Molecular dynamics simulation and 3D-pharmacophore analysis of new quinoline-based analogues with dual potential against EGFR and VEGFR-2

Article abstract of DOI:10.1016/j.ijbiomac.2019.09.077

Epidermal growth factor and vascular endothelial growth factor-2 are important targets of tyrosine kinase for the treatment of various cancerous diseases. Combination of inhibition of both targets to produce synergy in the signal pathway is a critical approach to identify novel tyrosine kinase inhibitors. In this study, a series of new compounds derived from the 4-aminoquinoline as dual inhibitors were synthesized. The obtained results of cytotoxicity assay against human carcinoma cell lines indicated 0.8 μM for 4c against A549 showing its high efficiency in comparison to erlotinib. Pharmacophore modeling as a structure-based method was investigated on dual inhibitors and 4c which was compared with co-crystallized in the active site of EGFR and VEGFR-2. They have shown the same binding orientation as vandetanib, erlotinib and sorafenib. Molecular dynamics simulation results approved that Met769, Lys721, Asp1046, and Lys868 are key residues in two binding sites for dual activity. Ala1050 and Pro968 were identified as new amino acid interaction sites for dual inhibition. 4c showed more favorable stability than vandetanib in VEGFR-2 receptor for a 50 ns dynamic simulation. The high correlation between essential pharmacophoric features of designed compounds and lead inhibitors interactions provided a deeper insight into the structural basis of 4-aminoquinoline inhibition.

Full text of DOI:10.1016/j.ijbiomac.2019.09.077

Journal Pre-proof
Molecular dynamics simulation and 3D-pharmacophore analysis
of new quinoline-based analogues with dual potential against
EGFR and VEGFR-2
Neda Fayyazi, Afshin Fassihi, Somayeh Esmaeili, Salman Taheri,
Jahan B. Ghasemi, Lotfollah Saghaei
PII:
S0141-8130(19)33964-9
DOI:
https://doi.org/10.1016/j.ijbiomac.2019.09.077
BIOMAC 13322
Reference:
To appear in:
International Journal of Biological Macromolecules
Received date:
Revised date:
Accepted date:
28 May 2019
3 August 2019
10 September 2019
Please cite this article as: N. Fayyazi, A. Fassihi, S. Esmaeili, et al., Molecular dynamics
simulation and 3D-pharmacophore analysis of new quinoline-based analogues with dual
potential against EGFR and VEGFR-2, International Journal of Biological
Macromolecules(2019), https://doi.org/10.1016/j.ijbiomac.2019.09.077
This is a PDF file of an article that has undergone enhancements after acceptance, such
as the addition of a cover page and metadata, and formatting for readability, but it is
not yet the definitive version of record. This version will undergo additional copyediting,
typesetting and review before it is published in its final form, but we are providing this
version to give early visibility of the article. Please note that, during the production
process, errors may be discovered which could affect the content, and all legal disclaimers
that apply to the journal pertain.
©
2019 Published by Elsevier.

Journal Pre-proof
Molecular dynamics simulation and 3D-pharmacophore analysis of new
quinoline-based analogues with dual potential against EGFR and VEGFR-2
a
a
c
d
b,
Neda Fayyazi , Afshin Fassihi , Somayeh Esmaeili , Salman Taheri , Jahan B. Ghasemi* and
Lotfollah Saghaei*a
a
Department of Medicinal Chemistry, School of Pharmacy and Pharmaceutical Sciences, Isfahan, Iran.
b
College of Sciences, Faculty of Chemistry, University of Tehran, Iran.
c
Traditional Medicine and Materia Medica Research center(TMRC), School of Traditional Medicine, Shahid Beheshti
University of Medical Sciences, Tehran, Iran
d
Chemistry and Chemical Engineering Research Center of Iran, Tehran, Iran.
Highlights



Novel 4-aminoquinoline derivatives as EGFR and VEGFR-2 dual inhibitors are presented.
Molecular modeling has confirmed the dual inhibitory property of derivatives.
New concepts for interaction modes of dual inhibitors were identified by computational and
experimental approaches.

Dual inhibitors are essential significant anti-proliferative activities in NSCLC.
1

Journal Pre-proof
Abstract
Epidermal growth factor and vascular endothelial growth factor-2 are important targets of tyrosine kinase for
the treatment of various cancerous diseases. Combination of inhibition of both targets to produce synergy in
the signal pathway is a critical approach to identify novel tyrosine kinase inhibitors. In this study, a series of
new compounds derived from the 4-aminoquinoline as dual inhibitors were synthesized. The obtained
results of cytotoxicity assay against human carcinoma cell lines indicated 0.8 µM for 4c against A549
showing its high efficiency in comparison to erlotinib. Pharmacophore modeling as a structure-based
method was investigated on dual inhibitors and 4c which was compared with co-crystallized in the active
site of EGFR and VEGFR-2. They have shown the same binding orientation as vandetanib, erlotinib and
sorafenib. Molecular dynamics simulation results approved that Met769, Lys721, Asp1046, and Lys868 are
key residues in two binding sites for dual activity. Ala1050 and Pro968 were identified as new amino acid
interaction sites for dual inhibition. 4c showed more favorable stability than vandetanib in VEGFR-2
receptor for a 50 ns dynamic simulation. The high correlation between essential pharmacophoric features of
designed compounds and lead inhibitors interactions provided a deeper insight into the structural basis of 4-
aminoquinoline inhibition.
Keyword: Growth factor receptors; Dual inhibition; Molecular dynamics simulation
2

Journal Pre-proof
Introduction:
Epidermal growth factor receptor (EGFR), a transmembrane glycoprotein is classified as an
important enzyme with intrinsic intracellular protein-tyrosine kinase activity [1]. This protein
catalyzes the transfer phosphate group from an ATP molecule to a tyrosine residue of the target
protein that leads to signal transduction. Activation of EGFR causes several signal transductions in
the normal process of DNA synthesis and cell proliferation [2, 3]. Vascular endothelial growth
factor receptor (VEGFR) is another protein with tyrosine kinases activity; thus, another member of
the signal transduction pathway in the process of cell proliferation. When DNA is mutated,
abnormal proliferation happens, which leads to a state which is called cancer [4, 5]. Mutations in
EGFR are shown as oncogenic reasons in different cancerous cell lines [6, 7]. To another
considerable interest in tumor cell growth, members of the EGFR pathway can contribute to the
expression of VEGF [8]. Recent investigations imply that the gene mutations of EGFR lead to
overexpression of proteins which regulate activates of several downstream pathways in an
undefined manner. These pathways increase diverse malignancies, specifically non-small cell lung
cancer (NSCLC) [9]. The NSCLC has reported as 75% of lung cancers and is the main reason for
malignant disease-associated deaths in the world. VEGFR is considered to play a critical role in the
process of tumor angiogenesis. Thus, both EGFR and VEGFR are regarded as targets in combating
such abnormal proliferation and in anti-angiogenic strategies exploited in cancer therapy. Recently
drugs, including afatinib, erlotinib, gefitinib, brigatinib and icotinib are used as efficient anticancer
agents for lung cancer exerting their therapeutic effects by inhibiting epidermal growth factor.
Among the available drugs, first-generation quinazoline-based ATP competitive inhibitors like
erlotinib have shown to be effective in NSCLC patients and act against L858R and G719S
mutations of EGFR [10, 11]. The development of inhibitory compounds with combined
3

