic Program (DTP) [34][35][36][37] . NCI cell lines include leukaemia, melanoma, and cancers of the breast, kidney, ovarian, colon, central nervous system (CNS), prostate, and nonsmall cell (NSC) lung. The growth inhibition percentage (GI%) representing the in vitro antiproliferative activity was illustrated in Table 1.
Derivative 6j with N'-(4-(piperidin-1-yl)benzylidene)-3-(4-chlorophenyl)-quinazolinone acetohydrazide showed antiproliferative activity with mean GI% of 34.19. It showed significantly potent antiproliferative activity against 5 cell lines including NSC lung cancer (HOP-62 and NCI-H460); renal cancer (786-0 and ACHN) and prostate DU-145 with GI% 81.63%, 90.26%, 92.8%, 82.62% and 70.89% respectively. It showed moderate antiproliferative activity against 23 cell lines with GI% range of 69.91-32.2%. Finally, derivatives 6 h, 6i, 6k and 6 l did not show activity against all cell lines. The most remarkable effects to be concluded in the structure variations of the synthesised derivatives and their antiproliferative activity are represented in Figure 4. The incorporation of thioacetyl linker at position 2 of the quinazolinone ring was found to be essential for the antiproliferative activity as most of the synthesised compounds with thioacetyl bridge in series 6a-l were found to have of high anticancer activity while compounds 3a-g, lacking the thioacetyl bridge, showed no promising results. This may be due to the increased flexibility introduced by thioacetyl spacer, which enables the compounds to fit better into the receptor 44 .
In series 6a-l, general better antiproliferative activity were displayed by compounds bearing chloro atom at the para-position of the quinazolinone 3-phenyl ring as demonstrated by the relatively high GI percentage values of compounds 6d (mean GI% ¼ 80.23%), 6f (mean GI% ¼ 42.46%) and 6j (mean GI% ¼ 34.19%) over their unsubstituted analogues 6a (mean GI% ¼ 37.29%), 6c (mean GI% ¼ 0%) and 6 g (mean GI% ¼ 4.28%), respectively. This observation was in accordance with the previously reported study in other related compounds 45 . It was also observed that the substituted phenoxy derivatives 6a-f have generally better GI% values than the corresponding piperazinyl (or its isostere) aryl group 6g-l.

Evaluation of in vitro antiproliferative activity of compound 6d at 5 dose concentrations
Compound 6d was found to be the most effective anticancer agent in this study based on preliminary single dose (10 lM) screening results. It demonstrated promising efficacy against a variety of cancer cell lines, with mean growth inhibition of 80.23% (Table 1). Accordingly, compound 6d was further subjected to five dose concentrations assay (0.01, 0.1, 1, 10 and   100 lM) to detect its dose-response behaviour and calculate the values of GI 50 (the dose that inhibit 50% of cell growth in comparison to control), TGI (the dose that completely inhibit growth), and LC 50 (the dose that kill 50% of the cells). The values are presented in Table 2. The obtained results showed that, compound 6d displayed superior sub-micromolar activity towards NSC lung cancer cell line NCI-H460 with GI 50 ¼ 0.789 mM, however it showed high lethality with this cell line (LC 50 ¼ 7.78 mM). Compound 6d exhibited also potent and broad-spectrum antiproliferative activity against most of the tested cancer cell lines, with GI 50 values in the range of 1.29-5.97 mM, in addition to moderate antiproliferative activity against leukaemia cell line HL-60(TB) with GI 50 ¼ 16.0 mM. Compound 6d also exhibited high cytostatic activity (TGI range: 2.59-9.55 mM) against 40 cancer cell lines. Moreover, 6d demonstrated good to moderate cytostatic activity against the rest of cell lines with TGI range of 11.1-48.5 mM. Compound 6d exhibited remarkable differences between its cytotoxic indicator (LC 50 ) and cytostatic markers (GI 50 and TGI) against colon cancer HT29; ovarian cancer OVCAR-8; prostate cancer PC-3 and breast cancer HS 578T, which indicates a wide therapeutic index.

In vitro EGFR kinase inhibitory assay
The most potent compounds 6a, 6b, 6d, 6f and 6j, with promising antiproliferative activity with mean inhibition range of 34.19-80.23%, were assayed for their EGFR inhibition. The results were summarised in Table 3 and Figure 5 as 50% inhibition concentration value (IC 50 ) calculated from the concentration inhibition response curve. In this assay, erlotinib is used as positive control. Compound 6d potently inhibited EGFR with IC 50 ¼ 0.069 ± 0.004 mM in comparison to erlotinib with IC 50 value of 0.045 ± 0.003 mM. It was obvious that there was non-significant difference between IC 50 values scored by both compound 6d and erlotinib.     Compound 6f with IC 50 ¼ 0.133 ± 0.008 mM, showed nearly half the potency of EGFR inhibition compared to compound 6d. Moreover, compound 6b was nearly equipotent in EGFR inhibition with compound 6j with IC 50 values of 0.487 ± 0.030 and 0.478 ± 0.029 mM, respectively. Compound 6d, showed nearly three times the potency of EGFR inhibition compared to compound 6a with IC 50 ¼ 0.202 ± 0.012 mM. The antiproliferative activity of the compounds correlated with their high ability to inhibit EGFR. These biological results indicate that compound 6d is a promising antiproliferative agent with EGFR inhibitory activity.

