Synthesis, pharmacological and structural studies of 5-substituted-3-(1-arylmethyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indoles as multi-target ligands of aminergic GPCRs

Article abstract of DOI:10.1016/j.ejmech.2019.07.050

Schizophrenia is a complex disease with not fully understood pathomechanism, involving many neurotransmitters and their receptors. This is why it is best treated with multi-target drugs, such as second generation antipsychotics. Here we present 5-substitu

Full text of DOI:10.1016/j.ejmech.2019.07.050

Accepted Manuscript
Synthesis, pharmacological and structural studies of 5-substituted-3-(1-
arylmethyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indoles as multi-target ligands of
aminergic GPCRs
Magda Kondej, Tomasz M. Wróbel, Andrea G. Silva, Piotr Stępnicki, Oliwia Koszła,
Ewa Kędzierska, Agata Bartyzel, Grażyna Biała, Dariusz Matosiuk, Maria I. Loza,
Marián Castro, Agnieszka A. Kaczor
PII:
S0223-5234(19)30671-3
DOI:
https://doi.org/10.1016/j.ejmech.2019.07.050
EJMECH 11547
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 31 May 2019
Revised Date: 3 July 2019
Accepted Date: 17 July 2019
Please cite this article as: M. Kondej, T.M. Wróbel, A.G. Silva, P. Stępnicki, O. Koszła, E. Kędzierska,
A. Bartyzel, Graż. Biała, D. Matosiuk, M.I. Loza, Mariá. Castro, A.A. Kaczor, Synthesis, pharmacological
and structural studies of 5-substituted-3-(1-arylmethyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indoles as
multi-target ligands of aminergic GPCRs, European Journal of Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.ejmech.2019.07.050.
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
Synthesis, pharmacological and structural studies of 5-substituted-3-(1-
arylmethyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indoles as multi-target
ligands of aminergic GPCRs
1
1,2
3
1
1
Magda Kondej , Tomasz M. Wróbel , Andrea G. Silva , Piotr Stępnicki , Oliwia Koszła ,
4
5
4
1
3
Ewa Kędzierska , Agata Bartyzel , Grażyna Biała , Dariusz Matosiuk , Maria I. Loza ,
3
1,6*
Marián Castro , Agnieszka A. Kaczor
1
Department of Synthesis and Chemical Technology of Pharmaceutical Substances, Faculty of
Pharmacy with Division of Medical Analytics, Medical University of Lublin, 4A Chodźki St.,
PL-20093 Lublin, Poland
2
Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences,
University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark
3
Department of Pharmacology, Universidade de Santiago de Compostela, Center for
Research in Molecular Medicine and Chronic Diseases (CIMUS), Avda de Barcelona,
E-15782 Santiago de Compostela, Spain
4
Department of Pharmacology and Pharmacodynamics, Faculty of Pharmacy with Division of
Medical Analytics, Medical University of Lublin, 4A Chodźki St., PL-20093 Lublin, Poland
5
Department of General and Coordination Chemistry, Maria Curie-Skłodowska University,
M. Curie-Skłodowskiej Sq. 2, PL-20031 Lublin, Poland
6
School of Pharmacy, University of Eastern Finland, Yliopistonranta 1, P.O. Box 1627,
FI-70211 Kuopio, Finland
E-mail: agnieszka.kaczor@umlub.pl

ACCEPTED MANUSCRIPT
Abstract: Schizophrenia is a complex disease with not fully understood pathomechanism,
involving many neurotransmitters and their receptors. This is why it is best treated with multi-
target drugs, such as second generation antipsychotics. Here we present 5-substituted-3-(1-
arylmethyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indoles (1-20) which are ligands of dopamine
D and serotonin 5-HT and 5-HT receptors and display affinity in the nanomolar range.
2
1A
2A
These compounds were designed as modifications of the virtual hit experimentally confirmed,
D2AAK1, and synthesized from indole or 5-alkoxyindoles and N-substituted piperidin-4-ones
in methanol in the presence of potassium hydroxide. Compound 9 was subjected to X-ray
studies and it crystallizes in the centrosymmetric monoclinic space group P2 /c with one
1
molecule in an asymmetric unit. Three most potent compounds (5, 9 and 17) turned out to be
antagonists of both D and 5-HT receptors what is beneficial for their potential application
2
2A
as antipsychotics. Compound 5 was subjected to behavioral studies and exhibited
antipsychotic, pro-cognitive and antidepressant activity in appropriate mice models. Structure-
activity relationships for compounds 1-20 were rationalized using molecular docking. It was
found that, in general, bulky C5-alkoxy substituents at the indole moiety are not favorable as
they direct towards aqueous environment of the extracellular vestibule. Keywords:
antipsychotics; behavioral studies, G protein-coupled receptors; indole derivatives; multi-
target compounds; schizophrenia.
Graphical abstract:

ACCEPTED MANUSCRIPT
Highlights:
•
•
•
•
•
Novel multi-target ligands of aminergic GPCRs.
Antagonists of dopamine D and serotonin 5-HT receptors.
2
2A
Compound 5 exhibits antipsychotic, pro-cognitive and antidepressant activity.
X-ray studies of compound 9 reveal energetically stable conformation.
Molecular docking enables to rationalize the observed SAR.
1
. Introduction

