First-in-class DAPK1/CSF1R dual inhibitors: Discovery of 3,5-dimethoxy-N-(4-(4-methoxyphenoxy)-2-((6-morpholinopyridin-3-yl)amino)pyrimidin-5-yl)benzamide as a potential anti-tauopathies agent

Article abstract of DOI:10.1016/j.ejmech.2018.10.057

Kinase irregularity has been correlated with several complex neurodegenerative tauopathies. Development of selective inhibitors of these kinases might afford promising anti-tauopathy therapies. While DAPK1 inhibitors halt the formation of tau aggregates and counteract neuronal death, CSF1R inhibitors could alleviate the tauopathies-associated neuroinflammation. Herein, we report the design, synthesis, biological evaluation, mechanistic study, and molecular docking study of novel CSF1R/DAPK1 dual inhibitors as multifunctional molecules inhibiting the formation of tau aggregates and neuroinflammation. Compound 3l, the most potent DAPK1 inhibitor in the in vitro kinase assay (IC₅₀ = 1.25 μM) was the most effective tau aggregates formation inhibitor in the cellular assay (IC₅₀ = 5.0 μM). Also, compound 3l elicited potent inhibition of CSF1R in the in vitro kinase assay (IC₅₀ = 0.15 μM) and promising inhibition of nitric oxide production in LPS-induced BV-2 cells (55% inhibition at 10 μM concentration). Kinase profiling and hERG binding assay anticipated the absence of off-target toxicities while the PAMPA-BBB assay predicted potentially high BBB permeability. The mechanistic study and selectivity profile suggest compound 3l as a non-ATP-competitive DAPK1 inhibitor and an ATP-competitive CSF1R inhibitor while the in silico calculations illustrated binding of compound 3l to the substrate-binding site of DAPK1. Hence, compound 3l might act as a protein-protein interaction inhibitor by hindering DAPK1 kinase reaction through preventing the binding of DAPK1 substrates.

Full text of DOI:10.1016/j.ejmech.2018.10.057

Accepted Manuscript
First-in-class DAPK1/CSF1R dual inhibitors: Discovery of 3,5-dimethoxy-N-(4-(4-
methoxyphenoxy)-2-((6-morpholinopyridin-3-yl)amino)pyrimidin-5-yl)benzamide as a
potential anti-tauopathies agent
Ahmed Karam Farag, Ahmed H.E. Hassan, Hyeanjeong Jeong, Youngji Kwon, Jin
Gyu Choi, Myung Sook Oh, Ki Duk Park, Yun Kyung Kim, Eun Joo Roh
PII:
S0223-5234(18)30926-7
DOI:
https://doi.org/10.1016/j.ejmech.2018.10.057
EJMECH 10843
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 1 August 2018
Revised Date: 15 October 2018
Accepted Date: 23 October 2018
Please cite this article as: A.K. Farag, A.H.E. Hassan, H. Jeong, Y. Kwon, J.G. Choi, M.S. Oh, K.D.
Park, Y.K. Kim, E.J. Roh, First-in-class DAPK1/CSF1R dual inhibitors: Discovery of 3,5-dimethoxy-N-
(4-(4-methoxyphenoxy)-2-((6-morpholinopyridin-3-yl)amino)pyrimidin-5-yl)benzamide as a potential
anti-tauopathies agent, European Journal of Medicinal Chemistry (2018), doi: https://doi.org/10.1016/
j.ejmech.2018.10.057.
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
DAPK1 substrate-
binding site
Compound 3l
CSF1R IC50 = 0.15 µM.
DAPK1 IC = 1.25 µM.
•
•
•
•
•
•
5
0
Non-ATP-competitive inhibition of DAPK1.
No inhibition of any of the DAPK or PDGFR families.
Inhibition of tau aggregation IC50 = 5.0 µM.
55% inhibition in NO production in LPS-stimulated
BV-2 microglial cells.
•
•
No inhibition for hERG.
PAMPA-BBB = high permeability.

ACCEPTED MANUSCRIPT
First-in-Class DAPK1/CSF1R Dual Inhibitors: Discovery of 3,5-Dimethoxy-N-
(4-(4-methoxyphenoxy)-2-((6-morpholinopyridin-3-yl)amino)pyrimidin-5-
yl)benzamide as a Potential Anti-Tauopathies Agent
a,b,1
c,d
e,f
g
Ahmed Karam Farag , Ahmed H. E. Hassan , Hyeanjeong Jeong , Youngji Kwon ,
h
g
b,i
b,e
a,b*
Jin Gyu Choi , Myung Sook Oh , Ki Duk Park , Yun Kyung Kim , Eun Joo Roh
a
b
Chemical Kinomics Research Center, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea.
Division of Bio-Medical Science &Technology, KIST School, University of Science and Technology, Seoul 02792, Republic of
Korea.
c
Department of Medicinal Chemistry, Faculty of Pharmacy, Mansoura University, Mansoura, 35516, Egypt.
d
Medicinal Chemistry Laboratory, Department of Pharmacy, College of Pharmacy, Kyung Hee University, Seoul 02447,
Republic of Korea.
e
Brain Science Institute, Convergence Research Center for Diagnosis, Treatment and Care System of Dementia, Korea Institute
of Science and Technology (KIST), Seoul, 02791, Republic of Korea.
f
Department of Life Science, Korea University, Seoul, 02841, Republic of Korea.
g
Department of Life and Nanopharmaceutical Sciences, Graduate School, Kyung Hee University, Seoul, Republic of Korea.
h
BK21 PLUS Integrated Education and Research Center for Nature-inspired Drug Development Targeting Healthy Aging,
Kyung Hee University, Seoul, Republic of Korea.
i
Convergence Research Center for Diagnosis, Treatment and Care System of Dementia, Korea Institute of Science and
Technology (KIST), Republic of Korea.
*
Corresponding author: Eun Joo Roh
Email address: r8636@kist.re.kr
Postal address: Korea Institute of Science and Technology, Materials and Life Science Research
Division, Chemical Kinomics Research Center, Hwarang-ro 14-gil 5, Seongbuk-gu, Seoul,
Republic of Korea, 02792.
1
Current affiliation and address: RI Translational Research Team, Division of Applied RI, Korea Institute of Radiological and
Medical Sciences (KIRAMS), 75 Nowon-ro, Nowon-gu, Seoul, Republic of Korea, 01812

