Structural-based design, synthesis, and antitumor activity of novel alloxazine analogues with potential selective kinase inhibition

Article abstract of DOI:10.1016/j.ejmech.2018.04.029

Protein kinases are promising therapeutic targets for cancer therapy. Here, we applied multiple approaches to optimize the potency and selectivity of our reported alloxazine scaffold. Flexible moieties at position 2 of the hetero-tricyclic system were inc

Full text of DOI:10.1016/j.ejmech.2018.04.029

Accepted Manuscript
Structural-based design, synthesis, and antitumor activity of novel alloxazine
analogues with potential selective kinase inhibition
Waleed H. Malki, Ahmed M. Gouda, Hamdy E.A. Ali, Rabaa Al-Rousan, Doaa
Samaha, Ashraf N. Abdalla, Juan Bustamante, Zakaria Y. Abd Elmageed, Hamed I.
Ali
PII:
S0223-5234(18)30363-5
10.1016/j.ejmech.2018.04.029
EJMECH 10378
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 7 February 2018
Revised Date: 12 April 2018
Accepted Date: 14 April 2018
Please cite this article as: W.H. Malki, A.M. Gouda, H.E.A. Ali, R. Al-Rousan, D. Samaha, A.N. Abdalla,
J. Bustamante, Z.Y. Abd Elmageed, H.I. Ali, Structural-based design, synthesis, and antitumor activity
of novel alloxazine analogues with potential selective kinase inhibition, European Journal of Medicinal
Chemistry (2018), doi: 10.1016/j.ejmech.2018.04.029.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Structural-Based Design, Synthesis, and Antitumor Activity of Novel
Alloxazine Analogues with Potential Selective Kinase Inhibition
1
2, 3
4
5
6,7
Waleed H. Malki , Ahmed M. Gouda , Hamdy E. A. Ali , Rabaa Al-Rousan , Doaa Samaha ,
1
4
4
4,7*
Ashraf N. Abdalla , Juan Bustamante , Zakaria Y. Abd Elmageed , Hamed I. Ali
1
3
4
5
,2
1
2
Departments of Pharmacology and Pharmaceutical Chemistry, College of Pharmacy, Umm Al-Qura University, KSA.
Department of Medicinal Chemistry, College of Pharmacy, Beni-Suef University, Beni-Suef 62514, Egypt.
Rangel College of Pharmacy, Health Science Center, Texas A&M University, Kingsville, Texas 78363, United States.
TheBenandMayteeFischCollegeofPharmacy,The University of Texas at Tyler, Tyler, Tx 75799, United States.
6
Institute of Chemistry, Humboldt-Universität zu Berlin, Brook-Taylor-Str. 2, Berlin 12489, Germany.
Department of Pharmaceutical Chemistry, College of Pharmacy, Helwan University, Cairo 11795, Egypt.
7
Abstract
Protein kinases are promising therapeutic targets for cancer therapy. Here, we applied multiple approaches to
optimize the potency and selectivity of our reported alloxazine scaffold. Flexible moieties at position 2 of the
hetero-tricyclic system were incorporated to fit into the ATP binding site and extend to the adjacent allosteric
site and selectively inhibit protein kinases. This design led to potential selective inhibition of ABL1,
CDK1/Cyclin A1, FAK, and SRC kinase by 30–59%. Cytotoxicity was improved by ∼ 50 times for the
optimized lead (10b; IC = 40nM) against breast cancer (MCF-7) cells. Many compounds revealed potential
5
0
cytotoxicity against ovarian (A2780) and colon carcinoma (HCT116) cells of ∼ 10–30 time improvement (IC50
-17nM). The results of the Annexin-V/PI apoptotic assay demonstrated that many compounds induced
5
significantly early (89–146%) and a dramatically late (556–1180%) cell death in comparison to the vehicle
control of MCF-7 cells. SAR indicated that 5-deazaalloxazines have a higher selectivity for Abl-1 and FAK
1

ACCEPTED MANUSCRIPT
kinases than alloxazines. The correlations between GoldScore fitness into FAK and SRC kinases and IC50
2
against MCF-7 and A2780 cells were considerable (R : 0.86-0.98).
Keywords: Alloxazine; Antitumor; Kinase inhibitor; Docking; Apoptosis.
1
. Introduction
Cancer is a major global health burden and the second most common cause of death in the United States. In
2
017, about 1.7 million new cancer cases are expected to be diagnosed in the United States and 600,000 of them
1
are expected to die from this disease. Anticancer drug development has shifted from traditional cytotoxic
drugs, which are associated with numerous nonspecific toxicities, to agents that target and modulate proteins,
2
such as kinases, whose activities are more specifically associated with neoplastic cells. Protein and lipid
kinases are the most important widely studied targets for drug discovery because of the pivotal roles they play
3
-4
in regulating cellular pathways.
Among these kinases, tyrosine kinases catalyze ATP-dependent
phosphorylation of tyrosine residues on target proteins to mediate signaling pathways responsible for cell
5
-6
proliferation, differentiation, metabolism, and death. Protein kinases (PKs) play a crucial role in molecular
7
-8
pathogenesis of different types of cancers, They are therefore considered as main targets for discovery of new
9
drugs. For instance, protein kinase inhibitors (PKIs) as anti-angiogenic agents can be successfully applied as a
1
0
new strategy for cancer therapy. Allosteric modulators are now being developed to target these kinases.
Imatinib (Gleevec ) is the first PTKI that has been approved by the FDA, which represents a revolutionary
1
1 12
targeted cancer therapy.
Type II inhibitors, such as imatinib, bind at the ATP site of bcr-abl and extend into
an adjacent site close to the αC helix therefore inhibiting the enzyme activity of the protein semi-competitively.
1
3
This binding mode pushes the DFG-motif of the activation loop (A-loop) towards the ATP binding site and
1
3
prevents ATP molecules from accessing the hinge region. Imatinib is a potent kinase inhibitor for BCR-ABL
1
4-16
and a stem-cell factor receptor (c-kit).
Selective PKIs, inhibit both of the receptor and non-receptor kinases
and are attractive avenues for the design of antitumor drugs that inhibit tumor development, including
proliferation, invasion, and angiogenesis.
2