ACCEPTED MANUSCRIPT
For a long time, the classical strategy of a drug action, consisting in one molecule
targeting one certain receptor, was in the lead of the pharmacological approaches to the
treatment of central nervous system diseases. This "magic bullet" paradigm was thought to be
superior to multi-target "magic shotgun" concept, which assumes several molecular targets
being simultaneously modulated by one molecule of a drug. The advantage of selectively
acting drugs was based on the assumption that there is lower possibility to cause adverse
effects comparing to drugs acting via a number of receptors. However, the complexity of
central nervous system diseases often constitutes an obstacle to effective treatment with
selective drugs. Therefore, the currently predominant approach in drug discovery is to search
for one molecule that would be able to affect several targets [1]. Multi-target agents exert
their effect through a wide range of receptors and involve several neurotransmitter pathways,
what may be beneficial to the treatment of such diseases, since they are usually associated
with impairment in transmission of more than one neuronal messenger. Moreover, applying
one multi-target drug instead of a few selective drugs may reduce the frequency of side effects
occurrence and allows to avoid drug-drug interactions [2].
One of the central nervous system diseases characterized by complex
pathomechanism, involving a number of neurotransmitters, is schizophrenia. Ranking as the
th
1
5 top leading cause of disability among the population worldwide [3], schizophrenia has a
major impact in health care systems and society. It is estimated that the disease affects 1% of
global population with its beginning observed usually in early adulthood [4]. This chronic,
severe mental disorder is characterized by dysfunctions in perception, thought processes and
behavior, as well as cognitive and memory impairment. Patients suffering from schizophrenia
experience symptoms such as hallucinations, delusions, disorganized speech or movement
disorders, anhedonia, flattening or social withdrawal [5].

ACCEPTED MANUSCRIPT
The etiology of schizophrenia involves environmental and genetic components, which
may constitute predisposing factors for the development of the disease. Nonetheless, the exact
causes of schizophrenia are still not satisfactorily understood. The main concept explaining
the basis of the disease is related to dysfunctions in dopaminergic transmission in the
mesolimbic pathway, resulting in positive symptoms, and in the mesocortical circuit,
associated with the occurrence of negative symptoms. Schizophrenia is treated with
antipsychotic drugs that fall into three generations. Classical first generation drugs are
dopamine D receptor antagonists and are effective against positive symptoms, with low
2
affinity to reduce negative symptoms. Newer antipsychotics possess multi-target profile of
action since they exhibit additional affinity for other molecular targets, and notably for
serotonin receptors. Drugs of second generation bind to serotonin 5-HT receptor with higher
2
A
affinity than to D receptor. Affecting serotonin neurotransmission may be beneficial to the
2
pharmacotherapy of schizophrenic patients. Activation of 5-HT1A receptor leads to increased
release of dopamine in the mesocortical circuit, what may contribute to enhancing efficacy
against negative and cognitive symptoms [6]. Antagonism at serotonin 5-HT2A receptors,
which are located on dopaminergic neurons, results in reducing negative symptoms of
schizophrenia and improving extrapyramidal side effects profile of atypical antipsychotics,
due to increased activity of nigrostriatal dopaminergic circuit [7]. Thus, multi-target drugs for
the treatment of schizophrenia are more potent agents with a wider range of symptoms that
are alleviated and with a safer profile.
In previous studies, we performed structure-based virtual screening in order to identify
novel antagonists of dopamine D receptor with potential application in the treatment of
2
schizophrenia [8]. As a result, D2AAK1 (Fig. 1) with K of 58 nM to D receptor has been
i
2
found. This compound is characterized by multi-target profile, since it possesses additional
affinity to D , D , 5-HT and 5-HT2A receptors (K of 1418, 614, 125 and 358 nM,
1
3
1A
i

ACCEPTED MANUSCRIPT
respectively). Functional studies confirmed that D2AAK1 is an antagonist of D and 5-HT2A
2
receptor and partial agonist of 5-HT1A receptor, what is beneficial for antipsychotic activity.
Subsequent behavioral studies of D2AAK1 confirmed its antipsychotic properties, evaluated
in a classical test consisting in reducing amphetamine-induced hyperactivity in mice.
Moreover, the studied compound exerts anxiolytic and pro-cognitive effect in vivo [9].
Fig. 1. Structure of virtual hit, D2AAK1.
In this work, as a continuation of the raised subject, we aimed at finding novel
potential antipsychotics with multi-target receptor profile, what has been achieved by
designing and synthesizing analogues of the lead structure D2AAK1. Performed
modifications include replacement of the substituent attached to nitrogen atom of 1,2,3,6-
tetrahydropyridin-4-yl moiety and removing methoxy group from position 5 of the indole
moiety or implementing ethoxy or isopropoxy group in this position.. Due to the promising
activity of D2AAK1 and the presence of a protonatable nitrogen atom that is crucial for
interactions with aminergic GPCRs, we decided to keep 3-(1,2,3,6-tetrahydropyridin-4-yl)-
1
H-indole scaffold, a new set-up with potential antipsychotic activity. The obtained
compounds were then subjected to molecular modeling, X-ray, and detailed in vitro and
behavioral studies to characterize their pharmacological multitarget profile and activity as
potential antipsychotics.
2
2
. Results
.1. Chemistry
For the synthesis of compounds 1-20 we adopted the method developed by Guillaume
et al. [10]. Shortly, indole or 5-alkoxyindole was heated with a variety of commercially

