Discovery of Multitarget Agents Active as Broad-Spectrum Antivirals and Correctors of Cystic Fibrosis Transmembrane Conductance Regulator for Associated Pulmonary Diseases

Article abstract of DOI:10.1021/acs.jmedchem.6b01521

Enteroviruses (EVs) are among the most frequent infectious agents in humans worldwide and represent the leading cause of upper respiratory tract infections. No drugs for the treatment of EV infections are currently available. Recent studies have also linked EV infection with pulmonary exacerbations, especially in cystic fibrosis (CF) patients, and the importance of this link is probably underestimated. The aim of this work was to develop a new class of multitarget agents active both as broad-spectrum antivirals and as correctors of the F508del-cystic fibrosis transmembrane conductance regulator (CFTR) folding defect responsible for >90% of CF cases. We report herein the discovery of the first small molecules able to simultaneously act as correctors of the F508del-CFTR folding defect and as broad-spectrum antivirals against a panel of EVs representative of all major species.

Full text of DOI:10.1021/acs.jmedchem.6b01521

Subscriber access provided by Fudan University
Article
Discovery of Multi-Target Agents Active as Broad-Spectrum
Antivirals and Correctors of Cystic Fibrosis Transmembrane
Conductance Regulator (CFTR) for Associated Pulmonary Diseases
Sabrina Tassini, Liang Sun, Kristina Lanko, Emmanuele Crespan, Emily Langron, Federico Falchi,
Miroslava Kissova, Jorge I. Armijos-Rivera, Leen Delang, Carmen Mirabelli, Johan Neyts, Marco Pieroni,
Andrea Cavalli, Gabriele Costantino, Giovanni Maga, Paola Vergani, Pieter LEYSSEN, and Marco Radi
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01521 • Publication Date (Web): 25 Jan 2017
Downloaded from http://pubs.acs.org on January 26, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of MultiꢀTarget Agents Active as
BroadꢀSpectrum Antivirals and Correctors of
Cystic Fibrosis Transmembrane Conductance
Regulator (CFTR) for Associated Pulmonary
Diseases
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
b
c
d
Sabrina Tassini, Liang Sun, Kristina Lanko, Emmanuele Crespan, Emily Langron,
e,f
c
c
b
Federico Falchi, Miroslava Kissova, Jorge I. ArmijosꢀRivera, Leen Delang, Carmen
b
b,*
a
e,f
a
Mirabelli, Johan Neyts, Marco Pieroni, Andrea Cavalli, Gabriele Costantino, Giovanni
c
d
b
a,*
Maga, Paola Vergani, Pieter Leyssen, and Marco Radi
a
P4T Group, Dipartimento di Scienze degli Alimenti e del Farmaco, Università degli Studi di
Parma, Viale delle Scienze, 27/A, 43124 Parma, Italy;
b
Laboratory of Virology and Experimental Chemotherapy, Rega Institute for Medical
Research, KU Leuven, Minderbroedersstraat 10, 3000, Leuven, Belgium;
c
Istituto di Genetica Molecolare, IGMꢀCNR, Via Abbiategrasso 207, 27100 Pavia, Italy;
d
Department of Neuroscience, Physiology and Pharmacology, University College London,
Gower Street, WC1E 6BT London, UK;
e
CompuNet, Istituto Italiano di Tecnologia, Via Morego 30, Iꢀ16163 Genova, Italy;
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
f
Department of Pharmacy and Biotechnology, University of Bologna, Via Belmeloro 6, Iꢀ
4
0126 Bologna, Italy.
KEYWORDS: PI4KIIIβ, F508delꢀCFTR, broadꢀspectrum antivirals, enterovirus, cystic
fibrosis, multiꢀtarget.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ABSTRACT
Enteroviruses (EVs) are among the most frequent infectious agents in humans worldwide
and represent the leading cause of upper respiratory tract infections. No drugs for the
treatment of EV infections are currently available. Recent studies have also linked
enterovirus infection with pulmonary exacerbations, especially in cystic fibrosis (CF)
patients, and the importance of this link is probably underestimated. The aim of this work
was to develop a new class of multiꢀtarget agents active both as broadꢀspectrum antivirals
and as correctors of the F508delꢀCFTR folding defect responsible for >90% of CF cases. We
report herein the discovery of the first smallꢀmolecules able to simultaneously act as
correctors of the F508delꢀCFTR folding defect and as broadꢀspectrum antivirals against a
panel of enteroviruses representative of all major species.
INTRODUCTION
Enteroviruses (EVs) are positiveꢀsense single stranded RNA viruses, classified into 12
species, including four human enterovirus species (EVꢀA to EVꢀD), three species of human
1
rhinoviruses (RVꢀA to RVꢀC) and five enterovirus species that only infect animals. EVs are
responsible for a great variety of clinical manifestations, especially in young children, which
may result in lifeꢀthreatening neurological complications (e.g. encephalitis, meningitis and
2
ꢀ4
poliomyelitisꢀlike paralysis). Furthermore, RV infections are now considered one of the
major causes of acute exacerbations in chronic pulmonary diseases like asthma, chronic
ACS Paragon Plus Environment
2