Journal Pre-proof
EGF/VEGF receptor in the same molecule is, therefore, an opportunity to increase the antitumor
efficacy and broaden application possibilities [12, 13].
Using computational approaches in the design of multi-target directed ligands (MTDLs)
have been developed recently [14]. Techniques such as molecular docking and molecular dynamics
simulations and structure-based pharmacophore screening have been used as rational strategies for
the design of multi-target directed ligands. Multi-target therapy is more effective than traditional
drug therapy [15]. Several investigations describing new molecular design based on the erlotinib
and pelitinib with the core structure of 4-aminoquinazoline and 4-aminoquinoline have been
reported. 2-Substituted-4-amino-6-halogenquinoline was shown to have anti-proliferative activity
against different cancer cell lines in vitro. The structure-activity relationship (SAR) studies have
shown that the variation in activity depends on the introduction of an ethylene linkage into the
4
-aminoquinazoline and 4-aminoquinoline scaffolds [16, 17].
VEGFR-2 a member of the VEGFR family appears to mediate almost all of the known
cellular responses to vascular epidermal growth factor [18]. Small-molecule inhibitors of this
receptor were developed for anticancer therapy by many pharmaceutical companies. Different
scaffolds such as benzimidazolyl and indolyl quinoline have been reported as VEGFR-2 inhibitors.
VEGFR-2, along with EGFR, are both important targets for cancer therapy. Simultaneous
inhibition of both EGFR and VEGFR-2 signaling pathways represents a synergistic effect in cancer
therapy [19-21]. Combined EGFR and VEGFR-2 inhibition are being evaluated in several
preclinical and clinical trials. Vandetanib, as a dual inhibitor, represents the benefits of this novel
therapeutic approach.
In this study, a series of EGFR and VEGFR-2 dual inhibitors with good anti-proliferative
activities were selected for pharmacophore generation. Based on the resulted pharmacophore, six
4

Journal Pre-proof
possible inhibitors were primarily designed. Considering the structures of the successfully reported
inhibitors, the designed compounds were decorated to reach three 4-aminoquinoline derivatives as
combined EGFR and VEGFR-2 inhibitors. They were synthesized then evaluated as cytotoxic
agents against A549 a non-small cell lung cancer (NSCLC), MCF7 (human breast cancer), HT29
(human colon adenocarcinoma) and AGO 1522 (non-cancerous cell line). Then other in silico
methods, molecular docking and molecular dynamics simulation, were exploited to show the details
of interactions of the best compound inside the EGFR and VEGFR-2 binding sites. Sorafenib as a
specific inhibitor of VEGFR-2, erlotinib as a selected EGFR inhibitor and vandetanib as a dual
inhibitor of both receptors was also involved in these studies. To have a knowledge of the drug-
likeness of the compounds pharmacokinetic properties of the designed dual inhibitors were also
predicted before preparation of the new compounds in the lab. Adsorption, distribution,
metabolism, excretion, and toxicity (ADMET) as the major pharmacokinetic features were the
drug-likeness criteria determined herein. The other biological analyses such as
immunocytochemistry (ICC) and apoptotic cells by annexin V/PI staining methods were carried out
for further confirmation of the theoretical results.
2
. Material and Methods
2
2
.1. Experimental section
.1.1.Chemistry
NMR spectra were recorded on a Bruker 500 MHz spectrometer (Bruker Bioscience, Billerica,
MA, USA). All melting points were obtained on a Büchi Melting Point B-540 apparatus (Büchi
Labor technic, Flawil, Switzerland). Mass spectra (MS) were taken in ESI mode using a Shimadzu
LC-MS 2010 spectrometer, FT-IR spectra were acquired on an Agilent spectrometer, using pellets.
Thin layer and column chromatography were carried out on silica gel plates GF254 (CAMAG,
5

Journal Pre-proof
Germany), and on Sephadex LH20 (Fluka, Germany) and silica gel 60 (70-230 mesh; Merck,
Germany), respectively. All chemicals were obtained from commercial suppliers (Merck, Sigma-
Aldrich, Fluka) and used without purification unless otherwise specified. Yields were not
optimized. General procedure for the preparation of the compounds is provided in Scheme 1.
[
Scheme 1. Near Here]
.1.1.1. (Chloromethyl)-5-hydroxy-4Hpyran-4-one (1)
This compound was prepared according to the procedure previously reported by Dehkordi et al.
22]
.1.1.2. 5-(Benzyloxy)-2-(hydroxymethyl)-4H-pyran-4-one (2a)
This compound was prepared according to the procedure previously reported by Dehkordi et al.
22,33]
MP:165-166°C ,Yield:74.2%, appearance:Light yellow needle crystals
2
[
2
[
2
.1.1.2. 5-Benzyloxy -2-piperidinomethylpyran-4(1H)-one (2b)
A solution of piperidine and triethylamine in N, N dimethylformamide (25 mL) was added to 2.2 g
of 2-(chloromethyl)-5-benzyloxy-4H pyran-4 one. The reaction mixture was allowed to stir at room
temperature for 48 hours. The mixture was diluted with water (200mL) and extracted with
dichloromethane. The residue was purified by column chromatography (methanol, dichloromethane
2
0:80) to afford the desired product.
1
MP: 90-92 °C, HNMR (500 MHz, DMSO-d
6
): δ = 8.01 (s, 1H, C6-H pyranon),7.26-7.42(m, 5H, -
CH
2
C
6
H
5
), 6.47 (s, 1H, C3-H pyranone), 4.91 (s, 2H, -CH
2
Ph), 3.30 (s, 2H, -CH
2
-N), 2.21-2.42 (m,
) 훿(ppm)=
1
3
4
1
2
2
H, piperidine ring), 1.25-1.50 (m, 6H, piperidine ring). CNMR (125 MHz, DMSO-d
6
75.2, 166.0, 146.9, 141.8,135.8, 130.8, 128.05, 127.8, 127.7, 127.4, 113.8, 71.0, 59.3, 54.2, 54.1,
3.4(2C), 21.6.
.1.1. 3. N-(3-aminopropyl) -7-chloroquinolin -4-amine (3)
6

