Discovery of New Heterocyclic Ring Systems as Novel and Potent V1AReceptor Antagonists

Article abstract of DOI:10.1021/acschemneuro.0c00486

Autism spectrum disorder is a neurodevelopmental disease with increasing occurrence. Recent studies focus on the development of novel V1A receptor antagonists which can influence the core symptoms of autism through the AVP pathway. In this study, we describe the synthesis of new heterocyclic ring systems. These are a novel class of brain-penetrating V1A antagonists with improved metabolic stability and in vivo potency. The efficacy of the compounds was strongly influenced by the position of the chlorine atom, suggesting halogen bond formation between the ligands and the V1A receptor.

Full text of DOI:10.1021/acschemneuro.0c00486

pubs.acs.org/chemneuro
Letter
Discovery of New Heterocyclic Ring Systems as Novel and Potent
V1A Receptor Antagonists
́
́
́
́
́
́
́
́
́
́
́
́
Ferenc Baska,* Gabor Szanto, Eva Bozo, Zoltan Szakacs, Miklos Dekany, Krisztina Szondine Kordas,
́
́
́
́
́
́
́
́
́
Katalin Domany-Kovacs, Dalma Kurko, Barbara Hodoscsek, Ildiko Magdo, Balazs Kramos, Zoltan Beni,
́
Monika Vastag, and Imre Bata
Cite This: https://dx.doi.org/10.1021/acschemneuro.0c00486
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Autism spectrum disorder is a neurodevelopmental disease with
increasing occurrence. Recent studies focus on the development of novel V1A
receptor antagonists which can influence the core symptoms of autism through
the AVP pathway. In this study, we describe the synthesis of new heterocyclic
ring systems. These are a novel class of brain-penetrating V1A antagonists with
improved metabolic stability and in vivo potency. The efficacy of the compounds
was strongly influenced by the position of the chlorine atom, suggesting halogen
bond formation between the ligands and the V1A receptor.
KEYWORDS: V1A receptor, CNS, ASD, imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]diazepine,
imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]diazocine, halogen bonding
when Roche demonstrated under clinical circumstances that
he arginine-vasopressin (AVP) hormone plays a key role
Tin the regulation of various physiological processes the inhibition of the V1A receptor may improve social
communication skills in adults with high-functioning
ASD.8,10 Their frontrunner compound balovaptan (2) showed
significant efficacy on the Vineland-II scale in ASD;11
furthermore, Roche initiated a phase 2 study in children and
adolescents in 2016. Up to June 2020, the molecule was being
tested in Phase 2, then the study was terminated due to
ineffectiveness in the studied population.12
We have previously identified a novel, urea-containing
compound family as V1A antagonists, with compound 3 as the
best representative of the series.13 Although it showed
remarkable potency at the V1A receptor, poor metabolic
stability limited its progression. Our recently published
tricyclic molecules showed much better physicochemical
properties, and compound 4 was also active in the in vivo
pharmacological tests14 (Figure 2, Table 1). Our initial SAR
studies revealed that compounds with a seven-membered
saturated azepine ring possess the best efficacy; therefore, we
including water reabsorption, hormone secretion, cardiovas-
cular modulation, and social behavior. AVP targets three
different receptors (V1A, V1B, V2) in the body, and a wide
variety of studies explored the role of these receptors in certain
diseases of the central nervous system (CNS).1 The V1A
receptor came into focus of CNS research when an association
between the receptor gene (AVPR1a) polymorphism and
autism was discovered.2,3 Autism is a spectrum disorder
(ASD), and it is considered to be one of the most common
neurodevelopmental diseases nowadays, affecting 1 in every 68
children in the United States.4 The increasing incidence of
ASD5 and the estimated cost of a lifetime treatment6
emphasizes the demand for research and development of
new therapies. So far, pharmacological therapy for ASD has not
focused on the core symptoms but mainly on the associated
symptoms of the disease. Recent studies revealed that V1A
receptor antagonists can regulate the effect of the AVP
hormone and thus influence the core symptoms of autism.7,8
The first selective V1A receptor antagonist with viable
metabolic stability and CNS penetration was reported by the
researchers of Pfizer.9 PF-184563 (1) was originally developed
for the treatment of dysmenorrhea; despite its published
benefits, however, the compound did not reach the clinical
phase. It was a milestone in the research of V1A antagonists
Received: July 29, 2020
Accepted: October 13, 2020
https://dx.doi.org/10.1021/acschemneuro.0c00486
© XXXX American Chemical Society
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX
A

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
10−6 cm/s, PDR = 0.8) according to the VB-Caco-2 assay
results.