ACCEPTED MANUSCRIPT
Many researchers have focused on 5-deazaflavins as potential PKIs and determined their promising antitumor
1
7-23
activities against different cancer cell lines.
Flavins, 5-deazaflavins, and their alloxazine analogues are
substructures in a number of essential naturally occurring redox cofactors such as flavin adenine
2
4
dinucleotide (FAD) and nicotinamide adenine dinucleotide (NAD). Owing to their important function in redox
2
5-26
reactions, as well as their roles in the repair of damaged DNA in biological systems,
5-deazaflavin
2
7
derivatives have received considerable attention in cancer research. We previously reported that 5-deazaflavin
derivatives have significant antitumor activities against NCI-H 460, HCT 116, A431, CCRF-HSB-2, KB, MCF-
1
7-20, 28-29
7
, and HeLa cancer cell lines.
of HMD2 and stabilize the wild-type p53. Our previous reports
demonstrated that 5-deazaflavins are promising antitumor candidates due to exhibiting good anti-proliferative
By assessing their biological activity, 5-deazaflavins inhibit E3 activity
3
0
17-20, 28
contributed to this growing field and
1
7
activity against different cancer cell lines. Interestingly, Ali et al. reported that 2-deoxo-10-methyl-2-phenyl-
5
-deazaflavin exhibited a reduction in tumor size of BALB/c nude mice that had been subcutaneously
transplanted with A431 xenografts. The size of the tumor was reduced by 53.4%, following 10 days of
intraperitoneal (i.p.) administration (100 mg/kg; i.p.). Consequently, these compounds are promising leads for
the design of potent and selective antitumor agents. The promising results of the flavins, 5-deazaflavins, and
their corresponding alloxazines analogues in addition to the current challenges of anticancer drug discovery
prompted us to design and develop new scaffolds with higher selectivity against specific kinase(s) in a wide
array of cancer cell lines. Thus, we thought to mimic the flexible structure of imatinib and its related flexible
analogues and to explore better binding affinity with a higher selectivity to different PKs. This design applies
multiple approaches to interfere with the hydrophobic pocket of the ATP binding site to enhance the polarity of
our potent antitumor alloxazine nucleus and to enhance the selective inhibition of PKs by fitting into the
allosteric site to selectively interact with the protein kinases containing glutamate residue (Glu286 as shown in
Figure 1), such as Abl, c-kit, and PDGF-R kinases. In the present study, we designed, synthesized, and
3

ACCEPTED MANUSCRIPT
determined different flexible derivatives of alloxazine that mimic the flexible structure of imatinib and its
analogues but with potentially better binding affinity and higher selectivity to PKs.
2
2
. Results and Discussion
.1. Structure-based design and lead optimization
The synthesis of compounds in this study was based on a structure-based design strategy, which was developed
based on the binding mode of different kinase inhibitors into their corresponding kinases. An initial approach to
drive potency was to design analogues, which can bind to the inactive conformation of different kinases, where
the DFG motif flips out to make a channel beyond the gatekeeper residue (Thr315) of Abl kinase. Regarding the
A-loop, the conformation of the DFG motif has the greatest effect on the binding pocket of the Abl kinase
3
1
domain. The side chain of the gatekeeper residue Thr670 in c-kit kinase should also be kept closer to the
backbone amide of Asp810 in the c-Kit DFG motif. Our drug design strategy was based on optimizing hits by
manipulating the lead compound I (7-methyl-2-(methylthio)pyrimido[4,5-b]quinolin-4(3H)-one) through the
following approaches:
a) To replace the flexible 4-(pyridin-3-yl)-N-o-tolylpyrimidin-2-amine moiety of imatinib by hetero-tricyclic
systems {benzo[g]pteridin-4(3H)-one or pyrimido[4,5-b]quinolin-4(3H)-one} to be capable of interfering with
hydrophobic pocket selectivity region 1 of the ATP binding site (Leu248, Tyr253, Val256, Leu370, Phe382).
b) To incorporate different analogues of the 4-(4-methylpiperazinylmethyl)benzylamine moiety of imatinib to
be able to form hydrogen bonds with Ile360 and His361, as well as to properly accommodate the allosteric site
to enhance selectivity against different cancer cell lines.
ry
c) To incorporate a 2-carbon spacer to allow the flexible cyclic 2 amine to protrude out of the active site to be
directed to the allosteric site, as shown in Figure 1.
A significant limitation was reported to be the lack of the aqueous solubility of different flavin analogues, which
3
0
limits their applications in preclinical studies. Accordingly, many of the designed derivatives contain a
4

ACCEPTED MANUSCRIPT
morpholine ring to enhance their water solubility, and the morpholine nitrogen allows the generation of water-
soluble amine salts.
Figure 1. The proposed binding mode of the designed compounds into the Abl domain to hinder ATP from
reaching its binding site. The hetero-tricyclic ring interacts hydrophobically with hydrophobic pocket 1, and the
flexible tail moiety at position 2 selectively accommodates into the allosteric site and hydrophobic pocket 2.
This approach improved IC of the lead compound {7-methyl-2-(methylthio)pyrimido[4,5-b]quinolin-4(3H)-
5
0
one}by 10 times, as shown in Figure 2 for the optimized lead 1 (unpublished data). In the present study, we are
going to further ameliorate IC by an additional four times to get optimized lead 2 (10b) of IC of 0.04 µM by
5
0
50
inserting the 4-((4-methylpiperazin-1-yl)methyl)phenyl)methanamine at position 2.
Figure 2. Two-phase lead optimization of 5-deazaalloxazine (lead compound I) to improve its IC50 against
MCF-7 from micromolar to nanomolar range.
Here under we described the structural optimization of the lead compound (I) by the design of 164 different
derivatives of 2-deoxo-alloxazines, 2-deoxo-5-deazaalloxazines, pyrimido[4,5-b]quinolin-4(10H)-one, and
5

ACCEPTED MANUSCRIPT
benzo[g] pteridine-2,4 (3H, 10H)-dione, which have different polar side chains at position 2 (see Figure 3, and
more details are shown in the supplementary data).
The small library of 164 compounds (1a-1dd, 2a-2dd, 3a-3mm, 4a-4i, 5a-5pp, 9a-9e, and 10a-10i; data are
1
2
3
presented in the supplementary data, Table 1S) was designed by changing the substituents R, R , R , R , and X
Figure 3) based on the literature database. Then the 3D structures of the designed compounds were energetically
minimized to be investigated as kinase inhibitors by docking into the crystal structure of Abl (PDB code:
(
3
1
32
33
34
35
2
hyy), c-kit (PDB code: 1t46), FAK (PDB code: 4q9s), Src (PDB: 4mxo), B-raf (PDB: 4rzv), and FLT1
3
6
(
PDB: 3hng) kinases in complex with their co-crystallized kinase inhibitors. GOLD software package version
3
7
5
.2.2 (Cambridge Crystallographic Data Centre, Cambridge, U.K.) was used in this study.
Our reported antitumor activities indicated that the alloxazine scaffold has a potential growth inhibitory activity
33
against MCF-7 breast cancer cells within IC50 of 0.04 µM. Therefore the target kinases FAK (PDB code: 4q9s)
3
4
and Src (PDB: 4mxo) were selected as templates for docking study because they are the expressed abundantly
3
8, 39-40, 39-41
in MCF-7 breast cancer cells
On the other hand our lead compounds were optimized by
incorporating the tail moiety: 4-(4-methylpiperazinylmethyl)benzylamine of imatinib at position 2 to improve
the anti-proliferative activities and selectivity of the optimized compounds. Therefore the other two kinases,
3
1
32
namely Abl (PDB code: 2hyy) and c-kit (PDB code: 1t46) were selected based on the fact that imatinib
3
5
targets specifically targets ABL and the receptor tyrosine kinases KIT. Other kinases B-raf (PDB: 4rzv), and
3
6
FLT1 (PDB: 3hng) were selected randomly for comparison.
The molecular docking study involving GOLD 5.2.2 as a powerful computational drug design tool (the detailed
results of these studies are included in the supplementary information) was applied, and the best candidates that
exhibited the highest binding affinities into the different kinases were selected for further chemical synthesis.
The short-listing process started by excluding the compounds that revealed the lowest GoldScore fitness. Out of
the 164 designed derivatives, 60 alloxazine, 5-deazaalloxazine, flavin, and 5-deazaalloxazines derivatives
6