ACCEPTED MANUSCRIPT
available N-substituted piperidin-4-ones in 1M potassium hydroxide methanol solution at
7
5 °C in a sealed vial (Scheme 1). This straightforward method proved to be superior in our
hands to the reaction under acidic conditions [11] and allowed for the rapid synthesis of the
desired compounds with minimal efforts invested in purification. Synthesis of compound 11,
under slightly different conditions was described by us before [13]. Precise structures of
obtained compounds are presented in Table 2.
Scheme 1. Synthesis of compounds 1-20.
2
.2. X-ray studies
The molecular structure of 3-(1-benzyl-1,2,3,6-tetrahydropyridin-4-yl)-1H-indole (9),
including the atom labelling, is shown in Fig. 2. Compound crystallizes in the
centrosymmetric monoclinic space group P2 /c with one molecule in an asymmetric unit. The
1
interatomic distances and angles agree with those described in the literature [12] and are
comparable with those observed for other closely related indole derivatives [13–16].
However, there is a fluctuation of bond distances involving carbon atoms in benzene. The
bond length C(17)-C(18) is slightly lower (1.333(5) Å) than that observed for C −ꢀ C bonds
ar
ar
(∼ 1.380 Å). The indole ring of the compound is planar with a maximum deviation of 0.019(2)
Å of (C8) atom from the best plane. The tetrahydropyridine ring is slightly inclined to the
indole ring system by 14.6(1)°. The r.m.s. atomic displacement for the nonhydrogen atoms in
indole ring-tetrahydropyridine moiety is 0.188(2) Å. The mean plane of the remaining part of

ACCEPTED MANUSCRIPT
compound i.e. the phenyl ring ((C15)–(C20)) is nearly perpendicular to the plane created by
the rest of the molecule, forming an angle of 86.79(6)°. Similar to the previously described
indole derivative (3-(1-benzyl-1,2,3,6-tetrahydropyridin-4-yl)-5-ethoxy-1H-indole, (11) [13]
the tetrahydropyridine ring system adopts a half-chair conformation [17–20]. In contrast to
compound 11, the N2 atom in compound 9 is below (-0.339(1) Å) and C5 atom lies above
(0.320(1) Å) the mean plane defined by non-hydrogen N2/C13/C12/C11/C10/C9 atoms (the
puckering parameters: Q = 0.506(1) Å, = 50.6(1) and = 333.8(1)°)
θ
°
ϕ
Fig. 2. A view of the molecular structure of 9, with the atom labeling and displacement
ellipsoids drawn at the 30% probability level.
The difference in conformation between 9 and 11 is shown in Fig. 3.

ACCEPTED MANUSCRIPT
Fig. 3. Comparison between the structure of 9 (blue) and 11 (red).
The crystal packing of compound 9 was determined by intermolecular hydrogen bonds
of the N‒H···N type (Table 1). The hydrogen bonds between the indole and
tetrahydropyridine rings of neighboring molecules form a one-dimensional network with a
C(8) infinite chain motif [21] (see Supplementary Information, Fig. S1). As a result, the
molecules are ordered into infinite ribbons extending along the [010] direction (Figs. S1B and
S2).
Table 1. Hydrogen bonding and C-H···Cg interactions geometry of compound 9.
Hydrogen bond [Å, º]
D-H···A
d(D-H)
d(H···A)
d(D···A)
∠ DHA
N(1)−H(1N)···N(2)i
0.84(2)
2.40(2).
3.195(2)
158(2)
C-H···Cg interactions [Å, º]
C-H ···Cg
d(D-H)
0.97
d(H···Cg) d(C···Cg) ∠ CHCg
C(12)–H(12A)···Cg(1)i
C(12)–H(12A)···Cg(3)i
3.00
2.92
3.621(2)
3.789(2)
123
149
0.97
Symmetry codes: (i) -x,y+1/2,-z+1/2; Cg1 and Cg3 are the centroids of the N1/C1/C2/C3/C8 and
C3/C4/C5/C6/C7/C8 rings, respectively.
2
.3.
Affinity of compounds at selected targets
The affinity of the compounds for dopamine D , and serotonin 5-HT and 5-HT2A
2
1A
receptors was evaluated in in vitro radioligand binding assays, on cell membranes from cell
lines stably expressing the cloned human receptors. Compounds were evaluated at a single
(10
µM) concentration and for those showing an inhibition of the specific radioligand binding
(%inh.) higher than 65%, concentration-response curves were constructed in order to
determine their dissociation constant (K ). This resulted in 18 compounds evaluated in
i
concentration-response curves at D , the 20 compounds evaluated at 5-HT receptors, and 6
2
1A
compounds evaluated at 5-HT2A receptors. Affinity (pK , K ) values or %inh. at 10 µM are
i
i