To evaluate the size of the middle ring we have prepared a
six- and an eight-membered derivative bearing the same
substituents (R1 and R2) as their corresponding imidazo[1,2-a]
[1,2,4]triazolo[4,3-c][1,3]diazepine analogue (Figure 3).
These tricycles (imidazo[1,2-a][1,2,4]triazolo[4,3-c]-
pyrimidine and imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]-
diazocine) are also novel heterocyclic ring systems and were
primarily evaluated using the same in vitro test as described in
Table 1.
Figure 1. V1A antagonists PF-184563 and balovaptan (RG7314).
Comparing the activities of the corresponding molecules it
was found that the ring- contracted pyrimidine derivative (11)
had moderate binding (Ki = 324 nM) to the human V1A
receptor but surprisingly the ring-enlarged diazocine analogue
(12) displayed significantly better results (Ki = 7 nM; IC50
=
25 nM). Although, the metabolic stability slightly decreased
with the higher clogP value (3.2), it did not cause considerable
changes in the permeability properties of the compound
(Table 2).
Subsequently, replacement of the imidazole ring with a
thiophene resulted in the thieno[2,3-f ][1,2,4]triazolo[4,3-
a]azepine derivatives (13−15) which are also novel tricycles
in the literature (Figure 4). Interestingly, the unsubstituted
variant (13) was found to be the most potent molecule in our
small fused ring library (Ki = 4.2 nM, IC50 = 23 nM).
By comparing compounds 9 and 13, it can be seen that the
imidazole-thiophene ring exchange resulted in more than 150-
fold improvement in affinity (Table 1 and 2). However, the
higher clogP (4.1) of compound 13 led to decreased metabolic
stability, limiting our further options.
Additionally, we also tried to prepare the chlorine containing
analogue, but during the synthesis we obtained two products.
After separation and analytical validation, we found that the
chlorine atom was introduced not only into the thiophene ring
(14) but into the oxidized azepine ring (15). As shown in
Table 2, while the previously synthesized imidazo-diazepine
and imidazo-diazocine derivatives had a much better effect
with chlorine on the heterocycle, the additional chlorine in the
Figure 2. V1A antagonists from Gedeon Richter.
focused on the anellated phenyl ring in the tricycle. No studies
have been published yet in the literature before of the
substitution of this phenyl ring and its V1A antagonist effect.
Formerly published medicinal chemistry publications stated
that, in some cases, the phenyl ring can be replaced by
heteroaromatic isostere rings without the loss of activity.15,16
Our aim was to increase the poor metabolic stability of
compound 4 in rodents by reducing its high clogP (4.8) value
and finding the best balance of potency and stability.
Our starting point was compound 5 (Figure 2), which is an
analogue of PF-184563 (Figure 1) and exhibited strong human
V
1A binding and antagonist activity with moderate human and
poor rodent metabolic stability.14 In order to lower the clogP,
the phenyl ring was replaced with an imidazole as shown in
Table 1. The imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]-
diazepine scaffold was prepared via a multiple-step synthesis
starting from 2-nitro-1H-imidazole (16, Scheme 1). The first
compound (6, Table 1) from this family demonstrated
moderate binding to the V1A receptor and interestingly, the
elimination of the chlorine from the imidazole ring (7; R2 = H)
resulted a completely inactive derivative.
thiophene ring decreased the antagonist activity (14, IC50
=
147 nM) of the compound. Notably, derivative (15) gave the
best result in the Ca2+ assay in these novel ring systems.