ACCEPTED MANUSCRIPT
revealed
Figure 3. The lead compound (I) and the structures of the designed derivatives of 2-deoxo-alloxazines, 2-deoxo
-deazaalloxazines, pyrimido[4,5-b]quinolin-4(10H)-one, and benzo[g] pteridine-2,4 (3H, 10H)-dione.
5
the highest GoldScore fitness into six different kinases (Table 1). There were 14 synthetically feasible
compounds (9a-e and 10a-i) out of the best 60 candidates, which were selected for further chemical synthesis in
this study. Interestingly, these candidate compounds revealed multiple hydrogen bonds into the assigned kinases.
They interacted hydrophobically through their hetero-tricyclic rings {benzo[g]pteridin-4(3H)-one or
pyrimido[4,5-b]quinolin-4(3H)-one}with the amino acids of hydrophobic selectivity region 1 of the ATP binding
site.
Moreover, the flexible 4-((4-methylpiperazinylmethyl) benzylamine, 4-(morpholinomethyl) benzylamine, 4-
(piperidinyl methyl) benzylamine, 4-(dimethylaminomethyl) benzylamine, 1-(4-methoxyphenethyl) piperidin-
4
- amine, and 1-phenylethylamine moieties were oriented to protrude out of the active site to be directed to
7

ACCEPTED MANUSCRIPT
the allosteric site, as shown clearly in Figure 4. Additional and detail docking studies are presented under
Molecular docking study.
Table 1. The docking results for the top 10 ranks of the compounds docked into six different kinases in
comparison to their native bound inhibitors according to GOLD5.2.2 program.
Gold
Score
82.41
81.90
81.78
80.47
79.94
78.58
78.41
78.16
77.32
76.50
Gold
Score
78.41
77.92
77.62
76.44
76.36
74.59
73.72
73.12
72.91
72.32
78.19
Gold
Score
74.07
73.32
73.16
71.32
71.23
70.51
69.90
68.39
65.70
65.50
35.05
PK (pdb)
Compd.
PK (pdb)
Compd.
PK (pdb)
Compd.
9e
9b
9a
10b
9d
10g
10f
3hh
9c
3m
3l
5aa
9c
10g
10e
9d
1
0b
9a
9d
a
9b
b
c
Abl kinase
PDB: 2hyy)
c-kit kinase
(PDB: 1t46)
FAK kinase
(PDB: 4q9s)
6n
(
1
0e
1
0d
9e
6o
9e
10d
9b
9c
3kk
STI-571
g
h
Imatinib 102.08
30G
3
z
82.326
80.326
79.65
79.37
78.44
76.74
76.14
75.34
72.35
71.37
9e
9g
10f
9b
10b
10e
10a
9a
5aa
3aa
88.28
82.44
81.28
79.06
78.32
10f
10g
9b
9a
2i
9e
9d
10e
10i
2y i
8ST
80.26
79.91
76.16
74.55
74.53
74.51
73.80
73.17
73.13
72.82
86.37
3
aa
bb
3
3
w
e
f
3
u
B-raf kinase
(PDB: 4rzv)
FLT1 kinase
d
Src kinase
3
t
78.30 (PDB: 3hng)
(PDB: 4mxo)
3
v
75.22
74.83
74.80
74.78
3
s
9
d
3ee
Bosutinib 77.55
Vemurafenib 94.81
a
Human Abl kinase domain in complex with imatinib (STI571, Gleevec).
c-kit tyrosine kinase in complex with STI-571.
Human Focal Adhesion Kinase (FAK)
Human Src kinase bound to kinase inhibitor bosutinib.
BRAF (R509H) kinase domain monomer bound to Vemurafenib.
VEGFR1 in complex with N-(4-chlorophenyl)-2-((pyridin-4-ylmethyl)amino)benzamide
b
c
d
e
f
g
h
i
4
3
-(4-Methyl-piperazin-1-ylmethyl)-N-[4-methyl- 3-(4-pyridin-3-yl-pyrimidin-2-ylamino)-phenyl]-benzamide (imatinib).
,5-Dihydro[1,2,4]triazino[3,4-c][1,4]benzoxazin- 2(1H)-one.
N-(4-Chlorophenyl)-2-((pyridin-4-ylmethyl)amino)benzamide.
2
.2. Chemistry
The 5-deazaalloxazines (substituted 2-aminopyrimido[4,5-b]quinolin-4(3H)-one; 9a-e) and alloxazines
(substituted 2-amino benzo[g]pteridin-4(3H)-one;10a-i) were prepared as shown in Scheme 1. The precursors of
8

ACCEPTED MANUSCRIPT
6
-(N-alkylanilino)-2-methylthiopyrimidin-4(3H)-ones (3a–d) were prepared from 6-amino-2-thiouracil (1) by an
amine exchange route with the appropriate anilines at high temperature. The corresponding anilinium chloride
salts were used as catalysts to afford the 6-anilino-2-thioxo-2,3-dihydropyrimidin-4(1H)-one analogues (2a-d).
These 2-thioxo analogues were subjected to S-methylation at position 2 by MeI in an alkaline solution under a
chilling temperature to afford the 6-(N-alkylanilino)-2-methylthiopyrimidin-4(3H)-ones (3a–d) in yields of 69–
9
0%. The 2-deoxo-2-methylthio-5-deazaalloxazines (4a-d) were prepared by the Vilsmeier-Haack cyclization
Figure 4. (A) The molecular overlay view of Abl (PDB code: 2hyy), c-kit (PDB code: 1t46), FAK (PDB
code: 4q9s), Src (PDB: 4mxo), B-raf (PDB: 4rzv), and FLT1 (PDB: 3hng) kinases showing the binding site in
yellow surface view. (B) The binding mode of compound 9e into Abl kinase (PDB code: 2hyy), in which its
hetero-tricyclic ring was oriented into hydrophobic pocket region 1 of the ATP-binding site and orients its
flexible 4-(morpholinomethyl)benzylamine to protrude out of the active site to be directed to the allosteric
site.
reaction from their corresponding open 6-anilino derivatives (3a–d) to afford yellow needles in 62-87% yields.
On the other hand, 2-deoxo-2-methylthioalloxazine-5-oxides (5a–d) were prepared from their corresponding
open structure of 6-anilino derivatives (3a–d) by nitrosative cyclization to afford the orange-red products in 43–
9