ACCEPTED MANUSCRIPT
given in Table 2. Unless otherwise indicated, compounds for which K values are given fully
i
displaced the specific radioligand binding at the highest concentrations assayed (either 10 µM
or 100 µM) (Supplementary Figures S3, S4 and S5 for D , 5-HT1A and 5-HT2A binding
2
assays, respectively). Five compounds (compounds 5, 9, 10, 13 and 17) displayed affinities at
D receptors within the low to moderate nanomolar range, three of them (compounds 5, 9 and
2
1
7) showing multitarget affinity profiles in the nanomolar range at 5-HT1A and 5-HT2A
receptors.
Table 2. Competitive radioligand binding data at human D , 5-HT and 5-HT receptors.
2
1A
2A
Compound
R1
R2
D2
58 [8]
5-HT1A 5-HT2A
125 [8] 358 [8]
D2AAK1
OMe
1
2
3
4
5
6
7
H
257 ± 752 ± 6 173 ± 1
60
OMe
OEt
OiPr
H
142 ± 195 ± 15
±
±
±
±
a
18
33
1%
62 ± 1% 486 ± 47
18 7%
51 ± 1% 431 ± 5
10%
± 305 ± 171
11 28
141 ± 157 ± 1888 ±
13 373
166 ± 139 ± 0 ± 2%
1
55
72
3.1
2
OMe
OEt
33

ACCEPTED MANUSCRIPT
a
66
33
8
9
1
1
1
1
1
1
1
1
1
1
2
OiPr
H
465 ± 214 ± 1 14
65 2%
26 ± 9.9 354 ± 125
35
± 139 ± 41
27 10%
±
±
±
1
45
0
1
2
3
4
5
6
7
8
9
0
OMe
OEt
OiPr
H
63
6.6
1
151 ± 220 ± 2099 ±
a
1
7 [13]
321 ± 203 ± 32
11 12%
± 243 ± 52
31 6%
148 ± 134 ± 7 48
1%
132 ± 194 ± 38
56 3%
165 ± 378 ± 36
76 1%
± 243 ± 194
17 34
112 ± 135 ± 24
19 6%
108 ± 160 ± 9 3 ± 3%
45
745
±
±
±
±
±
±
±
60
87
1.1
1
OMe
OEt
OiPr
H
26
35
28
27
0.1
1
OMe
OEt
OiPr
15
25
299 ± 290 ± 43
±
68
80
8%
Haloperidol
3.25
.18
±
0
5
-Carboxamidotryptamine
0.34 ±
0.02
Risperidone
0.37 ±
.09
0

ACCEPTED MANUSCRIPT
Data are expressed as %inh. at 10
µ
M (mean ± SEM of 1 -3 experiments performed in
duplicate) or K (nM) (mean ± SEM of 2 - 3 independent experiments performed in duplicate)
i
a
when full displacement was achieved. Full displacement of specific binding was not achieved
at the maximum concentration assayed (100 µM) so K value retrieved from the analysis
i
might be not accurate.
2
.4.
Structure-activity relationships
Compounds with C5 unsubstituted indole generally displayed enhanced affinity
towards D receptor. As the bulkiness of R alkoxy fragment increased the affinity dropped.
2
2
Size of the R substituent attached to the tetrahydropyridine fragment seemed to have less
1
pronounced effect. Compounds with benzyl (9) and 3-methoxybenzyl (17) group were found
to be the most potent, with their affinity towards D receptor being 26 nM and 27 nM
2
respectively. Interestingly, exchanging 3-methoxy (17) with 4-methoxy (13) decreased
activity three-fold. Within compounds containing indole C5 methoxy group, compound 10
showed two-fold increased affinity compared to 2, 6, 14 and 18. Compounds with
furanylmethyl substituent displayed decreased affinity, prominently highlighted by
compounds 3 and 4. This observation suggests the benzyl group contributes favorably towards
increased affinity while installation of furanylmethyl is detrimental do the D binding.
2
In case of 5-HT receptors, compounds with methoxy substituent displayed increased
1
A
affinity with K values of 195 nM, 157 nM, 139 nM, 134 nM and 135 nM for compounds 2, 6,
i
1
0, 14 and 18 respectively. This contrasts with D affinity where no indole C5 substitution
2
was beneficial for the potency. Increased bulkiness of R group mirrors the trend observed in
2
the case of D affinity, albeit to a lower extent. However, presence of benzyl group at R was
2
1
beneficial to the 5-HT1A binding, similarly to situation observed in the case of D binding.
2
Furanylmethyl substituent proved again to be the least favorable with the most pronounced
effect observed in compound 1 which showed the lowest 5-HT1A affinity (K = 752 nM)
i
among all compounds tested. This further corroborates polarizing effect of benzyl and
furanylmethyl groups in respect to their effect on affinity.

ACCEPTED MANUSCRIPT
All compounds were found to have significantly decreased affinity towards 5-HT2A
receptor with only 1, 5, 9 and 17 displaying potencies in the nanomolar range. These are the
compounds without indole C5 substituent, suggesting preference of unsubstituted indole
moiety for the increased affinity in this receptor type. All variations of R tested by us seemed
1
to have limited effect to enhance 5-HT binding.
2
A
2
.5. Receptor functional studies
2
.5.1. Functional study of selected multi-target compounds at dopamine D receptors
2
The functional behavior of compounds 5, 9 and 17 at dopamine D receptors was
2
investigated in cAMP assays of D -mediated inhibition of cAMP production stimulated by
2
forskolin, in CHO-K1 cells expressing the cloned human D receptor. None of the compounds
2
(0.1 nM – 100 µM) was able to inhibit forskolin-stimulated cAMP production, indicative of a
lack of D agonist efficacy, while dopamine (1 nM – 1 mM) inhibited forskolin-stimulated
2
cAMP production with an EC of 40 nM in the same conditions (data not shown). On the
5
0
contrary, compounds 5, 9 and 17 (0.1 nM – 100 µM), as well as haloperidol (0.1 nM – 10
µM) included as reference antagonist in the assays, fully antagonized the inhibition of cAMP
production elicited by a submaximal (1 µM) concentration of dopamine in a concentration-
dependent manner (Supplementary Fig. S6). The potency of the three selected compounds as
D antagonists was determined in cAMP assays by Schild analysis of concentration-response
2
curves of the D -like receptor full agonist quinpirole (10 pM – 1 mM) in the absence or
2
presence of three different concentrations of the compounds. Compounds 5, 9 and 17 elicited
a parallel rightward shift of the concentration-response curves of quinpirole in a
concentration-dependent manner (Fig. 4). Schild plots supported a behavior of competitive
antagonism for the three compounds, with fits to linear regression of slope close to unity, and
the compounds displayed the same order of potency (pK , K ) (Table 3) than of affinity (K ) at
b
b
i
D receptors.
2

