Targeting nuclear protein TDP-43 by cell division cycle kinase 7 inhibitors: A new therapeutic approach for amyotrophic lateral sclerosis

Article abstract of DOI:10.1016/j.ejmech.2020.112968

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease with no known cure. Aggregates of the nuclear protein TDP-43 have been recognized as a hallmark of proteinopathy in both familial and sporadic cases of ALS. Post-translational modifications of this protein, include hyperphosphorylation, cause disruption of TDP-43 homeostasis and as a consequence, promotion of its neurotoxicity. Among the kinases involved in these changes, cell division cycle kinase 7 (CDC7) plays an important role by directly phosphorylating TDP-43. In the present manuscript the discovery, synthesis, and optimization of a new family of selective and ATP-competitive CDC7 inhibitors based on 6-mercaptopurine scaffold are described. Moreover, we demonstrate the ability of these inhibitors to reduce TDP-43 phosphorylation in both cell cultures and transgenic animal models such as C. elegans and Prp-hTDP43 (A315T) mice. Altogether, the compounds described here may be useful as versatile tools to explore the role of CDC7 in TDP-43 phosphorylation and also as new drug candidates for the future development of ALS therapies.

Full text of DOI:10.1016/j.ejmech.2020.112968

Journal Pre-proof
Targeting nuclear protein TDP-43 by cell division cycle kinase 7 inhibitors: a new
therapeutic approach for amyotrophic lateral sclerosis
Elisa Rojas-Prats, Loreto Martinez-Gonzalez, Claudia Gonzalo-Consuegra, Nicole
F. Liachko, Concepción Perez, David Ramírez, Brian C. Kraemer, Ángeles Martin-
Requero, Daniel I. Perez, Carmen Gil, Eva de Lago, Ana Martinez
PII:
S0223-5234(20)30940-5
DOI:
https://doi.org/10.1016/j.ejmech.2020.112968
EJMECH 112968
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 July 2020
Revised Date: 15 October 2020
Accepted Date: 23 October 2020
Please cite this article as: E. Rojas-Prats, L. Martinez-Gonzalez, C. Gonzalo-Consuegra, N.F.
Liachko, C. Perez, D. Ramírez, B.C. Kraemer, Ángeles Martin-Requero, D.I. Perez, C. Gil, E. de
Lago, A. Martinez, Targeting nuclear protein TDP-43 by cell division cycle kinase 7 inhibitors: a new
therapeutic approach for amyotrophic lateral sclerosis, European Journal of Medicinal Chemistry, https://
doi.org/10.1016/j.ejmech.2020.112968.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2020 Elsevier Masson SAS. All rights reserved.

CDC7
P
P
TDP-43
P
P
TDP-43
P
TDP-43
mRNA
P
TDP-43

REVISED 2 14.10.2020
Targeting nuclear protein TDP-43 by cell division cycle kinase 7 inhibitors:
a new therapeutic approach for amyotrophic lateral sclerosis
1,¥
1,¥
2
Elisa Rojas-Prats, Loreto Martinez-Gonzalez, Claudia Gonzalo-Consuegra, Nicole
3,4
5
6
3,4,7,8
F. Liachko, Concepción Perez, David Ramírez, Brian C. Kraemer,
Ángeles
1,9
1
1,9
2,9,
Martin-Requero, Daniel I. Perez, Carmen Gil, Eva de Lago, * and Ana
1,9,
Martinez
*
1
Centro de Investigaciones Biológicas Margarita Salas-CSIC, Ramiro de Maeztu 9,
8040 Madrid (Spain)
2
2
Instituto Universitario de Investigación en Neuroquímica, Departamento de Bioquímica
y Biología Molecular, Facultad de Medicina, Universidad Complutense, Madrid (Spain)
3
Department of Medicine, Division of Gerontology and Geriatric Medicine, University of
Washington, Seattle, WA 98104 (USA)
4
Geriatrics Research Education and Clinical Center, Veterans Affairs Puget Sound
Health Care System, Seattle, WA 98108 (USA)
5
Instituto de Química Medica-CSIC, Juan de la Cierva 3, 28006 Madrid (Spain)
6
Instituto de Ciencias Biomédicas, Universidad Autónoma de Chile, Llano
Subercaseaux 2801 – piso 6, Santiago, Chile.
7
Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle,
WA 98195 (USA)
8
Department of Pathology, University of Washington, Seattle, Washington 98195, USA.
9
Centro de Investigación Biomédica en Red de Enfermedades Neurodegenerativas
(
CIBERNED), Instituto de Salud Carlos III, 28031 Madrid (Spain)
¥
Equal contribution
Correspondence to:
Prof. Ana Martinez
CIB-CSIC
Dra. Eva de Lago
Dpto. Bioquimica y Biologia Molecular
Facultad de Medicina, UCM
Plaza Ramon y Cajal s/n, Ciudad Universitaria
28040 Madrid, Spain
Ramiro de Maeztu 9
28040 Madrid, Spain
Tel.:+34 91 837 3112
e-mail: ana.martinez@csic.es
Tel.:+34 91 394 1454
e-mail: elagofem@ucm.es
1

Abstract
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease with no known
cure. Aggregates of the nuclear protein TDP-43 have been recognized as a hallmark of
proteinopathy in both familial and sporadic cases of ALS. Post-translational
modifications of this protein, include hyperphosphorylation, cause disruption of TDP-43
homeostasis and as a consequence, promotion of its neurotoxicity. Among the kinases
involved in these changes, cell division cycle kinase 7 (CDC7) plays an important role
by directly phosphorylating TDP-43. In the present manuscript the discovery, synthesis,
and optimization of a new family of selective and ATP-competitive CDC7 inhibitors
based on 6-mercaptopurine scaffold are described. Moreover, we demonstrate the
ability of these inhibitors to reduce TDP-43 phosphorylation in both cell cultures and
transgenic animal models such as C. elegans and Prp-hTDP43 (A315T) mice.
Altogether, the compounds described here may be useful as versatile tools to explore
the role of CDC7 in TDP-43 phosphorylation and also as new drug candidates for the
future development of ALS therapies.
Keywords: CDC7 inhibitors, TDP-43, ALS, FTLD, drug discovery
2