ACCEPTED MANUSCRIPT
5
6% yields. For the preparation of the polar amine moiety of our target compounds, the key intermediate
4
2
benzonitrile derivatives were prepared with a modification of the reported method by reacting 4-bromomethyl
Scheme 1. Design of new flexible scaffold from alloxazine and 5-deazaalloxazine analogues. Reagents and
o
conditions: (a) NaOEt, EtOH, rt, 1h, then reflux, 2h; (b) Fusion, 170 C, 9h; (c) MeI, 2N NaOH, ice bath, 30
o
o
min; (d) POCl , DMF, 90 C, 1-2h; (e) HOAc, NaNO , rt, 2-5h; (f) DMF, rt, 0.5-3 h, then 80-90 C, 2-6h; (g)
3
2
o
THF, LiAlH , 25 C, 16-20 h; (h) DMF, reflux 7-9h.
4
benzonitile (6) with an equivalent quantity of appropriate secondary amines to afford 4-
1
0

ACCEPTED MANUSCRIPT
(aminomethyl)benzonitrile analogues (7a-d) in 72–98% yields. Furthermore, these benzonitrile analogues were
subjected to
a
drastic reduction byLiAlH4 overnight at room temperature to afford (4-
(aminomethyl)phenyl)methanamine derivatives (8a-d) in 41–82% yields. The 5-deazaalloxazines (9a-e) were
prepared by coupling reactions for the nucleophilic replacement of 2-methylthio group of 2-deoxo-2-
methylthio-5-deazaalloxazines (4a-d) by 4-(aminomethyl) phenyl) methanamine derivatives (8a, b and c). Then
the crude product was purified by Combiflash chromatography (0-30% of methanol in ethyl acetate) to afford
the title compounds as yellow solids in 18–69% yields. Similarly, the alloxazines (10a-i) were synthesized by
coupling
2-methylthioalloxazine
(5a-d)
and
the
appropriate
amine
derivatives
;
4-
(
(aminomethyl)phenyl)methanamine (8a and 8c), 4-(aminomethyl)phenyl)-N,N-dimethylmethanamine (8d), the
commercially available 1-(4-methoxyphen- ethyl)piperidin-4-amine (8e), or 1-phenylethan-1-amine (8f). The
crude products were further subjected for Combiflash chromatography (0-30% of methanol in ethyl acetate) to
afford the title compounds as yellow-orange solids in 23–69% yields. The newly synthesized compounds were
1
13
structurally elucidated by different spectrometric methods, including H NMR (300 or 500 MHz), C NMR (75
1
35
and 125 MHz), DEPT C
75 or 125 MHz), HRMS (ESI), or MS (ES+), in addition to the quantitative CHN elemental analysis. The
detected peaks of the different spectra were assigned for the corresponding atoms. This was achieved according
(
1
7-18, 22-23, 28-29
to our previously reported considerable flavin analogs
and the scientific logical and probabilistic
1
13
methods for the chemical shifts and coupling constants, and it was supported by the predicted H NMR and C
NMR provided by ChemDraw Professional (PerkinElmer) and Mnova NMR 8.0 to be rigorous.
2
.3. In-vitro Protein kinase inhibition assay
PK target profiling is an essential step in drug discovery to assert the specificity of the designed lead
compounds. PK target profiling utilizes a large and diverse panel of highly active protein-kinase and protein-
phosphatase targets. The protein kinase profiling study was conducted by KINEXUS Bioinformatics
Corporation, Vancouver, British Columbia, Canada. KINEXUS used a radioactive assay for profiling
1
1

ACCEPTED MANUSCRIPT
evaluation of five of our synthesized compounds (9c, 9a, 9b, 10b and 10f) against a panel of 20 protein kinases.
The protein kinase assay was performed at a 10 µM concentration in a single measurement, including the
reference compound imatinib (also known as Gleevec) by employing the standardized assay methodology in the
experiment. The results reported as percentage activity and percentage inhibition changes compared to the
control are presented in Figure 5 and in the supplementary data Table 2S. The intra-assay variability was
determined to be less than 10%. Inhibition of target activity by the compound was assigned negative (-) values,
while activation of target activity was assigned positive (+) values. The values with greater than 25% change in
activity compared to control are considered to be potentially significant. The profiling data revealed inhibitions
and activations at low to considerable levels, ranging from 1–59% for inhibitions and 1–60% for activations.
Potent inhibition (ranging from 30–59%) was observed with the five compounds, compared to the control
against a select number of the kinases tested, including compound 9c: ABL1 and FAK; compound 9a: ABL1
and FAK; compound 9b: CDK1/CyclinA1 and FAK; compound 10b: CDK1/CyclinA1; compound 10f:
CDK1/CyclinA1 and imatinib: ABL1, BRAF, CDK1/CyclinA1, c-KIT, FAK, FLT1, and SRC. Overall, only
four of the 20 kinases tested, including ABL1, CDK1/CyclinA1, FAK, and SRC showed inhibitions ranging
between 30–50% at 10 µM concentrations with some of the test compounds. At the lower end of the inhibition
range (from 20–30%), a few additional kinases were inhibited, such as ACVR2A, CK1 alpha1, and FLT1.
Out of all the kinases tested, only three kinases were suppressed by all of the compounds, including ACVR2A
(
5–25% inhibition), FAK (10–59% inhibition) and SRC (17–21% inhibition). Compound 9c showed the
greatest inhibition (59%) with FAK, followed by 9a (48%) and 9b (42%). Compound 9c also inhibited ABL1
44%) and 9a (33%). In addition, three compounds, 9b, 10b, and 10f had inhibitory effects (34–40%) on
(
CDK1/CyclinA1activities. All the other inhibitions observed for the test compounds were less than 30%.
In comparison, the reference compound, imatinib, resulted in reasonably high inhibitions with the kinases
ABL1 (88%) and FAK (62%) and moderate inhibitions with BRAF (44%), CDK1/Cyclin A1 (31%), c-KIT
(47%), FLT (42%), and SRC (30%). The synthesized compounds in this study are highly soluble in aqueous
1
2