ACCEPTED MANUSCRIPT
Abstract
Kinase irregularity has been correlated with several complex neurodegenerative tauopathies.
Development of selective inhibitors of these kinases might afford promising anti-tauopathy
therapies. While DAPK1 inhibitors halt the formation of tau aggregates and counteract neuronal
death, CSF1R inhibitors could alleviate the tauopathies-associated neuroinflammation. Herein,
we report the design, synthesis, biological evaluation, mechanistic study, and molecular docking
study of novel CSF1R/DAPK1 dual inhibitors as multifunctional molecules inhibiting the
formation of tau aggregates and neuroinflammation. Compound 3l, the most potent DAPK1
inhibitor in the in vitro kinase assay (IC50 = 1.25 µM) was the most effective tau aggregates
formation inhibitor in the cellular assay (IC50 = 5.0 µM). Also, compound 3l elicited potent
inhibition of CSF1R in the in vitro kinase assay (IC = 0.15 µM) and promising inhibition of
5
0
nitric oxide production in LPS-induced BV-2 cells (55% inhibition at 10 µM concentration).
Kinase profiling and hERG binding assay anticipated the absence of off-target toxicities while
the PAMPA-BBB assay predicted potentially high BBB permeability. The mechanistic study and
selectivity profile suggest compound 3l as a non-ATP-competitive DAPK1 inhibitor and an
ATP-competitive CSF1R inhibitor while the in silico calculations illustrated binding of
compound 3l to the substrate-binding site of DAPK1. Hence, compound 3l might act as a
protein-protein interaction inhibitor by hindering DAPK1 kinase reaction through preventing the
binding of DAPK1 substrates.
Keywords
DAPK1; CSF1R; Tauopathies; Neuroinflammation; multifunctional molecules.

ACCEPTED MANUSCRIPT
1
. Introduction
The decline of mental capacities caused by chronic neurodegenerative disorders is the largest
global source of disabilities producing more than 20% of the “Years Lived with Disability
YLD)” which is a measurement for a disease burden [1]. Among neurodegenerative diseases,
(
tauopathies emerge as a large class of more than twenty-five diseases characterized by
neurodegeneration accompanied by the formation of filamentous aggregates of
hyperphosphorylated microtubule-associated protein (MAP) tau within the CNS [2, 3].
Alzheimer's disease (AD) is the most common tauopathy, while relatively less encountered
tauopathies include post-encephalitic parkinsonism, Pick’s disease, progressive supranuclear
palsy (PSP) and corticobasal degeneration (CBD) [4, 5]. The currently available therapies
provide only symptomatic treatments without halting or slowing down the progression of these
diseases. Certainly, there is an urgent need to develop an effective treatment for this class of
diseases that address their underlying causes.
Despite the high solubility of tau proteins, anomalously hyperphosphorylated tau protein
aggregates into neurofibrillary tangles (NFTs), the hallmark of tauopathies [6]. It is known that
this hyperphosphorylation converts the normally functioning tau protein into a neurotoxin that
triggers neuronal death. Accordingly, research to identify small molecule inhibitors of tau
hyperphosphorylation is a promising strategy towards the development of an effective therapy
[6, 7]. Death-associated protein kinase 1 (DAPK1) was recently discovered as a valid therapeutic
target for neurodegenerative diseases [8-10]. The role of DAPK1 as a novel regulator of tau
phosphorylation indicates that its upregulation might be involved in tauopathies. Recent reports
illustrated that DAPK1 regulates tau protein by direct phosphorylation of Thr231, Ser262, and

ACCEPTED MANUSCRIPT
Ser396 and indirectly via phosphorylating tau regulatory proteins such as peptidyl-prolyl cis-
trans isomerase NIMA-interacting 1 (PIN1) and microtubule affinity-regulating kinase 1/2
(MARK1/2) [8]. These actions are believed to mediate tau protein hyperphosphorylation and
subsequent formation of NFTs. In addition, DAPK1, being a strong regulator of apoptosis and
autophagy mediates neuronal cell death via phosphorylating N-myc downstream-regulated gene
2
(NDRG2) ultimately triggering neuronal cell death. In fact, anomalous deregulation and
hyperphosphorylation of tau protein are strongly correlated with DAPK1 activation [11-13].
Therefore, inhibition of DAPK1 might be beneficial for the treatment of tauopathies. To the best
of our knowledge, no clinically useful DAPK1 inhibitor has been realized yet despite the
validation of DAPK1 as a viable drug target [14].
In addition to NFTs of hyperphosphorylated tau protein, reactive gliosis; a continuous glial
cells activation is a hallmark that characterizes tauopathies. It establishes a chronic
neuroinflammatory process resulting in the production of inflammatory mediators that eventually
contribute to the progression of the disease [15-18]. In contrast to the previous belief that
neuroinflammation arises as a result of neurodegeneration, recent evidence showed that the
activation of microglia and neuroinflammation precede the aggregation and formation of tau
tangles [19]. In a repetitive cycle, the formed aggregates trigger a more inflammatory response
that produces more inflammatory mediators which contribute to the formation of more protein
aggregates through various mechanisms [20].
Microglia are innate immune cells residing in the brain. They exist in a dormant state under
non-pathological conditions [21]. However, the chronic activation of microglia results in a
reactive gliosis which contributes to the neurodegeneration observed in several tauopathies such
as AD, parkinsonism, Pick’s disease and chronic traumatic encephalopathy [22-24]. Culminated

