Neurogenic and neuroprotective donepezil-flavonoid hybrids with sigma-1 affinity and inhibition of key enzymes in Alzheimer's disease

Article abstract of DOI:10.1016/j.ejmech.2018.07.026

In this work we describe neurogenic and neuroprotective donepezil-flavonoid hybrids (DFHs), exhibiting nanomolar affinities for the sigma-1 receptor (σ₁R) and inhibition of key enzymes in Alzheimer's disease (AD), such as acetylcholinesterase (AChE), 5-lipoxygenase (5-LOX), and monoamine oxidases (MAOs). In general, new compounds scavenge free radical species, are predicted to be brain-permeable, and protect neuronal cells against mitochondrial oxidative stress. N-(2-(1-Benzylpiperidin-4-yl)ethyl)-6,7-dimethoxy-4-oxo-4H-chromene-2-carboxamide (18) is highlighted due to its interesting biological profile in σ₁R, AChE, 5-LOX, MAO-A and MAO-B. In phenotypic assays, it protects a neuronal cell line against mitochondrial oxidative stress and promotes maturation of neural stem cells into a neuronal phenotype, which could contribute to the reparation of neuronal tissues. Molecular modelling studies of 18 in AChE, 5-LOX and σ₁R revealed the main interactions with these proteins, which will be further exploited in the optimization of new, more efficient DFHs.

Full text of DOI:10.1016/j.ejmech.2018.07.026

Accepted Manuscript
Neurogenic and neuroprotective donepezil-flavonoid hybrids with sigma-1 affinity and
inhibition of key enzymes in Alzheimer's disease
Martín Estrada Valencia, Clara Herrera-Arozamena, Lucía de Andrés, Concepción
Pérez, José A. Morales-García, Ana Pérez-Castillo, Eva Ramos, Alejandro Romero,
Dolores Viña, Matilde Yáñez, Erik Laurini, Sabrina Pricl, María Isabel Rodríguez-
Franco
PII:
S0223-5234(18)30582-8
10.1016/j.ejmech.2018.07.026
EJMECH 10561
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 29 May 2018
Revised Date: 18 June 2018
Accepted Date: 9 July 2018
Please cite this article as: Martí. Estrada Valencia, C. Herrera-Arozamena, Lucí. de Andrés, Concepció.
Pérez, José.A. Morales-García, A. Pérez-Castillo, E. Ramos, A. Romero, D. Viña, M. Yáñez, E.
Laurini, S. Pricl, Marí.Isabel. Rodríguez-Franco, Neurogenic and neuroprotective donepezil-flavonoid
hybrids with sigma-1 affinity and inhibition of key enzymes in Alzheimer's disease, European Journal of
Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.07.026.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Neurogenic and neuroprotective donepezil-
flavonoid hybrids with sigma-1 affinity and
inhibition of key enzymes in Alzheimer’s disease
Dedicated to the memory of our dear colleague Dr. Elena Soriano
a
a
a
Martín Estrada Valencia , Clara Herrera-Arozamena , Lucía de Andrés , Concepción
a
b,c,d
b,c
e
Pérez , José A. Morales-García , Ana Pérez-Castillo , Eva Ramos , Alejandro
e
f
g
h
h
Romero , Dolores Viña , Matilde Yáñez , Erik Laurini , Sabrina Pricl , and María Isabel
a,
Rodríguez-Franco *
a
Instituto de Química Médica, Consejo Superior de Investigaciones Científicas (IQM-
CSIC), C/ Juan de la Cierva 3, 28006-Madrid, Spain.
b
Instituto de Investigaciones Biomédicas “Alberto Sols”, Consejo Superior de
Investigaciones Científicas (IIB-CSIC), C/Arturo Duperier 4, 28029-Madrid, Spain.
c
Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas
(CIBERNED), C/ Valderrebollo 5, 28031-Madrid, Spain.
d
Departamento de Biología Celular, Facultad de Medicina, Universidad Complutense de
Madrid, 28040-Madrid, Spain.
e
Departamento de Farmacología y Toxicología, Facultad de Veterinaria, Universidad
Complutense de Madrid, 28040-Madrid, Spain.
f
Center for Research in Molecular Medicine and Chronic Diseases (CIMUS),
Universidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain.
g
Departamento de Farmacología, Facultad de Farmacia, Universidad de Santiago de
Compostela, 15782-Santiago de Compostela, Spain.
h
Molecular Simulation Engineering (MOSE) Laboratory, Department of Engineering
and Architecture (DEA), University of Trieste, 34127 Trieste, Italy.
*
Corresponding Author
E-mail address: isabelrguez@iqm.csic.es
1

ACCEPTED MANUSCRIPT
Keywords:
Donepezil-flavonoid hybrids
Neurogenesis
Neuroprotection
Sigma receptors
Human acetylcholinesterase
Human 5-lipoxygenase
Human monoamine oxidases
Highlights:
•
•
•
•
•
Donepezil-flavonoid hybrids with high affinity for sigma-1 receptors
Inhibition of acetylcholinesterase, 5-lipoxygenase, and monoamine oxidases
Many donepezil-flavonoid hybrids are radical-scavenger and brain-permeable
Neuroprotection vs. mitochondrial oxidative stress in human neuroblastoma
Hybrid 18 stimulates differentiation of stem-cells to a neuronal phenotype in
vitro
2