ACCEPTED MANUSCRIPT
solutions due to their polar side chains at position 2, namely, 4-(4-methylpiperazinylmethyl)benzylamine, 4-
(morpholino methyl)benzylamine, 4-(piperidinyl methyl)benzylamine, 4-(dimethylaminomethyl)benzylamine,
1
-(4-methoxy
9
a
-33 -25 -7
-16 -16 -9
18
14
5
19
19
10
0
-11 -28
36 12
60 16
32 -3
-9 25
-7 -2
5
6
-48
-42
50 -25 -12
4
8
33
-2
-4
-5
-7
-7
18
9
0
2
7
-1
11
-5
3
5
4
-19
-17
-17
9
b
-6 -34 -21
37 -22
7
9
c
-44 -15
1
1
8
4
1
18 -6
-34 -1
-1 -59
7
8
-4 -15
-5
1
1
0b
0f
-4
8
-6 -17
-5 -12
-2 -44
-10
-21
8
1
1
12
-5
37
12 -7 -19
18
-21
-34
-31 11
9
46 -13 10
7
Imatinib -88
-47
9
-2 -62
-42 -16 -21 -11 10 10 10 -30
Figure 5. Protein kinase profiling for five compounds (9a, 9b, 9c, 10b, and 10f) at 10 µM concentration,
represented by percentage activity or percentage inhibition against 20 protein kinases in comparison with the
3
3
reference compound imatinib using the P-ATP radiometric assay method.
phenethyl)piperidin-4-amine, and 1-phenylethylamine. These polar moieties may be responsible for the
substantial potency and selectivity (30–59%). In contrast, some of the kinases were shown to elicit high
activations (30–60%) compared to the control values, with the following compounds: compound 9c: c-KIT and
1
3

ACCEPTED MANUSCRIPT
HER2; compound 9a: c-KIT and FLT2; compound 9b: c-KIT and FLT2; compound 10b: FLT3; and compound
1
0f: FLT3. Of the 20 tested kinases, c-KIT, FLT2, FLT3, and HER2 showed varying levels of activation at 10
µM concentration of the test compounds. Surprisingly, only one kinase, CAMK1, was activated (albeit at low
levels, 5–19%). Compound 9b had the highest activity of the tested compounds with a 60% activation above the
control substrate with c-KIT. Imatinib also showed slight activation with some of the tested kinases. Other listed
measurements for activations or inhibitions are considered insignificant. The activity of the c-Kit kinase is
tightly regulated in normal cells, whereas the deregulated c-Kit kinase activity is implicated in the pathogenesis
2
2
of human cancers. Based on PTK profiling data, compounds 9a, 9b, and 9c revealed activation of this c-kit
kinase by 36, 60, and 32%, respectively, as shown in Figure 5. Therefore, these compounds were studied for
their potential binding into c-kit kinase (PDB: 1t46). It was found that these compounds were bound to the
target kinase within RMSD < 2.5Å without a hydrogen bond detected. This binding mode may explain the
opposite activation profiling results of these tested compounds against c-kit kinase.
The PTK profiling study exhibited that the 5-deaza-alloxazine analogues (9a, 9b, and 9c) revealed the most
promising inhibitory effects of -48, -42, and -59%, respectively, against human focal adhesion kinase (FAK) in
comparison to 62% inhibition by imatinib. Highly correlated to these results, compound 9c, which has the
highest inhibitory effect against FAK kinase (-59%), exhibited the highest GOLD fitness score of 71.32 on its
docking into the corresponding FAK kinase (PDB: 4q9s) (data are presented in the supplementary data, Table
3
S).
2
.4. Cell proliferation assay:
Table 3. List of kinases and their expression in breast cancer and K562 cell lines
tissues
/
1
4

ACCEPTED MANUSCRIPT
The
Kinase
FAK
Other name
PTK2
Cell line (Reference)
MTT
3
8
MDA-MB-231 (Chen et al. )
MCF-7 (Li et al and Datta et al.
39-40
assay
)
3
9-41
SRC
FLT2
SRC
HCK
PKC
-
MCF-7, MDA-MB-231 (Li et al and Thakur et al
)
4
5-46
FGFR1
-
-
MCF-7 (Tarkkonen et al and Patel et al
)
4
7
K562 (Pene-Dumitrescu et al )
4
8
K562 (Pene-Dumitrescu et al )
4
9
K562 (Murray et al )
5
0-51
BCR-ABL
K562 (Karimiani et al and Luchetti et al
)
4
3
44
developed by Mosmann and modified by Miura was used to determine the inhibitory effects of test
compounds (9a-e, 10a,b, and 10d-i) against MCF-7, K562, A2780, HCT116, and WM793B cell lines. The selected
panel of cancer cells have different genetic backgrounds. MCF-7 breast cancer cells and K562 cell line were
selected because they express many of the target kinases in this study as shown in Table 3.
It was found that overexpression of receptor tyrosine-protein kinase HER2/neu occurs in 20% of invasive breast
cancers and has been found to predict a poorer prognosis than does normal expression. Several additional PKs
39-41
have been found to be expressed or overexpressed in breast cancer, including BRK, c-Src, and FGFR.
The
4
7
48
K562 cell line, derived from a chronic myeloid leukemia (CML) patient express SRC, , HCK, , PKC isotypes,
4
9
50-51
and B3A2 bcr-abl hybrid gene.
The ovarian cancer A2780 cells were utilized as conventionally known to
5
2
be used to study drug toxicity. Lapatinib is an FDA-approved treatment of HER2-positive metastatic breast
53
cancer, acting as a PTK-inhibitor , was used as a reference in this assay. Imatinib was also used as a refrence
5
4
compound by virtue of Kadivar et al. who demonstrated that imatinib has an antiproliferative effect and
induces
Table 4. The growth inhibitory activities (IC50; µM) against MCF-7,
K562, A2780, HCT116, and WM793B cell lines
1
5

ACCEPTED MANUSCRIPT
Compound MCF7
K562
A2780
HCT116 WM793B
9
a
0.53
0.62
11.61
2.34
3.13
1.49
0.04
0.20
0.55
0.16
0.80
0.16
5.73
7.12
6.80
13.54
0.32
0.82
0.72
13.00
7.23
5.22
1.53
0.05
0.31
1.30
0.17
0.93
0.14
5.78
19.66
10.12
10.54
> 50
38.98
11.01
49.09
> 50
ND
9
b
26.8
9.97
15.91
23.15
>50
0.54
10.11
1.52
2.89
2.01
0.05
0.19
0.70
0.20
0.90
0.34
4.83
8.90
10.40
9
c
9
d
9
e
1
0a
1
1
0b
0d
>50
>50
40.15
> 50
> 50
33.14
> 50
> 50
ND
1
0e
0f
0g
0h
0i
>50
1
>50
1
17.31
>50
1
1
>50
Imatinib
3.43
ND
Lapatinib
ND
ND: Not determined
apoptosis in breast cancer cell lines compared to normal cells. Our results demonstrated that 8 out of 13 (60%)
of the designed compounds showed IC less than 1.0 µM against MCF-7 breast cancer cells. The most active
5
0
compounds were compund 10b (IC =0.04 µM), compounds 10f and 10h (IC =0.16 µM), compound 10d
5
0
50
(
IC =0.2 µM), and compounds 9a and 10e (IC =0.53 and 0.55 µM), respectively, as depicted in Table 4.
50 50
Against the ovarian cancer A2780 cells, compounds 9a, 9b, 10b, 10e, 10f, 10g, and 10h exhibited
antiproliferative activity of less than 1.0 µM of IC50: 0.32, 0.54, 0.05, 0.19, 0.70, 0.2, 0.90, and 0.34 µM,
respectively (Table 4). Compound 10b revealed the most potent antiproliferative activity (IC50=0.05 µM)
against the same cell line in comparison to other compounds. Furthermore, we examined cell toxicity using
colon carcinoma HCT116 cells. Compounds 10b, 10f, and 10h revealed the most potent activities against the
assigned cells, with IC of 0.05, 0.17, and 0.14 µM, respectively. Comparatively, compunds 9a, 9b, 10d, and
5
0
1
0g still showed reasonable antiproliferative activities at concentrations less than 1.0 µM.
1
6