Page 3 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
obstructive pulmonary disorder (COPD) and cystic fibrosis (CF) in children and adults.
Physicians pay particular attention to patients that already suffer from respiratory diseases,
such as CF or asthma, as they could be particularly affected by an additional enterovirus
6
infection. An increasing number of studies also suggest that respiratory viruses, in particular
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
enterovirus and rhinovirus, contribute significantly to CF pulmonary exacerbations,
7
hospitalization, decreased lung function and predisposition to bacterial colonization. The
mechanistic link between viral infections and deterioration of CF lung function is not fully
8
understood and their impact is probably underestimated, especially in young children.
Despite their high clinical and socioeconomic impact, to date there is no approved antiviral
therapy for the prophylaxis and/or the treatment of enterovirus infections, and the
management of patients is currently limited to symptomatic treatment and supportive care.
Therefore, there is an unmet need for broadꢀspectrum antiviral drugs as a rapid defense
strategy against enterovirus infections and virusꢀrelated exacerbations.
Nowadays, host factors are considered as very attractive targets for the development of
antiviral drugs because they are unlikely to mutate and develop resistance in response to
9
therapy. Moreover, since viruses belonging to the same genus or family usually share the
same cellular pathways for replication, targeting a host factor may allow the development of
1
0
effective broadꢀspectrum antiviral compounds. Although some toxicity risks may be
expected from inhibiting a host factor, it should be kept in mind that most drugs currently
used in therapy target host proteins with excellent therapeutic outcomes and acceptable safety
profiles. In particular, it has been well documented that the host lipid kinase
phosphatidylinositol 4ꢀkinase IIIβ (PI4KIIIβ) is critical for RNA replication of several
1
1ꢀ14
enteroviruses.
PI4KIIIβ belongs to the phosphatidylinositol 4ꢀkinases (PI4Ks) that
synthesize phosphatidylinositol 4ꢀphosphate (PI4P) from phosphatidylinositol (PI). PI4P is
involved in signaling and cellular trafficking mainly at the Golgi and transꢀGolgi network
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(TGN), it contributes to defining the characteristics of plasma membranes and it activates a
1
5ꢀ18
variety of ion channels, including CFTR.
Four PI4K isoforms have been identified in
mammals, classified as type II (PI4KIIα and PI4KIIβ) or type III (PI4KIIIα and PI4KIIIβ)
19
based on their primary sequences and catalytic properties. Type III PI4Ks are hijacked by
several ss(+)RNA viruses (especially from Flaviviridae, Picornaviridae and Coronaviridae
families) to remodel cellular membranes and generate PI4P lipidꢀenriched organelles
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
20
specialized for viral replication.
A few PI4KIIIβ inhibitors with antiviral activity against a panel of picornaviruses have been
2
1ꢀ23
reported recently (Figure 1).
Generally, chemical inhibition of PI4KIIIβ does not
24
influence cell viability. One possible explanation might be that while the small amounts of
PI4P produced by other PI4K isoforms could be enough to support cell trafficking and
20
signaling, it would not be sufficient to sustain viral RNA synthesis. A major aim in the
development of PI4KIIIs inhibitors is to achieve selective inhibition of the α or β isoforms.
Among known PI4KIIIs inhibitors, compound 1 (PIK93) is about 100ꢀfold more potent
25,26
against the PI4KIIIβ isoform, although it also has detectable activity towards PI3ꢀkinases.
Figure 1. Representative PI4KIIIβ inhibitors.
ACS Paragon Plus Environment
4

Page 5 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Considering the growing need for novel broadꢀspectrum antivirals to fight emerging
epidemics and the link between respiratory viruses and pulmonary exacerbation in cystic
fibrosis patients, our aim was the development of a new class of multiꢀtarget agents active
both as broadꢀspectrum antivirals (by targeting PI4KIIIβ) and as correctors of the F508delꢀ
CFTR folding defect responsible for >90% of CF cases. We here report the discovery of the
first smallꢀmolecule compounds able to simultaneously act as moderately efficacious
correctors of the F508delꢀCFTR folding defect and broadꢀspectrum antivirals against a panel
of enteroviruses (linked to CF pulmonary exacerbations).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Drug repurposing and polypharmacology are two very attractive approaches in modern
drug discovery. The first offers the possibility of recycling known drugs or advanced drug
candidates developed for a different disease. The second results in simultaneous action on
27,28
different targets/diseases with a single, rationally designed drug.
In particular,
polypharmacology aims at producing multiꢀtarget agents whose interference with multiple
biochemical pathways offers an advantage ꢀ in terms of drug load, efficacy and safety ꢀ over
combination therapy. This approach is well suited to complex diseases that generally require
the simultaneous administration of many different drugs. Considering the increasing number
of reports on the connection between enterovirus infections and pulmonary exacerbations in
CF patients, we reasoned that an ideal drug candidate for such closely related diseases might
be a multiꢀtarget agent able to act, at the same time, on proteins/pathways implicated in
enterovirus replication (PI4KIIIβ) and on F508delꢀCFTR biogenesis. At the beginning of this
work, the Xꢀray structures of the above targets were not available for a structureꢀbased study.
We therefore developed a PI4KIIIβ homology model to be used for the design of PI4KIIIβ
inhibitors, selecting those whose chemical scaffolds resemble known CFTR
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
correctors/potentiators. The structure of the complex of PI3Kγ with compound 1 (PDB ID:
2
6
2
CHZ) has been used to build the homology model of PI4KIIIβ by using Prime software
(see methods): this structure shows an identity of 30%, a positive of 52% and a score of 322.
The presence of 1 in the structure of PI3Kγ allowed us to identify its likely binding site in
PI4KIIIβ and hypothesize its binding mode. A 10 ns molecular dynamics simulation on the
modeled PI4KIIIβ protein containing compound 1 was performed using the software
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29
Desmond. In the latter (equilibrated) part of the trajectory (last 2 ns) 100 frames were
extracted and clusterized on the basis of RMSD. Five clusters were obtained. All PI4Kβ
30
inhibitors available in Pubchem were docked in the compound 1 binding site of each cluster
and the frame with the best correlation between docking score and enzymatic activity was
selected for virtual screening. A highꢀthroughput docking (HTD) approach was then applied
to the compound 1 binding site in our PI4KIIIβ model to identify high affinity hits within the
3
1
Asinex database collection. Compound selection was based on the ranking score and visual
inspection of the PI4KIIIβ catalytic site, but also took into account the 2D similarity to
known CFTR correctors/potentiators. Thirteen commercially available compounds (5-17,
32,33
Figure 2), four of which (6, 11, 16, 17) resemble known CFTR correctors,
were selected
for biological investigation. These computational results were confirmed on the recently
released crystal structure of PI4KIIIβ coꢀcrystallized with compound 1, and this structure
34
(PDB ID: 4D0L) was used for all the following simulations.
These commercially available compounds were then tested both against the PI4KIIIβ
enzyme and in a virusꢀcellꢀbased replication assay. In particular these compounds were
evaluated for antiviral activity against a panel of enteroviruses that are representative of all
major species: enterovirus group A (EV71), group B (coxsackievirus B3, CVB3, and
echovirus 11, ECHO11), group C (poliovirus 1, PV1), group D (enterovirus 68, EV68),
rhinovirus group A (RV02) and rhinovirus group B (RV14). Among the selected compounds,
ACS Paragon Plus Environment
6