ACCEPTED MANUSCRIPT
Abbreviations: AChE, acetylcholinesterase; AD, Alzheimer’s disease; ATCh,
acetylthiocholine;
BBB,
blood-brain
barrier;
BOP,
(benzotriazol-1-
BuChE,
yloxy)tris(dimethylamino)phosphonium
hexafluorophosphate;
butyrylcholinesterase; CAS, catalytic anionic site of AChE; DCM, dichloromethane;
DMF, dimethylformamide; CDI, 1,1'-carbonyldiimidazole; CNS, central nervous
system; DAPI, 4',6-diamidino-2-phenylindole; DFHs, donepezil - flavonoid hybrids;
DTG, 1,3-di-o-tolylguanidine; HMBC, heteronuclear multiple bond correlation; HPLC,
high-performance liquid chromatography; HRMS, high resolution mass spectrum;
HSQC, heteronuclear single quantum correlation; K , binding constant; LB,
i
Lineweaver-Burk method; 5-LOX, 5-lipoxygenase; MAO-A and MAO-B, monoamine
oxidase-A and monoamine oxidase-B; MAP-2, microtubule-associated protein 2; MD,
Molecular Dynamics; MM/PBSA, Molecular Mechanics/Poisson-Boltzmann Surface
Area; MTDLs, multi-target-directed ligands; MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5-
diphenyl-tetrazolium bromide; mw, microwave; NDGA, nordihydroguaiaretic acid;
NDs, neurodegenerative diseases; NMDA, N-methyl-D-aspartate receptor; NMR,
nuclear magnetic resonance; NS, neurospheres; NSC, neural stem cells; ORAC, oxygen
radical scavenger capacity assay; OS, oxidative stress; PAINS, pan assay interference
compound; PAMPA-BBB, parallel artificial membrane permeability assay for the
blood-brain barrier; PAS, peripheral anionic site of AChE; P , permeability value;
e
PRBFED, per-residue binding free energy deconvolution; PTZ, pentazocine; ROS,
radical oxygen species; SEM, standard error of the mean; SGZ, subgranular zone; σ R
1
and σ R, sigma-1 and sigma-2 receptor; SVZ, subventricular zone; THF,
2
tetrahydrofuran; Tuj1, β-tubulin III marker.
3

ACCEPTED MANUSCRIPT
ABSTRACT
In this work we describe neurogenic and neuroprotective donepezil-flavonoid
hybrids (DFHs), exhibiting nanomolar affinities for the sigma-1 receptor (σ R) and
1
inhibition of key enzymes in Alzheimer’s disease (AD), such as acetylcholinesterase
(AChE), 5-lipoxygenase (5-LOX), and monoamine oxidases (MAOs). In general, new
compounds scavenge free radical species, are predicted to be brain-permeable, and
protect neuronal cells against mitochondrial oxidative stress. N-(2-(1-Benzylpiperidin-
4
-yl)ethyl)-6,7-dimethoxy-4-oxo-4H-chromene-2-carboxamide (18) is highlighted due
to its interesting biological profile in σ R, AChE, 5-LOX, MAO-A and MAO-B. In
1
phenotypic assays, it protects a neuronal cell line against mitochondrial oxidative stress
and promotes maturation of neural stem cells into a neuronal phenotype, which could
contribute to the reparation of neuronal tissues. Molecular modelling studies of 18 in
AChE, 5-LOX and σ R revealed the main interactions with these proteins, which will
1
be further exploited in the optimization of new, more efficient DFHs.
4

ACCEPTED MANUSCRIPT
1
. Introduction
Alzheimer’s disease (AD) is the most common of the age-related neurodegenerative
pathologies (NDs), causing devastating health, social, familiar, and economic problems
in world-wide countries [1]. AD is characterized by the occurrence of abnormal protein
aggregates in the brain (amyloid plaques and neurofibrillary tangles), the loss of
synaptic connections, and death of extensive neuronal populations, especially those
involved in the intellectual functions [2]. Currently, there are only four drugs in the
market for the palliative management of AD: three acetylcholinesterase (AChE)
inhibitors (donepezil, rivastigmine, and galantamine) and an antagonist of the N-methyl-
D-aspartate (NMDA) receptor, i.e., memantine [3].
Due to the highly complex pathophysiology of AD, several combinations of drugs
acting at different targets have been probed in clinical trials, such as the mixture of
donepezil and memantine that provided superior benefits over monotherapies [4]. In a
similar way, the multi-target-directed ligands (MTDLs) paradigm proposes the design
of single molecules that were active against several pharmacological targets, in the
quest of a superior therapeutic efficacy [5,6]. The MTDL approach would be
particularly useful when drugs hit targets involved upstream in neurotoxic cascades, as
they could really stop or delay neurodegeneration [7].
Nowadays, the causes that trigger AD are not completely known, but different
findings indicate that mitochondrial dysfunctions that increase oxidative stress (OS) and
neuroinflammation are at the vertex of many pathological cascades [8]. The implication
of OS in AD is supported by different post-mortem studies, which show a considerable
increase in the peroxidation of biomolecules (lipids, proteins, and nucleic acids) in the
most affected regions of the brain of patients with severe AD [9]. Moreover, it has been
proven that cellular oxidative damage is a fact that precedes the appearance of abnormal
5

ACCEPTED MANUSCRIPT
protein aggregates [10]. These observations suggest that protective agents against
mitochondrial oxidative stress may be useful in the prevention and treatment of AD
[
11].
On the other hand, several epidemiological studies have revealed that patients under
chronical treatment with non-steroidal anti-inflammatory drugs showed a reduction in
the risk of developing AD, although the protection mechanism is still a subject of debate
[
12]. 5-Lipoxygenase (5-LOX), a key enzyme involved in the inflammatory responses,
is expressed at higher levels in the hippocampus of patients affected by AD [13].
Contextually, overexpression of 5-LOX leads to memory deficiencies and increase of
protein aggregates in transgenic mice [14].
5
-LOX catalyses two early steps of the peroxidation of arachidonic acid to
leukotriene A4, which is further transformed to eicosanoids by several enzymatic
systems. In the central nervous system (CNS), eicosanoids perform numerous functions
of cellular communication, but they are also involved in the activation of inflammatory
cascades. Particularly, it has been shown that pro-inflammatory eicosanoids, including
leukotriene A4, increase the expression of amyloid precursor protein (APP), with the
consequent growth of pathogenic amyloid plaques [15]. For this reason, inhibitors of 5-
LOX have been proposed as potential drugs for the treatment of different diseases,
including AD [16].
Monoamine oxidases (MAO-A and MAO-B) are metabolic enzymes that are
located on the outer mitochondrial membrane. As for 5-LOX, the brain levels of these
proteins are increased in neurodegenerative pathologies, such as AD and Parkinson's
disease [17]. Abnormally high concentrations of MAOs enhance the expression of β-
and γ-secretases, promoting the formation of the pathogenic β-amyloid peptide from the
amyloid precursor protein, and increase the OS by the production of hydrogen peroxide
6