ACCEPTED MANUSCRIPT
Fig. 4. Functional assays of cAMP signalling at cloned D receptors. Concentration-response
2
(inhibition of forskolin-stimulated cAMP production) curves of quinpirole in the absence or
presence of compound 5 (A), 9 (B) or 17 (C) at the indicated concentrations in CHO-K1 cells
expressing human D receptors. The graphs show data (mean ± SEM) of one representative
2
experiment out of two independent experiments performed in duplicate. (D, E, F) Schild plots
2
from the experiments shown in A, B and C were fitted to linear regression (R = 0.999, 0.999,
and 1.000; slope = 0.972, 1.069, and 1.072, for compounds 5, 9, and 17, respectively).

ACCEPTED MANUSCRIPT
Table 3. Evaluation of potency and efficacy of selected compounds at the indicated targets in
a
in vitro functional assays.
a
b
c
Compound
hD2
h5-HT1A
h5-HT2A
pKb
Kb
%Emax
max.
%inh.
pIC50
(mean ±
SEM)
IC50
(
mean ± SEM)
(
mean ±
SEM)
[nM]
[nM]
(mean ±
SEM)
5
9
8.30 ±
5.06
3.51
4.95
67 ± 9%
89 ± 6%
76 ± 8%
96 ± 1%
5.62 ± 0.20
5.58 ± 0.27
5.10 ± 0.22
2415
2618
7980
0
.10
8.45 ±
.03
8.31 ±
95 ± 3%
98 ± 2%
0
1
5
7
0
.12
d
-Carboxamidotryptamine
ND
ND
ND
95 ± 3%
ND
ND
ND
ND
Risperidone
ND
99 ± 3%
9.15 ± 0.04
0.72
a
Antagonism of quinpirole-mediated inhibition of forskolin-stimulated cAMP
production in CHO-K1 cells expressing human D receptors. Antagonist potency (pK ,
2
b
-
logK ; K ) was determined by Schild analysis of concentration-response curves of
b b
quinpirole in the absence or presence of three different concentrations of the
b
compounds; agonist efficacy (%Emax, maximal % of inhibition of forskolin-stimulated
c
cAMP production) in HEK293 cells expressing human 5-HT1A receptors; antagonist
efficacy (max. %inh., maximal % of inhibition of 5-HT-stimulated IP production) and
potency (pIC , -log IC ; IC , concentration of the compound eliciting the 50% of
5
0
50
50
d
maximal compound effect) in CHO-K1 cells expressing human 5-HT receptors; not
2
A
determined. Data are mean ± SEM of 2 (D , 5-HT ) or 3 (5-HT ) independent
2
2A
1A
experiments performed in duplicate.
.5.2. Functional study of selected multi-target compounds at serotonin 5-HT and 5-HT
2A
2
1A
receptors
The three selected compounds (5, 9 and 17) were evaluated in functional studies at 5-
HT1A receptors in cAMP functional assays of 5-HT -mediated inhibition of forskolin-
1A
stimulated cAMP production, in HEK293 cells stably expressing the cloned human 5-HT1A

ACCEPTED MANUSCRIPT
receptor. The compounds, assayed at two concentrations (10 µM and 100 µM) inhibited the
cAMP production stimulated by forskolin to different extent, indicative of 5-HT1A agonist
efficacy. None of the compounds at the concentrations assayed reached the maximal response
of the full 5-HT agonist 5-CT (Table 3), while concentrations of the compounds higher than
1
A
1
00 µM were not evaluated in order to avoid potential artefacts or solubility issues.
In functional assays of 5-HT2A signalling, any of the three selected compounds showed
agonist efficacy at promoting inositol phosphates (IP) production at the concentrations
assayed (from 0.1 nM to 10 µM), while 5-HT promoted IP production with EC of 99 nM in
5
0
the same assays (data not shown). However, compounds 5, 9 and 17 completely antagonized
-HT-stimulated IP production in a concentration-dependent manner (Fig. 5) with IC values
5
5
0
shown in Table 3, resulting compounds 5 and 9 equipotent and compound 17 slightly less
potent as 5-HT2A antagonists. K values of the three selected compounds as 5-HT
b
2A
antagonists estimated following the equation reported by Leff and Dougall [22] (see
experimental section) were in the high nanomolar range (K (mean ± SEM) = 238 ± 103 nM,
b
2
79 ± 154 nM and 803 ± 379 nM for compounds 5, 9 and 17, respectively), and they indicated
potencies as 5-HT antagonists of around 50 times, 80 times and 200 times lower than as D
2
2
A
antagonists for compounds 5, 9 and 17, respectively.
Fig. 5. Functional assays of inositol phosphate signalling at cloned 5-HT2A receptors.
Concentration-response curves of compounds 5, 9, 17 and risperidone (as reference 5-HT2A
antagonist) on IP production stimulated by 1 µM 5-HT in CHO-K1 cells expressing human 5-
HT receptors. The graph shows data (mean ± SEM) of one representative experiment out of
2
A
2
independent experiments performed in duplicate.