Page 7 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
only the bithiazole 17 showed activity in cellꢀfree and cellꢀbased assays and possesses a
3
3,35
chemical scaffold (the bithiazole) of a known family of CFTR correctors (Figure 2).
Compound 17 was therefore selected as a promising starting point for further structureꢀbased
optimization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
H
N
HN
NH
O
O
S
N
N
N
O
O
H
2
N
N
N
NH
N
N
N
O
N
O
O
F
6
NH
7
HN
N
5
N
O
N
F
3
C
N
8
9
O
O
HO
NC
O
O
N
H
N
S
N
O
HN
F
N
N
HN
O
O
O
N
N
N
N
Cl
N
NH
O
N
N
N
O
H
N
10
11
O
12
13
S
S
O
S
N
O
O
N
N
NH
S
O
O
S
N
H
O
O
N
H
17
OH
N
O
O
O
HN
N
N
OH
F
PI4KIIIβ: IC50 = 0.48 ꢀM
N
O
O
N
N
HO
EV71 (EV group A): EC50 = 0.92 ꢀM
CVB3 (EV group B): EC50 = 2.17 ꢀM
PV1 (EV group C): EC50 = 3.09 ꢀM
NH
O
N
O
H
1
4
15
16
Figure 2. Chemical structure of compounds selected by virtual screening and activity profile
of the hit compound 17.
Chemistry
Compound 17 was initially resynthesized to validate the biological activity of the
commercial sample and to set up a synthetic protocol for its chemical diversification starting
from cheap and commercially available building blocks (Scheme 1).
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
.
2
Scheme 1. Reagents and conditions: (a) EtOH, reflux, 12 h, 95%; (b) 48% aqueous HBr, Br ,
1
,4ꢀdioxane, 60 °C, 3 h, 90%; (c) (i) benzoyl isothiocyanate, DCM, rt, 12 h, (ii) NaOH 1N,
3
THF, reflux, 3 h, 72%; (d) EtOH, reflux, 1 h, 84%; (e) acetyl chloride, Et N, DCM, rt, 15 h,
77%.
Thiourea 18 was condensed with 3ꢀchloroꢀ2,4ꢀpentadione 19 in refluxing ethanol to afford
36
1
ꢀ(2ꢀaminoꢀ4ꢀmethylthiazolꢀ5ꢀyl)ethanone 20 in nearly quantitative yield, followed by
bromination α to the carbonyl to give compound 21. The subsequent condensation of
intermediate 21 with 1ꢀ(3ꢀacetylphenyl)thiourea 23 gave bithiazole 24 that was finally Nꢀ
acetylated to obtain the desired compound 17. Thiourea 23 was synthesized by reaction of 3’ꢀ
aminoacetophenone 22 with benzoyl isothiocyanate, followed by a basic hydrolysis to
3
7
remove the benzoyl group.
Docking studies on compound 17 (see Molecular modeling and SAR section) showed a
pattern of interactions within the ATPꢀbinding pocket of PI4KIIIβ very similar to that of the
reference compound 1. The proposed binding mode of 17 suggested that two main portions of
this molecule could be functionalized to explore the biologically relevant chemical space: i)
ACS Paragon Plus Environment
8

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the 2ꢀamino group on the 4ꢀmethylthiazole ring (left part) and, ii) the phenyl ring (right part).
We first explored the chemical space around the left part of compound 17, introducing
bulkier groups and urea/thiourea functions in place of the acetamide moiety. The intermediate
24 represents in fact an advanced intermediate that could be easily functionalized on the 2ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
amino group to give a series of functionalized derivatives (25a-g and 26a-d) (Scheme 2).
1
Scheme 2. Reagents and conditions: (a) method A (for 25a-c), R COCl (for 25a,b) or
1
1
(
R CO)
2
O (for 25c), Et
3
N, DCM, rt, 12ꢀ15 h, 65ꢀ80%; method B (for 25d-f) R COCl, Et
3
N,
1
DCM, reflux, 15 h, 65ꢀ75%; method C (for 25g), (R CO)
2
O, Et
3
N, DMF, 50 °C, 12 h, 63%;
2
(b) R NCX, pyridine, reflux, 12ꢀ18 h, 52ꢀ69%.
Compound 24 was first reacted with different acyl chlorides or anhydrides to obtain
compounds 25a-g, while the urea/thiourea derivatives 26a-d were synthesized by reacting 24
with the appropriate isocyanates/isothiocyanates. We also decided to replace the acetamide
moiety of compound 17 with chainꢀextended ureidic groups for the SAR development.
Unfortunately, the synthesis of compounds 30a,b following the approach described above
would have required very expensive isocyanates. Thus, an alternative synthetic approach was
used for the synthesis of 30a,b (Scheme 3): starting from 1ꢀ(2ꢀaminoꢀ4ꢀmethylthiazolꢀ5ꢀ
yl)ethanone 20, reaction with diphenyl carbonate gave good yields of the desired phenyl
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
carbamate 27 that reacted readily with the appropriate amines to give the urea intermediates
38
2
8a,b. Similar to the synthesis of compound 17, the bromination α to the carbonyl and the
subsequent condensation of intermediates 29a,b with the 1ꢀ(3ꢀacetylphenyl)thiourea 23, gave
the desired compounds 30a,b.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Reagents and conditions: (a) diphenyl carbonate, NaH, DMF, rt, 30 min, 67%; (b)
4ꢀchlorobenzylamine (for 28a) or 1ꢀadamantylamine (for 28b), THF, 50 °C, 5ꢀ6 h, 61ꢀ93%;
(
c) 48% aqueous HBr, Br , 1,4ꢀdioxane, 60 °C, 3 h, 88ꢀ92%; (d) EtOH, reflux, 1 h, 78ꢀ83%.
2
We next explored the right part of compound 17, keeping the 2ꢀacetamido group on the left
part of the molecule unchanged and modifying the substitution pattern of the phenyl ring on
the right part. Since the acetamide moiety on the left part of the molecule was conserved,
intermediate 20 was conveniently acetylated before the bromination α to the carbonyl
(Scheme 4). Final compounds 35a-d were quickly obtained in good yields by reacting 32
with substituted thioureas 34a-d, previously synthesized from the corresponding amines 33a-
d.
ACS Paragon Plus Environment
1
0