ACCEPTED MANUSCRIPT
pieces of evidence showed that the inflammatory component is crucial for the initiation and
development of tauopathies [25-28]. The kinase known as colony stimulating factor 1 receptor
(CSF1R; also known as macrophage colony-stimulating factor receptor; M-CSFR) is a key
regulator of survival and proliferation of the microglial cells [29, 30]. Several reports showed
that inhibition of CSF1R results in a reduction of microglia-dependent neuroinflammation and
slowing down the progression of neurodegenerative diseases [31, 32]. In addition, the use of a
CSF1R inhibitor to deplete microglia was found to inhibit the transmission of tau protein from
neuron to neuron and reduce the microglial-assisted neurotoxicity [33]. Collectively, these
reports emphasize the significance of developing CSF1R inhibitors targeting microglia as
potential therapeutics for neurodegenerative diseases.
For successful evasion of the vicious cycle of neuroinflammation-protein aggregates
formation, it might be more useful to target both of the two major elements of the cycle. Towards
achieving this goal, two strategies might be possible: development of a combination therapy or
development of a multifunctional single molecule targeting both NFTs formation and
neuroinflammation. A multifunctional single molecule is believed to be a more favorable
strategy in terms of pharmacokinetic and pharmacodynamic advantages, as well as, lowering side
effects and toxicity [34, 35]. Bearing these aims and concepts in our minds, we sought the
development of a single small molecule inhibitor of both DAPK1 and CSF1R as first-in-class
inhibitors of tau phosphorylation and neuroinflammation. Herein, we report our approach and
promising results.
2
. Results and Discussion
.1. Ligands design
2

ACCEPTED MANUSCRIPT
Recently, an imidazopyridazine inhibitor of DAPK1 (1; Figure 1) that binds the ATP binding
site of DAPK1 was reported as a starting point for the development of DAPK1 inhibitors as
potential neuroprotective agents [36]. The co-crystal structure of compound 1 with DAPK1
(PDB code: 4TXC) showed that N-1 of the imidazopyridazine moiety anchors the Val96 of the
hinge region with a single hydrogen bond. The back hydrophobic pocket I of the ATP binding
site is occupied by the hydroxylmethoxyphenyl moiety of compound 1 showing several polar
hydrogen bonding interactions, while the hydrophobic pocket II of the ATP binding site is vacant
and the aliphatic chain extends toward the solvent exposure area.
On the other hand, we have previously reported a series of potent phenoxypyrimidine
scaffold-based inhibitors of CSF1R [37]. In the conducted molecular simulation study reported
for our previously developed CSF1R inhibitors, compound 2 showed binding mode within the
ATP binding site of CSF1R in which the 4-morpholinophenyl moiety was directed towards a
hydrophobic pocket and the substituent on the phenoxy moiety showed interaction with β3 near
the hinge region. Meanwhile, the 3,5-dimethoxyphenyl moiety of compound 2 was directed
towards the solvent exposure region. This might indicate that manipulation of the (3,5-
dimethoxybenzylidene)amino moiety at position 5 while retaining a substituted-phenoxy moiety
at position 4 and a substituted phenyl moiety attached to the amino group of the core pyrimidine-
2
-amine might maintain the CSF1R inhibitory profile. Accordingly, modified molecules
possessing the general structure 3 might act as CSF1R inhibitors (Figure 1). Based on this, we
have used compound 2 as a starting point to develop small molecules dual inhibitors of DAPK1
and CSF1R.
In this regard, we anticipated that the pyrimidine-2-amine core of the modified compounds (3)
might functionally replace the imidazopyridazine moiety of compound 1 in binding the hinge

ACCEPTED MANUSCRIPT
region, while the substituted-phenyl moiety on the amino group might occupy the vacant
hydrophobic pocket II of the ATP binding site of DAPK1. The imine linker of compound 2
might be replaced by an amide group to keep a proper distance between the hinge binder and the
hydrophobic pocket-binding moiety. As the hydroxylmethoxyphenyl moiety of compound 1
showed several polar hydrogen bonding interactions within the hydrophobic pocket I of the ATP
2
binding site of DAPK1, it was planned to set R to either a phenyl group substituted with groups
capable of forming polar interactions; or to a simple amino group. The latter was planned to
explore whether only polar hydrogen bonding interactions of the introduced amino group is
sufficient for activity or a hydrophobic group is also necessary. Additionally, in the proposed
design, it was planned to explore the impact on activity resulting from:
1
(
(
(
i)
Variation of the R substituent at the phenoxy moiety,
Replacement of the morpholino moiety with other groups,
ii)
iii) Incorporation of nitrogen atom into the phenyl moiety attached to the amino
group of the core pyrimidine-2-amine.

ACCEPTED MANUSCRIPT
Figure 1. The proposed design of DAPK1-CSF1R dual inhibitors.
2
.2. Chemistry
The targeted compounds were prepared according to Scheme 1. The required starting 2-
chloro-5-nitro-4-phenoxypyrimidine derivatives 4a–c were prepared as reported previously [37].
Nucleophilic aromatic substitution of the chlorine atom at position 2 of derivatives 4a–c with
various aromatic amines yielded 5-nitro-4-phenoxy-N-aryl pyrimidin-2-amine derivatives 5a–f.
Palladium-catalyzed hydrogenation of the nitro group of the derivatives 5a–f afforded the
corresponding amino derivatives 6a–f. The urea derivatives 3a–f were prepared by stirring an
acetic acid solution of compounds 6a–f with a concentrated aqueous solution of excess
potassium cyanate at room temperature. The reaction of the amino derivatives 6a–f with various
acid chlorides at low temperature followed by warming to ambient temperature afforded the
primary amides 3g–l. Addition of the acid chlorides at low temperature was crucial for

ACCEPTED MANUSCRIPT
minimizing the formation of the di-acylated products which were detected when only ice bath
was used resulting in a considerable reduction of the obtained yield.
Scheme 1. Synthesis of compounds 3a–l. Reagents and conditions: (a) appropriate amine,
pyridine, THF, 80 °C, 4 h; (b) H , 10% Pd/C, MeOH, rt, 12 h; (c) for compounds 3a-f: potassium
2
cyanate, glacial acetic acid, rt, 2 h; for compounds 3g-l: appropriate acid chloride, DIPEA, DCM,
–
78 °C to rt, 2 h.
The arylamine derivatives required for the preparation of compounds 5a–f were prepared as
depicted in Scheme 2 following the reported procedures [38-40]. Thus, the activated halogens
chloro or fluoro) at the para- position to the nitro group of compounds 7a,b were displaced with
(
secondary amines such as morpholine or piperidine in a nucleophilic aromatic substitution
reaction to provide nitro derivatives 8a–c. Catalytic hydrogenation of the nitro derivatives
afforded the required arylamine derivatives 9a–c.