ACCEPTED MANUSCRIPT
Compounds 9a, 9c, 9d, and 10g exhibited the lowest IC50 of 13.54, 9.97, 15.91, and 17.31 µM, respectively
against The chronic myeloid leukemia K562 cell line in comparison to imatinib (IC50 = 3.43 µM) which is used

clinically (Gleevec ) to treat chronic myelogenous leukemia (CML). On the other hand, the melanoma cell line
WM793B was screened against our compounds, but the results declared that WM793B cells are resistant to the
tested compounds. Nonetheless, coumpunds 9a and 9d exhibited the maximum inhibitory effects (IC50=10.54
and 11.01 µM), which are considered as the highest activities compared to other compounds (Table 4). The
results of compound 9a against WM793B cells are compatible with its superior inihibtory effects revealed
against other cancer cell lines.
2
.5. Annexin V PI/FITC apoptosis assay
2
4, 55
Based on previous reports
which showed that flavin analogues are capable of activating and stabilizing the
tumor suppressor p53, and consequently induces both cell cycle arrest and apoptosis; we have conducted
annexin V Fluorescein Isothiocyanate/Propidium Iodide (FITC/PI) apoptosis assay. Compounds 10b, 10d, 10e,
1
0f and 10h were tested for their possible apoptosis inducing activities. MCF7 cells were labeled by annexin-V
PI/FITC, followed by flowcytometry analysis (BC, FC500). The compounds were selected for this assay
depending on their potent antiproliferative activities and selectivity as depicted in Figure 5 and Table 4,
respectively. Each of compounds 10e, 10f and 10d induced the early apoptotic MCF-7 cell populations in a
dose dependent manner by
1
7

ACCEPTED MANUSCRIPT
Figure 6. Annexin V FITC/PI apoptosis assay for MCF7 cells treated separately with compounds 10b, 10d,
1
0e, 10f, and 10h at 0, 1, 5 and 10µM (72h). X axis is annexin V and Y axis is PI for each compound. C1:
necrotic cells (PI+/ annexin V-); C2: late apoptotic cells (PI+/ annexin V+); C3: live cells (PI-/annexin V-); C4:
1
8

ACCEPTED MANUSCRIPT
early apoptotic cells (PI-/ annexin V+). Data shown is mean % cell number ± SD (n=3). Experiment was
repeated 3x.
two folds compared to the control as shown in Figure 6. Meanwhile, compound 10h at its lowest concentration
(1.0 µM) showed a significant early drug apoptosis of more than two folds compared to the control, whereas,
this activity decreased to 14% early apoptotic events at 5 and 10 µM. Regarding the late apoptotic changes,
compound 10f and 10h showed the best late apoptotic effects at 10µM concentration of 12 and 13 times more
than the control, respectively (Figure 6). Compound 10b showed no evidence of early apoptotic activity, and
that result is less consistent with its MTT assay and PK profiling activity against the panel 20 kinases.
Despite that some of the aforementioned apoptotic results are more or less consistent with the MTT assays,
which was conducted for a longer time, we cannot speculate to use the MTT assay to analyze the apoptosis
assay. MTT assay measures the rate of surviving cells and cannot distinguish the manner and stages of cell
5
6
death. Additionally, it can be suggested that compounds generally may reveal their antitumor effects by
downregulating any of the different hallmarks of cancer without affecting apoptosis.
2
.6. Structure activity relationship (SAR)
In this study, three main structural features of our compounds are discussed. These include the polar aryl amine
moiety at position 2, the substituents at position 5, and the electronic effects of substituents at position 7, as
depicted in Figure 7. Regarding MTT assay, compounds 9a, 9b, 10b, 10d, 10f, and 10h reveled the highest
activities against MCF-7, A2780, and HCT116 tumor cells. These results indicate that alloxazine skeleton (in
compounds 10b, 10d, 10f, and 10h) revealed ∼ 3–13 times more potent antiproliferative activities than 5-
deazaalloxazine skeleton (in compounds 9a and 9b). The alloxazine derivatives having 4-(4- methyl piperazinyl
methyl)benzylamine moiety (10b) at position 2 demonstrated up to five times more potent activities than
compounds with 4-(morpholinomethyl)benzylamine (10d), 1-(4-methoxyphenethyl) piperidin-4-amine (10f),
and 1-phenylethylamine (10h). In addition, compounds 9a and 9b with 4-(4-methylpiperazinylmethyl)
benzylamine moiety at position 2 exhibited a range of 6–20 times more potent antiproliferative activities than
1
9

ACCEPTED MANUSCRIPT
compounds with other moieties at the same position. The substituents with different electronic effects at
position 7, such as H, Me, OMe, and Cl, have less significant effects on the antiproliferative activity.
Nevertheless,
7-OMe
and
7-Me
Figure 7. The structural features of alloxazines and 5-deazaalloxazines manipulated to improve the selectivity,
antitumor activity, and apoptotic cell death.
substituents revealed the lowest IC in compounds 10b and 10f against MCF-7 cells (IC = 0.04 and 0.16,
5
0
50
respectively).
Regarding PK inhibition assay, compounds 9a and 9c similarly have the highest inhibition of FAK kinase by
8%
4
and 59% and ABL1 by 33 and 44%, respectively and compound 9b inhibits FAK by 42%. Compounds 9b, 10b,
and 10f inhibited CDK1/CyclinA1 equally by 34%. The PK inhibitory activities of the title compounds confirm
that at position 2, the following moieties: 4-(4-methylpiperazinylmethyl)benzylamine (9a, 9b, and 10b), 4-
(
piperidinyl methyl)benzylamine (9c), or (4-methoxyphenethyl) piperidin-4-amine (10f) of alloxazines or 5-
deazaalloxazines are responsible for the selective inhibition of FAK, ABL1, and CDK1/CyclinA1 kinases. The
-deazalloxazines (CH moiety at position 5) revealed a selectivity for inhibition of Abl-1 and FAK kinases
more than the alloxazines. In regard to the apoptotic effect, compound 10d with 4-
5
2
0