Introduction
Amyotrophic lateral sclerosis (ALS) is a fatal and devastating neurodegenerative
disease that features the progressive loss of cortical and spinal motor neurons, leading
1
to muscle weakness and spasticity as main clinical symptoms. Classified as an orphan
disease, ALS affects more than 800,000 patients worldwide with an estimated
incidence of 1.5-2.7 per 100,000 people in Europe and North America. Despite
important advances in the understanding of ALS molecular pathobiology, the etiology
remains unknown. Today, ALS is considered a complex disease where a multifactorial
2
combination of environmental and genetic factors may be involved in its origin. There
is no cure for ALS and only riluzole and more recently edaravone are approved for their
palliative treatment with limited results for the patients; thus, there is an urgent need to
3
find an effective treatment.
Pathologically, several distinct aggregating proteinopathies have been described in
patients with ALS. Aggregates of superoxide dismutase 1 (SOD1) were first identified
4
in patients with mutations in this detoxifying enzyme. In fact, the discovery that the
SOD1 G93A mutation causes ALS in the beginning of the 90’s allowed the
5
development of the first transgenic mouse used in drug discovery for ALS. However,
only between the 5 to 10% of the ALS cases are familiar (genetically inherited) and the
SOD1 mutations account for only about 10 to 20% of these cases. Unfortunately, drugs
effective in the SOD1 G93A mice have failed in clinical trials, which may be attributed
6
to differences in pathology between SOD1-driven and other types of ALS. More
recently, the nuclear RNA-binding proteins, called fused in sarcoma (FUS) and
transactive response DNA binding protein of 43 kDa (TDP-43) have been identified in
7
aggregates present in motor neurons of some cases of ALS. In fact, TDP-43 pathology
is present in nearly 95% of all cases of ALS, and has been recognized as the most
common proteinopathy in both familial and sporadic cases of ALS. TDP-43 aggregates
3

are also the hallmark of frontotemporal lobular dementia (FTLD) patients with TDP-43
8
inclusions (FTLD-TDP), another neurodegenerative proteinopathy.
In healthy neurons, TDP-43 resides mainly in the nucleus where it regulates the
splicing patterns of transcripts and the expression of several important genes.. In
disease-affected neurons, the dynamic equilibrium of TDP-43 between nucleus and
cytoplasm is altered. As a result, TDP-43 concentration increases in the cytoplasm,
possibly as a mixed consequence of genetic susceptibility, aging, prolonged stress,
9
chronic inflammation, etc. The loss of TDP-43 from the nucleus prevents its normal
functions, such as the regulation of RNA metabolism. While in the cytoplasm, TDP-43
experiences
different
post-translational
modifications
including
truncation,
ubiquitination, phosphorylation, acetylation, etc. that leads to the formation of toxic
10
aggregates within the motor neurons, and prevents TDP-43 gene regulation activity.
Of these modifications, phosphorylation is the most consistent marker of pathological
TDP-43 and has been implicated in its neurotoxicity. In the search for new therapeutic
agents for ALS and others neurodegenerative diseases, TDP-43 emerges as a
promising drug target. Restoration of TDP-43 homeostasis by using small molecules
11
may be one of the potential approaches.
Protein kinases are well established drug targets, with more than forty kinase inhibitors
12
on the market as drugs for the treatment of oncological diseases, and other
pathologies such as rheumatoid arthritis, pulmonary fibrosis, myelofibrosis and quite
13
recently, glaucoma. Kinases, which are responsible for protein phosphorylation, are
involved in many different cellular signaling events and also play crucial roles in post-
translational modifications of numerous proteins. In its pathological state, TDP-43 can
be phosphorylated by several kinases. These include casein kinase 1 (CK-1) delta and
14
epsilon. In fact, benzothiazole-based CK-1 inhibitors are able to decrease TDP-43
15-16
phosphorylation both in cellular and animal models.
More recently, they have been
demonstrated to promote recovery of TDP-43 homeostasis, i.e. protein phosphorylation
4

decrease and nuclear localization recovery, in human based cell models from FTLD
17-18
and ALS patients.
These data indicate the promise of kinase inhibitors broadly, and
CK-1 inhibitors specifically, as compounds of interest for further development as
potential ALS and FTLD therapies.
Recent research has identified three additional kinases controlling TDP-43
phosphorylation in vivo. These are cell division cycle kinase 7 (CDC7) and tau tubulin
19-20
kinase 1 and 2 (TTBK1 and TTBK2).
Thus, inhibitors of these kinases may be
innovative drug candidates for the future effective treatment of ALS and FTLD. In fact,
CDC7, a serine/threonine kinase that is essential for regulating DNA replication in all
21
eukaryotes, has been an attractive target for cancer therapy. Several inhibitors have
22-23
been synthetized
and used as versatile chemical tools to decipher the role of
pharmacological CDC7 inhibition in different cancer types, some of them reaching
24
clinical trials. Among them, the CDC7 inhibitor PHA-767491, that failed in clinical trial
phase 1 for solid tumors, has recently shown to decrease TDP-43 phosphorylation in
19
C. elegans but it is not able to cross the blood brain barrier in vertebrates, and thus it
19
is not possible to use it to study central nervous system pathologies. Moreover, PHA-
25
7
67491 also inhibits CDK9, making it difficult to decipher the main kinase involved in
its therapeutic potential over TDP-43 alteration.
Considering this back-ground, a medicinal chemistry program focused on finding
selective and brain permeable small molecules CDC7 inhibitors able to modulate TDP-
43 pathology was started. In the present manuscript, we describe the discovery,
synthesis and optimization of a new family of selective and ATP-competitive CDC7
inhibitors based on 6-mercaptopurine scaffold. Moreover, we test their ability to reduce
TDP-43 phosphorylation both in cultured cells and in two transgenic animal models of
ALS. Altogether, compounds here described could be considered as new drug
candidates for the future therapy of ALS and/or FTLD or as versatile tools to explore
the role of CDC7 in different neurological disorders where TDP-43 is involved.
5