ACCEPTED MANUSCRIPT
Scheme 2. Synthesis of compounds 9a–c. Reagents and conditions: (a) appropriate amine,
2
THF, 70 °C, 2 h (b) H , 10% Pt/C, MeOH, rt, 12 h.
2
.3. In vitro kinase assay
In an initial attempt to evaluate the efficacy of the prepared compounds on the kinase reaction
of CSF1R and DAPK1, the compounds were assayed in a single-dose mode (10 µM except for
compounds 3a and 3b) in a duplicate radioisotope-based assay (Table 1). In all cases, ATP
concentration of 10 µM was used and the in vitro kinase assay was conducted according to
SM
Reaction Biology Corporation’s HotSpot methodology. This assay method is a radiometric
assay that directly measures the phosphorylated substrate that is the direct outcome of the kinase
catalytic activity [41]. Compounds eliciting high percent of inhibition of the kinase reaction were
subsequently evaluated for their respective IC50 values to measure their potencies.
Table 1. Substitution pattern, Percent inhibition values and IC50 values for the newly
synthesized compounds over CSF1R and DAPK1.
CSF1R
DAPK1
% Inhibition ±
1
2
Compd.
R
X
Y
R
%
Inhibition ±
IC ± SD
IC ± SD
50
5
0
a
b
a
b
SD values
values (μM)
SD values
values (μM)
c
d
3
a
b
4-CH O
CH
CH
NH2
NH2
65.02 ± 0.70
59.61 ± 0.29
4.42 ± 0.08
10.79 ± 7.10
44.91 ± 3.10
ND
3
d
c
3
4-CF3
ND
14.30 ± 6.31

ACCEPTED MANUSCRIPT
d
d
d
3
c
3- CH O
CH
CH
N
NH2
NH2
77.58 ± 0.64
20.99 ± 6.86
35.43 ± 0.56
1.29 ± 0.02
15.83 ± 2.63
17.11 ± 0.08
19.71 ± 2.04
ND
ND
ND
3
d
3
d
e
4- CH O
F
ND
3
d
3
4- CH
3
O
NH
2
ND
d
d
3
f
4- CH O
N
NH2
28.44 ± 2.99
ND
31.71 ± 4.25
ND
3
c
c
3
3
g
4- CH O
CH
74.70 ± 1.28
4.89 ± 0.18
0.12 ± 0.003
0.78 ± 0.04
1.72 ± 0.04
69.70 ± 1.19
2.61 ± 0.31
2.77 ± 0.41
7.44 ± 2.81
2.69 ± 0.21
3
h
4- CH O
CH
CH
N
95.34 ± 1.28
91.85 ± 1.31
75.99 ± 0.49
65.95 ± 2.21
75.06 ± 1.26
67.28 ± 2.95
3
3
i
j
4- CH O
3
3
4- CH O
3
d
3
k
4- CH
3
O
CH
N
F
35.20 ± 6.99
ND
65.00 ± 0.70
66.22 ± 0.26
4.29 ± 0.23
3
l
4- CH O
77.94 ± 0.52
0.15 ± 0.006
1.25 ± 0.35
3
a
b
c
d
The mean % inhibition of 10 µM concentration of two independent experiments.
The mean values of IC50 calculated from duplicate 10-dose IC50 assay with 3-fold serial dilution starting at 20 µM.
The mean % inhibition of 20 µM concentration of two independent experiments.
ND, not determined.
2
First, the urea compounds 3a–c in which R was set to amino group and the arylamine moiety
was set to 4-morpholinoaniline were prepared and evaluated. As shown in Table 1, the 10 µM
dose of these compounds were capable of inhibiting the kinase reaction of CSF1R by around
6
0% or more. Among these compounds, the best inhibition of CSF1R kinase reaction was
elicited when the substitution on the phenoxy moiety was set to 3-methoxy (compound 3c;
inhibition=77.58%), while the 4-trifluoromethyl substitution resulted in the least effective
CSF1R inhibitor (compound 3b; inhibition=59.61%). The 10 µM concentration of compound 3a
which possesses a 4-methoxy substituent on the phenoxy moiety elicited average inhibition of

ACCEPTED MANUSCRIPT
CSF1R (inhibition=65.02%). However, it elicited micromolar IC50 value for CSF1R inhibition.
Compound 3c showed higher IC value (1.29 µM) which is consistent with its higher inhibition
5
0
percent. When these three compounds were tested for inhibition of DAPK1 kinase reaction,
compound 3c that elicited the highest potency against CSF1R showed a very low inhibition
percent against DAPK1. Accordingly, compounds 3a and 3b were tested for DAPK1 inhibition
at a concentration of 10 µM. Nevertheless, both compounds 3a and 3b elicited low inhibition of
DAPK1 with a measured IC50 value of 14.30 µM for compound 3b, which elicited the highest
DAPK1 inhibition among these three compounds. Because compound 3b, which possesses a 3-
trifluoromethyl substituent on the phenoxy moiety was the least effective CSF1R inhibitor
among these compounds, while compounds bearing methoxy substituents were more effective
CSF1R inhibitors, all further synthesized compounds possessed 4-methoxy substituent on the
phenoxy moiety. Thus, compounds 3d–f were prepared via modification of the 4-
morpholinoaniline moiety. In vitro kinase assay showed that replacement of the morpholine
moiety by fluorine atom afforded compound with reduced efficiency for inhibition of CSF1R,
yet did not significantly enhance the measured inhibition percent of DAPK1 (compound 3d;
Table 1). Similarly, replacement of the aniline moiety by the heterocyclic aminopyridine alone
(
compound 3e; Table 1) or when combined with changing the morpholine moiety into piperidine
compound 3f; Table 1) did not result in compounds eliciting any promising inhibition percent
(
against both of DAPK1 and CSF1R.
2
Next, amide compounds in which R was set to a substituted-phenyl moiety were prepared
and evaluated. Because we found that the 4-morpholinoaniline, the 4-methoxyphenyl as
arylamine, and the phenoxy moieties relatively afforded compounds with viable inhibitory
activity against CSF1R, these moieties were retained in the prepared amide compounds 3g–l,