Based on these experimental results, a molecular modeling
study was initiated aiming to explain the differences between
the obtained binding affinities and to gain at least a qualitative
picture of the ligand−receptor complex. First, the lipophilic
ligand efficiency (LLE, determined as pKi -−clogP) values were
analyzed to clarify the role of the halogen atom. In general, the
binding affinity of ligands forming the same interaction pattern
in the binding site is increasing in line with the lipophilicity
while the LLE value stays quasi constant.17 Through the
comparison of LLE values one might identify structural motifs
where higher binding enthalpy, and thus stronger ligand-
protein interactions could be presumed.
In our case the lowest LLE values (Figure 5) were obtained
mostly for compounds (7 (inactive), 9, 13 and 15) bearing no
aromatic chlorine atom. This suggests that the aromatic
halogen atom plays an important role in the ligand−receptor
binding mechanism. However, compounds 11 and 14 which
containan aromatic chlorine atom showed similarly low LLE
values as well. Without structural models this could not be
satisfactorily explained.
Moving forward, additional compounds were synthesized
using our previous experience in the field. The replacement of
the pyridinyl-piperidine side chain (R1 group, Table 1) of
compound 6 to an oxazolyl-cyclohexyl group (compound 8)
provided a more prominent binding value (Ki = 50 nM) with
an antagonist effect (IC50 = 2.1 μM) in the micromolar range.
As we assumed, the moderate drop in the clogP value (2.6)
resulted in an excellent microsomal stability in all three species
(Table 1).
After changing the oxazolyl-cyclohexyl group to the side
chain of balovaptan (R1 = pyridinyloxy-cyclohexyl), the
chlorine-free compound (9, Table 1) displayed moderate
binding which could be further increased with the incorpo-
ration of the chlorine to the appropriate position. Compound
10 possessed not only improved binding (Ki = 14 nM) but
significant antagonist activity (IC50 = 178 nM). It was also
observed that the relatively low clogP (2.8) value of the
molecule resulted in favorable metabolic stability in all species
and it had excellent permeability (VB-Caco-2 PappA‑B = 41 ×
B
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
Table 1. In Vitro Pharmacological Characterization of the Reference Compounds and Substituted Imidazo[1,2-
a][1,2,4]triazolo[4,3-c][1,3]diazepines
a
Calculator Plugins were used for structure property prediction and calculation, Marvin 18.24.0, 2018, ChemAxon (http://www.chemaxon.com).
b
Kinetic solubility was determined using 1*, 5#, and 10 mM DMSO stock solutions (buffer: PBS pH= 7.4, incubation time: 2 h, RT, instrument:
c
3
HPLC-MS). Values are the mean of three independent measurements with SD of 5%. Displacement of H-labeled arginine vasopressin from
d
human vasopressin V1A receptor. In vitro human V1A IC50 determined in a Ca2+ mobilization assay using a fluorescent plate reader. IC50 values are
e
the mean of at least three independent measurements with SD of 20%. Metabolic stability performed in human, rat and mouse microsomes in
the presence of NADPH at 37 °C. Permeability in VB-Caco-2 cell assay and efflux or permeability directional ratio (PDR).
f
Figure 3. Changing the size of the middle ring in the imidazo[1,2-a][1,2,4]triazolo[4,3-c] ring system.
In the next step a common pharmacophore model (Figure
6) was built based on the known reference compounds. The
obtained model contained five pharmacophore sites: an
acceptor (A), two aromatic (R) and two hydrophobic (H)
sites where H2 is a halogen atom. The alignment of
compounds 3−15 with the model (Figure 7) showed that
the orientation of the aromatic chlorine is somewhat different
in 11 and 14 from that in the other aromatic chlorine
containing ligands. These deviations could explain the lower
LLE values calculated for these compounds and corroborate
the idea that the aromatic chlorine plays an important role in
the interactions with the V1A binding site.
In order to better understand the role of this halogen
interaction and to gain a qualitative picture of the ligand
receptor complex, a docking study was undertaken. In the
absence of published V1A receptor−ligand crystal structure,
homology models available in the GPCR database were
utilized for this study. Three structural models, representing
C
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
Table 2. In Vitro Pharmacological Results of Additional New Heterocyclic Ring Systems
a
b
c
d
e
f
Ex
clogP
Solubility (μM)
hV1A (Ki, nM)
hV1A (IC50, nM)
CLint (h/r/m), μL/min/mg
VB-Caco-2 PappA‑B × 10−6 cm·s−1/PDR
11
12
13
14
15
2.5
3.2
4.1
4.9
4.1
98.8
97.0
100
91.1
85.0
324
7
4.2
11.2
5.7
25
23
147
15
10/29/87
60/1.0
49/0.8
13/225/302
10/168/128
16/243/128
a
Calculator Plugins were used for structure property prediction and calculation, Marvin 18.24.0, 2018, ChemAxon (http://www.chemaxon.com).