Results and Discussion
Hit identification
A forward chemical genetic approach was used to identify new CDC7 inhibitors. This
methodology has two main steps: first, identifying molecules able to interact with the
target of interest and second, confirming target engagement activity in a phenotypic
assay. Therefore, we first selected an in vitro biological assay to evaluate CDC7
2
6
activity. A set of chemically diverse compounds were chosen from our in-house MBC
chemical library, which has been recognized as a valuable tool to provide drug-like hit
27
molecules for screening. We initially assayed the potential inhibition of CDC7 by each
of these molecules at a fixed concentration of 10 µM. For compounds where the
percentage of kinase inhibition was greater than the 50%, a dose-response curve was
calculated and the IC50 value determined. Following this approach, the benzoyl-
mercaptopurine derivative 1 was identified with an IC value of 2.12 µM emerging as a
50
new chemical prototype for CDC7 inhibition (Figure 1).
Hit optimization. Chemical synthesis
A medicinal chemistry program was then designed to produce a focused set of
compounds with different structural modifications around the central heterocyclic
scaffold, allowing us to study the influence of different substituents in CDC7 inhibition
(
Figure 1). Using different synthetic strategies previously described, a set of more than
0 compounds was synthetized, purified and characterized. Changes in the nature of
3
heteroatoms, the nature and length of the linker joined to the mercaptopurine core, the
nature and position of substituents in the phenyl ring have been made to optimize the
biological potency of the hit compound building a consistent structure-activity
relationship.
6

1
R
S, O, NH
H, 2-(Cl, Br, I, CF
-(F, Cl, Br, I, NO
CF , CN, OCH
-(F, Cl, Br, I, NO
CF , CN, OCH
SCH , CH OH, NHAc)
3
)
3
2
,
,
3
3
)
N, CH
4
2
3
3,
3
2
COCH ,
2
(
1
)
NH, S, NHCH
NCH CH Ph
2Ph,
CH
2
CH ,
2
IC50= 2.1 µM
2
2
CH
2,
-
Figure 1. Chemical structure of the new CDC7 inhibitor identified,
benzoylmercaptopurine 1, and chemical modifications proposed.
Initially, substituents in the phenyl ring of the benzoyl-mercaptopurine derivative 1 were
varied using 6-mercaptopurine and different bromoacetophenones as reagents. In that
case, nucleophilic substitution in basic aprotic medium yielded derivatives 1-3. The
same reaction conditions were applied to obtain 6-benzylmercaptopurines (4-25),
where the carbonyl group is eliminated from the linker, and 6-
phenethylmercaptopurines (compounds 26-29) where the carbonyl group is substituted
by a methylene group. Many diverse benzylhalides were used as reagents in basic
medium and in some cases a mixture of two compounds was obtained that after
isolation provided the 6-benzylmercaptopurines substituted at N-9 as minority
compound (7b, 15b, 23b-25b). When phenethylhalides were used, disubstituted
derivatives (26b-29b) were isolated every time (Scheme 1).
7

Scheme 1. i) K CO , DMF, r.t.
2
3
Derivatives without any spacer between the sulfur atom of the mercaptopurine-central
core and the phenyl substituent were obtained by aromatic nucleophilic substitution of
6-chloropurine with different thiophenols under basic catalysis in n-butanol.
Compounds 30 and 31 were isolated in good yields. The aromatic nucleophilic
substitution using different benzyl alcohols or benzylamines in basic medium or under
microwave irradiation respectively was also used to yield substituted oxo- and amino-
purines, derivatives 32-35, where the sulfur atom of the hit 1 is replaced by its
bioisosters such as nitrogen or oxygen ones (Scheme 2).
8

o
Scheme 2. i) Et N, K CO , n-BuOH, reflux; ii) MW, 100 C, 5 min; iii) NaOH, reflux; iv) NaH,
3
2
3
DMF, r.t.
Finally, modifications of the heterocyclic core were proposed and 7H-pyrrolo[2,3-
d]pyrimidines and thieno[2,3-d]pyrimidines (compounds 36-39) were synthesized by
nucleophilic substitution of the chlorine atom of 4-chloro-7H-pyrrolo[2,3-d]pyrimidine or
4-chlorothieno[2,3-d]pyrimidine with the appropriate benzyl mercaptane derivative in
basic polar aprotic medium (Scheme 2).
Although some of the compounds were previously described, we fully characterized
here the structures of all the synthetized compounds unequivocally based on their
1
13
spectroscopical ( H and C-NMR) and analytical data. They arecollected in the
experimental part and supporting information.
Biological evaluation of CDC7 inhibition
The purine-based prepared heterocyclic derivatives (2-39) were evaluated in a
commercially available CDC7 activity assay to determine their potential inhibition.
CDC7 is only active when it is associated with its activation unit Dbf4, and interaction
9

28
with Dbf4 is necessary for CDC7 ATP recognition and substrate binding. Thus, the
methodology utilizes a fluorescein-labeled substrate of CDC7/Dbf4 and an associated
specific antibody, which results in an increased TR-FRET value when this substrate
and ATP are allowed to react. The amount of antibody that is bound to the tracer is
directly proportional to the amount of phosphorylated substrate present and in this
manner, kinase activity can be detected and measured by an increase in the TR-FRET
value. All the compounds were initially tested at a fixed concentration of 10 µM and the
IC50 was calculated when the percentage of inhibition was greater than 50%. Results
are collected in table 1 showing a large collection of CDC7 inhibitors with IC values in
50
the submicromolar range, more potent that the initial hit 1. For comparative purposes,
the well-known CDC7 inhibitor calledPHA-767491 is also included in Table 1 as
standard reference.
Table 1. CDC-7 inhibition and BBB permeability prediction of synthetized
mercaptopurine derivatives (1-35) and heterocyclic analogues (36-39).
1
Comp.
no.
A
X
Z
Y
R
B
CDC-7
IC50 (µM)
BBB
Prediction
n.t.
a,b
1
2
3
4
5
6
7
N
N
N
N
N
N
N
N
N
N
N
N
N
N
S
S
S
S
S
S
S
S
S
S
S
S
S
S
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
2
2
2
2
2
2
2
2
2
2
2
2
2
2
C=O
C=O
C=O
p-Br
p-Cl
p-NO
p-Br
p-Cl
p-NO
H
H
p-I
p-CF
NH
NH
NH
NH
NH
NH
NH
2.12 ± 1.50
4.15 ± 2.64
3.11 ± 1.68
0.44 ± 0.11
0.18 ± 0.08
0.13 ± 0.01
0.20 ± 0.04
n.t.
n.t.
2
2
-
-
-
-
-
-
-
-
-
-
-
CNS +
CNS +
CNS +
CNS +
n.t.
c
7
b
NCH
2
Ph
0%
8
9
NH
0.30 ± 0.03
0.36 ± 0.07
0.29 ± 0.07
0.14 ± 0.01
0.47 ± 0.06
0.30 ± 0.02
CNS +
CNS +
CNS+/CNS-
CNS +
CNS -
3
NH
NH
NH
NH
NH
1
0
1
2
3
p-CN
p-SMe
p-NHAc
p-
1
1
1
CNS+/CNS-
2
CH OH
1
1
4
5
N
N
N
N
N
S
S
S
S
S
CH
CH
CH
CH
CH
2
2
2
2
2
-
-
-
-
-
m-Br
m-Cl
m-Cl
m-NO
m-I
NH
NH
0.13 ± 0.03
0.14 ± 0.03
30%
0.19 ± 0.03
0.08 ± 0.02
CNS -
CNS +
n.t.
c
1
5b
NCH
2
(3ClPh)
1
1
6
7
2
NH
NH
CNS+/CNS-
CNS +
10