ACCEPTED MANUSCRIPT
(morpholinomethyl)benzylamine at position 2, induced late cell death significantly at 10µM by ∼ 7 times more
the early death. Furthermore, compound 10f and compound 10h (with 1-(4-methoxyphenethyl) piperidin-4-
amine) showed the best late apoptosis at 10µM concentration of 12 and 13 times more than the control,
respectively. Whereas, the structure features of 1-phenylethylamine at position 2 (10h) is correlated with the
early cell death.
2
2
.7. Molecular docking study of the best alloxazine candidates into PKs
.7.1. Validation of the accuracy and performance of GOLD docking program
The most straightforward method for validation of the used scoring function is to examine how closely the best-
59
docked conformation resembles the native co-crystallized ligands in the experimental biological method.
This study was initially performed by re-docking of the co-crystallized ligands, namely imatinib, STI-571
imatinib), 30G (3,5-dihydro [1,2,4] triazino[3,4-c][1,4]benzoxazin-2(1H)-one, bosutinib, vemurafenib, and
ST (N-(4-chlorophenyl)-2-((pyridin-4-ylmethyl)amino)benzamide) into their corresponding kinases of Abl, c-
(
8
kit, FAK, Src, B-raf, and FLT1, respectively. Almost all of the docked poses were superimposed on the natively
co-crystallized bound ligands within a root mean square deviation (RMSD) of 0.21, 1.30, 4.17, 0.92, 0.22, and
0
.37 Å, respectively, as shown in Figure 8 A-D (FAK; PDB: 4q9s and Src; PDB: 4mxo are not represented).
These results similarly emulate the experimental results, revealing successful docking protocol and confirming
the accuracy of the GOLD docking program for further docking of the designed hits.
2
.7.2. GoldScore fitness of the designed alloxazine candidates into PTKs
The Goldscore was implemented in this study as a search algorithmic function for “drug-like” substances
because it has been reported to afford superior docking results to the Chemscore as a scoring function for
3
7
GOLD. All the docking output results, including GoldScore fitness, protein-ligand intermolecular hydrogen
bonding S(hb-ext), and protein-ligand intermolecular Van der Waals interactions S(vdw-ext) for the top five
candidates of 9a-e and 10a-i are represented in the supplementary data, Table 3S. The docked compounds into
human Abl kinase domain (PDB: 2hyy) revealed Gold fitness scores ranging from 63.55–82.41. Compounds 9a,
2
1

ACCEPTED MANUSCRIPT
9
b, 9d, 9e, and 10b exhibited the most noteworthy GoldScore fitness of 79.94–82.41 in comparison to the
reference compound imatinib of 102.08. Many of these compounds were docked proximately superimposed on
the native co-crystallized ligand within 1.17-1.55Å. They docked into the binding site through up to two
hydrogen bonds to a higher extent with Thr315 (OH), Glu286 (C=O), and Met318 (NH) amino acids. In this
regard, compound 9a, which exhibited a high remarkable GoldScore fitness into the Abl kinase of 81.78, was
docked overlapping on
Figure 8. Validation of the accuracy of GOLD 5.2.2 for docking into Abl; (PDB: 2hyy; A), c-kit; (PDB: 1t46;
B), B-raf; (PDB: 4rzv; C), and FLT1; (PDB: 3hng; D) kinases. The docked poses are represented by ball and
stick view, colored by elements, and the native co-crystallized ligands are shown as yellow sticks. The
hydrogen bonds are depicted by green dashed lines.
2
2

ACCEPTED MANUSCRIPT
the native bound imatinib within RMSD of 1.24 Å. It was bound to the binding site of the titled kinase by one
hydrogen between its C=N moiety and OH of Thr315 amino acids, as shown in Figure 9. Compound 9c was
Figure 9. Docking of compound 9a (ball and stick, colored by element) into chain B of Abl kinase (PDB:
hyy). It exhibited one hydrogen bond (green dotted lines) with Thr315. It was docked within RMSD of 1.24 Å
2
from the native co-crystallized imatinib (yellow stick).
docked within RMSD of 1.58 Å of the co-crystallized 30G ligand and bound into the binding site by two
hydrogens between its carbonyl group at position 4 and its terminal piperidine-N atom with NH moieties of
Cys502 and Lys454 amino acids, respectively as shown in Figure 10.
Figure 10. Compound 9c is docked into FAK kinase (PDB:4q9s). It reveals two hydrogen bonds with Cys502
and Lys454. Its planar pyrimidoquinoline ring interacts hydrophobically with Ile428, Val436, and Ile553 and
the 2-benzyl moiety interacts by a π-cation electrostatic bond with the quaternary ammonium moiety of Lys454.
2
3

ACCEPTED MANUSCRIPT
Figure 11. LigPlot view of compound 10b docked into SH2 domain of Src kinase (pdb: 4mxo) within RMSD
of 2.14 Å. It binds into the binding site via two hydrogen bonds with Tyr340 and Ala390 and interacts
hydrophobically with the key amino acids Leu393, Val281, Lys295, and Leu273 of the binding site.
In addition, the hydrophobic interaction plays a crucial role to binding the pyrimidoquinoline ring of compound
9
c with Ile428, Val436, and Ile553. Moreover, the 2-benzyl moiety interacts by a π-cation electrostatic bond
with the quaternary ammonium moiety of Lys454 and interacts hydrophobically with the aliphatic hydrocarbon
moiety of Lys454. Accordingly, compound 9c demonstrated a Gold score (external vdw) of 46.43. Many of our
compounds exhibited mild human Src kinase inhibition of -17 to -21% using the radiometric PK-profiling
assay. Further molecular docking studies were conducted to determine their binding affinities and the mode of
their interactions into the target kinase. The best docked compounds were embedded into the intracellular Src
SH2 domain target within RMSD of 1.59–2.53Å. They bound through up to two hydrogen bonds, mainly with
Met341 and Tyr340 amino acids. Into the Src SH2 domain, compound 10b was docked to exhibit a GoldScore
fitness of 69.05, comparable to the score of the natively bound bosutinib of 77.55. Compound 10b was
2
4

ACCEPTED MANUSCRIPT
embedded into the binding domain within RMSD of 2.14 Å from the native co-crystallized bosutinib ligand. It
showed two hydrogen bonds between its 2-NH and 7-O moieties moieties and Tyr340 and Ala390 amino acids
and interacted hydrophobically with Leu393, Val281, Lys295, and Leu273, as shown in the LigPlot view of
Figure 11.
2
.9. Correlation between the GoldScore fitness and IC50
The analysis of the docking results revealed a high correlation between the antiproliferative activities against
the human breast adenocarcinoma (MCF-7) cells for compounds 9a, 9b, 9d, 10a, 10b, 10d, 10f, and 10i and
2
their GoldScore fitness into FAK kinase. The obtained correlation coefficient (R ) was 0.979, as illustrated in
Figure 12-A. Likewise, the docking results of compounds 9a, 9b, 10a, 10b, 10d, and 10i into SRC kinase
exhibited considerable correlations between their Gold fitness scores and the antiproliferative activities (IC )
5
0
against breast adenocarcinoma (MCF-7) and ovarian cancer cells (A2780). The obtained correlation coefficients
2
(
R ) were 0.8811 (not represented) and 0.8595 (Figure 12-B), respectively. These results indicated the high
consistency and correlation between the biological results and the GoldScore for compounds revealing the best
antiproliferative results against the used tumor cell lines and the best Po inhibitory activities against the target
kinases.
2
5