Journal Pre-proof
A mixture of 1,3-diaminopropane (50 mmol) and 4,7-dichloroquinoline (11.6 mmol) was heated at
o
o
8
0 C for 1.5 hours under Argon atmosphere and subsequently at 145 C for 5 hours with
continuous stirring. After cooling, NaOH (2 M, 10 mL) was added and stirred to precipitate the
desired product. The compound was recrystalized in ethyl acetate, ethanol and methanol 3:1:1.
o
1
MP: 78 C, HNMR (500 MHz, D
Hz, 1H, quinoline ring), 7.87 (d, J=2.4 Hz, 1H, quinoline ring), 7.52 (dd, J=2.4, 8.8Hz, 1H,
quinoline ring), 6.57 (d, J=5.2 Hz, 1H, quinoline ring), 3.56 (t, J=6.8 Hz, 2H, -NHCH CH CH
CH -NH
) 휹(ppm) = 152.6, 152.3, 149.5, 136.3, 127.5, 126.0, 124.2, 118.7,
2
O-d
6
): δ = 8.43 (d, J=5.2 Hz, 1H, quinoline ring), 8.05 (d, J=8.8
2
2
2
-
NH
2
), 3.13 (t, J=6.8 Hz, 2H, -NHCH
2
CH
2
CH
2
-NH
2
), 2.07-2.09 (m, 2H, -NHCH
2
CH
2
2
2
).
13CNMR (125 MHz, DMSO-d
6
9
9.65, 41.9, 39.7, 30.4.
.1.1.4.1-(3-(7-Chloroquinolin-4-ylamino)propyl)-5-(benzyloxy)-2-(hydroxymethyl)pyridine-4(1H)-
2
one (4a)
To a solution of compound 2a (4.3 mmol, 1eq) in ethanol/water was added a proper primary
amine(N-(3-aminopropyl)-7-chloroquinolin-4-amine) (6.9 mmol, 1.5 eq). pH of the solution was
o
adjusted to 11 by NaOH (2N ). Then this solution was heated at 80 C under nitrogen atmosphere
for 30 hours. Completion of the reaction was monitored by TLC. The pH of the reaction mixture
was adjusted to 1 by using concentrated hydrochloric acid. After evaporation of the solvent and
washing with 100 mL dichloromethane, the solid was dried under vacuum to give the crude
product. The isolated product was purified by Sephadex column (eluent: methanol) and silica gel
(dichloromethane, methanol). The procedure is illustrated in scheme 1.
o
-1
Mp:135-137 C, yield 38.5%, IR, cm in KBr: 3414(O-H-str), 3220 (NH), 2884 (=C-H,
aromatic,str) 2353 and 2322 (C-H, aliphatic str), 1610 (C=O), 1453 and 1215 (C=C,aromatic,str),
1HNMR (500 MHz, CD
OD-d ): δ = 8.32 (d, J=5.9Hz, 1H, quinoline ring), 8. 02 (d, J=9Hz, 1H,
3
6
quinoline ring), 7.81 (s, 1H, quinoline ring), 7.46 (s, 1H, C6-H, pyridinone), 7.20-7.45 (m, 5H,
7

Journal Pre-proof
-
CH
2
C
6
H
5
), 7.14 (d, J=9Hz, 1H quinoline ring,), 6.53 (s, 1H, C3-H, pyridinone), 6.51 (d, J=5.9,
1
H, quinoline ring), 4.60 (s, 2H, -O-CH
2
-C
6
H
5
), 4.42 (d, J=2.6Hz, 2H, CH
N), 3.35 (m, 2H, -NHCH CH CH N), 2.3 (m, 2H, -
OD-d ) 휹(ppm)= 170.6, 155.11, 149.2(2), 148.0,
2
-OH), (OH,NH
exchange), 3.65-3.75 (m, 2H, -NHCH
2
CH
2
CH
2
2
2
2
NHCH
2
CH
2
CH
2
N). 13CNMR (125 MHz, CD
3
6
1
7
45.2, 138.4, 137.4, 129.3, 129.06, 128.8, 128.6, 127.3, 127.0, 125.1, 124.3, 117.6, 116, 112, 99.7,
2.3, 61.3, 50.2, 41.2, 30.2. LC ESI-MS (+) m/z (%): 450.0.
2
.1.1.6.1-3-(7-Chloroquinolin-4-ylamino)propyl)-5-(benzyloxy)-2-((piperidine-1-yl)methyl)
pyridine-4(1H)-one (4b)
This compound was prepared and purified in a manner that 4a was, the primary amine in the
preparation of 4b was N-(3-aminopropyl)-7-chloroquinoline-4-amine. The obtained product was
further purified by column chromatography on Sephadex to afford final pure products which were
then was recrystallized with ethyl acetate, dichloromethane(3:1 ) (Scheme 1).
o
-1
MP: 179 C, yield 29.2%, FTIR, cm in KBr: 3376 (N-H-str), 3063 (=C-H, aromatic,str), 2929
(C-N), 2855 (C-H, aliphatic str), 1610 (C=O), 1463 and 1334 (C=C,aromatic, str).
1HNMR (500 MHz, DMSO-d
): δ = 8.40 (d, J=5.7 Hz, 1H, quinoline ring), 8.10 (d, J=9Hz, 1H,
6
quinoline ring), 7.81 (d, J=2Hz, 1H, quinoline ring), 7.57 (s, 1H, C6-H, pyridinone ring), 7.45-7.43
dd, J=2Hz, J=9Hz, 1H, quinoline ring), 7.38 (s, 1H, quinoline ring), 7.20-7.45 (m, 5H, -CH ),
.56 (d, J=5.7 Hz, 1H, quinoline ring), 6.40 (s, 1H, C3-H, pyridinone ring,), 5.25 (s, 2H, O-CH
), 4.9 (s, 2H, CH -N), 4.49 (m, 2H, NHCH CH CH N), 3.35-3.45 (m, 4H, NHCH CH CH N),
.2-2.3 (br, s, 4H, CH -N-piperidine ring), 1.2-1.3 (br, m, 6H, CH -piperidine ring). CNMR (125
(
2
C
6
H
5
6
2
-
C
6
H
5
2
2
2
2
2
2
2
1
3
2
2
2
MHz, DMSO-d
6
) 훿(ppm)= 170, 152.5, 149.1, 149.6, 148.8, 147.4, 137.1, 135.2, 129.4, 129.18,
8

Journal Pre-proof
1
4
29.02, 127.7, 126.30, 124.2, 119.30, 117.9, 116.8, 99.8, 72.4, 60.9, 55.2, 54.7, 52.1, 48.9, 41.9,
0.7, 30.8, 30.6, 26.8, 25.02 . LC ESI-MS (+) m/z (%): 517.
2
.1.1.5.1-(3-(7-Chloroquinolin-4-ylamino)propyl)-5-(benzyloxy)-2-(chloromethyl)pyridine-4(1H)-
one (4c)
Compound 4a was dissolved in thionyl chloride followed by stirring for 48 hours at room
temperature and at the end of this time hexane was added to collect the product. It was purified
with column chromatography (Sephadex), then recrystallized in ethyl acetate and methanol
(Scheme 1).
o
-1
Mp: 211-215 C, yield 55.2%, FTIR, cm in KBr 3436 (N-H-str), 3018 (=C-H, aromatic,str), 2870
and 2802(C-H, aliphatic str), 1617 (C=O), 1463 and 1334 (C=C, aromatic,str), 675-760 (C-Cl):
1HNMR (500 MHz, DMSO-d
): δ=9.85 (d, NH, 1H), 8.84 (d, J=8.5Hz, 1H, quinoline ring), 8.59
d, J=7 Hz, 1H, quinoline ring), 8.50-8.57 (m, 1H, quinoline ring), 8.13 (s, 1H, quinoline ring), 7.70
d, J=8.5 Hz,1H, quinoline ring), 7.50 (s, 1H, C6-H, pyridinone), 7.40 (s, 1H, C3-H, pyridinone),
.2-7.8 (m, 5H, -CH ), 6.94 (d, J=7Hz, 1H, quinoline ring), 5.12 (s, 2H, CH2-Cl), 4.59 (s, 2H,
CH CH N), 2.10-2.30 (m, 2H, NHCH CH CH N).
) 훿(ppm)= 170.6, 155.0, 154.6, 145.3, 144.3, 143.0, 138.6, 138.4,
6
(
(
7
2
C
6
H
5
O-CH
2
C
6
H
5
), 3.50-4.10 (m, 4H, NHCH
2
2
2
2
2
2
13CNMR (125 MHz, CD
OD-d
3
6
1
4
37.4, 129.4, 129.09, 128.8, 127.3, 125.2, 125.1, 124.6, 123.6, 118.2, 116.8, 99.8, 72.5, 61.5, 42.9,
2.2, 30.2. LC ESI-MS(+)m/z(%):468.
2
2
.2. Biological and SAR evaluation
.2.1. Cytotoxic assay
2
.2.1.1. Cell lines and cell culture
9