1
1
2
2
2
2
8
9
0
1
2
3
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
NH
O
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
-
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
-
-
-
-
-
-
-
-
-
m-CF
3
NH
NH
NH
NH
NH
NH
0.32 ± 0.06
0.34 ± 0.06
0.24 ± 0.06
0.36 ± 0.05
0.27 ± 0.06
0.09 ± 0.02
CNS +
CNS+/CNS-
CNS +
CNS +
CNS +
CNS +
n.t.
m-CN
m-OMe
m-F
o-Br
o-Cl
o-Cl
o-I
o-I
c
2
2
2
2
2
2
2
3b
2
4b
2
5b
2
6b
2
7b
2
8b
2
NCH
2
(2ClPh)
NH
NCH (2IPh)
7%
4
0.58 ± 0.10
CNS +
n.t.
CNS +
n.t.
CNS +
n.t.
CNS +
n.t.
CNS +
n.t.
CNS +
n.t.
n.t.
n.t.
a
2
4%
5
-
-
o-CF
o-CF
p-Cl
p-Cl
p-F
3
3
NH
(2CF
NH
0.25 ± 0.07
c
NCH
2
3
Ph)
9%
6
CH
CH
CH
CH
CH
CH
CH
CH
-
2
2
2
2
2
2
2
2
0.14 ± 0.03
c
N(CH
N(CH
2
)
2
(4ClPh)
NH
(4FPh)
NH
(3ClPh)
NH
(2ClPh)
3%
7
0.39 ± 0.05
c
p-F
2
)
2
12%
8
m-Cl
m-Cl
o-Cl
o-Cl
p-Br
p-OMe
m-Cl
m-Cl
0.30 ± 0.04
c
N(CH
2
)
2
8%
9
0.19 ± 0.03
c
9b
N(CH
2
)
2
4%
3
3
3
3
0
1
2
3
NH
NH
NH
NH
1.38 ± 0.43
1.72 ± 0.26
3.06 ± 1.00
0.19 ± 0.05
-
-
-
-
CH
CH
2
n.t.
CNS +
2
3
3
3
3
3
3
4
5
6
7
8
9
N
N
CH
CH
CH
CH
O
O
S
S
S
S
CH
CH
CH
CH
CH
CH
2
2
2
2
2
2
-
-
-
-
-
-
H
p-Br
H
m-Br
H
NH
NH
S
S
NH
NH
0.17 ± 0.04
0.91 ± 0.24
CNS+/CNS-
CNS +
n.t.
c
18%
11%
c
n.t.
CNS +
n.d.
0.23 ± 0.04
0.39 ± 0.11
m-Br
a
b
Reference compound PHA-767491: IC50=0.01 µM; IC50Values ± Standard deviation (ten different
c
concentrations per compound, in two independent experiments) Origin 9.0 software used; this value
represents the % of CDC7 inhibition at a fix compound concentration of 10 µM; n.t.= not tested; n.d.: not
possible to determine in the experimental conditions.
In view of the results collected in table 1, some preliminary structure-activity
relationships could be established. Firstly, the presence of the heterocyclic NH in the
imidazole-fused ring is crucial for activity and totally inactive compounds are obtained
when any substituent is attached to this nitrogen atom such is the case of compounds
7b, 15b, and 23b-29b or when this NH is replaced by a sulfur atom (thieno[2,3-
d]pyrimidines 36 and 37).
The length and the nature of the linker attached to the exocyclic sulfur atom also
influence the potency of CDC7 inhibition. While kinase inhibition is similar in
mercaptopurines 1-3, in the micromolar range, it increases one order of magnitude
when the carbonyl group is eliminated from the spacer between both of the aromatic
fragments (derivatives 4-6). However, CDC7 inhibition decreases when the linker
disappears and the phenyl moiety is directly attached to the sulfur atom of
11

mercaptopurine scaffold (compounds 30 and 31). It is important to note that inhibitory
potency also increases when the carbonyl group present in the linker of the hit
compound 1 is changed to a methylene group (purines derivatives 26-29).
Finally, it is worthwhile to mention that the exocyclic sulfur atom may be replaced by its
bioisostere, the oxygen atom (compounds 15 and 7 versus 33 and 34, respectively)
without affecting the kinase inhibition, but the inhibitory potency significantly decreases
when the sulfur is substituted by an NH group (derivative 32).
In general, the nature and the substitution position of different groups in the phenyl ring
do not seriously affect the inhibition of CDC7 by this family of mercaptorpurines, having
a great number of compounds (derivatives 4-25) with IC values in the submicromolar
50
range, between 0.1 and 0.9 µM.
To study the binding mode of these new CDC7 inhibitors, the kinetic mechanism of
CDC7 inhibition was investigated by varying the concentrations of both ATP and the
tested inhibitors in the kinase reaction. Three of the most potent compounds and firstly
available, derivatives 7, 15 and 22 were selected for this assay with the goal of
extrapolating the result to all the purine-based CDC7 inhibitors here described. For this
TM
goal, we chose a substrate phosphorylating assay based on ADP-Glo technology
checking the previous inhibitory activity found. The Lineweaver-Burk or double-
reciprocal plots of 1/v versus 1/[ATP] at fixed concentrations of these selected
inhibitors showed an intersecting pattern consistent with an ATP-competitive inhibition
as all lines converge at or near the y axis (Figure 2). These results show that
mercaptopurine-based derivatives, and specifically compounds 7, 15 and 22, are ATP
competitive CDC7 inhibitors.
12