ACCEPTED MANUSCRIPT
Figure 12. Correlation between the antiproliferative activities (IC ; µM) and the GoldScore fitness for (A)
5
0
2
compounds (9a, 9b, 9d, 10a, 10b, 10d, 10f, and 10i) of (R ) was 0.979 between their fitness into FAK kinase
2
and IC against MCF-7 cells, (B) compounds (9a, 9b, 10a, 10b, 10d, and 10i) of (R ) was 0.8595 between their
5
0
fitness into SRC kinase and IC50 against A2780 cells.
. Conclusion
3
In this study, derivatives of 2-deoxo-5-deaza alloxazines (9a-e) and 2-deoxo-alloxazines (10a-i) were selected
and synthesized from different side chains at position 2 and substituents at position 7 with different electronic
properties. The corresponding precursors (4a-d) and (5a-d) were subjected for coupling reactions for the
nucleophilic replacement of their 2-methylthio groups by the appropriate amine, namely 4-((aminomethyl)phenyl)
methanamine derivatives (8a-c), 4-(aminomethyl)phenyl)-N,N-dimethylmethanamine (8d), and the commercially
available 1-(4-methoxyphenethyl) piperidin-4-amine (8e), or 1-phenylethan-1-amine (8f). Combiflash
chromatography was utilized for purification of the polar products in 18–69% and 23–69% yields, respectively.
Furthermore, compound inhibition profiling was undertaken to measure the effect of five test compounds (9a,
®
9
b, 9c, 10f, and 10f) in addition to the reference compound imatinib (Gleevec ) against 20 proteins kinases.
2
6

ACCEPTED MANUSCRIPT
The profiling data showed potent inhibition (ranging from 30–59%) with ABL1, CDK1/Cyclin A1, FAK, and
SRC. Compound 9c exerted the highest inhibition of 59% for FAK and a 44% inhibition for ABL1. Compound
9
a inhibited FAK and ABL1 by 48% and 33%, respectively. Similarly, compound 9b inhibited FAK by 42%.
Compounds 9b, 10b, and 10f inhibited CDK1/CyclinA1 equally at the 34% range. The ACVR2A, FAK, and
SRC kinases were inhibited at different levels by all compounds. Not surprisingly, the reference compound,
imatinib, had the highest inhibitions with ABL1 (88%) and with FAK (62%) and moderate inhibitions with
BRAF (44%), CDK1/Cyclin A1 (31%), c-KIT (47%), FLT1 (42%), and SRC (30%). Some of the compounds
exhibited low to moderate activations for a few of the tested kinases, including c-KIT, FLT2, FLT2, and HER2.
Compound 9b was observed to have the greatest activation of c-KIT at 60%, followed by compounds 9c and 9a,
with 32–36% inhibitions. FLT2 kinase was activated by 50% by compound 9a and 37% by 9b. Compound 10f
was recorded with a 46% activation of FLT3 followed by 37% with 10b. Lastly, compound 9c had a 33%
activation of HER2. All the other activations noted were less than 20%. Many of our designed and synthesized
compounds demonstrated significant apoptotic changes at different concentrations. Notably, compound 10h
induced early death of MCF-7 cells (145.6%) at the lowest concentration (1.0µM). Likewise, compounds 10e
and 10f induced early cell death at 5.0µM (88.9 and 100.0%) and 10µM (145.6 and 130.0%), respectively.
Regarding the late apoptotic effects, compounds 10f and 10h exhibited dramatic cell death of 1100 and 1180%,
respectively, at 10.0µM concentration. In addition, compound 10d at 10µM induced late cell death by 556% in
comparison to the control. Even though compound 10b exhibited high antiproliferative activities (IC : 0.04-
5
0
0
.05µM) and kinase inhibition activities of -34% (CDK1/CyclinA1), it has a less significant anti-apoptotic
activity using Annexin V/FITC assay. The antiproliferative activity, measured by MTT assay, may result from
5
7
cell death and/or inhibition of proliferation; however, cell apoptosis is only one type of cancer hallmark.
Therefore, evaluating the anti-tumor activity of chemical compounds by MTT assay may not be enough, and
further biological tests are needed. Hence, further studies are needed to address the mechanism by which our
compounds reveal their antitumor effects other than the apoptotic pathway.
2
7

ACCEPTED MANUSCRIPT
In summary, our current drug design strategy was initiated to improve the antitumor activity of our lead
compound—7-methyl-2-methylthio-pyrimido[4,5-b]quinolin-4(3H)one and its derivatives—by replacing the 2-
2
9
methylthio moiety by the more polar 2-hydroxyethylamino moiety. This approach improved IC by 10 times,
5
0
for the optimized lead (I). In the present study, the IC was improved by an additional four times to get the
5
0
optimized lead
2
(10b) of IC50 of 0.04 µM by inserting the 4-((4-methylpiperazin-1-
yl)methyl)phenyl)methanamine in position 2. In addition, the overall results suggest that the SAR for inhibition
of PK and antiproliferative activity in substituted alloxazines and 5-deazaalloxazines derivatives are
multifactorial events. The moieties at position 2 have the most significant influence on antiproliferative and PKI
activities, whereas the substituents at position 7 have a less significant effect on these activities. In addition, the
5
-deazaalloxazines have a higher selectivity for Abl-1 and FAK kinases, more than the corresponding
alloxazines.
The comparison with the biological activity, along with the docking binding affinities, and interactions of the
synthesized compounds afford more insight into the mode of interaction of the designed compounds into the
target kinases. This study justified the differential activation and inhibition of the target kinases. For instance,
the compounds exhibiting an activation of the c-kit kinase did not show any hydrogen bonding with the target
kinase, whereas the inhibitors revealed hydrogen bonding specifically with Glu640 amino acids. Moreover, the
correlation between the docking GoldScore fitness and the antiproliferative activity (IC ; µM) were conducted.
5
0
2
The obtained correlations were considerable (R : 0.979), especially for docking of compounds 9a, 9b, 9d, 10a,
1
4
0b, 10d, 10f, and 10i into FAK kinase.
. Experimental
Materials quality control & reagents
The mass spectra were recorded on GCMS-QP 1000 EX Shimadzu Gas Chromatography MS spectrometer,
E.I.70 ev. Microanalyses were measured by Automatic CHN analyzer, Vario E1
Ш, Elementary-Germany at the
Microanalytical unit. All reagents were of commercial quality and were used without further purification.
2
8