Page 11 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4. Reagents and conditions: (a) acetyl chloride, pyridine, THF/DCM, 0 °C, 3 h,
2
87%; (b) Br , 1,4ꢀdioxane, 50 °C, 22 h, 84%; (c) (i) benzoyl isothiocyanate, DCM, rt, 12 h,
(
ii) NaOH 1N, THF, reflux, 2 h, 72ꢀ78%; (d) EtOH, reflux, 1h, 71ꢀ85%.
Then we decided to modify the central bithiazole scaffold of the hit compound 17, to get
additional SAR information. As described in Scheme 5, we first introduced an imidazole ring
by reacting intermediate 21with the 1ꢀ(3ꢀacetylphenyl)guanidine 36, obtained by treating 3’ꢀ
39
aminoacetophenone 22 with cyanamide. Compound 38 was thus synthesized by acylation of
the intermediate 37 with acetyl chloride.
4
0
A scaffold hopping approach (FAFꢀdrugs2 server) was also employed to identify
alternatives to the bithiazole scaffold: among the molecules proposed by the software, the
asymmetrical N,N’ꢀdiarylthiourea scaffold was considered the most promising on the basis of
its synthetic accessibility and the antiviral activity of some closely related analogues reported
41
in the literature. Intermediate 41 was easily obtained by addition of 3’ꢀaminoacetophenone
22 to 3ꢀnitrophenyl isothiocyanate 39, followed by reduction of the nitro group with iron
powder in acidic ethanol (Scheme 6). The subsequent acylation of the amino group led to
final compounds 42a,b.
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 5. Reagents and conditions: (a) cyanamide, HNO
3 2
, EtOH/H O, reflux, 24 h, 73%; (b)
Et N, EtOH, reflux, 12 h, 82%; (c) acetyl chloride, Et N, DCM, rt, 8 h, 47%.
3
3
Scheme 6. Reagents and conditions: (a) DCM, rt, 18 h, 88%; (b) Fe, HCl, EtOH, reflux, 2 h,
5%; (c) acetyl chloride (for 42a) or trimethylacetyl chloride (for 42b), pyridine, THF, rt, 2
h, 67ꢀ69%.
7
ACS Paragon Plus Environment
1
2

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Moreover we noted high chemical similarity between compound 17 and compound 46, a
known inhibitor of DCꢀSIGN (dendritic cell (DC)ꢀspecific intercellular adhesion moleculeꢀ3
4
2
grabbing nonintegrin). The role of DCꢀSIGN in the binding and transmission of different
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
3
pathogens, including enteroviruses, has been well investigated. So we decided to evaluate
the antiviral effect of compound 46 in a virusꢀcellꢀbased assay and its activity on PI4KIIIβ.
As described in Scheme 7, compound 46 was synthesized following the procedure previously
reported for compound 17.
Scheme 7. Reagents and conditions: (a) (i) benzoyl isothiocyanate, DCM, rt, 12 h, (ii) NaOH
1N, THF, reflux, 3 h, 80%; (b) EtOH, reflux, 30 min, 77%; (c) propionyl chloride, Et
3
N,
DCM, rt, 8 h, 68%.
Finally, we decided to prepare two compounds (related to the hit 17) known in the CFTR
field to evaluate the potential role of the CFTR channel in viral replication: i) compound 47a,
which is active in correcting the F508delꢀCFTR defect and, ii) compound 47b, which is
inactive in correcting the F508delꢀCFTR defect (Figure 3). Compound 47a,b were
44
synthesized following the procedure reported in literature.
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Chemical structures of compounds 47a,b.
Biology
All the synthesized compounds were initially evaluated for their inhibitory potency against
PI4KIIIβ kinase in vitro and for their cellꢀbased antiviral activity: EV71 was used as the
primary target for SAR exploration since compound 17 revealed the best and most
reproducible antiviral activity against this virus. In particular, the antiviral activity against
EV71 was evaluated in EV71ꢀinduced CPEꢀreduction assay in rhabdosarcoma (RD) cells.
Both the EC values and the CC values were measured. Uninfected, treated cells were also
50
50
inspected under the microscope to evaluate whether the compounds altered normal cell
morphology. The EC50 and CC50 values allowed us to calculate the selectivity index (SI),
defined as CC50/EC50. Compound 3 was used as a positive control. Results are summarized in
Table 1.
A close correlation between the antiviral activity measured in the cellꢀbased assay and the
inhibitory potency of the PI4KIIIβ kinase was observed, with only a few exceptions. The best
results were obtained via modifying the left part of the molecule. In particular compounds
25a,b, bearing respectively a propanamide and a pivalamide moiety instead of the acetamide
function of compound 17, showed a very promising antiviral activity.
ACS Paragon Plus Environment
1
4