ACCEPTED MANUSCRIPT
2
.6. Molecular modeling
In order to study the ligand-receptor interactions at the molecular level, compounds 1-
2
0 were docked to the X-ray structure of the human dopamine D receptor in the inactive state
2
(PDB ID: 6CM4) [23] and previously reported homology models of the human serotonin
5
-HT1A and 5-HT2A receptors, also in inactive conformations [9]. In case of all the studied
receptors the main ligand-receptor contact was the interaction between the protonatable
nitrogen atom of the ligand and the conserved Asp (3.32) (numbering according to
Ballesteros-Weinstein nomenclature) [24] of the receptor as typical for orthosteric ligands of
aminergic GPCRs [8,9].
The results of molecular docking of most potent compounds 5, 9 and 17 to the human
dopamine D receptor are presented in Fig. 6. The position of ligands in the orthosteric pocket
2
of the D receptor is in accordance with previously reported docking pose for the virtual hit
2
D2AAK1 [8,9] and docking pose of compound 11 (D2AAK1_3) [13]. Indole moiety is
directed towards extracellular part of the receptor while the arylmethyl group is placed deeper
in the receptor cavity. In addition to interaction with the conserved Asp (3.32) compounds 5,
9
and 17 interact via π-π stacking interactions with aromatic residues: Trp (6.48), Phe (6.50),
Phe (6.51) or Trp 100 from the first extracellular loop, ecl1 or Phe (6.50), Phe (6.51),
respectively.

ACCEPTED MANUSCRIPT
Fig. 6. Compounds 5 (A, B), 9 (C, D) and 17 (E, F) in the binding pocket of human dopamine
D receptor. A, C, E – 3D view of the binding site. Ligands represented as sticks with
2
magenta carbon atoms. Protein represented as wire with grey carbon atoms, main interacting
residues shown as sticks. Hydrogen bonds shown as red dashed lines. Non-polar hydrogen
atoms not shown for clarity. B, D F– 2D view of the binding site.
The results of molecular docking of most potent compounds 5, 9 and 17 to the human
serotonin 5-HT receptor are presented in Fig. 7. The ligand poses in this receptor are similar
1
A

ACCEPTED MANUSCRIPT
to the poses at the D receptor and are in agreement with the previously reported pose of the
2
virtual hit D2AAK1 [9]. Compound 9 interacts only with the conserved Asp (3.32) while
compounds 5 and 9 form hydrogen bond with Cys 187 from the second extracellular loop,
ecl2 and interacts via
additional hydrogen bond with Lys 191 from the ecl2 and
3.28).
π
-
π
stacking interaction with Tyr (2.63) and compound 17 forms
π-π
stacking interaction with Phe
(

ACCEPTED MANUSCRIPT
Fig. 7. Compounds 5 (A, B), 9 (C, D) and 17 (E, F) in the binding pocket of human serotonin
5
-HT1A receptor. A, C, E – 3D view of the binding site. Ligands represented as sticks with
magenta carbon atoms. Protein represented as wire with grey carbon atoms, main interacting
residues shown as sticks. Hydrogen bonds shown as red dashed lines. Non-polar hydrogen
atoms not shown for clarity. B, D F– 2D view of the binding site.

ACCEPTED MANUSCRIPT
The results of molecular docking of most potent compounds 5, 9 and 17 to the human
serotonin 5-HT2A receptor are presented in Fig. 8. The docking poses of the studied
compounds at this receptor follow the docking poses at the D receptor and 5-HT receptor
2
1A
and are in accordance with previously reported docking poses for the virtual hit, D2AAK1
9]. In addition to interactions with the conserved Asp (3.32), compound 5 forms a hydrogen
bond with Cys 227 from the ecl2 and stacking interaction with Trp (3.28) and Phe (6.52),
compound 9 interacts via stacking with Trp (3.28) and Phe (6.52) while compound 17
forms a hydrogen bond with Cys 227 from the ecl2 and stacking interaction with Trp
3.28).
[
π-π
π-π
π-π
(
Molecular docking partially enabled to explain the observed structure activity
relationships of the studied compounds 1-20 at the studied receptors. The rationale for a bulky
substituent at C5 of the indole moiety was to reduce affinity at dopamine D receptor while
2
keeping the affinity at the serotonin 5-HT2A receptor. From the Glide scoring function it
seemed that the effect of unfavorable interactions of alkoxy groups with aqueous environment
of the extracellular receptor vestibule is observed to the lesser extent at the studied serotonin
receptors. It was only partially confirmed by in vitro tests as 5-HT receptor affinity was also
2
A
significantly diminished for compounds with bulky substituents at C5 of the indole moiety.
Regarding arylmethyl substituents on the nitrogen atom, the binding pockets of the studied
receptors can accommodate all the substituents studied so the simplification of the substituent
at this position in comparison to the virtual hit D2AAK1 has minor effect on the ligand
affinities at the studied receptors.