ACCEPTED MANUSCRIPT
[
18]. Thus, MAOs inhibition could bring beneficial effects in neurotransmission by
lowering catabolism of several monoamine neurotransmitters and by reducing the
generation of amyloid plaques and radical oxygen species (ROS) [19,20].
Sigma receptors, and in particular the subtype-1 (σ R), are expressed in CNS areas
1
related to emotional and cognitive functions, such as hippocampus and neocortex [21].
Activation of σ R evokes neuroprotection by increasing levels of endogenous
1
antioxidant proteins [22] and by maintenance of mitochondrial respiration and ATP
synthesis [23], among other mechanisms. Thus, σ R agonists are potential drugs for
1
AD, such as anavex2-73 that is currently in phase 2a of clinical trials in patients with
mild to moderate AD [24].
Although questioned for decades, adult neurogenesis is now recognized as an
important mechanism of neuron self-renewal along the life of the individual [25]. In the
adult human brain, there are two niches of stem-cells located in the subventricular and
the subgranular zones (SVZ and SGZ, respectively), which are capable of generating
new neurons [26]. As the neuronal self-repair mechanisms decrease with aging and
dramatically fall in NDs and other mental pathologies, the pharmacological stimulation
of neurogenic niches might counteract the neuronal loss observed in such diseases [27].
Many signalling pathways involved in neurogenic processes have been identified
and different drug types have been recently tested in neuronal plasticity. For instance,
antioxidants and σ R agonists have positive effects in the generation and maturation of
1
new neurons from stem-cell niches and also promote new neuronal connections by
stimulating neurite outgrowth [28,29].
Therefore, neurogenesis and neuroprotection have recently emerged as innovative
and complementary strategies for battling AD [27]. The development of new active
7

ACCEPTED MANUSCRIPT
agents capable of protecting the more vulnerable neuronal populations or even replacing
damaged neurons with new functionally competent cells, could provide disease-
modifying therapeutic regimens for AD [30]. In recent years, our group has reported the
synthesis and biological evaluation of several families of MTDLs of potential interest
for the treatment of AD, endowed with neuroprotective [31,32] or neurogenic properties
[
33-36], or even with the combination of both actions [37,38].
Along the latter line of dual activity, in the present effort we developed 29 new
donepezil – flavonoid hybrids (DFHs). On one hand, we selected the N-
benzylpiperidine fragment from donepezil, due to its well-known σ R-agonism and
1
AChE-inhibition [39]. On the other hand, we selected the 4-oxo-4H-chromene scaffold
present in numerous flavonoids (e.g. apigenin and luteolin), which showed inhibition of
MAO [40] and 5-LOX [41,42], as well as antioxidant, anti-inflammatory and
neurogenic properties [43]. Indeed, the C-glycoside flavonoid spinosin was found to
enhance adult hippocampal neurogenesis in mice, linked to an improvement in
cognitive performance [44] (Figure 1).
In these new DFHs, the N-benzylpiperidine fragment was unchanged due to its
well-known excellent fit in the catalytic anionic site of AChE (CAS) [45]. In relation to
the potential modifications in the flavonoid scaffold we planned to: (i) change the 4-
oxo-4H-chromene skeleton by related heterocycles, such as quinolin-4(1H)-one or
quinoline rings; (ii) introduce methoxy groups to improve interactions with the
peripheral anionic site of hAChE (PAS), as previously described [46]; (iii) insert
hydroxyl groups to achieve radical capture capacity and 5-LOX inhibition, as described
for phenolic hydroxyl derivatives [47]. Finally, we thought connecting N-benzylpiperine
and flavonoid fragments by a linker of different nature (amide or ester) and length (n =
0
– 3).
8

ACCEPTED MANUSCRIPT
OCH3
O
O
OH
OH
Apigenin (R = H)
Luteolin (R = OH)
OH
OCH3
R
N
O
Donepezil
HO
HO
OH
OH
O
O
O
O
N
R
OH
O
X
n
H
O
Z
O
OH
OH
O
Donepezil - flavonoid hybrids (1-29)
HO
Spinosin
(
n = 0 - 3; X, Z = NH, O; R = OCH3, NO , NH , OH)
OH
2
2
Figure 1. Structures of donepezil, flavonoids (apigenin, luteolin, and spinosin), and
donepezil – flavonoid hybrids (DFHs, 1-29)
Accordingly, in what follows we describe the synthesis of these new DFHs (1-29)
and their biological evaluation that includes σ R binding, antioxidant properties, in vitro
1
CNS penetration, inhibition of 5-LOX, AChE and MAOs, and phenotypic assays of
neurogenesis and neuroprotection. Finally, molecular modelling studies have been
carried out to elucidate the binding mechanism of DFHs on the three main target
proteins, that is AChE, 5-LOX and σ R.
1
9

ACCEPTED MANUSCRIPT
2
. Results and discussion
2
.1. Synthesis of DFHs
Scheme 1 depicts the synthetic routes for the synthesis of the new hybrids (1-29).
Some amine and alcohol intermediates derived from N-benzylpiperidine are
commercially available and others were obtained following previously reported
methods [35,38]. 4-Oxo-4H-chromene-2-carboxylic acids and methyl quinoline-2-
carboxylates were synthesized according to well-known routes [46,48] (see Supporting
Information for further details).
The synthesis of chromone-based hybrids 1, 3-11, 13, 14, 16, 18, 20 and 22 was
developed by a coupling reaction between the corresponding N-benzylpiperidine
alcohol or amine and different 4-oxo-4H-chromene-2-carboxylic acids, using two
different methods: i) a microwave (mw)-assisted reaction with 1,1'-carbonyldiimidazole
(
CDI) as activating agent, and ii) a coupling reaction at room temperature (rt) with
benzotriazol-1-yloxy)tris(dimethylamino)phosphonium hexafluorophosphate (BOP)
(
and triethylamine (Et N). In both cases, dimethylformamide (DMF) was selected as
3
solvent due to the low solubility of the 4-chromone carboxylic acids in less polar
solvents. However, when activation is carried out with BOP, other solvents could be
used, i.e. tetrahydrofuran (THF) or dichloromethane (DCM), because the Et N concurs
3
in acids solubilisation. Hydroxy-chromone hybrids 2, 12, 17, 19 and 21 were obtained
via deprotection of the corresponding methoxylated derivative by overnight treatment
with boron tribromide (BBr ) at rt. It is worth-mentioning that, to ensure good reaction
3
yields (70-90%), 1 equivalent of BBr for each ether group to be cleaved, plus an
3
additional 1 equivalent for each heteroatom present in the molecule, must be used [49].
The amino-chromone hybrid 15 was obtained with an almost quantitative yield (98%)
by the catalytic hydrogenation of the corresponding nitro derivative 14.
1
0