ACCEPTED MANUSCRIPT
2
while R was varied. The results of the in vitro kinase assay showed that all of these amide
compounds retained the inhibitory activity against CSF1R but gained a promising inhibitory
activity against DAPK1. While the 3,5-dimethoxyphenyl moiety bestowed compound 3g with
excellent inhibitory activity against CSF1R (IC = 0.12 µM; Table 1) and good inhibitory
5
0
activity against DAPK1 (IC = 2.77 µM; Table 1), the 3-dimethoxyphenyl moiety rendered
5
0
compound 3i less potent DAPK1 inhibitor (IC = 7.44 µM; Table 1) and also slightly diminished
5
0
its inhibitory activity against CSF1R (IC50= 0.78 µM; Table 1). On the other hand, the 3-
trifluoromethylphenyl moiety in compound 3g resulted in severe deterioration of the inhibitory
activity against CSF1R (IC50= 4.89 µM; Table 1) despite retaining good inhibitory activity
against DAPK1 (IC = 2.61 µM; Table 1). Therefore, another three compounds 3j–l were
5
0
2
prepared so that the R group is fixed to 3,5-dimethoxyphenyl moiety while the variations in the
arylamine moiety were explored. The results showed that 4-fluoroaniline as the arylamine
moiety in compound 3k almost doubled the IC value of DAPK1 and resulted in the loss of
5
0
potential inhibitory activity against CSF1R. Relative to compound 3h, compound 3l in which the
-morpholinoaniline was changed to 4-morpholino-pyridinamine showed improvement of the
4
IC50 value against DAPK1 while the potency against CSF1R was retained (Table 1). However,
when the isosteric 4-piperidino-pyridinamine was used for compound 3j in place of the 4-
morpholino-pyridinamine moiety of compound 3l, the potency decreased against CSF1R (Table
1
) while the IC value against DAPK1 switched back to a similar level of compound 3h.
50
In summary, we could successfully develop small molecules optimized for activity against
both DAPK1 and CSF1R despite the known inherent difficulties in the optimization of a single
molecule for activity against more than one target. The results show that the amide compounds

ACCEPTED MANUSCRIPT
possessed promising dual inhibitory activity against both DAPK1 and CSF1R. Considering both
targets, compound 3l was the most potent compound achieved within this series.
2.4. Mechanism of DAPK1 and CSF1R inhibition
To identify the mode of action of the newly developed compounds as DAPK1/CSF1R dual
inhibitors, a mechanistic study has been conducted for the most active member (compound 3l)
using three different concentrations of ATP (1, 10, and 100 µM).
2
.4.1.Mechanism of DAPK1 inhibition
The impact of compound 3l; the most potent DAPK1/CSF1R inhibitor on the phosphorylation
of 20 µM concentration of the synthetic peptide substrate (KKLNRTLSFAEPG) by recombinant
human DAPK1 (amino acids sequence 1-363) in presence of logarithmically increasing
concentrations of ATP (1, 10, and 100 µM concentrations) was studied. The lowest ATP
concentration was selected to be 1 µM which is close to the ATP K value for DAPK1 which is
m
1
.24 µM concentration [36].
In contrast to the expected behavior of the typical ATP competitive kinase inhibitors whose
elicited inhibition potency decreases with increasing the concentration of ATP, the dose-response
curves (Figure 2) did not show a reduction of the measured inhibition for the kinase reaction
upon increasing the ATP concentration. In addition, a slight enhancement of the measured IC50
was observed when ATP concentration was increased (IC values of 2.89, 1.25, and 0.92 µM
5
0
when ATP concentrations were 1, 10, and 100 µM respectively). In other words, there was a
proportional slight increase in the potency of compound 3l with increasing the ATP
concentration. The calculated K for inhibition of DAPK1 by compound 3l was 0.581
µM.
i

ACCEPTED MANUSCRIPT
Considering that the ATP K
m
value for DAPK1 is 1.24 µM, compound 3l might be a viable
DAPK1 inhibitor at the cellular concentrations of ATP.
Compound IC50 Data for DAPK1
1
1
20
00
3
I
(1 u M A T P )
(1 0 u M A T P )
I (1 0 0 u M A T P )
3
I
8
6
4
2
0
0
0
0
0
3
-
9
-8
-7
-6
-5
-4
L o g [3 I] (M )
3
I (1 uM ATP)
3I (10 uM ATP)
3I (100 uM ATP)
-0.2589
HillSlope
-0.3537
-0.4471
E CI C5 050
2.894e-006
1.248e-006
9.218e-007
Figure 2. Dose-response curves for the impact of compound 3l on the kinase reaction of
DAPK1 at 1, 10, and 100 µM concentrations of ATP.
As no competitive effect between ATP and compound 3l was observed from the conducted
study, it might be anticipated that compound 3l is a non-ATP competitive inhibitor that binds to
a site other than the ATP binding site of DAPK1. Furthermore, it seemed that ATP binding shifts
conformational ensemble of DAPK1 to the conformational structure favored for binding of
compound 3l. Since no previous DAPK1 inhibitor has been reported to show a similar mode of
inhibition, to the best of our knowledge, compound 3l would be the first in its class as a non-ATP
competitive DAPK1 inhibitor.