1
0
9
8
7
6
5
4
3
2
1
0
7
6
5
4
3
2
1
1
1
1
4
2
0
8
6
4
2
0
[
7] 10 µM
[22] 10 µM
[15] 10 µM
[7] 5 µM
[22] 5 µM
[
15] 5 µM
control
0,9
control
0,9
control
0,9
0
0,1
-
0,1
0,1
0,3
0,5
0,7
1,1
-
0,1
0,1
0,3
0,5
0,7
1,1
-
0,1
0,3
0,5
0,7
1,1
1
/[ATP] µM-1
1/[ATP] µM-1
1/[ATP] µM-1
Figure 2. Kinetic data determined for derivatives 7, 15 and 22. ATP concentrations in
the reaction mixture varied from 1 to 100 µM. Compound concentrations used are
depicted in the plot and the concentration of the substrate was kept constant at 12.5
µM. Each point is the mean of two different experiments, each one analyzed in
duplicate.
As there is conservation between ATP binding sites among the kinome, kinase
inhibitors targeting the ATP binding site may have a lack of selectivity and potential
undesirable side effects precluding further drug development. To start the assessment
of kinase selectivity in this new family of compounds, the three previous selected
compounds were tested against a limited set of kinases based on structural
assessment. In all cases, compounds were tested at 10 µM in forty-three CDC7
structurally related kinases such as cyclin dependent kinases (CDKs), proto-oncogene
serine/threonine-protein kinases (PIMs), casein kinases (CSNKs) and dual specificity
tyrosine-phosphorylation-regulated kinases (DYRKs). The three compounds tested
were very selective in the kinase panel evaluated, as no special potent inhibition was
observed for the evaluated kinases besides CDC7 (Figure 3). The final selectivity score
or “S” score, a quantitative measure of compound selectivity calculated by dividing the
number of kinases that compounds bind to by the total number of distinct kinases
29
tested, is 0.023 for the three compounds as they only interacts with 1 out of 43
13

kinases evaluated with a percentage of activity below 5%. These data point to a
selective new family of CDC7 inhibitors despite being ATP-competitive.
Compound 7
Compound 15
Compound 22
1
00
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
Figure 3. Kinase profiling for CDC7 inhibitors 7, 15 and 22.
In silico binding mode studies
The high degree of kinase selectivity of our purine-based inhibitor for CDC7 compared
to structurally similar kinases such as casein kinase family members, led us to explore
their potential binding mode by computational techniques. In fact, CDC7 has unique
30
features in its structure with respect to other kinases. To gain insights into the
interaction between CDC7 and the purine derivatives here described, a molecular
modeling study was performed using the published crystal structure of the CDC7
31
kinase in complex with PHA-767491 (PHA) (PDB entry: 4F9B). All purine derivatives
with an IC50 below 10 µM were docked into the ATP-binding site of the protein using
Glide software (see Experimental Section) based in our experimental ATP competition
results. To assess the reliability of the results, the same docking protocol was applied
to adenosine diphosphate (ADP) and PHA, the known CDC7 inhibitor, both crystallized
into the protein. No differences with the X-ray experimental complex (Figure S1) were
shown, validating the docking procedure. Then, MM-GBSA calculations were
14

performed to postprocess and re-score the ligand docking solutions and to correlate
their structural binding mode with the experimental activity against CDC7. The
predicted binding free energies values (∆Gbind) against the experimental activities
(
pIC ) reveal a moderate relationship between these two variables (correlation
50
2
coefficient R = 0.72) (Figure S2). The ∆Gbind values for CDC7 inhibitors vary between -
9.57 and -41.41 kcal/mol. This computational protocol (docking + MM-GBSA) has
been successfully used to study correlations between theoretical affinities and
4
32-33
experimental activities,
and here we used it to further study the structure-activity
relationship of thirty four different purine derivatives against CDC7. In this way,
compounds 2, 15, 26, 30, 32, and 33 were selected in order to explore the effect of the
nature and the length of the linker in CDC7 inhibition.
Computational analysis suggests that there are two key interactions for the activity of
the new CDC7 inhibitors: two hydrogen bonds between the purine central core and
Pro135 and Leu137 (Figure 4). The lack of one of these two interactions is translated in
a loss of the activity confirming the experimental results obtained for purines N-
substituted such as 7b, 15b, and 23b-29b or thieno[2,3-d]pyrimidines 36 and 37, all of
them inactive compounds.
Among active compounds, the computational study has allowed us to rationalize the
relationship between the chemical structure and biological activity. Regarding the
length of the linker between the purine moiety and the phenyl ring, the activity
increases when the carbonyl group is suppressed (i.e. compound 15). In contrast, the
activity of the inhibitors decreases when linker disappears and the phenyl moiety is
directly bonded to the sulfur atom (i.e. compound 30). The reason could be that the
phenyl moiety in 15 make more extensive van der Waals contacts with the catalytic
30
residues Lys90, Asp196, Ser181 and Asn182. In fact, a clear π-cation interaction
between compound 15 and Lys90 stabilizes the binding of this compound. In contrast,
15

the less potent compound 30 interacts with His139 through a π-π stacking interaction,
being less exposed to other important residues as mentioned above.
SER70
SER70
LYS90
15
2
PRO135
PRO135
MET134
MET134
LYS90
LEU137
LEU137
ASP196
ASP196
HIS139
HIS139
ASN182
SER181
SER181
ASN182
SER70
LYS90
SER70
3
0
26
PRO135
PRO135
MET134
MET134
LYS90
LEU137
LEU137
ASP196
ASP196
HIS139
HIS139
SER181
ASN182
SER181
ASN182
Figure 4. Binding modes of compound 15 vs. 2, 26, and 30 in the catalytic site of
CDC7.
Considering the linker nature, the activity of purine-based CDC7 inhibitors increases
when a methylene moiety substitutes the carbonyl group (i.e. compound 26 vs 2). This
difference in the biological activity can be explained by analyzing the components of
the ∆Gbind energy calculated by MM-GBSA (Table 2). Compounds 15 and 26 have
similar binding energies, with relative ∆Gbind between -48 and -49 kcal/mol. However,
compound 2 has a relative ∆Gbind = -42 kcal/mol. The major difference in the binding
energies of these compounds lies in the polar contribution of the solvation energy
(
∆G ) where compound 2 presents a difference of 6.5 kcal/mol in comparison with
GB
compounds 15 and 26. Although this difference is not so significant, it makes the
molecule 2 more difficult to be desolvated prior to enter in the binding site. In this way,
16