Journal Pre-proof
A549 as human lung carcinoma[23], HT29 as human colonic adenocarcinoma, MCF7 as human
breast cancer and AGO1522 (non-cancerous, human skin fibroblast-like cells) were purchased from
Pasture Institute of Iran (Tehran). A549, HT29, MCF7 were obtained from American Tissue
Culture Collection (ATCC; CCL-185, HTB-38, HTB-22, respectively). The in vitro cytotoxic
activities of the prepared compounds were evaluated against A549, MCF7, HT29, and AGO1522
cell lines. A549 and AGO1522 cell lines were cultured in the suitable medium of RPMI1640
(
Sigma Aldrich, Germany), HT29 and MCF7 were cultured in DMEM (Sigma Aldrich, Germany)
and FBS (Gibco, Italy) 10% in a humidified incubator (Binder, Germany) at 37 °C in an atmosphere
of 5% CO
2
.
2
.2.2.2. Antiproliferative assay
Antiproliferative assays were performed by MTT assay according to the method proposed by Alley
et al. with some modifications. Cells were seeded into a 96-well plate at the following cell density:
3
2
3
2
3
2
3
2
9
×10 cell/cm (MCF7), 5×10 cell/cm (HT29), 5×10 cell/cm (A549), 8.5×10 cell/cm
(AGO1522). The stock solutions of compounds were prepared in DMSO. After 24 hours incubation
period, the supernatants in 96-well plates were replaced with culture media containing different
concentrations (0.05-10 µM) of the tested compounds. Three wells for each concentration of three
independent experiments were used for each cell line. Then, the cells were incubated in a CO
2
incubator with the target compounds. Test samples of erlotinib (Sigma, Germany) at the same
concentrations as the test compounds were added to six wells of three independent experiments as a
positive reference. The cells were treated with different concentrations of the compounds and
incubated for 48 h. After a 48-hour exposure period, 25 μL of MTT reagent [3-(4, 5-
dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (Sigma, Germany)] (final concentration 0.5
-1
mg.ml ) to each cell was added to each sample. After 4 hours, the media was replaced by 200 μL
1
0

Journal Pre-proof
DMSO (Merck, Germany) to dissolve the purple formazan crystals produced. The absorbance at
5
70 nm of each well was measured with an ELISA plate reader (Epoch, BioTeK, USA ). IC50
values in micromolar (the concentration at which 50% of the cells could survive) were determined.
The obtained data represented the means of three independent experiments in triplicate and were
expressed as means ± SEM.
2
.2.2.3.Analysis of Immunocytochemistry (ICC)
For inhibition assessment of EGFR and VEGFR-2 tyrosine kinase activity, synthetic drug
(4c) and erlotinib (Sigma, Germany) were used at 0.5 µM, which were prepared in DMSO. Growth
medium with DMSO was used as a control instead of target compounds. This concentration was
applied for growth inhibition. The cells were treated with the same concentration of the compounds
in three independent experiments and incubated for 48 h. After a 48-hour exposure period, all
cultured cells (treated and untreated) were fixed in 4% paraformaldehyde solution (Merck,
Germany). After fixation, the samples were washed with PBS (Gibco, Italy) 4 times every 5
minutes. Hydrochloric acid (2N) was used for 30 minutes on cultured cells and neutralized the acid
with borate buffer for 5 minutes then the samples were permeablized with 0.3% triton x-100
(
Sigma-Aldrich, Denmark) for 30 minutes and blocked with goat-serum albumin for one hour.
Primary antibody for EGFR(sc-373746) Santa Cruz Biotechnology (Santa Cruz, CA) and VEGFR-
( Cat. No. ab39256), Abcam (Cambridge, MA, USA) were incubated overnight in 2-8ºC at a
concentration of 0.01 mg/ml. Secondary antibody for EGFR is the goat anti-mouse IgG H&L
FITC) (Cat. No. ab6785, Abcam) and for VEGFR-2, is the goat anti-rabbit IgG H&L (FITC) (Cat.
2
(
No. ab6717, Abcam) at a concentration of 6µg/ml, were followed by a nuclear stain (DAPI, (4′,6‐
diamidino‐ 2‐ phenylindole),-Life Technologies). Samples were incubated in 37 ºC in the dark
incubator for 1.5 hours. Incubated samples were transferred to the darkroom and rinsed 4 times,
1
1

Journal Pre-proof
DAPI was added and immediately removed and then the samples that covered with PBS. The cells
were visualized with the fluorescence microscope. Light immunofluorescence microscopy images
were taken with LABOMED BX40 microscope equipped with a LABOMED microscope.
2
.2.2.4. Analysis of apoptotic cells by annexin V/PI staining
Apoptotic cell death of synthetic drug (4c) and erlotinib was determined by flow cytometry using
an annexin V/fluorescein isothiocyanate (FITC) kit (Abcam, Cambridge, United Kingdom). Briefly,
3
the A549 cells were seeded in six‐ well plates for every sample at a concentration of 5×10 cells
per well and were incubated overnight. The seeded cells were treated with the one concentration
(0.5 µM) of erlotinib, 4c, and DMSO as a control for 48hr. For an analysis of the level of apoptosis,
cells were harvested and washed twice with cold PBS. After discarding the supernatant, cells were
suspended in 500 μl binding buffer and subsequently stained with 5μl annexin V‐ FITC for 15 min
on ice in the dark. Subsequently, 3μl of PI (50μg/ml) was added. The cells were then analyzed by
flow cytometer using FACS Calibur flow cytometer (BD Bioscience, San Jose, CA, USA, Flowcyt
Science-Based Company). Cells were determined as viable cells (the LL quadrant, annexin
V−/PI−), early apoptotic cells (the LR quadrant, annexin V+/PI−), late apoptotic cells (the UR
quadrant, annexin V+/PI +), and necrotic cells (the UL quadrant, annexin V−/PI+).
2
2
.3. Theoretical sections
.3.1. Molecular docking simulations
2
.3.1.1. Inside EGFR active site
1
2