ACCEPTED MANUSCRIPT
2
.4.2.Mechanism of CSF1R inhibition
As compound 3l was found to be non-ATP-competitive DAPK1 inhibitor, we addressed
verification of its mechanism of inhibition of CSF1R. Accordingly, compound 3l was assayed
against CSF1R kinase at three different ATP concentration and the IC values were calculated.
5
0
As shown in Figure 3, measured IC values of compound 3l decreased as the ATP concentration
5
0
increased (IC50 values of 0.069, 0.153, and 0.273 µM when ATP concentrations were 1, 10, and
00 µM respectively) and the calculated K for inhibition of DAPK1 by compound 3l was 0.129
M. This behavior might indicate that compound 3l competes with ATP for the ATP pocket of
1
i
µ
CSF1R kinase. Accordingly, we might conclude that compound 3l acts as ATP-competitive
inhibitor of CSF1R although it might act as a non-ATP-competitive inhibitor for DAPK1.
Compound IC50 Data for CSF1R
1
20
00
3
3
3
I (1 uM ATP)
1
I (10 uM ATP)
I (100 uM ATP)
80
60
40
20
0
-
9
-8
-7
-6
-5
-4
Log [3I] (M)
3
I (1 uM ATP) 3I (10 uM ATP) 3I (100 uM ATP)
HillSlope -0.7701
EC 6.873e-008
-0.6986
-0.6879
I
5
C
0
1.533e-007
2.728e-007
5
0

ACCEPTED MANUSCRIPT
Figure 3. Dose-response curves for the impact of 3l on the CSF1R kinase reaction at 1, 10,
and 100 µM concentrations of ATP.
2
.5. In vitro cellular assays
.5.1.Tau aggregation inhibition assay
As discussed in the introduction section, DAPK1 is involved in tau hyperphosphorylation,
2
which is an essential step in the formation of tau aggregates. An in vitro active DAPK1 inhibitor
should also be able to elicit a cellular activity manifested in the inhibition of the formation of tau
aggregates to be considered a promising useful molecule for the targeting of tauopathies.
Therefore, the urea compound 3b and amide compounds (3g–l), which showed promising
DAPK1 inhibitory activity, were evaluated at decreasing concentrations of 50, 10, and 1 µM
using “tau Bimolecular Fluorescence Complementation assay” (tau-BiFC cell assay), which
vitalizes tau aggregation in living HEK293 cell line [42, 43]. In the course of conduction of the
assay, we followed the reported protocol using the same materials employing 30 µM
concentration of forskolin as tau aggregation activator.
As shown in Figure 4, the results of the cellular assay showed that compound 3b, which is the
least potent DAPK1 inhibitor, as well as compound 3g, which incorporate the lipophilic 3-
(trifluoromethyl)phenyl moiety were ineffective at all of the used concentrations in the cellular
assay for protection against the forskolin-induced tau aggregation. This cellular outcome for
compound 3g could possibly arise from a pharmacokinetic impact. For all other compounds (3h–
l) possessing the relatively more hydrophilic methoxylated phenyl moieties, a variable degree of
protection at 50 µM concentration was elicited. Among them, compound 3l was the most
effective in eliciting a complete protection against forskolin-induced tau aggregation at 50 µM

ACCEPTED MANUSCRIPT
concentration. This matches with the in vitro conducted kinase assay, which revealed compound
3
l among all of the prepared compounds to possess the most potent IC against DAPK1. On the
50
other hand, compound 3i, which incorporates 3-methoxyphenyl moiety, showed the least
efficient protection at 50 µM concentration which is also consistent with it low potency revealed
from the conducted in vitro kinase assay against DAPK1. In addition to the potent compound 3l,
compounds 3h and 3k produced excellent protective activity at 50 µM concentration while
compound 3j produced an average protective activity. However, when the test concentration was
decreased to 10 µM, only compound 3l retained its excellent potency while compounds 3h and
3
k showed almost a similar moderate protective activity. Consequently, compound 3l was
selected for further evaluation in 5-doses IC50 mode employing 1, 2, 4, 8 and 16 µM
concentrations. The assay showed that compound 3l elicited a cellular IC50 value of 5.0 µM
while the 16 µM dose concentration almost completely protected against the forskolin-induced
tau phosphorylation restoring it to the basal level. These results might indicate that the DAPK1
inhibitor compound 3l could serve as a promising lead compound for the development of
inhibitors of tau aggregates formation, and thus the development of promising tauopathies
therapeutics.

ACCEPTED MANUSCRIPT
Figure 4. Effects of DAPK1 inhibitors on tau aggregation. HEK293 tau-BiFC cells were
treated with each compound in various concentrations in the presence of forskolin (30 µM) for
4
8 h. Data are mean ± SD. The intensity of BiFC fluorescence was quantified by using a
TM
Harmony 3.1 software (PerkinElmer ).
2
.5.2.In vitro anti-neuroinflammatory evaluation
The increased nitric oxide (NO) production is a major event in microglial neuroinflammation.
In addition to its contribution to the increased hyperphosphorylation and formation of tau
aggregates [44, 45], it also contributes to the neurodegenerative events by several mechanisms
[46, 47]. Furthermore, it is interesting that nitration of tau protein has been identified in
tauopathies [48, 49]. Moreover, NO produced from microglial neuroinflammation could lead to
the formation of neurotoxic NO-dependent reactive nitrogen species (RNS) such as peroxynitrite
[50, 51]. As we sought the development of multifunctional agents preventing both of formation
of tau NFTs and neuroinflammation, it was important to evaluate the anti-neuroinflammatory
activity of the compounds that showed promising protective activity against the formation of tau
aggregates. Consequently, compound 3l which showed the most potent activity in the tau-BiFC

ACCEPTED MANUSCRIPT
cellular assay, as well as, compound 3h which also elicited a good protective activity were
selected for evaluation of anti-neuroinflammatory effects in the BV-2 microglial cell line. In
addition, the in vitro kinase assay showed that these two compounds possess a submicromolar
IC values activity against CSF1R. Accordingly, the lipopolysaccharides (LPS)-stimulated BV-
5
0
2
cell line was used as a model for microglia-induced neuroinflammation to evaluate the effects
of these two compounds [52].
As shown in Figure 5A, the 100 ng/mL concentration of LPS resulted in an abrupt increase of
NO production in BV-2 cells. However, the presence of 10 µM concentration of compound 3l
attenuated the increased production of NO by around 55% of the production elicited by the
positive control. This inhibition of LPS-induced NO production in microglia was highly
significant (p< 0.001). As shown in Figure 5B the viability of BV-2 cells in the presence of both
of compound 3l and LPS was more than 92% of the control indicating that this reduction of NO
production is due to specific inhibition of BV-2 cells by compound 3l. However, a low inhibition
percent for NO production was elicited by compound 3h (Figure 5A) which was significant only
at p=0.1023. It is worthy to mention that the structural difference between both compounds is
only the N atom on the aromatic ring on position 2 of the pyrimidine ring. This structural
difference resulted in the dramatic difference in cellular activity. Possibly, the introduction of the
N atom to compound 3l improved its physicochemical properties and allowed for a better cellular
activity.