the carbonyl group of the molecule 2 locates in the outer part of the cavity, being more
exposed to the solvent (Figure 4).
Table 2. Binding energies (kcal/mol) calculated by MM-GBSA of compounds 2, 15, 26,
30, 32 and 33.
Comp.
2
IC50
ΔGbind
ΔEMM
ΔGsol
ΔGGB*
29.83
23.39
(
µM)
4.15
0.14
ΔEinternal*
4.332
1.369
Hbondcor*
-0.334
0.102
P
ackingcor*
S
elfContcor*
0.003
0.299
ΔECoulomb
-24.424
-24.221
*
ΔEvdW*
-41.48
-37.62
ΔGSA*
-10.03
-10.55
-42.27
-48.04
-0.172
-0.810
15
26
30
32
33
0.14
1.38
3.06
0.19
-49.57
-45.01
-41.65
-48.71
3.030
-0.044
1.249
0.185
-0.703
-0.613
-0.499
-0.607
-0.703
-0.613
-0.499
-0.607
0.012
0.008
0.116
0.055
-24.942
-25.036
-26.940
-24.844
-39.09
-31.76
-29.71
-33.13
23.32
21.94
23.99
19.05
-10.58
-8.82
-9.67
-9.17
*Contribution to the total energy from different components: ΔEinternal=Covalent binding energy; Hbondcor.=Hydrogen-
bonding correction; Packingcor.=ꢀ- ꢀ packing correction; SelfContcor=Self-contact correction; ΔECoulomb=Coulomb energy;
ΔEvdW=Van der Waals energy; ΔGGB=Generalized Born electrostatic solvation energy; ΔGSA=Lipophilic energy.
Regarding the heteroatom of the linker, when the sulfur (compound 15) is replaced by
an oxygen (compound 33), the activity as well as the ∆Gbind does not change. However,
when it is replaced by an amine group (compound 32), both CDC7 inhibition activity
and relative ∆Gbind decreases. This variation may be also explained by analyzing the
components of the ∆Gbind energy. Compound 15 has a relative ∆Gbind of -48 kcal/mol,
while compound 32 has a relative ∆Gbind = -41 kcal/mol. The major difference between
these two compounds lies in the van der Waals interaction energy, 8 kcal/mol between
them (Table 2).
Finally, it is worth mentioning that the main deviation in the binding mode of purine-
based CDC7 inhibitors compared to the well-characterized PHA is the presence of a
hydrogen-bond interaction in the hinge region between the backbone carbonyl group of
Pro135 and the heterocyclic NH and, the lack of H-bonding with catalytic residues such
as Lys90 or Asp196 (Figure 4 and S1). Therefore, PHA binds not only to CDC7 but
34
also to different structural related kinases such as CDK1 and CDK2, being a dual
35
inhibitor of CDK9. Somehow, the new described purine-based compounds, less
potent but much more selective than PHA inhibitor, point to a new well-conserved
binding mode that gives new hints for further potency improvements. One of the main
17

reasons that can explain this selectivity may be that purine-based compounds here
described form van der Waals contacts with Ser70, Met118 and His139 which are
some of the most variable residues among protein kinases. Even more, compounds 15
and 26 present halogen bond interactions with Ser70 (Figure 4).
Blood brain barrier (BBB) permeability
Brain penetrant chemical probes and/or drug candidates are needed when a central
nervous system disease, such as ALS, is going to be studied in animal models. Here,
we have used a parallel artificial membrane permeability assay (PAMPA) using porcine
lipid to emulate the human blood brain barrier (BBB) in order to determine this essential
drug-like characteristic for our compounds of interest. We calculated the predicted
brain permeability for all the compounds with an IC value for CDC7 inhibition below 1
50
µM. In this assay, ten reference commercial drugs used clinically with known brain
permeability are included as controls. A good correlation between experimental-
2
described values was obtained Pe (exptl) = 0.849 (bibl) - 0.1282 (R = 0.9828) (Figure
S3). From this equation and following the pattern established in the literature for BBB
36
permeation prediction we could classify compounds as CNS+ when they present a
-
6
-1
permeability > 3.26 x 10 cm.s . Based on these results we can consider that a large
number of compounds, specifically 4-9, 11, 15, 17, 18, 20-29, 33, 35, and 38, are
predicted to be able to cross the BBB by passive permeation (Figure 5, Table 1 and
Table S3) and thus they can be considered for further pharmacological studies in
neurological diseases. Compounds 10, 13, 16, 19, and 34 are borderline for PAMPA
permeability (CNS+/CNS-), and finally compounds 12 and 14 are predicted to not be
able to cross the BBB (CNS-). Compound 39 precipitates under the experimental
conditions employed; therefore, there was not a clear UV spectrum after filtration
making it impossible to determine the BBB permeability.
18

SNC+
SNC+/SNC-
SNC -
Figure 5. PAMPA in vitro permeability (P ) plot of CDC7 inhibitors (green squares) and control
e
−
6
−1
drugs (black diamonds). CNSꢀ+ꢀ(green) for compounds with P ꢀ>ꢀ3.26ꢀ×ꢀ10 ꢀcmꢀs , CNS- (red)
for P ꢀ<ꢀ1.56ꢀ×ꢀ10 ꢀcmꢀs and CNS+/CNS- (orange) for 1.56ꢀ×ꢀ10 < P < 3.26ꢀ×ꢀ10 cmꢀs .
e
–6
−1
–6
−6
−1
e
e
Data represent the meanꢀ±ꢀSD from two independent experiments.
TDP-43 modulation by CDC7 inhibitors in cellular and in vivo models
Neuroprotective effect of CDC7 inhibitors in a neuronal cell model of induced TDP-43
phosphorylation by ethacrynic acid treatment
The next step was to determine if small molecule inhibition of CDC7 kinase activity, by
the new purine-based inhibitors may prevent TDP-43 phosphorylation in vitro. In order
to be sure that results of interest are produced by CDC7 inhibition, we tested a large
set of 23 compounds with brain permeability values compatible with the ability to cross
the BBB. The selected cell model was a neuronal cell-based model of induced TDP-43
phosphorylation driven by glutathione depletion. Human neuroblastoma cell line SH-
SY5Y cells were pre-treated with CDC7 inhibitors for 1 h and phosphorylation of
endogenous cellular TDP-43 was then induced with exposure to ethacrynic acid (EA)
19