Page 15 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Activity of synthesized derivatives in PI4KIIIβ inhibition assay and in virusꢀcellꢀ
based EV71 assay.
d
d
Compd
PI4K IIIβ EV71
EV71
EV71
SI
SI
c
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
IC50
EC50
(ꢁM)
CC50
(ꢁM)
CC (ꢁM)
5
0
a
b
(ꢁM)
1
2
2
7
0.48
0.27
0.32
0.92±2.75 16.5±9.04
0.38±0.10 10.1±4.82
0.27±0.05 7.94±1.24
9.73±0.87
6.37±2.46
8.83±0.89
30.1±9.93
11.8±4.56
ND
17.9 10.6
26.6 23.6
29.4 32.7
12.9 15.0
11.3 23.1
ND ND
ND ND
5a
5b
25c
21.89
18.85
>50
2.0±0.95
25.8±5.06
2
2
2
5d
5e
5f
0.51±0.14 5.75±2.12
e
>44.6
>55.7
ND
ND
>50
ND
25g
4.67
7.69
>50
1.42±0.04 42.3±4.94
101.0±15.8 29.7 71.1
f
2
2
2
2
6a
6b
6c
6d
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
64.2
ND
ND
142.0±62.8
ND
ND
ꢀ
ND ND
ND ND
NA
ND
1.82
>50
2±0.04
NA
8.87±3.85
ND
4.4
ND ND
4.4 ND
ND
30a
3.95
12.40
2.48
2.63
1.55
3.71
>50
4.77±0.28 20.9±6.97
NA ND
1.93±0.81 18.6±6.01
3
3
3
0b
5a
5b
ND ND
9.64 ND
ND ND
NA
ND
35c
1.2±0.16
9.17±0.85
5.9
8.2
ND
ND
3
3
4
5d
8
0.68±0.04 5.59±0.32
>90.2
NA
85.7
ND
ND
ND ND
ND ND
ND ND
2a
NA
42b
NA
NA
4
4
4
3
6
50.00
NA
8.58±0.77 51.6±27.1
6
ND
7a
NA
NA
0.73
ND
ND ND
ND ND
7b
NA
ND
g
0.06
>125
ꢀ
ꢀ
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
b
Values are the mean of at least three independent experiments. CC50 values were assessed by MTS method.
c
d
CC50 values were determined by microscopically detectable alteration of cell morphology. Selectivity index
SI = CC50/EC50). ND = not determined. NA = not active. Reference 13
e
f
g
(
Compounds 25a,b inhibited PI4KIIIβ and exhibited a significant antiviral effect at subꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
micromolar concentrations, demonstrating a better activity than compound 17. Compound
5g, characterized by the Boc amino group, proved to be the most interesting compound of
2
the entire series showing the highest selectivity index in the EV71 cellꢀbased assay. Also
changing the right portion of hit compound 17 gave interesting results (compounds 35a,c,d).
The central bithiazole scaffold proved to be essential for antiviral activity, as changing it gave
inactive compounds (compounds 38, 42a,b). Finally, the reported compounds 46, 47a,b were
devoid of antiviral activity and PI4KIIIβ inhibition activity.
Based on these activity data, and considering that the SI of a promising antiviral candidate
should be at least greater than 10, compounds 17, 25a-d, 25g were selected for further
studies. The broadꢀspectrum activity of the six selected compounds was evaluated against a
panel of enteroviruses representative of all major groups: enterovirus group B
(
coxsackievirus B3 and echovirus11, ECHO11), group C (poliovirus 1), group D (enterovirus
6
8), rhinovirus group A (RV02) and rhinovirus group B (RV14) (see the Experimental
Section for details). Results are reported in Table 2. The selected compounds showed
micromolar and subꢀmicromolar activity against different enteroviruses within the tested
panel. In addition, the antiviral activity of the less toxic compound 25g was confirmed
against a representative panel of EV71 clinical isolates. As shown in Table 3, we could
confirm the activity of compounds 25g against the clinically relevant EV71 specimens. Only
the (sub)genogroup B5 appeared to be less sensitive. Furthermore we evaluated the lipid
kinase isoform selectivity of our best PI4KIIIβ inhibitors by testing them in an in vitro
inhibition assay on the related enzyme PI4KIIIα and PI3Kꢀα/p85α (Table 4).
ACS Paragon Plus Environment
1
6

Page 17 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Evaluation of the broadꢀspectrum antiviral activity of the most potent derivatives
against a representative panel of enteroviruses.
Compd
CVB3
ECHO11
PV1
EV68
RV14
RV02
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EC50
CC50
EC50
CC50
EC50
CC50
EC50
CC50
EC50
CC50
EC50
CC50
(ꢁM) (ꢁM) (ꢁM) (ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
(ꢁM)
a
b
b
17
2.17
2.16
ND
101±
3.6
1.57± >268
0.23
3.09
>269
NA
ND
>268
ND
NA
ND
3
2
5a
5b
89.1± 0.97
16.2
12±4. <1.52 5.49± NA
07 1.42
ND
3.09
ND
ND
ND
NA
ND
>259
ND
2
20.9± 0.72± 5.07± <1.41 5.07± 1.38
.83 0.05 1.24 1.3
4.85± ND
1.09
2.01± ND
0.05
3
25c
3.87± 58.8± 3.51
.23 7.14
29.3± 2.86± 30.3± ND
1.78 0.36 3.66
10.6± ND
0.7
10.6± ND
0.2
0
25d
2.19± 45.1± 1.77± 145±2 2.23± 145±2 NA
0.23 4.44 0.13 7.6 0.4 7.6
>230
ND
2.05± ND
0.36
25g
2.74± 80.1± 2.93
124±7 13.3± 124±7 ND
.94 1.9 .94
>232
ND
ND
ND
0.17
6.97
a
b
NA = not active. ND = not determined
Results showed a higher specificity of the tested bithiazole derivatives for the PI4KIIIβ
isoform with poor inhibition of both PI4KIIIα and PI3Kꢀα/p85α at 100 ꢁM concentration of
each compound. The specificity of compound 25g was also tested on a small panel of
unrelated kinases: it shows only a low inhibitory effect on Src and CDK6. Despite the latter
enzymes being involved in cell cycle regulation and representing common targets of
antitumor compounds, 25g did not show any toxicity or morphology alteration at antiviral
concentration in the tested cell lines. In addition, recent studies indicated that Src inhibitors
45
have no effect on EV71 replication while CDK6 seems to be downꢀregulated in response to
46
EV71 infection. Finally, compounds reported in Table 4 were evaluated for their CFTR
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 3. Evaluation of the antiviral activity of compound 25g against EV71 clinical isolates.
a
EC (ꢁM)
5
0
Genogroup
Strain
Genbank
Compd 25g
<1.39
21.00
3.58
B2
B5
11316
AB575927
GQ231942
GQ231936
ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
TW/96016/08
TW/70902/08
H08300 461#812
TW/1956/05
C2
C4
0.97
GQ231926
GQ231927
<1.39
1.17
TW/2429/04
a
All values are based on at least three independent doseꢀresponse curves.
Table 4. Inhibitory effect of selected compounds against members of PIK family and
profiling of compound 25g against a small panel of unrelated kinases.
PI3Kꢀ
PI4KIIIβ
IC50(ꢁM)
PI4KIIIα
α/p85α
% Residual activity
Compd
Compd
kinase
a
a
% Residual activity
at 100 ꢁM
a
at 100 ꢁM
17
0.48
0.27
0.32
21.89
18.85
4.67
2.48
1.55
3.71
52
72
71
73
81
69
58
71
73
37
40
93
85
54
64
81
51
100
25g
Src FL
GSK3β
32
79
100
82
88
53
59
62
57
49
25
70
2
5a
5b
5c
5d
25g
5a
5c
5d
2
Hck FL
2
FAK
2
DYRK1A
ABL FL
FLT3
3
3
CDK2/cA2
CDK9/cT1
CDK9/cK
CDK6/cD1
Pim1
3
ACS Paragon Plus Environment
1
8