ACCEPTED MANUSCRIPT
Figure 5. Effects of compounds 3h and 3l on LPS-induced BV-2 cell line. (A) Inhibition of
NO production in LPS-stimulated BV-2 cell line by compounds 3h and 3l. Values are the means
of five independent experiments ± SD. (B) the impact of for compounds 3h and 3l on the
viability of BV-2 cell line. Values are the means of five independent experiments ± SD.
2.6. Selectivity profile
A considerable homology of the ATP-binding site between different kinases exists because
kinases utilize the same natural molecule (ATP) as a ligand. As a result, the developed small
molecule kinase inhibitor most likely inhibits more than one kinase. Consequently, selectivity is
a major issue in developing kinase inhibitors to avoid off-target effects that may arise from the
inhibition of kinases other than the targeted kinases. Accordingly, compound 3l, which was
revealed as a promising molecule targeting both of tau aggregation and of neuroinflammation,
was subjected to kinase profiling to test its selectivity in inhibiting CSF1R and DAPK1 rather
than other kinases. Even though the mechanistic study showed that 3l does not compete with
ATP for DAPK1, there is a possibility that it might bind the ATP pocket of other kinases as it
was developed starting from ATP competitive inhibitors of CSF1R. In this regard, the selectivity
was evaluated by profiling compound 3l against a panel of 47 kinases. The panel included both
DAPK1 and CSF1R family members in addition to other kinases representing diverse kinase

ACCEPTED MANUSCRIPT
families. The profiling was performed via evaluating the percent inhibition produced by 10 µM
concentration of compound 3l.
1
00.00
8
6
4
2
0.00
0.00
0.00
0.00
0
.00
-
-
-
-
20.00
40.00
60.00
80.00
Figure 6. Selectivity profile elicited by 10 µM concentration of compound 3l against a panel
of 47 kinases including CSF1R and DAPK1. These data are the mean of two independent
experiments ± standard deviation.
As shown in Figure 6, compound 3l did not significantly inhibit most of the diverse kinases
from the TK or STK families represented in the panel except for the expected inhibition of the
src family members such as Lck, Lyn, and c-Src. In addition, compound 3l did not inhibit other
members of the PDGFR family arguing for a highly selective CSF1R inhibitor. Intriguingly,
compound 3l did not show any activity over DAPK2, DAPK3, and DRAK1 kinases despite the
high sequence homology between the kinase domains of DAPK1, DAPK2, and DAPK3 (more

ACCEPTED MANUSCRIPT
than 80% sequence homology in their kinase domains). It also did not show any activity against
other related kinases such as PIM1, LRRK2 or CaMK2a suggesting that compound 3l is highly
selective towards CSF1R and DAPK1.
To the best of our knowledge, none of the known DAPK1 inhibitors reached this level of
selectivity within the DAPK family with preferential selectivity for DAPK1 over the other
family members. The results obtained from the mechanistic study and the observed subtype
selectivity within the DAPK family suggest that compound 3l binds to another binding site other
than the ATP-binding site arguing for a potential first-in-class non-ATP-competitive inhibitor of
DAPK1. In summary, compound 3l showed high selectivity for DAPK1 and CSF1R against
other members of the DAPK or PDGFR families.
2.7. Cardiac safety profile (In vitro Predictor™ hERG Fluorescence Polarization
Assay)
Early assessment of the potential cytotoxicity is important to minimize the risk of failure of
drug discovery programs [53]. The human Ether-a-go-go-Related Gene (hERG aka KCNH2) is a
gene that encodes for two voltage-gated potassium channels [54]. The hERG potassium channels
are involved in action potential repolarization in the heart [55]. Both inhibition and activation of
hERG channels by non-cardiovascular drugs are considered as a severe toxic effect that should
be eliminated early in the drug development process [56]. In fact, several drugs have been
dropped after marketing because of their significant cardiotoxic effects due to their off-target
effect on the hERG channels [57]. Consequently, the promising compound 3l was evaluated
against the hERG channel in a 10-dose IC50 mode starting at 100 µM with 3-fold serial dilution
employing Predictor™ hERG Fluorescence Polarization Assay. The assay uses a membrane

ACCEPTED MANUSCRIPT
containing hERG channel protein and a high-affinity fluorescent hERG channel ligand (Tracer).
The ability of the test compound to compete with the tracer for the hERG binding reflects the
ability of the test compound to inhibit the hERG channel. Compound 3l showed good cardiac
safety profile with an in vitro IC value of 17.1 µM against hERG channel (Table 2). This result
5
0
suggests that compound 3l does not elicit significant cardiotoxicity and therefore, a suitable lead
compound for further development.
Table 2. hERG channel affinity data (as IC50 value) of compound 3l.
a
-6
Compd.
IC50 of tracer binding (
µM) against hERG
P
e
(10 Cm/sec) ± SD
b
3
l
17.1
0.01
ND
57.0 ± 3.3
c
d
E-4031
ND
e
f
Progesterone
Theophylline
33.3 ± 0.7
ND
0.26 ± 0.06
a
The compound was tested in a singlicate 10-dose IC50 mode starting at 100 µM with 3-fold serial dilution.
The compound was tested using 12.5 µM concentration and the results shown are the means of three independent
b
experiments ± standard deviation.
c
E-4031 is the positive control agent for the hERG channel assay.
ND: Not determined
Progesterone (50 µM) was used as the positive control agent for the PAMPA assay.
Theophylline (50 µM) was used as the negative control agent for the PAMPA assay.
d
e
f
2.8. Blood-Brain Barrier (BBB) penetration evaluation (PAMPA assay)
The ability of CNS-directed therapeutic agents to cross the BBB determines the effectiveness
of these agents in vivo and minimize the potential peripheral side effects that result from poor
BBB penetration. Since our designed compounds target DAPK1 and CSF1R for management of
tauopathies, it was essential to assess their ability to cross the BBB. For that, we have selected
the most potent member of our series (3l) to be evaluated using the parallel artificial membrane