ACCEPTED MANUSCRIPT
The 4-quinolone-based hybrids bearing methoxy groups 23, 25, 26, and 28 and the
trimethoxyquinoline 29 were obtained by trimethylaluminium [Al(CH ) ]-mediated
3
3
amide formation between the corresponding 4-oxo-1,4-dihydroquinoline or quinoline
methyl ester and 2-(1-benzylpiperidin-4-yl)ethan-1-amine (Scheme 1). The use of mw
irradiation at 120 ºC allowed obtaining the desired products in 1.5 min with medium to
good yields (50-90%). Finally, hydroxy-substituted compounds 24 and 27 were
obtained by treatment of methoxy derivatives with BBr in THF at rt as previously
3
mentioned.
All DFHs 1-29 were purified in silica gel cartridges using an automatic
chromatographic equip (IsoleraPrime, Biotage) and were characterized by their
1
13
analytical (HPLC, HRMS) and spectroscopic data ( H NMR, C NMR). Complete
1
13
NMR assignment of their hydrogen and carbon atoms were made by H – C two-
dimensional diagrams, mainly HSQC (heteronuclear single quantum correlation) and
HMBC (heteronuclear multiple bond correlation). In the case of the 6-nitrochromone
hybrid 14 the CH group at position 3 of the 4-chromone system was not observed either
1
3
1
in C or H NMR spectra, although its exact mass by high resolution mass spectroscopy
matched the one of the desired product. Moreover, the amino derivative 15, obtained by
hydrogenation of 14, exhibited all the expected signals and exact mass, allowing us to
confirm the structure of the nitro derivative.
1
1

ACCEPTED MANUSCRIPT
Scheme 1. Reagents and conditions: (a) i) Acid, CDI, DMF, mw, 120 °C, 10 min, ii)
amine, mw, 150 ºC, 10 min; (b) BOP, Et N, DMF, overnight, rt; (c) BBr , DCM,
3
3
overnight, 0 °C → rt; (d) H /Pd-C (5%), EtOH, overnight, rt; (e) Al(CH ) , THF, mw,
2
3 3
1
20 °C, 90 sec.
1
2

ACCEPTED MANUSCRIPT
Unlike the 4-chromone scaffold, the 4-oxo-1,4-dihydroquinoline structure can exist
in two tautomeric forms. In the keto tautomer, the hydrogen at position 3 possesses
lower aromatic character and consequently, lower chemical shifts should be observed
compared with the enol form. Given the lower chemical shifts observed in hybrids 23-
2
8 compared to the full aromatic derivative 29, we inferred that the keto form is the
favoured tautomer in 23-28. Probably due to this tautomeric equilibrium, some
quaternary carbons in the 4-quinolone scaffold were not detected and the multiplicity of
some aromatic protons was not well defined, appearing as broad signals. However, high
resolution mass spectroscopy confirmed the exact mass of the desired products.
2
.2. Inhibition of human cholinesterases, study of inhibitory mechanisms, and
prediction of CNS permeation
All new DFHs 1-29 were evaluated as inhibitors of human cholinesterases (hAChE
and hBuChE). Following this initial screening, a selection of the most potent inhibitors
was studied in Lineweaver-Burk experiments to determine their inhibition mechanisms.
Next, in order to establish their ability to reach the CNS all derivatives were also
evaluated in the in vitro parallel artificial membrane permeability assay for the blood-
brain barrier (PAMPA-BBB). All results are gathered in Table 1.
1
3

ACCEPTED MANUSCRIPT
Table 1. Inhibition of human cholinesterases (IC , nM) and prediction of the CNS-
5
0
a
permeation by DFHs (1-29).
hAChE
IC50 (nM)
hBuChE
hAChE b PAMPA-BBB
CNS
Compd
n
X
Z
R
-6
-1
IC50 (nM) selectivity P (10 cm s ) prediction
e
1
2
3
4
5
6
7
8
9
0
0
0
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
2
2
2
2
2
2
2
NH
NH
NH
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
6-OMe
6-OH
970 ± 40
2360 ± 140
660 ± 50
7490 ± 120
300 ± 25
3610 ± 90
1460 ± 70
56 ± 5
>10000
>10000
>10
4.2
7.7 ± 0.7
1.9 ± 0.6
8.1 ± 0.3
6.1 ± 0.3
3.2 ± 0.5
6.8 ± 0.3
n.d.
cns +
cns -
>
6,7-diOMe
6-OMe
6,7-diOMe
6-OMe
6,7-diOMe
6,7-diOMe
H
>10000
>15
>1.3
>33
>2.8
>6.8
>179
46
cns +
cns +
cns +/-
cns +
>10000
O
>10000
NH
NH
O
>10000
>10000
>10000
5.6 ± 0.3
6.7 ± 0.5
n.d.
cns +
cns +
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
110 ± 10
870 ± 14
170 ± 10
91 ± 8
5090 ± 360
>10000
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
5-OMe
6-OMe
6-OH
>12
38
6500 ± 500
>10000
10.0 ± 0.2
2.8 ± 0.4
7.6 ± 0.7
8.2 ± 0.1
7.6 ± 0.2
8.8 ± 0.5
3.7 ± 0.5
5.9 ± 0.8
3.2 ± 0.3
10.6 ± 0.1
4.5 ± 0.3
5.0 ± 0.3
12.0 ± 0.5
5.1 ± 0.4
10.9 ± 0.6
15.3 ± 1.5
4.9 ± 0.1
5.3 ± 0.4
13.0 ± 0.8
n.d.
cns +
cns +/-
cns +
cns +
cns +
cns +
cns +/-
cns +
cns +/-
cns +
cns +
cns +
cns +
cns +
cns +
cns +
cns +
cns +
cns +
>110
300
443
105
20
7-OMe
33 ± 2
9900 ± 900
9300 ± 700
4300 ± 350
6300 ± 450
>10000
6-NO
6-NH
21 ± 2
2
2
41 ± 3
5,7-diOMe
5,7-diOH
6,7-diOMe
6,7-diOH
6-OMe
320 ± 20
580 ± 40
46 ± 4
>17
>217
>2.4
16
>10000
4130 ± 190
370 ± 13
23 ± 2
>10000
5800 ± 450
>10000
6-OH
>435
>91
>11
33
6,7-diOMe
6-OMe
110 ± 10
880 ± 40
300 ± 20
450 ± 40
1640 ± 70
1150 ± 90
170 ± 10
620 ± 50
10 ± 1
>10000
NH NH
NH NH
NH NH
NH NH
NH NH
NH NH
>10000
6-OH
9880 ± 770
>10000
7-OMe
>22
>6.1
>8.7
>59
14
8-OMe
>10000
8-OH
>10000
6,7-diOMe
>10000
NH NH 4,6,7-triOMe
8790 ± 500
2500 ± 70
Donepezil
250
a
b
Results are given as mean ± SEM of three independent experiments. Selectivity index
for hAChE was calculated as IC (hBuChE) / IC (hAChE). n.d., not determined.
5
0
50
1
4