Page 19 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Values are the mean of two independent replicates
corrector/potentiator activity, to identify molecules that may be endowed with dual
antiviral/CFTR modulator activity. As shown in Figure 4A, some of the compounds (25a,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5d, 25g) acted as CFTR correctors, increasing steadyꢀstate levels of F508delꢀCFTR at the
plasma membrane after chronic (24 hour) incubation. Compound 48 (Lumacaftor), the
47
leading corrector drug, was used as a benchmark. This increased CFTR plasma membrane
density was measured with a recently developed assay exploiting a CFTR fusion to a pHꢀ
48
sensitive protein. The improvement in biogenesis also lead to increased anion permeability,
ꢀ
estimated from fluorescence quenching of a CFTRꢀfused YFP probe following extracellular I
addition (Figure 4B). None of the compounds acted as "potentiators" rapidly increasing anion
ꢀ
permeability, when added only immediately prior to I addition (Figure 4C). The approved
4
9
potentiator drug 49 (Ivacaftor) was used as a comparison. Overall, the drugꢀinduced
changes in the iodide entry rate and in membrane density followed similar patterns,
suggesting that the chemically corrected molecules of F508delꢀCFTR that reached the plasma
membrane displayed an ionꢀchannel function similar to those corrected by treatment with 48.
However, compound 25d appears to increase CFTR membrane density more than expected
from its effect on anion permeability (Figure 4D). Further studies will be required to
understand the underlying mechanism. Overall, the collected biological data indicates that a
fine chemical tuning of the bithiazole substituents is needed to generate compounds able to
specifically inhibit the PI4KIIIβ kinase and block the replication of different enteroviruses
while also correcting the F508delꢀCFTR folding defect. It is interesting to note that the most
promising CFTR correctors (25a, 25d, 25g) are also the most active broadꢀspectrum
antivirals and represent the first example of multiꢀtarget agents for tightly associated
pulmonary diseases like enterovirus infection and cystic fibrosis.
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Effects of selected compounds on CFTR biogenesis and function. All treatments
were carried out alongside low temperature incubation, known to improve F508delꢀCFTR
membraneꢀlocalization and used to increase the fluorescence signal. A) F508delꢀCFTRꢀ
ACS Paragon Plus Environment
2
0

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pHTomato present at the plasma membrane was quantified following 24 h incubation in 10
ꢀ
M of each drug. Incubation with 48 (Lumacaftor) was assessed in parallel, as a positive
control. B) Anion permeability quantified using a YFPꢀF508delꢀCFTR probe following 24 h
treatment as in A). C) Compounds (10 M) do not cause an immediate change in anion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
permeability. Potentiator compound 49 (Ivacaftor) was used as a positive control. D) For
most compounds, there is a similar ratio describing increase in membrane density over anion
permeability as caused by 48 (Lumacaftor).
Molecular modelling and SAR
50
The hit compound 17 was docked with the Glide software (SP) in the ATP binding site of
3
4
the PI4KIIIβ crystal structure (PDB ID: 4D0L) centering the grid on compound 1. The
predicted binding mode and interaction profile of compound 17 is very similar to that of
compound 1 (Figure 5). The NHꢀacetamide moiety of 17 is hydrogen bonded to VAL598,
which also interact with the thiazole nitrogen. The thiazole ring is involved in a PiꢀPi
stacking with TYR583, while the phenyl ring is involved in a Piꢀcation interaction with
LYS549. The O carbonyl moiety is hydrogen bonded to LYS377. Moreover, the binding
mode is completed by a series of hydrophobic interactions involving LEU383, ALA602,
VAL599, VAL602, LEU663, ILE595, TYR583, ILE671, ILE673, PRO381 and LEU374.
The first series of derivatives of the hit compound 17 encompasses different substitution on
the left part of the molecule (right part as per Figure 5ꢀ7 representation) by replacing the
methyl group of the acetamide moiety with different groups. Changing the methyl of the
acetamide with an ethyl or tertꢀbutyl group (25a,b) caused a 2ꢀfold increase of potency, but
the substitution with more hydrophobic and therefore more sterically bulky groups first
reduced the potency (25d) and then led to a complete loss of activity (25e,f). The limit seems
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to be a tertꢀbutoxycarbonyl group (25g) which still maintains low micromolar activity
(Figure 6).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. a) Predicted binding mode of the hit 17 (magenta sticks) superimposed to
compound 1 (green sticks) into the binding site of PI4KIIIβ. b) 2D ligand interaction diagram
of 17.
The substitution of the acetamide methyl with a CF (25c) weakened the inhibition (45ꢀfold
3
decrease) not due to steric reasons, but probably because of the electron withdrawing
3
properties of the CF group. Changing the acetamide moiety with a ureidic moiety caused a
slight decrease in potency (26a,c and 30a,b). The ureidic portion seems to interact with a
double hydrogen bond to VAL598 but at the same time, the hydrophobic NH substituent
moves away from the protein resulting in a solvation penalty (Figure 6). Finally, modifying a
ureidic group with a thioureidic moiety caused a complete loss of activity (26b,d).
ACS Paragon Plus Environment
2
2