Journal Pre-proof
Molecular docking of compounds into the three dimensional X-ray structure of EGFR (PDB
code: 1M17) was carried out using the Discovery Studio software (version 4.1) as implemented
through GOLD protocol. The 3D structure of EGFR (1M17) for docking investigation was
downloaded from Protein Data Bank. Three-dimensional structures of the designed compounds,
as well as erlotinib and vandetanib, were constructed using MarvinSketch software, and then
they were energetically minimized in Discovery Studio by CHARMm with 5000 iterations and
minimum RMS gradient of 0.01. All bound waters and ligands were eliminated from the protein,
and the polar hydrogen was added to the protein. Each compound was retained in erlotinib active
site for 10 poses. The docking poses were ranked by GOLD score fitness.
2
.3.1.2. Inside of the VEGFR-2 active site
The structure of VEGFR-2 receptor was taken from protein data bank with the ID code of
ASD. This structure has atomic coordinates of sorafenib in complex with VEGFR-2 active
4
conformation. The complex protein structure containing all ions, ligands, and all water
molecules was used by protein protocol in Discovery Studio client 4.1 to prepare the protein
needed for the docking experiment. Ligand preparation protocol with 0.01 RMS gradient,
Momaney-Ron partial charge estimation, smart algorithm minimizer with 10000 steps and
CHARMm force field were used for preparing sorafenib, vandetanib, and three synthesized
compounds. Gold algorithm of Discovery Studio Accelrys client software was used to carry out
molecular docking computations using GOLD default parameters in active sorafenib site.
2
.3.2. Molecular Dynamics Simulations of EGFR and VEGFR-2
1
3

Journal Pre-proof
Molecular dynamics (MD) simulation was performed in two steps. In the first step, the minimum
energy with the best interactions of compound 4c as the most efficient compound in the
biological assay, complexed with each protein was subjected to molecular dynamics simulation.
vandetanib, a dual EGFR and VEGFR-2 inhibitor, as the positive control complexed with these
proteins was also studied in this phase. MD simulations were processed in explicit aqueous
solution for 30 ns and 50 ns. The topology parameters of vandetanib and 4c were obtained from
Dundee PRODRG server [24, 25]. Each complex was placed in a cubic box with an edge of 1.5
nm; then the box was solvated with SPC water molecules. Two ions were added to neutralize the
whole system of VEGFR-2 (4ASD) and six ions for EGFR (1M17) proteins. Energy
minimization was performed on the complex species followed by 100 ps NVT and 100 ps NPT,
then MD simulations were performed at 310 K. Finally, MD simulations were carried out on the
whole system for 30 ns with a van der Walls cutoff of 1.2 nm. The Parrinello-Rahman method
was used for temperature and pressure coupling. The reference temperature was set as 310 K.
Time step was 2fs, and the trajectory was saved every ten ps. All the simulations were carried
out using the GROMACS 5.1.4 package with GROMOS9643a1 force field [26].
2
.3.3. 3D Pharmacophore modeling
Six known EGFR and VEGFR-2 inhibitors with 4-aminoquinazoline scaffold listed in Table 1,
including vandetanib as a standard compound with approved efficacy were used in
pharmacophore model generation [13, 17, 19, 27 and 28]. The obtained pharmacophore model
was searched among the six designed dual inhibitors in the present study. HipHop algorithm
built-in Discovery Studio Accelry client 4.1 was used in the generation of pharmacophoric
features [29]. HipHop pharmacophore models are generated usually from a set of active molecules.
1
4

Journal Pre-proof
Based on the atom-types in the active molecule, Hip Hop generates key common chemical features
for 3D-pharmacophore models creation. The generation of the HipHop common feature hypothesis
uses different classes of molecules from the most active ones and is independent of the biological
activity values. Pharmacophore model generation is qualitative in nature and should have the ability
to identify essential features for binding interaction to show the biological activity for both
receptors [30]. This research method was used to identify the 3D arrangement of chemical
functional groups (hypothesis) in dual inhibitor ligands complexed with EGFR and VEGFR-2
proteins and then was extended to cover two receptors with entire regions. The key pharmacophoric
features were in geometric complementary to the corresponding features of the EGFR and VEGFR-
2
active sites enabling favorable interactions with both active sites [31].
The binding information obtained by Gold algorithm with the flexible amino acid
docking was compared and analyzed for all these compounds emphasizing the binding mode of
erlotinib for EGFR protein and sorafenib for VEGFR-2. Information about the best-docked
conformation including the Gold score and protein interacting residues for the three synthesized
ligands as well as erlotinib as a selective EGFR, sorafenib as a selective VEGFR-2 and
vandetanib as a dual inhibitor of EGFR and VEGFR-2 are provided in the results and discussion
section.
[
Tables 1 near here]
2
.3.4. ADMET and properties predictions
ADMET prediction and TOPKAT modules of the Discovery Studio (4.1 Accelry clients) were used
for evaluating the pharmacokinetics and metabolic profiles of the ligands used for pharmacophore
modeling. Molecular descriptors were used for calculating the drug pharmacokinetics and toxicity
for ADMET prediction [32]. The prediction results are provided in the corresponding section.
1
5

Journal Pre-proof
3
3
. Results and Discussion
.1. Rational Design
In 2011, Vandetanib with a 4-aminoquinazolin scaffold was approved as a dual inhibitor of
EGFR and VEGFR-2 by AstraZeneca Pharmaceutical Company. Before approval, it had been
reported by ChEMBL533849 ID [33] in an antimalaria screening by GSK Company.
The 4-Aminoquinoline scaffold is the most common pharmacophore used for malaria growth
inhibition. Sorafenib is the famous VEGFR-2 selective inhibitor, and erlotinib is another drug
approved to inhibit EGFR selectively. By evaluation of the docking interaction modes, the
biological activity of quinoline structure in literature such as cabozantinib with quinoline scaffold,
and structure-based pharmacophore modeling by virtual screening in the ChEMBL database[33],
EGFR and VEGFR-2 were selected as targets for the design of some novel dual inhibitors of these
enzymes. The rational design is illustrated in Figure 2.
(Figs.2, 3, 4. Near Here)
Three moieties in the structures of EGFR and VEGFR-2 inhibitors can be pointed out according
to the interactions they establish with the corresponding targets. A chemical moiety located in
the middle of the molecule which provides the hydrogen bond acceptors and hydrogen-bond
donors demanded by the EGFR or VEGFR-2 active sites. Two other hydrophobic moieties with
determined radii which determine the selectivity of the growth factor inhibitors for EGFR or
VEGFR-2 enzymes. The discussed structural features are color-coded in Fig. 2. The three
moieties were considered in the structure of the designed compounds. The propylene diamine
linker and pyridinone can be regarded as a good replacement for the quinazoline ring to provide
1
6

Journal Pre-proof
the hydrogen bond acceptor and hydrogen bond donor groups (shown in red, Fig. 2). On the
other hand, chloromethyl, hydroxymethyl, 1-methyl piperidine, and chloroquinoline groups were
placed in the designed structures to improve the lipophilicity. In hybrid drug design, identified
safe scaffolds in terms of pharmacokinetic profiles are selected to incorporate into the hybrid
structure to reduce the time and cost These theoretical with experimental protocols were used in
some literature especially for mutant protein and resistant disease. [34].
3
.2. Structure-activity relationship and biological assay
All of the target compounds were evaluated for their anti-proliferative effects by MTT assay. A549,
MCF7, HT29, and AGO1522 cell lines were used in this assay. The IC50 values were calculated
from the sigmoidal dose-response curves (variable slope) that were generated by Graphpad prism.
These values are provided in Table 3. Piperidine and Cl as R substitutions showed different effects
1
on antiproliferative activity against MCF7, HT29, and A549. 4c with a Cl substitution showed the
lowest IC50 values among the studied compounds against all the studied cell lines, but had higher
confidence intervals (CI), in the range of CI values for erlotinib as a positive control.
Biological activity evaluation of the studied compounds against the three cancerous cell
lines (A549, MCF7, and HT29) showed that Cl substitution on R1 position increased the activity in
comparison with other substitutes. But, the piperidine moiety substituted on 4-aminoquinoline
scaffold decreased the inhibitory activity against A549 cell line compared with MCF7 and HT29
cell lines. Interestingly, Cl substitution increased the potency and induced a good inhibitory activity
considering erlotinib as a positive control. OH, substitution on the scaffold led to high potency
against A549 and moderated towards the other cell lines. All compounds had shown good potency
1
7