ACCEPTED MANUSCRIPT
Inhibition of hAChE and hBuChE were performed following the Ellman method
[
50], using donepezil as reference drug. Both series of DFHs, derived either from 4-
chromone or 4-quinolone, were found to be selective inhibitors of hAChE with IC50
values from the low micromolar to the nanomolar range, in some cases very close to the
donepezil IC50 value (10 nM). Inhibition of hBuChE was substantially less effective,
with IC above 10 µM for the majority of new compounds (Table 1). As a result, many
5
0
hybrids showed a high preference for hAChE with selectivity indexes comparable to or
even exceeding the one of donepezil, which currently is the most used drug in the
treatment of AD.
By comparing the two classes of DFHs, 4-chromone-derived compounds (11-13
and 18, with IC s = 33-170 nM) were more potent hAChE inhibitors than their 4-
5
0
quinolone counterparts (23-25 and 28, with IC s = 170-880 nM). This fact highlights
5
0
the relevance of the endocyclic oxygen and/or the stability of the exocyclic carbonyl
group in the 4-chromenone ring in the interaction with the enzyme, since its substitution
with a heterocycle susceptible of tautomerization (4-quinolinone) was detrimental to
activity.
In general, 4-chromone-based hybrids with an aliphatic linker of 2 or 3 methylene
groups were more potent and selective hAChE inhibitors than their counterparts with
shorter linkers (n = 0 or 1). Replacement of an amide by an ester bond did not affect
significantly the inhibitory potency, as evidenced by comparing IC values for the ester
5
0
derivative 8 and its amide counterpart 18 (56 nM and 46 nM, respectively).
Regarding amide derivatives, the nature and position of substituents in the
heterocyclic moiety yielded interesting effects on the hAChE inhibition. Best results
were obtained when substituents were located at positions 6 and / or 7 of the 4-oxo-4H-
chromene ring, whereas position 5 appeared to be detrimental for activity. The most
1
5

ACCEPTED MANUSCRIPT
potent hAChE inhibitors were 14 (6-NO ), 21 (6-OH), 13 (7-OMe), 15 (6-NH ), 18
2
2
(
6,7-diOMe), and 12 (6-OH) with IC values comprised between 21 and 91 nM.
5
0
A set of potent hAChE inhibitors including further different structural features (8,
, 14, 15, 18, 21, 22, and 28) were selected for studying reaction kinetics between the
9
enzyme and these 4-chromone- or 4-quinolone-based hybrids. Following the
Lineweaver-Burk (LB) method, the initial velocity of enzymatic inhibition was
measured at four concentrations of the substrate acetylthiocholine (ATCh), in absence
and presence of increasing concentrations of the inhibitor. For each inhibitor
concentration, the plot of the reciprocals of both velocity and ATCh concentration (1/V
vs. 1/[ATCh]) gave straight lines that were fitted by least-squares analysis, allowing the
determination of the corresponding kinetic parameters (i.e., the Michaelis constant Km
and the maximum velocity Vmax). For all DFH molecules studied, both 1/Vmax (y
intercept) and -1/K (x intercept) increased with increasing inhibitor concentration.
m
Figure 2 shows the LB analysis for the 4-chromone – N-benzylpiperidine hybrid 18 as
an example (see Supporting Information for the full set of LB graphs). This common
trend let us speculate that both competitive and non-competitive mechanisms are at play
in the enzymatic inhibition by the present DFHs, suggesting the simultaneous
interaction of the hybrids with two significant sites of hAChE, CAS and PAS. For each
tested compound, replot of slopes vs. inhibitor concentration gave a straight line that
was also fitted by least-squares analysis and whose intersection on the negative x-axis
provided the calculated inhibition constants (K ) listed in Table 2.
i
1
6

ACCEPTED MANUSCRIPT
Figure 2. Kinetic study on the mechanism of hAChE inhibition by hybrid 18. (A)
Overlaid Lineweaver-Burk reciprocal plots of hAChE initial velocity at increasing
substrate concentration (ATCh, 0.2 – 1.6 mM) in the absence of inhibitor and in the
presence of 18 (0 – 0.20 µM) are shown. Lines were derived from a least-squares
analysis of the data points. (B) Replot of slopes vs. inhibitor concentration for
calculating K as the intersection on the x-axis.
i
a
Table 2. hAChE inhibition constants (K , nM) of selected DFHs .
i
Compd.
n
X
Z
R
K (nM)
i
8
9
2
2
2
2
2
3
3
2
O
O
O
O
O
O
O
O
6,7-diOMe
H
11
57
12
62
39
11
21
60
NH
NH
NH
NH
NH
NH
14
15
18
21
22
28
6-NO
2
2
6-NH
6,7-diOMe
6-OH
6,7-diOMe
NH NH 6,7-diOMe
a
Calculated from LB experiments by replotting
slopes vs. inhibitor concentration.
1
7