37
for 24 h. Compound dose was selected as 10 µM after evaluating cell viability in the
presence or absence of CDC7 inhibitors at 5 and 10 µM in SH-SY5Y (Figure S4). As
seen in Figure 6, the addition of EA to the SH-SY5Y cells resulted in significant cell
death, which is partially prevented by several of the CDC7 inhibitors treatment. As
reference control, we used tideglusib, a GSK-3 inhibitor with neuroprotective activity by
decreasing TDP-43 phosphorylation.
Figure 6. Neuroprotective effect of CDC7 inhibitors at a fixed concentration of 10 µM in
neuroblastoma SH-SY5Y cells after being exposed to 20 µM of EA for 24 h.
Commercial GSK-3β inhibitor (Tideglusib 5 µM) was used as internal control. Cell
viability was measured by MTT assay. Data represent the mean ± SEM of 5 different
experiments (*p <0.05, **p <0.01, ***p <0.001, ****p <0.0001 significantly different from
+
+++
SH-SY5Y EA-treated cells,
p <0.0001 significantly different from Control).
To check if the neuroprotective effect of CDC7 inhibitors is produced by a decrease in
TDP-43 phosphorylation state, Western-blot analysis was done using a phospho-
specific (S409/410) anti-TDP-43 antibody. In Figure 7 the increase of TDP-43
phosphorylation caused by EA is shown together with the reduction of this aberrant
phosphorylation produced with the CDC7 inhibitors treatment. CDC7 inhibitors had no
effect on total cellular levels of TDP-43.
20

EA
Control
EA
7
15
17
22
pTDP-43
α-Tubulin
TDP-43
Figure 7. Effects of CDC7 inhibitors on TDP-43 phosphorylation in neuroblastoma SH-
SY5Y cells after being exposed to EA. Representative immunoblots showing pTDP-43
protein levels. Cells were exposed to EA (20 µM) for 24 hours in the presence or
absence of CDC7 inhibitors (10 µM). Data represent the mean ± SEM of 3 different
experiments. (***p <0.001, ****p <0.0001 significantly different from SH-SY5Y EA-
+
+
treated cells, p <0.01 significantly different from Control).
Furthermore, to fully check whether the inhibition of TDP-43 phosphorylation here
reported in SH-SY5Y by CDC7 inhibitors treatment reduces the amount of TDP-43
cytoplasmic localization, an immunofluorescence analysis was performed with
derivative 15. Results depicted in Figure 8 showed the decreased of cytosolic TDP-43
after the treatment with the CDC7 inhibitor 15, which is in agreement with previous
17
results. Moreover, two of the compounds described here, derivatives 7 and 15, have
been proved to be capable to restore the normal nuclear TDP-43 levels in immortalized
38
lymphocytes derived from either sporadic ALS or FTLD-TDP patients.
21

A)
Nucleus (HCS)
TDP-43
Merge
B)
Figure 8. Subcellular localization of TDP-43 detected by immunofluorescence after
compound 15 treatment in ethacrynic acid treated SH-SY5Y cells. (A) Cells were
exposed to EA (20 µM) in the presence or absence of compound 15 (10 µM) for 24
hours. TDP-43 localization was assessed by confocal laser scanning microscopy. (B)
Quantification of fluorescence intensity of TDP-43 inside and outside the nucleus was
determine in at least 30 different cells from 3 separate fields. Data represent the mean
±
SEM (magnification 63x). (*p <0.5 significantly different from SH-SY5Y EA-treated
+
cells, p <0.05 significantly different from Control).
22

Taken together, these results may point to CDC7 inhibition as a potential viable
strategy to reduce neurotoxic TDP-43 phosphorylation and cellular localization.
CDC7 inhibitor testing in transgenic C. elegans
In order to determine whether inhibition of CDC7 kinase activity by novel CDC7
inhibitors will prevent TDP-43 phosphorylation in vivo, we utilized as a starting point
transgenic Caenorhabditis elegans expressing human familial ALS mutant TDP-43
39
pan-neuronally (TDP-43 tg). C. elegans are a tractable invertebrate animal for
studying neurodegenerative disease, as they have a well-characterized differentiated
40
nervous system that includes the major neuronal types found in humans. TDP-43 tg
C. elegans exhibit age-dependent motor dysfunction, progressive neurodegeneration,
39
and accumulate robust amounts of phosphorylated TDP-43. The C. elegans cuticle is
41
a highly impervious barrier that prevents many small molecules from entering;
therefore, the concentration of small molecule required to see an effect in C. elegans is
typically high and may not reflect the effective dose that penetrates into the body of the
42
animals. However, despite these limitations, C. elegans have been successfully used
19, 43
to demonstrate in vivo activity of small molecules including kinase inhibitors.
To
assay CDC7 inhibitors in C. elegans, TDP-43 tg animals were raised on solid media in
the presence of the CDC7 inhibitors, grown for 6 days, and harvested for immunoblot
analysis. We tested the well characterized purine-based CDC7 inhibitors 7, 15 and 22
and found that levels of phosphorylated TDP-43 were significantly reduced following
treatment with inhibitor 7, with no apparent effect on total levels of TDP-43 (Figure 9).
Compounds 15 and 22 had more variable effects on TDP-43 phosphorylation, which
may be a consequence of a differential ability to effectively cross through the cuticle of
C. elegans. Nevertheless, we show that compound 7 is able to reduce TDP-43
phosphorylation in vivo.
23

1
.5
A)
0
10 25 50 75 µM[7]
pTDP-43
[7]
49kDa
1.0
4
7kDa
0
.5
.0
totalTDP-43
tubulin
*
0
55kDa
2.5
2.0
1.5
1.0
0.5
0.0
B)
0
10 25 50 75 µM [15]
[15]
49kDa
pTDP-43
totalTDP-43
tubulin
47kDa
55kDa
C)
4
3
2
1
0
0
10 25 50 75
µM [22]
[22]
pTDP-43
4
9kDa
totalTDP-43
tubulin
4
7kDa
5kDa
5
Figure 9. Effects of CDC7 inhibitors on TDP-43 phosphorylation in C. elegans. Stage-
matched populations of transgenic C. elegans expressing human TDP-43 (M337V)
were grown in the presence of control (DMSO) or the indicated concentration of
inhibitor for 6 days. (A) C. elegans treated with compound 7 had significantly
decreased levels of TDP-43 phosphorylation at 75 µM. Data represent the mean ±
SEM of 4 different experiments. p-values, one-way ANOVA with Dunnett’s multiple
comparisons test versus Control: 10 μM, 0.867; 25 μM, 0.896; 50 μM, 0.490; 75 μM,
0.024. (B) C. elegans treated with compound 15 had no significant changes to the
24