Kinetic solubility was determined using 10 mM DMSO stock solution from each sample (buffer: PBS pH= 7.4, incubation time: 2 h, RT,
b
c
3
instrument: HPLC-MS). Values are the mean of three independent measurements with a SD ≤ 5%. Displacement of H-labeled arginine
d
vasopressin from human vasopressin V1A receptor. In vitro human V1A IC50 determined in a Ca2+ mobilization assay using a fluorescent plate
e
reader. IC50 values are the mean of at least three independent measurements with a SD ≤ 20%. Metabolic stability performed in human, rat and
mouse microsomes in the presence of NADPH at 37 °C. Permeability in VB-Caco-2 cell assay and efflux or permeability directional ratio (PDR).
f
Figure 4. Thieno[2,3-f ][1,2,4]triazolo[4,3-a]azepine derivatives.
explain the collected SAR data. Similar poses were found in all
three models, but in the case of the active model, this was
predicted as the most favorable one (having the lowest docking
score). According to the model, the receptor’s Trp-204, Phe-
207, and Tyr-214 as well as Trp-304, Phe-307, and Phe-308 on
the “other side” form two apolar pockets that are ideally suited
to accommodate the pyrimidine and imidazole rings (the R1
and R2 aromatic groups in the pharmacophore model) of 10.
A hydrogen bond is formed between Gln-131 and the triazolo
nitrogen (the acceptor site (A) of the common pharmaco-
phore). In addition, the diazepine ring of 10 occupies a place
in a large apolar pocket that has enough room to “grow” the
ligand further by apolar groups (as it is in 3). Finally, this
model could explain the role of the aromatic chlorine atom as
well. The chlorine is close to two amide oxygens of the protein
backbone and it is ideally placed to form a halogen bond either
with the Thr-333 or with the Cys-303 residues.
Being able to identify new heterocycles (compound 10 and
12) with favorable physicochemical properties, remarkable V1A
potency and improved metabolic stability we submitted them
for additional profiling. Neither of the compounds showed
affinity or inhibition on the human V2 receptor and hERG ion
Figure 5. LLE diagram of compounds 1−15.
channel (Table 3).
Inspired by these results, we evaluated compounds 10 and
12 in oxytocin induced scratching tests in mice. It is well-
known that oxytocin (OXT) induces scratching behavior
through central V1A receptors.19 Using this target specific
response, V1A antagonist in vivo effect of the compounds could
be determined.
Compound 10 showed significant scratching inhibition after
intraperitoneal (i.p.) treatment (Table 3); however, it was not
active after oral (p.o.) administration despite its acceptable
brain penetration (Table 3.) at 1 h after 30 mg/kg dosing. In
the case of compound 12, this could be explained with the
poor metabolic stability in mice, however compound 4 was
orally active as well, despite its relatively poor metabolic
stability in rodents.14 We assumed a difference between the
human and rodent V1A binding affinity therefore we tested the
compound on the mouse V1A receptor (mV1A). The results
different functional states of the receptor, built on different
starting templates (see Experimental for details) were found in
the database. Since no published data were available for the
V1A receptor, a binding site (Figure 8) described for other
GPCRs18 was used in the docking calculations. This pocket is
close to the extracellular site and thus to the flexible loops. In
homology models, the reliability of the sequence alignment’s
result for flexible parts is always questionable. Considering
these, even if the investigated ligands were antagonists, induced
fit docking calculations for compound 10 were carried out
using all three homology models. Ten to twenty binding poses
were obtained for different models. Taking the number of
specific ligand−protein interactions, the ligand conformation
and the docking scores into account, after visual inspection the
binding pose shown in Figure 8 was selected that could nicely
D
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
Figure 6. Common pharmacophore generated using the reference compounds . Scripps ref.: WO2014/127350 Ex. 27.