ACCEPTED MANUSCRIPT
Since a favourable CNS penetration is indispensable for effective drugs against AD
and other cerebral pathologies, the in vitro evaluation of the capacity of these new
molecules to cross the blood-brain barrier (BBB) is of particular interest in early stages
of pharmaceutical research. In this work, we used the in vitro PAMPA-BBB assay
originally described by Di et al. [51], and partially modified by us for testing molecules
with limited water-solubility [37,52-54]. Accordingly, the passive CNS-permeation (Pe)
of the new DFH compounds through a lipid extract of porcine brain was measured at
room temperature, and these results are reported in the last column of Table 1. In the
same assay, 11 commercial drugs were also tested for comparison, and their
permeability values were normalized to the reported PAMPA-BBB data (see Supporting
Information for further details). According to the established literature values [51],
-6
-1
compounds with P exceeding 4·10 cm s should be able to cross the blood-brain
e
-6
-1
barrier (cns+), whereas those displaying P lower than 2·10 cm s will not reach the
e
CNS (cns-). Between these values, CNS permeability is uncertain (cns +/-).
With the exception of some hydroxyl derivatives, which according to their Pe
values could experience some difficulties in passing the BBB (i.e., 2, 12, 17, and 19),
the majority of new 4-chromone- and 4-quinolone-based hybrids synthesized in this
-6
-1
work were characterized by permeability values greater than 4·10 cm s (Table 1).
Thus, according to this in vitro BBB model, it is expected all these compounds can
penetrate the CNS by passive diffusion and thus, they could interact with their
biological targets.
2
.3. Inhibition of human 5-lipoxygenase (5-LOX)
The inhibition of human 5-LOX was performed following the fluorescence-based
method described by Pufahl et al. [55], using the two well-known inhibitors (R,S)-
1
8

ACCEPTED MANUSCRIPT
zileuton and nordihydroguaiaretic acid (NDGA) as reference drugs. From the results
shown in Table 3, some structure – activity relationships could be deduced. In general,
4
-chromone-based hybrids resulted to be better inhibitors than the 4-quinolone
derivatives, reaching IC50 values between one- and two-digit micromolar ranges. The
most active 5-LOX inhibitors were 19 (n = 2, R = 6,7-diOH), 14 (n = 2, R = 6-NO2),
and 21 (n = 3, R = 6-OH) with IC values comprised between 4.2 and 15.1 µM.
5
0
The nature of the substituents in the 4-chromone ring was reflected in major
differences in the 5-LOX inhibition, leading to the following potency rank: hydroxyl >
nitro > amino > methoxy groups. The amount and position of substituents in the
heterocycle also exerted some influences on the activity of the relevant compounds. In
general, dihydroxyl compounds were endowed with higher inhibitory potency than
chromones bearing one hydroxyl group. Moreover 6,7-positions were preferred over
other positions for activity. These observations are evident when comparing the
performance of 19 (n = 2, R = 6,7-diOH), which is the best inhibitor of the series (IC50
=
4.2 µM), with 17 (n = 2, R = 5,7-diOH, IC = 95.1 µM) and 12 (n = 2, R = 6-OH,
50
IC > 100 µM). Apparently, the linker length (methylene number = 0, 2 or 3) between
5
0
the 4-chromone moiety and the N-benzylpiperidine scaffolds could also influenced 5-
LOX inhibition, as 6-hydroxychromone hybrids with 0 or 2 methylene linkers (2 and
1
2, respectively) were inactive, whereas their counterpart with a 3-methylene chain 21
showed a good IC value (15.1 µM). It is noteworthy that hybrid 18 (n = 2, R = 6,7-
5
0
diOMe) showed a moderate inhibition of 5-LOX (IC = 74.3 µM).
5
0
2
.4. Human MAOs inhibition and antioxidant properties
The next step in the assessment of the multi-target profile of these new DFH
compounds was the evaluation of their ability to inhibit both isoforms of human
1
9

ACCEPTED MANUSCRIPT
recombinant MAO, expressed in baculovirus infected BTI insect cells. (R)-Deprenyl,
iproniazid, and moclobemide were also tested as reference compounds. All results are
summarized in Table 3. Several of the DFHs tested were moderate h-MAO inhibitors
with 2-digit micromolar IC s, i.e., in the same range of values displayed by iproniazid
5
0
and moclobemide. In both MAO subtypes, the most potent inhibitor of the series was
the 6,7-dimethoxy-4-oxo-4H-chromene hybrid 18, which proved to be 3-times more
active in inhibiting hMAO-B (IC50 = 5.2 µM) with respect to hMAO-A (IC50 = 15.3
µM). Regarding selectivity, the 6-nitro-4-oxo-4H-chromene hybrid 14 showed an
interesting preference toward hMAO-A (>5-fold), whereas the 8-hydroxy-4-oxo-1H-
quinoline derivative 27 was about 7-fold more active in hMAO-B than in hMAO-A. It
is worth-mentioning that high in vitro potencies for the inhibition of MAOs are not
required to achieve therapeutic effects in vivo. For instance, iproniazid and
moclobemide that are currently used in the clinics for the treatment of psychiatric and
neurological disorders, are characterized by IC (MAO-A) = 6.67 and 361 µM,
50
respectively [56].
The antioxidant activities of newly synthesized DFHs were evaluated using the
oxygen radical scavenger capacity assay (ORAC) and trolox as reference compound.
Trolox is a water-soluble analogue of vitamin E, which contains the aromatic part
responsible for the radical capture capacity of this vitamin. Results were expressed as
trolox equivalents (µmol of trolox / µmol of tested compound) in a comparative scale
where ORAC (trolox) = 1.0 (Table 3, last column). As expected, compounds bearing
methoxy or nitro groups in the heterocyclic scaffold were inactive, while hybrids with
amine or hydroxyl groups exhibited interesting antioxidant capacities, ranging from 0.9
to 1.7-fold the trolox value.
2
0