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. a) Superimposed binding modes of compounds 17 (purple sticks), 25a (orange
sticks), 25b (green sticks) and 25g (pink sticks) into the binding site of PI4KIIIβ. b)
Superimposed binding modes of compounds 26a (orange sticks), 26c (green sticks) and 30a
(pink sticks) into the binding site of PI4KIIIβ. c) 2D ligand interaction diagram of 25b. d) 2D
ligand interaction diagram of 30a.
The second series of derivatives of the hit compound 17 comprises different substitutions on
the right part of the molecule (left part as per Figure 5ꢀ7 representation) by changing the
substitution pattern of the phenyl ring. A change in the position of the acetyl group reduces
the potency by about 10ꢀfold (35d), and the same happens by introducing hydroxy and
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
Page 24 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
methoxy groups in the meta position (35b,c). The precise positioning of the hydrogenꢀbond
acceptor on the phenyl ring seems therefore important in improving the affinity for the
enzyme. Surprisingly, the deletion of the ketonic group reduces but does not abolish activity
(35a). The lack of the substituent on the phenyl ring deletes a hydrogen bond, but allows
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
optimization of the other interactions, in particular the Piꢀcation with lysine 549.
The last series of derivatives of the hit compound 17 encompasses modifications of the
bithiazole scaffold and closely related analogues reported as DCꢀSIGN inhibitors and CFTR
correctors. Replacement of the thiazole group of compound 17 with an imidazole (38) or
conversion of the bithiazole scaffold into a N,N’ꢀdiarylthiourea (42a,b) changes the binding
mode and abolishes the activity. Finally, also the DCꢀSIGN inhibitor 46 and CFTR correctors
47a,b presented a suboptimal interaction profile with PI4KIIIβ and resulted in a complete
5
1
loss of activity. As reported by Warrem et al. , docking programs and scoring functions
present a few limitations in correlating subtle structural differences of active ligands with
their enzymatic activity. This error is quite limited within homologous series of compounds
52
but it can be very important for structurally unrelated compounds. Our docking studies were
in fact able to distinguish between active and inactive compounds but it is no coincidence that
the reference compound 3, whose scaffold is very different from those of our series, showed a
docking score that is not in line with its enzymatic potency (Table 5). The most active
compounds 17, 25a and 25b also showed the best specificity for PI4KIIIβ over PI4KIIIα,
which seems to depend on the steric hindrance of the acetamide substitution. In fact,
considering compounds 17, 25a, 25b, 25g and 25d that differ only for the bulkiness of the
amide substituents in position C2 of the thiazole (methyl, ethyl, tertꢀbutyl, tertꢀbutoxy and
phenyl, respectively), the affinity for PI4KIIIα decreases going from 17 to 25d (see Table 4).
ACS Paragon Plus Environment
2
4

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 5. Correlation between IC50 and docking score for the synthesized compounds.
Compd PI4K IIIβ Docking Compd PI4K IIIβ
Docking
score
b
b
score
a
IC50
50
IC (ꢁM)
a
(ꢁM)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
3
5d
0b
3.71
ꢀ8,297
ꢀ8,222
ꢀ8,172
ꢀ8,108
ꢀ8,035
ꢀ8,012
ꢀ7,983
ꢀ7,97
25e
26d
25f
38
>50
>50
>50
>50
>50
>50
>50
21.90
0.06
>50
>50
>50
ꢀ7,431
ꢀ7,4
12.40
7.69
1.82
0.48
0.27
18.85
3.95
2.63
50.00
4.67
1.55
2.48
0.32
26a
ꢀ7,356
ꢀ7,355
ꢀ7,251
ꢀ7,144
ꢀ7,052
ꢀ6,678
ꢀ6,553
ꢀ6,424
ꢀ6,148
ꢀ7,431
2
1
2
6c
7
42a
26b
42b
25c
5a
25d
3
3
4
2
0a
5b
6
c
ꢀ7,726
ꢀ7,644
ꢀ7,63
3
47b
47a
25e
5g
35c
ꢀ7,605
ꢀ7,6
3
2
5a
5b
ꢀ7,484
a
b
Values are the mean of at least three independent experiments. Docking score was calculated by the software
c
Gold and expressed as Kcal/mol. Reference 13
The higher the bulkiness in C2 of the thiazole, the lower the affinity for PI4KIIIα over
PI4KIIIβ with the phenyl substituent (25d) representing the highest tolerated hindrance after
which the inhibition of PI4KIIIβ is also compromised. The reason for the specificity of these
compounds towards PI4KIIIβ seems to depend on the different opening (compared to the α
isoform) of a specific loop that in the β isoform goes from ILE595 to ILE604 (Figure 7). In
fact in this portion of protein, the sequence of PI4KIIIβ resembles more the sequence of PI3K
(
α, γ end δ) than the sequence of PI4KIIIα. In particular, the presence of a cysteine residue in
position 30 of PI4KIIIα (corresponding to proline 597 in PI4KIIIβ) could change the fold of
this “selectivity loop”.
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Observed binding mode of 25a (yellow sticks) superimposed to compound 1
(magenta sticks) into the binding site of PI4KIIIβ (green ribbons). The orange ribbons
represent the modelled structure of PI4KIIIα.
CONCLUSIONS
An increasing number of reports suggest a causal link between enterovirus infections and
pulmonary exacerbations in CF patients. We report the discovery of a new class of multiꢀ
target agents active as broadꢀspectrum antivirals and correctors of the F508delꢀCFTR folding
defect. To identify these drug candidates, we first carried out a virtual screening on the
PI4KIIIβ (a host protein involved in enterovirus replication) catalytic site to select
commercially available compounds: our choice was based on the bestꢀpredicted affinity for
the target kinase and 2D similarity (in a few cases) to known CFTR correctors/potentiators.
Among the selected compounds, hit 17 showed activity in cellꢀfree PI4KIIIβ inhibition assay
and cellꢀbased enterovirus replication assays and was therefore considered a promising
starting point for further structureꢀbased optimization. A small collection of analogues of
compound 17 was then designed, synthesized and biologically evaluated for their i) activity
against panel of enteroviruses representative of all major groups; ii) inhibition of lipid kinases
PI4KIIIβ, PI4KIIIα and PI3Kꢀα/p85α; iii) corrector/potentiator activity on F508delꢀCFTR.
ACS Paragon Plus Environment
2
6