Figure 7. Representative examples of Gedeon Richter’s compounds in the common pharmacophore model. Left: 4 (green), 11 (pink), and 14
(cyan). Right: 5 (orange), 10 (dark blue), and 12 (purple).
imidazole ring to other heterocycles. Key compounds from these
series were synthesized as shown in Schemes 1−5.
showed significantly lower potency (Table 3) for mV1A (Ki =
1300 nM) and that could be the reason for the much weaker
effect.
The 6,7-dihydro-5H-imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]-
diazepine derivatives (6−10) were prepared according to the
synthetic route outlined in Schemes 1 and 2. Beginning with the
commercially available 2-nitro-1H-imidazole (16) and ethyl 4-
bromobutanoate, the desired N-substituted ethyl 4-(2-nitro-1H-
imidazol-1-yl)butanoate (17) was prepared in the presence of
K2CO3 in DMF. The reduction of the nitro compound was carried
out by catalytic hydrogenation at room temperature (RT) utilizing
10% palladium on carbon. The formed amino compound (18) was
then reacted with potassium tert-butoxide to obtain the corresponding
19 cyclic amide as key intermediate.
In summary, several of the prepared novel heterocyclic ring
systems turned out to be potent V1A receptor antagonists with
appropriate substituents. Compound 10 was showed to be
effective in mice in vivo and exhibited improved selectivity
against the human V2 receptor and hERG ion channel, as well.
Optimization of the core ring revealed that the chlorine atom
has a much higher influence on the potency than the other
substituents. Therefore, we assume a halogen bond between
the receptor and the ligands. Pharmacophore modeling and
docking studies confirmed this proposition. Based on these
results, we continued the research of selective, brain penetrant
The appropriate chlorinated and nonchlorinated derivatives were
obtained then by two different routes shown in Scheme 2.
Chlorination of the unsubstituted bicycle with N-chlorosuccinimide
(NCS) resulted in the desired isomer (3-chloro-6,7-dihydro-5H-
imidazo[1,2-a][1,3]diazepin-8(9H)-one; 20) in 50% yield and no
further optimization was necessary.
After the successful chlorination step, Lawesson’s reagent was used
to prepare the thioamide for the triazole synthesis. Two methods were
applied for the reaction: both were successful, but when the reaction
was carried out in THF at room temperature, it resulted in much
better yields. In the final step, the prepared thioamides (21 or 22)
V
1A antagonists and further details will be reported in due time.
METHODS
■
PF-184563 and balovaptan were synthesized based on the original
patents.20 Chemical methods for compound 3 and compounds 4 and
5 were published recently.13,14 In order to determine some structure−
activity relationships, the following moieties were replaced: the side
chain of the triazole (R1), the size of the saturated ring and the
E
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
Figure 8. Binding pose of compound 10 in the active V1A homology model obtained from induced-fit docking calculation. The 3 Å environment of
the ligand is enlarged (right) showing the observed receptor−ligand interactions.
Table 3. Summarized In Vitro and In Vivo Data For 10, 12, 2 (Balovaptan), and 4
parameter
10
12
2
4
hV1A Ki (nM)
hV1A IC50 (nM)
mV1A Ki (nM)
14
178
1300
7
25
0.7
3
52
0.8
1.0
hV2 binding Ki (nM)
hERG IC50 (μM)
CLint (h/r/m) (μL/min/mg)
VB-Caco-2 PappA‑B × 10−6
cm·s−1/PDR
brain to plasma ratio (p.o.) at 30 mg/kg
OT-induced scratching inhibition
at 30 mg/kg
0% at 1 μM
17% at 10 μM
1/12/15
1% at 1 μM
>30
10/29/87
1200 nM
21 μM
0.5/22/17
560 nM
7/124/94
29/1.3
41/0.8
0.54
(i.p.)
98%
60/1.0
43/1.1
1.01
(p.o.)
80%
(i.p.)
41%
36%
(p.o.)