ACCEPTED MANUSCRIPT
Fig. 8. Compounds 5 (A, B), 9 (C, D) and 17 (E, F) in the binding pocket of human serotonin
5
-HT2A receptor. A, C, E – 3D view of the binding site. Ligands represented as sticks with
magenta carbon atoms. Protein represented as wire with grey carbon atoms, main interacting
residues shown as sticks. Hydrogen bonds shown as red dashed lines. Non-polar hydrogen
atoms not shown for clarity. B, D F– 2D view of the binding site.

ACCEPTED MANUSCRIPT
2
.7. Behavioral studies
All compounds studied are characterized by favorable lipophilicity values and are able
to cross blood-brain barrier (see Table S1 in Supplementary Information). Compound 5 was
selected for the behavioral studies due to most desired multi-receptor profile.
2
.7.1. Motor coordination and chimney tests
Compound 5 at the dose investigated (50 mg/kg) did not induce coordination
impairments as evaluated in the chimney test (Fig. 9).
Fig. 9. Influence of the compound 5 on motor coordination in mice evaluated in chimney test.
Investigated compound was injected i.p. 60 min before the test. Data are expressed as mean ±
SEM values. The t-test did not show significant changes in the time in chimney (p = 0.2253).
2
.7.2. Amphetamine-induced hyperactivity in mice
The effect of 5 on the spontaneous locomotor activity in mice is presented in Fig. 10.
Statistical analysis revealed significant changes in the locomotor activity after 5 treatment in
combination with amphetamine (two-way ANOVA: pretreatment [F(1,28)=8.1, p<0.01] and
interaction effect between pretreatment and treatment [F(1,28)=7.59, p<0.05]). The
administration of amphetamine increased the locomotor activity of mice vs. saline-treated
group (p<0.01). Moreover, the post hoc Bonferroni test showed that 5 (50 mg/kg) co-

ACCEPTED MANUSCRIPT
administered with amphetamine (5 mg/kg) decreased amphetamine-induced hyperactivity
when compared to the amphetamine-treated group (p<0.05) (Fig. 10).
Fig. 10. Influence of 5 on the amphetamine-induced hyperactivity in mice. Appropriate
groups of mice received 5 [50 mg/kg, (n = 8); i.p.], amphetamine [5 mg/kg, (n = 8); s.c.], 50
mg/kg of compound 5 and amphetamine (n = 8), and vehicle (n = 8, indicated as 0). Data are
presented as mean ± SEM of the distance travelled by mice during 30 min. The results from
the Bonferroni test analyses indicate: ** p < 0.01 vs. the vehicle-treated group and # p < 0.05
vs. amphetamine-treated group.
2
.7.3. Effect of acute administration of
the forced swim test (FST) in mice.
5 (50 mg/kg) on the total duration of immobility in
The influence of 5 (50 mg/kg) on the total duration of immobility was determined in
the forced swim test, FST (Fig. 11). The t-test revealed the statistically significant decrease in
immobility time (p<0.001, Fig. 11A) as well as increase in latency time to the first episode of
immobility (p<0.01) after acute administration of 5 (50 mg/kg; Fig. 11B).

ACCEPTED MANUSCRIPT
Fig. 11. The influence of the investigated compound 5 (50 mg/kg) on the total duration of
immobility in the forced swim test in mice. The investigated compound was administered i.p.
6
0 min before the test. Data are expressed as mean ± SEM values. The t-test showed
significant decrease in the immobility time (p<0.001) as well as increase in latency time to the
first episode of immobility (p<0.01) after acute administration of 5 (50 mg/kg).

ACCEPTED MANUSCRIPT
2
.7.4. Effect of acute administration of
5
on memory consolidation in mice
The influence of 5 (50 mg/kg) on memory consolidation (Fig. 12) was determined
during the retention trial of the passive avoidance (PA) task. T-test indicated that the acute
administration of studied 5 increased the IL values (p<0.01) indicating its pro-cognitive
effect.
Fig. 12. Compound 5 effects on memory consolidation in mice after an acute administration.
Appropriate groups of mice received injections of 5 [50 mg/kg (n = 10) and vehicle (n = 10);
i.p.] on Day 1 immediately after the test and the rodents were retested 24 h later (i.e., on Day
2
). Data are presented as mean ± SEM. The results from the t-test analyses indicate: * p <
0
.01 vs. the vehicle-treated group.
3
. Discussion and conclusions
Schizophrenia is a mental illness of not fully understood pathomechanisms, involving
many neurotransmitter systems [4]. This is why selective, one-target drugs are less efficient in
the treatment of this disease than second-generation or atypical drugs acting on a number of
aminergic GPCRs. Indeed, clozapine which is a gold standard for the treatment-resistant
schizophrenia has a nanomolar affinity to several GPCRs [25]. This is a paradigm change
from dirty drugs (not beneficial due to possible side effects) to modern multi-target drugs,
which are efficient in the treatment of diseases with complex pathomechanism. In accordance
to the complex pathomechanism of schizophrenia, the compounds reported here were