Journal Pre-proof
to inhibit the cell growth in all three cell lines, but 4c was the best compound and was used for
further investigations by molecular dynamics simulations.
(Table 3 Near Here)
Interestingly in a previously reported investigation on 4-anilino-quinazoline derivatives, the
presence of a quinoline amino fragment at C4 position and N-linker of the quinazoline core led
strong EGFR inhibition potency [17]. Synthesis and biological evaluation of four classes of
aminoquinazoline derivatives in another report indicated that functionalization of the aniline moiety
with further benzene rings have improved the antiproliferative activity [28]. The impact of
hydrogen bond formation by substituents on the quinazoline moiety has not been extensively and
specifically reported in EGFR inhibition by dual inhibitors. But the role of ATP binding pocket in
interaction with urea moiety in VEGFR-2 inhibitors is identified. The hydrogen bond donating
substituent in para position of the hydrophobic site in 4-aminoquinazoline moiety has increased the
activity for dual inhibition [27, 28].
3
.3. Molecular docking
Computer-aided molecular docking is a familiar method for explaining the receptor-ligand
interactions in drug discovery. The docking routine in Discovery Studio client Accelry 4.1.
Software with GOLD algorithm was used to dock the newly synthesized inhibitors into the
binding sites of EGFR and VEGFR-2 proteins. The docking reliability was validated using the x-
ray structures of erlotinib and sorafenib co-crystallized in EGFR and VEGFR-2 proteins,
respectively by redocking them into the corresponding binding sites. The previously reported
docking investigations of both inhibitors also confirmed the accuracy of the applied docking
protocol [35, 36]. The obtained RMSDs were 1.57 and 0.90 for EGFR and VEGFR-2 proteins,
1
8

Journal Pre-proof
respectively. All compounds were successfully docked into the active sites of VEGFR-2 and
EGFR, and the types of nonbonded interactions were compared with the interactions of vandetanib,
erlotinib, and sorafenib with their protein binding sites. At the end of the docking procedure, the
conformations of ligands were ranked by GOLD-fitness scores and Ligand-protein interactions
visualized in the Discovery Studio. The results of for the best-docked conformation according to
the highest total scores and minimum energies are shown in Table 2. The potential compounds
were also selected based on the interaction mode compared with the binding modes of sorafenib
and erlotinib in protein-complex crystals. The docking simulations of 4c inside EGFR showed
H-bond interactions with Met769 and Lys721 and hydrophobic binding with Pro770. These
binding modes were similar to those with erlotinib as a selective EGFR inhibitor and vandetanib
as a dual EGFR/VEGFR-2 inhibitor. The results of the binding interactions of 4c are
summarized in Table 2, and the docking modes for the best-docked conformation are illustrated
in Figures. 5 and 6. The docking simulations of 4c inside VEGFR-2 showed H-bonding to
Asp1046, Lys868, and Glu885, and hydrophobic interactions to Cys919, Leu889, Ileu888 and
Phe1047. We also assigned the essential residues of EGFR interacting with 4a. They were the
same as 4c, but different from 4b. All compounds have the same interaction binding modes
within VEGFR-2 receptor. GOLD fitness scores of 4c and 4b are higher than 4a for interaction
with this receptor, but the biological activity of 4b against A549 as an EGFR receptor containing
cell line was the lowest value in the series.
These docking results also showed that Asp1046 in VEGFR-2 and Met769, Lys721 and
Pro770 in EGFR are critical amino acids for inhibition activity. In the final step, vandetanib as
an FDA-approved dual inhibitor with acceptable biological activity was used in the docking
procedure inside the EGFR and VEGFR-2 active sites. The docking interaction mode of 4c, as
1
9

Journal Pre-proof
the most potent compound in the anti-proliferative assay, was compared with that of vandetanib.
The results are shown in Table 2 and Figures. 5 and 6.
(
Table 2 and Figures. 4, and 5, 6 Near Hear)
3
.4. Molecular dynamics simulation of the protein-ligand interactions
Molecular dynamic (MD) simulation is a computer simulation technique for showing the physical
position of atoms and molecules. Molecular docking describes only a single position of ligand-
protein complex but molecular dynamic simulation is usually performed to show the changing of
the ligand binding in terms of conformational changes. In MD simulation the ligands and residues
of the protein are allowed to interact for a fixed period.
Molecular dynamics simulation was used to investigate the type of interactions of 4c, the
best compound according to the experimental and docking results. The best-docked poses of this
compound, as well as vandetanib with both receptors of VEGFR-2 and EGFR, were chosen for MD
simulation. Molecular dynamics simulations jobs were performed using GROMACS 5.1.4
software. MD simulations were processed for 30 ns in an aqueous solution, and the protein-ligand
complex was placed in a cubic box. The stability of the system under simulation was evaluated by
root-mean-square deviation (RMSD) of the backbone structure related to the starting point. RMSD
of the trajectories from their initial structures were used to examine whether the structures of the
ligand-protein complex remained stable under the simulation conditions [25]. The obtained RMSDs
of 4c and vandetanib complexed to EGFR and VEGFR-2 separately are compared, and the
corresponding plots are provided in Figures. 11A and 12F. All systems reached at equilibrium near
1
5 ns simulation run. Root mean square fluctuations (RMSF) of the individual amino acid residues,
2
0

Journal Pre-proof
hydrogen bond formation, and gyration radius (Rg) were analyzed for all complexes and are shown
in Figure.11, 12.
Analysis of RMSD revealed that 4c-VEGFR-2 complex and vandetanib-VEGFR-2
complexes reached the equilibrium state within ten ns of simulation. The backbone RMSD varied
between 0.0-0.3 nm and 0.1-1.2 nm through the simulation runs for EGFR and VEGFR-2 of the
two ligands, respectively. It can be found that the RMSD profiles of these compounds are probably
stable in comparison with vandetanib-EGFR and vandetanib-VEGFR-2 complexes. The 4c-
VEGFR-2 complex is stable, and its variation is very close to the vandetanib complexes as the
positive ligand, and for approved this result, the MD simulation runs continued within 50 ns.
RMSD profiles of two ligands 4c and vandetanib complexes were analyzed. The dynamics
behaviors of VEGFR-2 and two ligands according to RMSD variations, were evaluated and
assessed during a 50 ns run and are shown in Fig.12 L.
RMSF studies are extremely suitable in measurements of the average atomic mobility of the
Cα atoms of the protein in the MD simulations. The separate and superimposed RMSF traces are
illustrated in Figs. 11B, 12G. The fluctuations of 4c complexes with EGFR and VEGFR-2 were
approximately close to the variations of vandetanib complexes with these proteins. The results of
RMSF analysis of VEGFR-2 showed that the residues in the binding site, which are interacting
with the ligands, are approximately stable, and they showed very little fluctuations in the period of
simulation runs.
The key residues of EGFR (Lys721, Met769) and the key residues in the active site of
VEGFR-2 (Asp1046, Lys868, and Ala1050) have the approximately lowest amount of RMSF in
the total MD simulations. The ranges of RMSF in 4c-EGFR and 4c-VEGFR-2 complexes were
2
1