94%
42%
at 10 mg/kg
32%
69%
a
Scheme 1. Synthesis of Compound 19
a
Reagents and conditions: (a) ethyl 4-bromobutanoate, K2CO3, DMF, 100 °C, 5 h, 86%; (b) Pd/C, H2, EtOH, RT, 1 h, 91%; (c) KOtBu, THF, 4
°C to RT, 3 h, 60%.
were suspended in xylene, and after addition of appropriate
hydrazides the reaction mixtures were refluxed for 24 h to yield the
desired title compounds (6−10).
formation led to the desired product (11) with similar results
compared to those of shown in Scheme 2.
To produce the imidazo-diazocine scaffold (Scheme 4), the
synthesis route had to be modified again, because the ring closure
with potassium tert-butoxide was not suitable in this case. Under these
conditions, the reaction mixture contained only traces of the bicycle
and mainly degradation products were identified. Before the reduction
step, ester hydrolysis was carried out (step b) and the cyclic amide
(30) could be prepared with intramolecular ring closure using HBTU
as coupling reagent. An analogous approach to the previously
described methods (Schemes 2 and 3) could be used for the
chlorination and triazole formation steps, leading to the final product
(12).
In the case of imidazo-pyrimidine derivative (Scheme 3) the
previously applied alkylation (shown in Scheme 1) of 2-nitro-1H-
imidazole (16) with ethyl 3-bromopropionate resulted in poor yield
(∼6%), therefore Michael addition with methyl prop-2-enoate was
utilized to obtain the corresponding alkyl derivative. Although the
reaction time was much longer, the methyl 3-(2-nitro-1H-imidazol-1-
yl)propanoate (23) was formed in acceptable yield (31%). Ring
closure has already occurred during the reduction step and the bicycle
(24) was isolated by flash chromatography. Chlorination and triazole
F
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
a
Scheme 2. Synthesis of 6,7-Dihydro-5H-imidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]diazepine Analogues (6−10)
a
Reagents and conditions: (a) NCS, DMF, RT, 24 h, 50%; (b) Lawesson’s r., pyridine, reflux, 3 h, 35−72% or Lawesson’s r., THF, RT, 24 h, 76−
100%; (c) appropriate hydrazide, xylene, reflux, 24 h, 24−54%
Scheme 3. Synthesis of the 3-chloro-5,6-
Scheme 5. Synthesis of 5,6-Dihydro-4H-thieno[2,3-
f ][1,2,4]triazolo[4,3-a]azepine Derivatives
a
dihydroimidazo[1,2-a][1,2,4]triazolo[4,3-c]pyrimidine
a
derivative (11)
a
a
Reagents and conditions: (a) NCS, DMF, RT, 1.5 h, 40% (2
Reagents and conditions: (a) methyl prop-2-enoate, DIPEA, DMF,
products formed); (b) Me3O−BF4, K2CO3, DCM, RT, 3 h; (c)
hydrazide, TFA, MeCN, 75 °C, 18 h, 13 + 14% (2 products formed);
(d) hydrazide, TFA, 1,4-dioxane, 75 °C, overnight, 79%.
RT, 7 days, 31%; (b) Pt/C, H2, EtOH, RT, 3 h, 33%. (c) NCS, DMF,
RT, 48 h, 23% (d) Lawesson’s reagent, pyridine, reflux, 3 h, 51% (e)
hydrazide, xylene, reflux, 24 h, 53%.
a
Scheme 4. Synthesis of the 3-Chloro-5,6,7,8-tetrahydroimidazo[1,2-a][1,2,4]triazolo[4,3-c][1,3]diazocine Derivative
a
Reagents and conditions: (a) ethyl 5-bromopentanoate, K2CO3, DMF, 100 °C, 5 h, 70%; (b) lithium hydroxide hydrate, THF, EtOH, water, RT,
overnight, 70%; (c) Pd/C, H2, EtOH, RT, 3 h, 99%; (d) HBTU, TEA, DMF, RT, 48 h, 35%; (e) NCS, DMF, RT, 24 h, 46%; (f) Lawesson’s
reagent, pyridine, reflux, 3 h, 84%; (g) hydrazide, xylene, reflux, 24 h, 51%.