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Three compounds (25a, 25d, 25g) were finally identified as novel multiꢀtarget agents able to
act as broadꢀspectrumꢀantivirals (enterovirus family) and as correctors of F508delꢀCFTR
folding defect. These compounds represent a valuable starting point to develop a novel
polypharmacological approach for the treatment of closelyꢀrelated pulmonary diseases such
as cystic fibrosis and enterovirus infections with a single pill.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Molecular modeling
Homology modeling
53
The structure of PI4KIIIβ was built with the Prime 38013 software on the basis of the
crystal structure 2CHZ using ClustalW for sequence alignment and knowledgeꢀbased as
5
4
building method. The structure of PI4KIIIα was built with the online server 3DꢀJIGSAW on
the basis of the crystal structure 4D0L.
Molecular dynamics
The structure of modelled PI4KIIIβ was aligned to the 2CHZ structure. Compound 1 was
extracted from 2CHZ and was manually introduced in the structure of PI4KIIIβ. A molecular
dynamics simulation of the resulting complex was performed using Desmond v40013. The
complex was neutralized using sodium counter ions. The complex and the counter ions were
immersed in a orthorhombic periodic SPC water bath that extended about 10 Å in each
direction. After an initial default relaxation protocol, a MD production run was performed for
10 ns with a time step of 2 fs.
Virtual screening
From the last 2 ns of the dynamics simulation 100 frames were extracted and clustered on
the basis of RMSD. Five clusters were generated. The protein representative of each cluster
55
was processed with the Schrödinger Suite 2014ꢀ3 Protein Preparation Wizard tool. On each
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
structure a grid was generated with the software Glide 65013 centering the grid on compound
1
, then all PI4KIIIβ ligands available from the Pubchem database were docked with the SP
protocol. Structures were selected for virtual screening on the basis of their enrichment
56
factor. The Mid Asinex database was extracted from the ZINC database and docked in the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
binding site of compound 1 using compound 1 as the center of the grid. The software Glide
65013 with the SP protocol was used for High throughput docking. The best 1000
compounds in terms of the docking score were selected and submitted to one more docking
cycle of docking with the XP protocol. On the basis of the docking score and a visual
inspection 25 compounds were selected and 13 compounds were purchased.
Ligand preparation
5
7
Ligands were prepared with the LigPrep tool available in the Schrödinger Suite 2015ꢀ4.
Ionization states were generated at pH 7.0 ± 2.0 with Epik.
Ligand docking
The Xꢀray coordinates of PI4KIIIβ in complex with compound 1 were extracted from the
Protein Data Bank (PDB code 4D0L). The structure was then processed with the Schrödinger
58
Suite 2015ꢀ4 Protein Preparation Wizard tool. The A Chain was selected, water molecules
were removed, and an exhaustive sampling of the orientations of groups, whose hydrogen
bonding network needs to be optimized, was performed. Finally, the protein structure was
59
refined to relieve steric clashes with a restrained minimization with the OPLS3 force field
until a final RMSD of 0.30 Å with respect to the input protein coordinates.
60
Docking studies were performed using Glide 69017 with the SP protocol. The protein
structure, prepared as described above, was used to build the energy grid. The enclosing box
was centered on the cocrystallized ligand. All parameters were set to their default value. The
docking protocol was validated by redocking the cocrystallized ligand (compound 1).
ACS Paragon Plus Environment
2
8

Page 29 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemistry
General. All commercially available chemicals were purchased from both Sigmaꢀ Aldrich
and Alfa Aesar and, unless otherwise noted, used without any previous purification. Solvents
used for workꢀup and purification procedures were of technical grade. Dry solvents used in
the reactions were obtained by distillation of technical grade materials over appropriate
dehydrating agents. Reactions were monitored by thin layer chromatography on silica gelꢀ
coated aluminium foils (silica gel on Al foils, SUPELCO Analytical, SigmaꢀAldrich) at 254
and 365 nm. Where indicated, products were purified by silica gel flash chromatography on
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
13
columns packed with Merck Geduran Si 60 (40ꢀ63 µm). H and C NMR spectra were
recorded on BRUKER AVANCE 300 MHz and BRUKER AVANCE 400 MHz
1
spectrometers. Chemical shifts (δ scale) are reported in parts per million relative to TMS. Hꢀ
NMR spectra are reported in this order: multiplicity and number of protons; signals were
characterized as: s (singlet), d (doublet), t (triplet), q (quadruplet), m (multiplet), bs (broad
signal). ESIꢀmass spectra were recorded on an API 150EX apparatus and are reported in the
form of (m/z). Elemental analyses were performed on a PerkinꢀElmer PE 2004 elemental
analyzer. Melting points were taken using a Gallenkamp melting point apparatus and were
uncorrected. All final compounds showed chemical purity ≥95% as determined by elemental
analysis data for C, H, and N (within 0.4% of the theoretical values).
Synthesis of Nꢀ(2ꢀ((3ꢀacetylphenyl)amino)ꢀ4'ꢀmethylꢀ[4,5'ꢀbithiazol]ꢀ2'ꢀyl)acetamide (17).
3
Et N (84 ꢁL, 0.60 mmol) was added to a stirred suspension of intermediate 24 (100 mg, 0.30
mmol) in dry DCM (4 mL) at 0 °C. After 15 minutes acetyl chloride (32 ꢁL, 0.45 mmol),
diluted in dry DCM (0.5 mL), was added dropwise. The resulting solution was warmed to
room temperature and stirred for 15 h. Next, H O and DCM were added and the aqueous
2
phase was extracted twice with DCM. The combined organic phases were washed with brine,
ACS Paragon Plus Environment
2
9