levels of TDP-43 phosphorylation. p-values, one-way ANOVA with Dunnett’s multiple
comparisons test versus Control: 10 μM, 0.985; 25 μM, 0.988; 50 μM, 0.999; 75 μM,
0.977. (C) compound 22 had no significant changes to the levels of TDP-43
phosphorylation. p-values, one-way ANOVA with Dunnett’s multiple comparisons test
versus Control: 10 μM, 0.387; 25 μM, 0.922; 50 μM, 0.999; 75 μM, 0.913.
Treatment of TDP-43 (A315T) transgenic mice model with a CDC7 inhibitor
The encouraging obtained results prompted us to test the mercaptopurine-based CDC7
inhibitor 22 to a mouse model of ALS, following our working hypothesis that in vivo
modulation of TDP-43 proteinopathy by CDC7 inhibitors may be a new therapeutic
approach for ALS. Selection of compound 22 was done based in efficacy shown
decreasing TDP-43 phosphorylation in human cell line (Figure 7).
We have used the TDP-43 (A315T) transgenic mice, which is one of the first
44
experimental models of ALS based on mutations in TDP-43 protein. The transgene,
driven by mouse prion promoter, is expressed in all neurons and glia in the spinal cord.
Due to TDP-43 overexpression, transgenic mice develop motor neuron disease. We
designed a chronic treatment in which compound 22 was administered intraperitonially
once a day for 30 days. Four groups each with 8 animals were studied: wild type +
vehicle, wild type + compound 22, transgenic + vehicle and transgenic + compound 22.
The trial was started with males of 65 days old. Weight was daily logged and no toxicity
was observed across the experiment (Figure S5). Behavioral tests such as rotarod and
hindlimb clasping test were weekly performed to evaluate motor coordination and
muscle strength in the animals. An amelioration in the disease progression and muscle
impairment was observed in the treated group in both tests (Figure 10). Both tests
measure pathological symptoms, such as muscle weakness, balance disturbance and
spasticity, that resemble those present during the human disease. The treatment with
the CDC7 inhibitor was highly effective in attenuating clasping behavior present in
TDP-43 (A315T) mice from the third week of treatment onwards (Figure 10B).
Additionally, the treatment with compound 22 produced a recovery in rotarod
25

performance after five weeks of treatment (Figure 10C). CDC7 inhibition had no effects
in wild type animals, as measured during rotarod and hindlimb clasping tests (Figure
10B and 10C).
Compound 22
i.p. 5 mg/Kg daily
Prp-TDP43A315T
A)
CDC7 INHIBITOR
P0
P65
P95
B)
C)
Wt vehicle
Wt+comp. 22
TgTDP-43 vehicle
TgTDP-43+comp. 22
Figure 10. Experimental design (A) and behavioral studies (B, C). (B) Hindlimb
clasping and (C) rotarod test results of wild-type and transgenic TDP-43 (A315T) mice
treated with compound 22 or vehicle. Values are expressed as means ± SEM; N ≥ 6
animals in each group. Two-way ANOVA test was used for statistical analysis followed
by the Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001 vs. WT-
Veh group; &p < 0.05, &&p < 0.01, &&&p < 0.001 vs.TDP-43-Veh group).
Moreover, after Western-blot analysis in the spinal cord tissue of TDP-43 (A315T)
mice, we observed a decrease in p-TDP-43 (Figure 11). These results show that
compound 22 is able to in vivo penetrate in the central nervous system confirming
previous PAMPA predictions and, thus, it is able to delay the muscle strength
deterioration produced in these TDP-43 (A315T) transgenic mice when it is measured
by rotarod and hindlimb clasping test. Finally, all these data correlate with in vivo
26

decrease of TDP-43 phosphorylation in spinal cord after the treatment with the CDC7
inhibitor.
WT
TDP-43
Veh.
C22
Veh.
C22
pTDP-43
$
TDP-43
GAPDH
WT
TDP-43
Veh.
C22
Veh.
C22
pTDP-43
TDP-43
GAPDH
Figure 11. Phosphorylated TDP-43 levels in the spinal cord of wild-type and transgenic
TDP-43 (A315T) mice treated with compound 22 or vehicle. Representative
immunoblots are shown. Data represent the mean ± SEM of 8 observations per group.
Data were assessed by one-way ANOVA followed by the Bonferroni test ($p < 0.05 vs.
WT-Veh group, *p < 0.05 vs. TDP43-Veh group).
Conclusions
CDC7, a kinase involved in different oncologic processes, has emerged as a new
target for different neurological conditions where the nuclear protein TDP-43 is involved
9
including ALS, FTLD and other rare diseases such as Alexander’s or Pick’s diseases.
Recently, a new Alzheimer’s-like dementia disorder has been described characterized
by pathological TDP-43 and termed LATE (Limbic-predominant age-related TDP-43
45
encephalopathy). In this context, prevention of TDP-43 phosphorylation through the
design and development of brain-permeable selective CDC7 inhibitors emerges as a
potential therapeutic strategy for these severe and unmet diseases. In this manuscript
27

we have discovered, using a chemical genetic approach, a new compound able to
target CDC7. After a hit-to-lead optimization process, a family of small molecules with a
central purine core was obtained as completely new CDC7 inhibitors. These
compounds have shown an ATP competitive binding mode but with a wide selectivity
over more than 40 structurally related kinases. Furthermore, these CDC7 inhibitors are
predicted to cross the BBB using a PAMPA methodology and are able to reduce TDP-
4
3 phosphorylation both in human cell lines and in vivo C. elegans model with
overexpression of hTDP-43. Finally, all these data have been confirmed in a TDP-43
A315T) transgenic mice model, where chronic administration of one of
(
mercaptopurine-based CDC7 inhibitors here described, compound 22, produced a
decrease in the severity of the clinical symptoms of the disease and an in vivo
reduction of TDP-43 phosphorylation in the spinal cord of the animals.
Altogether, purine-based CDC7 inhibitors here reported are valuable pharmacological
tools for the study of the role of CDC7 in neurological conditions and promising
candidates to be developed as innovative drugs for the pharmacotherapy of ALS and
others TDP-43-pathies
28