G
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
Synthesis of the bicycle containing anellated thiophene ring was
started with the commercially available 4,6,7,8-tetrahydro-5H-thieno-
[3,2-b]azepin-5-one (33, Scheme 5). The novel heterocyclic ring
system was prepared in two steps: the amide oxygen was selectively
methylated with trimethyloxonium tetrafluoroborate followed by a
solvent exchange. The active imino ether was then reacted with the
appropriate hydrazide to provide the corresponding thieno[2,3-
f][1,2,4]triazolo[4,3-a]azepine derivative (13). Similar to what was
described formerly the chlorination step was carried out with NCS.
Surprisingly, a chlorinated byproduct (35) also formed and both
reaction products had almost the same retention time. According to
the NMR measurements, the ratio was 2:1 for the desired product,
but the separation at this level was not possible. Therefore, we
continued the synthesis with the mixture and after the final ring
closure the two products (14 and 15) could be separated by flash
chromatography (Scheme 5).
were responsible for in vitro and in vivo pharmacology. I.M.,
B.K., and Z.B. accomplished the molecular modeling, and M.V.
was responsible for the ADME experiments. I.B. designed the
compounds, supervised the project and prepared the paper
with F.B.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
́
́
́
We thank Judit Mu
̈
ller, Anna Kancz, Beata Lizak, Katalin
́
́
́
Krisztina Futo, and Janos Koti for their contribution and
́
́
analytical support on this work. We also thank Reka Mohacsi
́
́
for the VB-Caco-2 assay, and Marta Than for the hERG data.
REFERENCES
■
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acschemneuro.0c00486.
■
(1) Koshimizu, T., Nakamura, K., Egashira, N., Hiroyama, M.,
Nonoguchi, H., and Tanoue, A. (2012) Vasopressin V1a and V1b
receptors: from molecules to physiological systems. Physiol. Rev. 92,
1813−1864.
sı
(2) Yirmiya, N., Rosenberg, C., Levi, S., Salomon, S., Shulman, C.,
Nemanov, L., Dina, C., and Ebstein, R. P. (2006) Association between
the arginine vasopressin 1a receptor (AVPR1a) gene and autism in a
family-based study: mediation by socialization skills. Mol. Psychiatry
11, 488−494.
Synthetic procedures, compound characterization, mo-
lecular modeling and assay description (hV1A binding
and functional, hV2 binding, in vivo pharmacology,
microsomal stability and VB-Caco-2 permeability assay)
(PDF)
(3) Wassink, T. H., Piven, J., Vieland, V. J., Pietila, J., Goedken, R. J.,
Folstein, S. E., and Sheffield, V. C. (2004) Examination of AVPR1a as
an autism susceptibility gene. Mol. Psychiatry 9, 968−972.
(4) Christensen, D., Bilder, D. A., Zahorodny, W., Pettygrove, S.,
Durkin, M., Fitzgerald, R., Rice, C., Kurzius-Spencer, M., Baio, J., and
Yeargin-Allsopp, M. (2016) Prevalence and Characteristics of Autism
Spectrum Disorder Among 4-Year-Old Children in the Autism and
Developmental Disabilities Monitoring Network. J. Dev Behav Pediatr.
37, 1−8.
AUTHOR INFORMATION
Corresponding Author
Ferenc Baska − Gedeon Richter Plc, Budapest H-1475,
Hungary; orcid.org/0000-0003-0708-2224;
Email: baskaf@richter.hu
■
Authors
(5) McDonald, M. E., and Paul, J. F. (2010) Timing of increased
autistic disorder cumulative incidence. Environ. Sci. Technol. 44,
2112−2118.
́
́
́
Gabor Szanto − Gedeon Richter Plc, Budapest H-1475,
Hungary
́
́
Eva Bozo − Gedeon Richter Plc, Budapest H-1475, Hungary
(6) Buescher, A. V., Cidav, Z., Knapp, M., and Mandell, D. S. (2014)
Costs of autism spectrum disorders in the United Kingdom and the
United States. JAMA Pediatr. 168, 721−728.
́
́
Zoltan Szakacs − Gedeon Richter Plc, Budapest H-1475,
Hungary
́
(7) Slusarz, M. J. (2015) Vasopressin V1A and V1B receptor
́
́
́
Miklos Dekany − Gedeon Richter Plc, Budapest H-1475,
modulators: a patent review (2012 - 2014). Expert Opin. Ther. Pat. 25,
711−722.