ACCEPTED MANUSCRIPT
designed to have affinity for the dopamine D and serotonin 5-HT and 5-HT2A receptors,
2
1A
targets of great importance in schizophrenia. In particular compounds 5, 9 and 17 display a
multi-target profile with nanomolar affinity to all the receptors studied and they can be
considered as the post promising compounds in the series.
All antipsychotics currently present at the market are antagonists or partial agonists of
the dopamine D receptor. In agreement with this, compounds 5, 9 and 17 are antagonists of
2
the dopamine D receptor, a mechanism of action well validated as beneficial for the
2
treatment of schizophrenia. Blocking dopamine D receptors enables to treat the positive
2
symptoms of schizophrenia that result from overactivity of D receptors in the mesolimbic
2
dopaminergic pathway (the neuronal projection from the ventral tegmental area (VTA) to the
nucleus accumbens, amygdala and hippocampus) [4]. There were attempts to design
antipsychotics beyond the dopaminergic hypothesis, however none of these compounds
entered the market [1].
Compounds 5, 9 and 17 are antagonists of the serotonin 5-HT2A receptor, a
mechanistic component that has been related with a better clinical performance of atypical
antipsychotics at managing negative symptoms of schizophrenia while still being able to
control positive symptoms [26]. Indeed, atypical antipsychotics combine high-affinity
antagonism of 5-HT2A receptors with D antagonist activity, and in the clinic these drugs
2
resulted superior than typical antipsychotics particularly concerning extrapyramidal side-
effects [27], the most serious adverse effects of first generation drugs. The functional data for
compounds 5, 9 and 17 and 5-HT receptor are less clear, although they indicate agonism or
1
A
partial agonism at this receptor, which would be also beneficial for schizophrenia treatment
28]. Detailed discussion of the involvement of dopamine and serotonin receptors in the
pathomechanism of schizophrenia is available in our review papers [1,4].
[

ACCEPTED MANUSCRIPT
Compounds 1-20 were obtained from the reaction of 5-substituted indoles and N-
methylarylpperidin-4-ones which is a most convenient method to obtain 3-(1-arylmethyl-
1
,2,3,6-tetrahydropyridin-4-yl)-1H-indole system [10]. Other approaches to obtain
compounds with this skeleton involve the reaction of 1H-indolo-2,3-dione with 1-substituted
derivatives of piperidin-4-one [29], from the reaction of 3-(piperidin-4-yl)-1H-indole with
substituted benzyl halides [30] or from the reaction of methyl ester of 2-propenoic acid with
substituted benzylamines and indole [31]. Our approach is a simple and convenient one-step
procedure and allowed us to obtain high-purity compounds.
One of the most potent compounds, compound 9, was selected for X-ray studies to get
knowledge about its energetically stable conformation in the solid state. The interatomic
distances and angles for this compound are in agreement with those described in the literature
[
12] and are similar to those observed for the other closely related indole derivatives [13–16].
However, the conformation of compound 9 differs from the previously reported conformation
of compound 11 (D2AAK1_3), see Fig. 3 [13].
The structure-activity relationship is clear in the reported series of compounds. The
bulkiness of 5-alkoxy substituent is not favorable for activity while the effect of N-methyl
aryl substituent is less important. The affinities of compounds 1-20 to the studied receptors
are comparable to previously reported N-alkyl-3-(1-arylmethyl-1,2,3,6-tetrahydropyridin-4-
yl)-1H-indoles [10,32].
Molecular docking simulations enabled to study ligand-receptor interactions at the
molecular level. The docking poses of compounds 1-20 in the orthosteric pockets of
dopamine D and serotonin 5-HT and 5-HT2A receptors are comparable and correspond to
2
1A
the previously reported bonding mode of the virtual hit D2AAK1 [8,9,13] and compound 11
D2AAK1_3) [13].
(

ACCEPTED MANUSCRIPT
Compound 5 with most favorable multi-receptor profile was subjected to in vivo
investigations. Prior to those the lipophilicity and blood-brain barrier permeation were
calculated for all the compounds (see Supplementary Information, Table S1) and all the
compounds were found to cross blood-brain barrier and to have beneficial lipophilicity values.
Compound 5 decreased amphetamine-induced hyperactivity in mice which confirms its
antipsychotic potential. Moreover, compound 5 increased memory consolidation (pro-
cognitive activity) in the passive avoidance test and decreased immobility time in the forced
swim test predictive of antidepressant activity. These additional pharmacological activities of
compound 5 are extremely relevant due to persistent negative and cognitive symptoms of
schizophrenia, which are often retained even between acute schizophrenia episodes. Most
probably, antidepressant and pro-cognitive activity of compound 5 is connected with its
affinity for serotonin receptors.
In summary, the reported series of multi-target compounds contributes to elaboration
of a better treatment for schizophrenia. Effort will be now made to obtain compounds with
extended atypical multi-receptor profile and with diminished affinity to off-targets.
4
. Experimental section
4
.1. Chemistry
All reagents used for synthesis were purchased from commercial suppliers and were
used without further purification. NMR spectra were recorded on a Bruker AVANCE III 600
MHz spectrometer equipped with a BBO Z-gradient probe. Spectra were recorded at 25 °C
using DMSO-d6 and CDCl as a solvent with a non-spinning sample in 5 mm NMR-tubes.
3
Chemical shifts were expressed in parts per million (ppm) using the solvent signal or TMS as
an internal standard. High resolution mass spectra (HRMS) were acquired on a Bruker
microTOF-Q II mass spectrometer with electrospray ionization (ESI). Data were processed