Cell cycle analysis
Most cytotoxic drugs exert their antiproliferative activity by disturbing some checkpoints in the cell cycle. These checkpoints are distinct stages in the cell cycle whose disruption causes growth termination 46 . The cell cycle distribution was measured by DNA flow cytometric analysis, upon incubation of HS578T cells treated with compound 6d for 24 h at its IC 50 concentration (2.17 mM).
The effect of compound 6d on the cell population in different cell phases was recorded and presented in Figure 6. The results were compared to the cell cycle analysis of breast cancer HS 578 T cells as untreated control. The proportions of cells in both the G0-G1 phase and the S phase were significantly increased by 1.12-fold. Moreover, significant decrease in the cell population occurred at the G2/M phase with 52%. In comparison to control, the cell population in the pre-G1 phase significantly increased by 16.74fold. These results showed that, compound 6d caused induction of cell cycle arrest in breast cancer HS 578 T cells at the G1/S phase.

Apoptosis assay
Annexin-V flow cytometry assay is an effective technique to determine apoptosis and for discrimination of programmed apoptosis and non-specific necrosis 26,47 . Annexin V is a protein with high affinity for phosphatidylserine (PS), a cell membrane component that translocate from the plasma membrane to the surface of the cell during apoptosis. Annexin-V and propidium iodide (PI) dual staining allows for the differentiation of live cells, early/late apoptotic cells, and necrotic cells. Fluorescent Annexin V conjugate can be used to identify PS on the cell surface. PI only enters dead cells and stains DNA 48 . The effect of compound 6d on cell apoptosis was examined in breast cancer HS 578T cell line using a dual staining assay. The percentage of apoptotic cells increased significantly in early (from 0.37% to 19.24%) and late (from 0.22% to 13.45%) apoptotic phases (Table 4 and Figure 7). Compound 6d increased total apoptosis significantly by 16.74-fold compared to the control.

Evaluation of caspase-3 activation
Apoptosis is triggered by the activation of caspases, particularly caspase-3, which is an initiator caspase responsible for apoptosis 49 . Therefore, the induction of apoptosis by compound 6d was examined by the evaluation of caspase-3 activation in breast HS 578T cancer cell line. The effect of erlotinib as well as untreated cell line was used as positive and negative control, respectively. Compound 6d and erlotinib significantly increased caspase-3 level by 10.75-and 8.78-fold, respectively when compared to negative control Figure 8. These results revealed that  compound 6d might induce apoptosis via a caspase-dependent mechanism.

Molecular docking study
Docking study was carried out for the most potent synthesised compounds into the EGFR protein complexed with erlotinib (PDB: 1M17) 50 using MOE (2022.09) software 51 . The aim of this study was to investigate the interactions between the EGFR-TK and the highly active derivatives, 6a, 6b, 6d, 6f and 6j. The study began by redocking the co-crystallised ligand, erlotinib, starting from a two-dimensional structure and using the same protocol for preparation and analysis. The most active derivatives were then docked, and the resulted scores and the binding interactions are included along with the Lipinski parameters in Table 5. Erlotinib fits into the gorge of the active site of EGFR-TK and binds to Cys751 via a water bridge, the N1 atom of the erlotinib accepts a H-bond from the Met769 amide nitrogen. Erlotinib C-2 hydrogen is in close proximity and interacts with Gln767. The anilino ring is coplanar with the quinazoline ring and shows arene-cation interaction with Lys721 ( Figure 9). Compounds 6a, 6b, 6d, 6f and 6j parameters were aligned with the Lipinski parameters, except for the logP parameter. Generally, compounds 6a, 6b, 6d, 6f and 6j were able to overlay as erlotinib with the quinazolinone ring inhabiting the same space of the quinazoline ring of erlotinib. The docking simulation showed that the least active compound 6b is less binding to the active site with only 2 interactions: [HB (Cys751) and Arene-H (Val702)], while erlotinib has 4 interactions: [HB (Cys751), HB (Gln767), HB (Met769) and Arene-cation (Lys721)]. On the other hand, the docking of the most active compound 6d ( Figure 10) showed a comparable binding profile with the active site as the co-crystalized drug (erlotinib). Compound 6d forms hydrogen bond with Cys751 via the water bridge and H-p interactions with Val702 which help stabilising the quinazolinone ring conformation. The p-chloro phenyl was able to overlay with the anilino ring of erlotinib and also interacts by H-p stacking with Lys721. The chlorine atom forms halogen bond 52 with the carboxylate group of Glu738 which might be the rationale behind the higher in vitro EGFR enzymatic inhibition of 6d when compared to its dechlorinated congener 6a. (Docking poses of compounds 6a, 6b, 6f and 6j are included in supplementary data)

Conclusion
Nineteen quinazolin-4-one derivatives were synthesised to mimic the third generation EGFR TKIs. The antiproliferative activity were assessed for all the prepared compounds. The results showed the significance of the thioacetyl linker for the antiproliferative activity. The compounds showed promising results with 6d exhibiting high inhibition of EGFR with IC 50 ¼ 0.069 ± 0.004 mM which is comparable with the positive control. In addition, 6d demonstrated an excellent in vitro antiproliferative activity with mean growth inhibition of 80.23% and was chosen by the National Cancer Institute (NCI), Maryland, USA for further investigation in which compound 6d exhibited potent and broad-spectrum antiproliferative activity against most of the tested cancer cell lines, with GI 50 values in the range of 1.29-5.97 mM. DNA-flow cytometric analysis demonstrated the inhibition of cell proliferation by compound 6d at G1/S phase. Docking study showed that compound 6d interacts similarly to erlotinib and highlights the role of the halogen bond in enhancing the binding to the EGFR binding site for further developments.