Journal Pre-proof
similar to the variations of vandetanib-EGFR, and vandetanib-VEGFR-2 complexes during MD
simulation runs.
The obtained Rgs were between 2 to 2.25 nm for EGFR complexed with 4c and vandetanib.
The radius of gyration was between 1.94 to 2.04 for vandetanib complexed with VEGFR-2 and
1
.99 to 2.06 for the 4c-VEGFR-2 complex. Rg is a critical parameter in the evaluation of the
stability of MD trajectories that was calculated for atoms and ligands of the receptors.
The hydrogen bonds in VEGFR-2 complexes with both ligands in the binding site were
analyzed in the entire simulation runs. Hydrogen bonds and hydrophobic interaction in VEGFR-2
are most specific interactions in biological recognition. The binding modes of 4c with VEGFR-2
from MD simulation results are close to the docked structures. During the period of MD simulation,
these compounds have revealed dynamic hydrogen bonding interactions with important binding site
residues in the active site of VEGFR-2. Figure 12I, K shows the numbers of hydrogen bonds
formed in the complex of the total compounds in the two receptors. The quinoline moiety was
inserted into the hydrophobic pocket of VEGFR-2. Hydrogen bonds of two compounds with the
binding site were analyzed during the entire simulation. There was a good correlation between the
type of hydrogen bonds between 4c and vandetanib, and the VEGFR-2 active site residues. The
hydrogen bonding analyses of the EGFR and VEGFR-2 complexes of 4c and vandetanib are shown
in Figures 7, 8, 9, 10. This easily shows the binding stability of ligands inside the EGFR and
VEGFR-2 binding sites. Hydrogen bonding for Lys721 and Met769 amino acids in the 4c-EGFR
complex was monitored during the 30 ns MD simulations run. Hydrogen bonding of vandetanib as
the positive control was observed in Lys721 and Met769 during all simulation time (Fig.7.)
Carbonyl group in pyridinone ring of 4c showed hydrogen bond interactions to Lys721. Met769
and Leu 694 were other residues that showed hydrogen bond interactions. The range of distance
2
2

Journal Pre-proof
between Lys721 and Met769 with the corresponding atoms of the 4c-EGFR complex was 0.3-0.4
nm. In addition, the average hydrogen bond length of vandetanib to Lys721 and Met769 was 0.4
nm (Fig. 7.). The hydrophobic interaction of Pro770 is important in EGFR active site, and they
were shown for both ligands in the active site (Fig.8). A new and very important hydrophobic
interaction between Pro968 of EGFR active site and vandetanib was found by molecular dynamics
simulation, which is shown in Fig. 8. In a previous MD simulation study on the interaction between
erlotinib and EGFR binding site, the contacts between the residues Met769, Lys721, and Pro770
and erlotinib were identified and reported [37,38]. The hydrogen bond between 4c and VEGFR-2
was observed through Lys868, Asp1046, and Cys919 during MD simulation run. In hydrogen
bonding between vandetanib and VEGFR-2, Lys868 and Ala1050 were involved (Figures 9 and
1
0). In the previously mentioned investigations, it was observed that sorafenib could form
hydrogen bonds with ASP1046 and GLU885 [35, 39, and 40]. Besides, the nitrogen atom of the
pyridine ring in sorafenib can form a hydrogen bond with CYS919. In these MD simulation
investigations, it was found that ASP1046, Lys868, and CYS919 act as key residues for dual
inhibitory action, and it was verified that Ala1050 has an interaction in favor of dual activity.
(Fig.7, 8, 9, 10, 11, 12, Near Here)
3
.5. 3D Pharmacophore modeling
The HipHop pharmacophore models were generated by considering the principal value 2 for all
ligands and the maximum omitted feature was 0. The final HipHop models, according to common
3
5
D pharmacophore features HHHDA, were ranked, and the corresponding scores were 56.00,
4.46, 54.08, 54.03, 53.86 and 53.26. The pharmacophoric fit values were obtained from the
ligands pharmacophore mapping, including D2204=4.998, D2014=4.5, D2016=3.7,
vandetanib=2.7, D2013=0.8, and AEE=0.6. The interaction patterns generated by 3D
2
3

Journal Pre-proof
pharmacophore models were compared to the interaction modes extracted by molecular docking
and molecular dynamics simulations. In this investigation, it was shown that the contribution of
hydrophobic features was determinant of the activity or inactivity of a typical ligand. The radius of
hydrophobic features is an important mode of interaction in the variation of biological activity. R
1
substituent in the synthesized compounds was considered as a hydrophobic feature with a radius of
2
.11. 4b with a piperidine substitution (3a) showed lower activity in biological assay than 4a and
c. In the case of 4b in the in vitro anticancer experiment, the increase of the hydrophobic feature
4
radius reduced the activity against A549 cancerous cell line. As a result, the lower the hydrophobic
radius, the higher the activity. This can be observed for Cl (2a) as the substitution as illustrated in
Figure 13. The hydrogen acceptor (HA) and hydrogen donor (HD) features have shown the
essential interactions in binding modes with EGFR and VEGFR-2. The pharmacophore analysis for
VEGFR-2 inhibitors in previous reports has shown five essential features, including H bond
donor/acceptor and hydrophobic properties. This pharmacophore model, along with the 3D-QSAR
model, approved that van der Waals interactions and hydrophobic site, play critical roles in the
active site of VEGFR-2 [41]. In a previously reported investigation, three-dimensional
pharmacophore models of EGFR were generated that included seven 3D pharmacophoric points
with three chemical features. One hydrogen bond feature, three hydrogen bond acceptors and three
aromatic rings are essential features for activity and erlotinib, and gefitinib has the highest affinity
based on these features [42]. A number of pharmacophore studies for VEGFR-2 with CoMFA and
CoMSIA approaches proved the significance of hydrophobic, hydrogen bond donor, and acceptor
interactions. The general approach from the studies was that steric bulky group as hydrophobic
moieties are located in the proximity of hydrophobic pocket of VEGFR-2. The aminoquinazoline
ring is a desired general scaffold in QSAR and pharmacophore-based studies of EGFR inhibitors.
In these studies, the nitrogen atom of the quinazoline ring and the amine group of the aniline
2
4