ACCEPTED MANUSCRIPT
Table 3. Inhibition of human 5-LOX and MAOs (IC , µM); oxygen radical absorbance
5
0
a
capacity (ORAC, trolox equiv.)
5
-LOX
MAO-A
MAO-B
ORAC
(trolox equiv.)
Compd.
n
X
Z
R
(
IC , µM)
(IC , µM)
(IC , µM)
5
0
50
50
b
b
b
b
b
b
2
7
9
0
1
2
2
2
2
2
2
2
2
2
3
3
3
2
2
2
2
2
2
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
O
O
O
O
O
O
O
O
O
O
O
O
O
O
6-OH
>100 (48%)
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
6,7-diOMe >100 (45%)
H
>100 (24%)
>100 (33%)
>100 (36%)
>100 (31%)
9.8 ± 0.7
n.d.
n.d.
n.d.
11
12
13
14
15
17
18
19
20
21
22
23
24
25
26
27
28
(
6-OMe
6-OH
7-OMe
> 100
> 100
n.d.
39.3 ± 2.8
> 100
51.2 ± 4.3
> 100
1.2 ± 0.1
n.d.
6-NO
6-NH
20.5 ± 1.3
> 100
> 100
n.a.
2
28.9 ± 2.5
95.1 ± 4.1
74.3 ± 3.2
4.2 ± 0.9
> 100
0.9 ± 0.1
0.8 ± 0.1
n.a.
2
5,7-diOH
6,7-diOMe
6,7-diOH
6-OMe
n.d.
n.d.
15.3 ± 1.3
70.5 ± 8.6
> 100
5.2 ± 0.3
37.9 ± 4.2
> 100
1.7 ± 0.2
n.a.
82.4 ± 3.2
15.1 ± 2.9
6-OH
> 100
> 100
1.3 ± 0.1
n.a.
b
b
6,7-diOMe >100 (48%)
> 100
> 100
NH NH
NH NH
NH NH
NH NH
NH NH
6-OMe
6-OH
>100 (31%)
n.d.
> 100
> 100
n.a.
51.3 ± 6.2
72.7 ± 6.5
52.4 ± 3.1
> 100
59.9 ± 6.7
56.4 ± 4.5
71.6 ± 5.4
13.6 ± 0.9
> 100
1.4 ± 0.1
n.a.
b
b
7-OMe
8-OMe
8-OH
>100 (31%)
>100 (47%)
77.6 ± 4.6
n.a.
1.0 ± 0.1
n.a.
b
NH NH 6,7-diOMe >100 (31%)
R,S)-Zileuton 0.15 ± 0.03
Nordihydroguaiaretic acid (NDGA)
R)-Deprenyl
> 100
n.d.
n.d.
n.d.
0.10 ± 0.02
n.d.
n.d.
n.d.
n.d.
(
68.7 ± 4.2
6.67 ± 0.8
361.4 ± 19.4
0.017 ± 0.002
7.5 ± 0.36
> 1000
n.d.
Iproniazid
n.d.
n.d.
Moclobemide
n.d.
n.d.
a
b
Results are given as mean ± SEM from five experiments. In brackets, inhibition
percentage at 100 µM. n.d.: not determined. n.a.: not antioxidant at 10 µM.
2
1

ACCEPTED MANUSCRIPT
2
.5. Affinity studies on σ and σ receptors
1 2
A selection of 4-chromone and 4-quinolone hybrids displaying different degree of
inhibition of human AChE, MAO-A/B, and 5-LOX, and covering also different
structural motifs, was investigated for their affinity on sigma receptors of animal origin,
obtained from guinea pig brain (σ ) and rat liver (σ ), respectively. The affinities of the
1
2
DFHs for both receptors were determined in competition experiments with the
3
appropriated radioligands. Specifically, [ H]-(+)-pentazocine (PTZ) served as the
3
radioligand for the σ assay, while [ H]-1,3-di-o-tolylguanidine (DTG) was employed in
1
the σ assay [57].
2
All tested compounds showed good affinities for the σ R, with K values between
1
i
the sub-nanomolar and the two-digit nanomolar range. Affinity values for the σ R were
2
in the three-digit nanomolar range, ultimately yielding a σ /σ preference of at least 6-
1
2
times (in the worst case) (Table 4).
The best σ R affinity values were provided by compounds bearing a hydrogen-bond
1
acceptor group in the position 6 of the heterocyclic scaffold, which can be ranked as
follows: NO > NH > OH > OCH The substitution of the 4-oxo-chromene or 4-oxo-
2
2
3.
quinoline ring at different positions or the presence of a double substitution led to a
decrease of the σ R binding capability, although the K values remained in the
1
i
nanomolar range. No significant differences were detected between the 4-oxo-chromene
Z = O) or 4-oxo-quinoline (Z = NH) aromatic moieties. Also, the length of the alkyl
(
spacer between the amide group and the N-benzylpiperidine moiety did not seem to
affect the σ R affinity.
1
2
2

ACCEPTED MANUSCRIPT
Table 4. Affinity and selectivity of a set of DFHs for σ and σ receptors
1
2
a
K (nM)
i
Selectivity
Compd. n
X
Z
R
b
for σ R
1
σ1R
σ2R
1
1
1
1
1
2
2
2
2
2
2
4
5
8
9
1
3
5
7
8
2
2
2
2
2
3
2
2
2
2
NH
NH
NH
NH
NH
NH
O
O
O
O
O
O
6-OH
42.1 ± 9.3
0.98 ± 0.32
345 ± 78
287 ± 43
8
293
290
6
6-NO
6-NH
2
1.45 ± 0.57 421 ± 106
2
6,7-diOMe
6,7-diOH
6-OH
37.4 ± 6.3
27.0 ± 6.1
3.48 ± 1.26
7.2 ± 2.4
45.2 ± 10.7
47.4 ± 7.5
36.3 ± 5.0
15.0 ± 3.0
-
239 ± 36
278 ± 33
301 ± 59
459 ± 83
376 ± 64
290 ± 42
487 ± 78
-
10
87
64
8
NH NH
NH NH
NH NH
6-OMe
7-OMe
8-OH
6
NH NH 6,7-diOMe
13
-
Pentazocine
DTG
54 ± 8
-
a
Ki values are given as mean ± SEM of three independent
b
experiments. Selectivity vs. σ R was calculated as K σ R/K σ R.
1
i
2
i 1
2
3

ACCEPTED MANUSCRIPT
2
.6. Study of theoretical medicinal chemistry alerts in free databases
Before evaluating the therapeutic potential of the new DFHs in cellular assays, and
in order to make an adequate selection of candidates, we studied the potential medicinal
chemistry alerts of 1-29 in two free databases, namely ZINC15
(http://zinc15.docking.org/) [58] and SwissADME (http://www.swissadme.ch/) [59].
According to the ZINC15 web site, none of the hybrids were highlighted as pan assay
interference compound (PAINS) or aggregator (see Table S1 in the Supplementary data
file). However, in the SwissADME platform three compounds were marked with
structural alerts, according to the Brenk method [60]. These compounds were hybrids
1
4 (nitro group), 15 (aniline derivative) and 19 (catechol), which were consequently
discarded for further cellular assays.
2
.7. Cell viability and neuroprotective properties
Cell viability and neuroprotective capacity of the new compounds against
mitochondrial oxidative stress were evaluated using the human neuroblastoma SH-
SY5Y cell line. A toxic cocktail of rotenone (30 µM) and oligomycin A (10 µM) was
used as a model of mitochondrial oxidative stress. This mixture of toxic molecules
blocks complexes I and V of the respiratory electron transport chain, inducing high
concentration of mitochondrial free radicals and cell death, confirming its validity as a
good experimental model of mitochondrial oxidative stress [31,38,61-63].
Accordingly, DFH compounds at increasing concentrations (0.1 - 3.0 µM) were co-
incubated with the toxic mixture for 24 h; then, cell survival was quantified by the 3-
[
[
4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide (MTT) colorimetric assay
64]. MTT reduction in basal conditions was considered as 100 % viability of cells
2
4