ACCEPTED MANUSCRIPT
permeability assay (PAMPA) [58]. The PAMPA assay uses BBB-PAMPA lipid membrane and
measures the ability of a certain chemical to cross this membrane and therefore measures its
effective permeability (P ) (detailed experimental protocol is discussed in the experimental
e
section). In order to evaluate the obtained P values, certain standards were applied. If the P is
e
e
-6
more than 0.4 (10 cm/sec), the tested compound is considered as highly permeable and if it is
-6
more than 10 (10 cm/sec), the compound is considered as highly CNS bioavailable.
Progesterone (high BBB-penetrating drug) was used as a positive control and showed P (± SD
values) value of 33.3 ± 0.7 while theophylline (low BBB-penetrating drug) was used as negative
control and showed P value of 0.26 ± 0.06. When compound 3l was assayed at the same
e
e
conditions, it showed P value of 57.0 ± 3.3 accounting for around 71% higher penetration ability
e
than the positive control (Table 2). These results suggest that compound 3l could have high BBB
penetration and high CNS bioavailability and is a suitable lead for further development.
2.9. In silico molecular simulation study
In order to identify the binding mode and receptor interactions of this new class of
DAPK1/CSF1R inhibitors, a molecular docking study was performed. The crystal structures of
the catalytic kinase domain of DAPK1 (PDB code: 4TXC; Figure 7) and CSF1R kinase domain
(PDB code: 3KRJ) were used for the docking study employing SwissDock which is based on the
docking software EADock DSS running on the Vital-IT cluster [59]. Docking was performed in
the accurate mode as a blind docking study where binding modes are generated in the vicinity of
all cavities of the used structure. Simultaneously, their CHARMM energies were estimated on a
grid. The binding modes with the most favorable energies were evaluated with FACTS and
clustered and visually inspected [60].

ACCEPTED MANUSCRIPT
2
.9.1.Docking results to DAPK1
The results of the conducted in silico blind docking showed that this class of compounds
might dock within two pockets only in the employed structure of DAPK1, the ATP-binding
pocket and another pocket which was identified as the protein substrate-binding site [61].
Figure 7. The ATP-binding site and the substrate-binding sites of the DAPK1 protein. A
hydrophobic surface was generated around the protein using Discovery studio program standard
protocol for surface generation.
In the case of compound 3l, which is the most potent compound among the prepared series,
the calculations predicted that the best-docked pose is within the substrate-binding pocket
(
Figure 8A). This binding mode for compound 3l showed a high overall favorable free energy
∆G = -9.21 kcal/mol). Several polar and hydrophobic interactions of compound 3l within the
(
substrate-binding site contributed to the stabilization of this binding mode. Thus, the 3,5-

ACCEPTED MANUSCRIPT
dimethoxybenzoyl group docked into the hydrophobic pocket eliciting interactions with Leu111,
Lys108, and Ala106. The carbonyl group of the amide established a hydrogen bond with the
backbone NH group of Leu111. The NH group at position 2 of the pyrimidine core interacted
with the Glu113 side chain. The morpholine was directed to the solvent exposure region while
the p-methoxyphenyl interacted with Asn243 and Glu239. On the other hand, the best scoring
pose of compound 3l within the ATP-binding site showed a relatively lower score (∆G = -8.36
kcal/mol) and, more importantly, was unable to establish the crucial interactions with key
residues within the hinge region (Figure 8B). It showed a flipped binding mode in which the
dimethoxyphenyl moiety docked into the hydrophobic pocket showing interactions with Leu19,
Val27, Ala40, Glu94, and Ile160. The pyrimidine ring interacted with Lys42 and Asp161. While
the morpholine was directed to the solvent and the phenyl group carrying the morpholine
interacted with Gln23. The methoxyphenyl moiety interacted with the peptide bond between
Gly20 and Ser21. Together, these results indicate that compound 3l might act through binding to
the substrate-binding site rather than the ATP binding pocket of DAPK1.

ACCEPTED MANUSCRIPT
Figure 8. In silico calculated binding modes of compound 3l within DAPK1. (A) Calculated
binding mode of 3l within the substrate-binding site of DAPK1. (B) Calculated binding mode of
3
l within the ATP-binding site of DAPK1.
Considering the case of compound 3g, which is a less potent DAPK1 inhibitor, a score lower
than that for compound 3l was calculated for the docked pose within the substrate-binding site
∆G = -8.92 kcal/mol for compound 3g while ∆G = -9.21 kcal/mol for compound 3l). This could
(
explain the relative potencies of these compounds. The calculated binding mode within the
substrate-binding site showed slight deviation in the orientation of compound 3g relative to that
of compound 3l (Figure 9A). In this binding mode of compound 3g, the 3-trifluoromethylphenyl
moiety docked into the hydrophobic pocket showing interactions with Phe102, Leu105, Ala106,
Ile209, and Leu210. Meanwhile, the pyrimidine ring interacted with Leu111, Ala116, and
Leu210. The morpholine was directed to solvent exposure. The amide NH interacted with
Leu210. Considering the in silico prediction for possible binding of compound 3g within the
ATP-binding site, the calculated pose showed a higher score (∆G = -9.23 kcal/mol) than that
within the substrate-binding site but missed the crucial interactions with key residues of the
hinge region (Figure 9B). In this calculated pose, the 4-methoxyphenyl ring fitted into the
hydrophobic pocket showing interactions with Val27, Lys42, Glu64, Leu93, Ile160, and Asp161,
while the pyrimidine ring interacted with Val27. The morpholine was directed to solvent
exposure while the phenyl group carrying the morpholine interacted with Leu19 and Met146.
The trifluoromethyl interacted with Phe24 and Ala25 while the phenyl group carrying the
trifluoromethyl moiety interacted with Ala25 and Asp161. Despite the higher score of the pose
of compound 3g within the ATP pocket than within the substrate-binding pocket, lacking