Journal Pre-proof
moiety play significant roles in hydrogen bond formation and increase in activity. The
hydrophobicity of ligands is also the main pharmacophore feature in the ring system of both
quinoline and aniline moieties. Quinoline and aniline moieties show essential hydrophobic
interactions in the EGFR active site [43]. These observations have illustrated the essential
pharmacophore features and evaluation of the binding modes in dual inhibition characteristics shed
light to new aspects of the design of novel dual inhibitors based on 4-aminoquinoline derivatives.
(Fig.13. Near Here)
3
.6. ADMET evaluation
Most drugs candidates fail in clinical trials because of their poor ADMET profiles. The
pharmacokinetic profiles supplemented with biological activity are “must” for a molecule to be
considered as a drug candidate. There are various in silico methods for predicting the
pharmacokinetics of a molecule. Drug-likeness parameters were evaluated, taking into
consideration Lipinski's "Rule of Five" [44, 45]. All calculated parameters for the novel synthesis
molecules were well within the prescribed ranges. These compounds were considered as drug-like
candidates and were prepared by the methods given in Section 2.1 and Table 1, but D2016 was
outside of the range. Results of pharmacokinetic screening revealed that compounds were
generated for pharmacophore features showed lipophilic natures according to their logP values. The
ADME screening results are summarized in Fig. 14. The blood-brain barrier penetration (BBB) and
2
D PSA are inside the 95% and 99% confidence ellipses. The screening results of ADME model
showed that seven compounds possess a 99% confidence level for human intestinal absorption and
blood-brain barrier (BBB). Similarly, the other active derivatives (D2016) also showed intestinal
absorption level falls outside the ADME model ellipses. The compound D(2014) for the filter that
2
5

Journal Pre-proof
indicates its poor intestinal absorption and BBB penetration ability. The plot of polar surface area
(PSA) and AlogP are represented in Fig 14 and showed in Table 4.
The A logP was associated with absorption or penetration; therefore, the value should be
less than 5. The PSA is another item property correlated to drug bioavailability that should be
smaller than 140. The linker substitutions of D2016 that placed in quinazoline ring caused the poor
ADMET parameters.
(Table 4 Near Hear)
4
.1. Analysis of Immunocytochemistry (ICC)
Immunocytochemistry (ICC) is a common technique for the visualization of proteins that is used to
anatomically visualize by biomolecule capable of binding specific primary antibody and the
localization of a specific protein or antigen in cells. To identify the effect of inhibition of synthetic
drug on EGFR and VEGFR-2, immunocytochemical analysis was conducted on A549 as lung
cancer cell lines. In this procedure, the expressions of EGFR and VEGFR-2 were evaluated as the
targets for bioactivity[46,47]. In EGFR (Figure 15), and VEGFR-2 (Figure 16) cell lines, treatment
with a synthetic drug (4c) and erlotinib led to decrease of expression of EGFR (green) in
comparison to untreated cell line. Reduction of expression of EGFR and VEGFR-2 as biomarkers
of target receptors in treated cell lines compared to the expression of both receptors in untreated cell
line and it is showed; there is an evidence of the interaction of synthetic compound (4c) and
erlotinib. Furthermore, treatment with erlotinib as positive control resulted in decreasing of the
expression of EGFR in comparison to 4c and untreated cell line, but the variation between the
expression of EGFR in a treated cell line with 4c and erlotinib is not significant with P value< 0.05.
2
6

Journal Pre-proof
VEGFR-2 expression in treated cells with 4c in comparison to the treated cells with erlotinib
showed a significant reduction with P value<0.05. These results indicated that molecular dynamics
simulations of interactions of the two receptors with target ligands and expression of two growth
factor proteins have same direction and correlation which approved the mechanism of compounds
action to inhibit the activity of two receptors. All results are shown in Fig.15, 16.
(Fig.15, 16. Near Here)
4
.2. Flow cytometry
To determine whether the growth inhibitory effect of the synthetic drug (4c) and erlotinib being
employed as a positive control was associated with the induction apoptotic cell [48], we evaluated
the apoptosis of A549 cell line after 48hr of treatment with 4c and erlotinib (0.5µM). As illustrated
in Figure 17, bruising with annexin V/PI showed the induction of cellular apoptosis in the A549
cell line. The percentage of total apoptotic cells, early apoptotic cells (the LR quadrant, annexin
V+/PI−), and late apoptotic cells (the UR quadrant, annexin V+/PI +) was approximately 76% in
treated cell line with synthetic drug (4c), 43.7% in treated cell line with erlotinib as positive control
and 22.04% in untreated A549 cell line. Our data indicated that 4c with 0.5 µM concentration has
increased early and late apoptosis in comparison to untreated cell line and the same concentration
of treated cell line with erlotinib. Besides, we detected the mechanism of compound 4c inhibition
activity by flow cytometry and found that the compound can induce the apoptosis of activated
A549 cells. All result is shown in Figure 17. The 4c as a synthetic compound increased the
percentage of apoptosis by Annexin V-FITC/PI staining at the same concentration. This result for
the apoptosis rate, showed the high potency of synthetic compound (4c) to induce apoptosis in the
treated cell line. The high potency to the growth inhibitory effects maybe returns to the synergy of
growth factor receptors that have induced the apoptosis signals. VEGFR-2, a tyrosine kinase, is the
2
7

Journal Pre-proof
major mediator of the antiapoptotic [49] and the binding of EGFR to the extracellular domain
induces receptor homo- or heterodimerization and activates the kinase domain. This may lead to the
autophosphorylation of the intracellular tyrosine residues at the C-terminal and the subsequent
induction of downstream signals. Inhibition of this pathway by EGFR tyrosine kinase domain
inhibitors such as erlotinib induced apoptosis signal [50]. These results possibly indicated the
synergy of both growth factor receptors that have induced the apoptosis.
(Fig.17. Near Here)
Conclusion
In summary, 4-aminoquinoline derivatives were synthesized and screened for biological activity as
dual EGFR and VEGFR-2 inhibitors. The preliminary investigation showed that two compounds
displayed favorable effects in the anti-proliferative assay against A549 (human lung carcinoma),
HT29 (human colonic adenocarcinoma), MCF7 (human breast cancer) and AGO1522 (non-
cancerous, human skin fibroblast-like cells) cell lines. Besides, molecular docking and MD
simulations revealed that the binding poses of 4c with EGFR are similar to that of vandetanib and
erlotinib. Lys721, Met769 are key residue for H-bonding and important for potential activity, but
variations of hydrophobic interactions may cause dual-action potency. Docking and MD
simulations of the 4c-VEGFR-2 complex and comparison of the hydrogen bond interaction modes
showed that the possible mechanism is similar to that of sorafenib and vandetanib. The
hydrophobic interactions are essential features for dual inhibition. Pro770 and Pro968 are critical
residues with hydrophobic interactions in the EGFR receptor, important for the design of a new
dual inhibitor. Furthermore, the radius of the hydrophobic ligand substitution is very important.
Ala1050 in VEGFR-2 receptor was identified as an important residue for hydrogen bond
2
8

Journal Pre-proof
interaction with vandetanib in the present research. Asp1046 and Lys868 have previously known as
hydrogen bond-forming residues for sorafenib. In this research, they were indicated to be a
participant in hydrogen bond formation with vandetanib, a dual EGFR and VEGFR-2 inhibitor. To
confirm these binding modes of inhibition in both receptors with 4c, we made
immunocytochemistry and apoptosis assay. The experimental results have showed a favorable
correlation to the theoretical results. The proposed study is anticipated as a potential platform to
understand the mechanism of association between two growth factor receptors. The obtained results
may help make structural modifications to design and prepare novel dual EGFR and VEGFR-2
inhibitors in future studies.
Acknowledgment
We want to show our gratitude to the Iranian pharmaceutical company and biotechnology research
center of Kashan for providing material and advising in this project. We would appreciate the
medical university of Shahid Beheshti and traditional medicine research center for their assistance
during this research.
2
9