ACCEPTED MANUSCRIPT
without any treatment. In all experiments, melatonin and donepezil were used as
reference drugs, which gave the following percentages of neuroprotection: melatonin
(10 nM) = 42.1 ± 3.01 % and donepezil (5 µM) = 38.4 ± 4.72 %.
For these studies, we selected hybrids with differences in the heterocycle core and
in the nature and length of the linker, and without any structural alert (Table 5). To a
greater or lesser extent, all tested compounds showed neuroprotective properties against
mitochondrial oxidative stress, in especial hybrids 5 (n = 1, X = O, Z = O, R = 6,7-
diOMe), 12 (n = 2, X = NH, Z = O, R = 6-OH), 18 (n = 2, X = NH, Z = O, R = 6,7-
diOMe), and 24 (n = 2, X = NH, Z = NH, R = 6-OH). The few differences found
between ORAC-active and ORAC-inactive compounds (hydroxyl and methoxy hybrids,
respectively), suggested that the neuroprotective properties of the new series of DFHs
are mainly due to the activation of the endogenous antioxidant response, rather than to
the direct capture of radical oxygen species.
2
5

ACCEPTED MANUSCRIPT
Table 5. Neuroprotection (%) of selected DFHs at the indicated concentrations in the
human neuroblastoma cell line SH-SY5Y against the cocktail of rotenone (30 µM) and
oligomycin A (10 µM).
Compd. n
X
Z
R
0.1 µM
0.3 µM
1.0 µM
3.0 µM
1
3
5
0
0
1
2
2
2
NH
NH
O
O
O
O
O
O
6-OMe
9.8 ± 1.1
13.9 ± 3.1
14.6 ± 1.2
12.3 ± 2.5
6.8 ± 1.2
8.6 ± 3.0
8.6 ± 1.7
12.6 ± 3.1
12.4 ± 3.6
15.2 ± 4.9
6,7-diOMe
5.7 ± 3.6
18.0 ± 3.1
6,7-diOMe 19.4 ± 0.7* 19.6 ± 3.0*
6-OH 15.0 ± 2.9 18.8 ± 3.8
1
2
8
4
NH
NH
1
2
6,7-diOMe 11.5 ± 1.5* 18.1 ± 1.6** 17.1 ± 1.9* 22.4 ± 2.5**
6-OH
18.6 ± 1.2* 13.7 ± 5.1* 17.4 ± 1.3* 29.7 ± 2.8**
NH NH
Results are given as mean ± SEM of 4 independent experiments in triplicate. * p < 0.05,
*
4
*p<0.01 respect to control. Neuroprotection values of reference drugs: melatonin (10 nM) =
2.1 ± 3.01 %; donepezil (5 µM) = 38.4 ± 4.72 %.
2
.8. Neurogenic studies
The brain-permeable 6,7-dimethoxy-4-oxo-4H-chromene hybrid 18, without any
structural alert in the online ZINC15 and SwissADME databases, and which displays an
interesting in vitro profile, namely affinity of σ R, inhibition of human AChE, 5-LOX,
1
MAO-A and MAO-B, and a good neuroprotective activity against mitochondrial
oxidative stress, was further evaluated as neurogenic agent. Adult mice neural stem cells
(NSC) were isolated from the subgranular zone of the hippocampal dentate gyrus and
cultured as neurospheres (NS), which were grown for 7 days in culture with or without
compound and then, they were adhered on a substrate to allow differentiation for 3 days
[
(
65-67]. Afterward, β-III-tubulin (clone TuJ1) and microtubule-associated protein 2
MAP-2) antibodies were used to visualize early proliferation and neuronal maturation,
2
6

ACCEPTED MANUSCRIPT
respectively. As shown in Figure 3, derivative 18 clearly promoted the differentiation of
NSCs and their further maturation to a neuronal phenotype. Moreover, an important
population of these new cells has reached a consolidated neuronal stage, which is indeed
observable in their distinctive morphology.
Inner part
Outer part
Control
18
Figure 3. In vitro neurogenic effect of hybrid 18 (10 µM) on mice hippocampal SGZ-
derived spheres. NSC enriched spheres were grown for 7 days in culture in the presence
of compound (10 µM). Later on, neurospheres were adhered on a substrate to allow
differentiation for 3 days in the presence of compound. Representative images show the
expression of β-III-tubulin (TuJ clone; green) and MAP-2 (red) inside the neurosphere
(inner part) and in the distal area (outer part). DAPI was used for nuclear staining. Scale
bar, 200µm.
2
7

ACCEPTED MANUSCRIPT
2
.9. In silico toxicity and metabolism of hybrid 18
To study the potential therapeutic success of hybrid 18, we performed an in silico
prediction of its toxicity and metabolism in humans using the Nexus knowledge-based
system [68]. The toxicity predictions obtained with the Derek Nexus program rely on
the comparison of the structural features of a given compound with one or more
toxicophore patterns (structural alerts) in the Lhasa's knowledge base 2018 (species:
human). There were 55 toxicity endpoints for compound 18, analyzed at the minimum
reasoning level of “impossible” [69] (Chart S1 in Supplementary Information). No
alerts where fired for hybrid 18.
Despite in silico structure-activity relationship toxicity models include implicit
consideration of metabolism, some metabolic relationships may be missed. Therefore,
an in silico prediction of the compound metabolism can add value to its toxicological
assessment. Hence, we performed a prediction of the most likely phase-I metabolites of
hybrid 18, obtaining 12 structures. Potential toxicity was also assessed for these
metabolites with Derek Nexus analyzing 55 endpoints (Chart S1). Again the minimal
likelihood to consider a toxic outcome in the analysis was “impossible”. Results showed
that the metabolites were not associated with any structural alerts for toxicity by Derek
(Table 6).
2
8