Hungary
́
́
Krisztina Szondine Kordas − Gedeon Richter Plc, Budapest H-
(8) Umbricht, D., Rubido, M., Del, V., Hollander, E., McCracken, J.
M., Shic, F., Scahill, L., Noeldeke, J., Boak, L., Khwaja, O., Squassante,
L., Grundschober, C., Kletzl, H., and Fontoura, P. (2017) A single
dose, randomized, controlled proof-of-mechanism study of novel
vasopressin Ia receptor antagonist (RG7713) in high-functioning
adults with Autism Spectrum Disorder. Neuropsychopharmacology 42,
1924.
1475, Hungary
́
́
Katalin Domany-Kovacs − Gedeon Richter Plc, Budapest H-
1475, Hungary
́
Dalma Kurko − Gedeon Richter Plc, Budapest H-1475,
Hungary
Barbara Hodoscsek − Gedeon Richter Plc, Budapest H-1475,
(9) Johnson, P. S., Ryckmans, T., Bryans, J., Beal, D. M., Dack, K. N.,
Feeder, N., Harrison, A., Lewis, M., Mason, H. J., Mills, J., Newman,
J., Pasquinet, C., Rawson, D. J., Roberts, L. R., Russel, R., Spark, D.,
Stobie, A., Underwood, T., Ward, R., and Wheeler, S. (2011)
Discovery of PF-184563, a potent and selective V1A antagonist for
treatment of dysmenorrhea. The influence of compound flexibility on
microsomal stability. Bioorg. Med. Chem. Lett. 21, 5684−5687.
(10) Ratni, H., Rogers-Evans, M., Bissantz, C., Grundschober, C.,
Moreau, J.-L., Schuler, F., Fischer, H., Sanchez, R. A., and Schnider, P.
(2015) Discovery of highly selective brain-penetrant vasopressin 1a
antagonists for potential treatment of Autism via chemogenomic and
scaffold hopping approach. J. Med. Chem. 58, 2275−2289.
(11) Schnider, P., Bissantz, C., Bruns, A., Dolente, C., Goetschi, E.,
Hungary
́
́
Ildiko Magdo − Gedeon Richter Plc, Budapest H-1475,
Hungary
́
́
Balazs Kramos − Gedeon Richter Plc, Budapest H-1475,
Hungary
́
́
Zoltan Beni − Gedeon Richter Plc, Budapest H-1475, Hungary
́
Monika Vastag − Gedeon Richter Plc, Budapest H-1475,
Hungary
Imre Bata − Gedeon Richter Plc, Budapest H-1475, Hungary
Complete contact information is available at:
https://pubs.acs.org/10.1021/acschemneuro.0c00486
Jakob-Roetne, R., Kunnecke, B., Mueggler, T., Muster, W., Parrott, N.,
̈
Author Contributions
F.B., G.Sz., and E.B. designed and/or synthesized the
compounds in this paper. Z.Sz. and M.D performed the
spectroscopic characterization. K.Sz.K., K.D.-K., D.K., and B.H.
Pinard, E., Ratni, H., Risterucci, C., Rogers-Evans, M., von Kienlin,
M., and Grundsschober, C. (2020) Discovery of Balovaptan, a
Vasopressin 1a Receptor Antagonist for the Treatment of Autism
Spectrum Disorder. J. Med. Chem. 63, 1511−1525.
́
H
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience
pubs.acs.org/chemneuro
Letter
(12) Hoffman-La Roche (2016)A Study to Investigate the Efficacy
and Safety of Balovaptan (RO5285119) in Participants With Autism
Spectrum Disorder (ASD). ClinicalTrials.gov. https://clinicaltrials.
gov/ct2/show/NCT02901431 (NCT 02901431).
́
́
́
́
́
́
́
(13) Szanto, G., Mako, A., Baska, F., Bozo, E., Domany-Kovacs, K.,
́
́
́
́
́
Kurko, D., Cselenyak, A., Mohacsi, R., Szondine Kordas, K., and Bata,
I. (2020) New V1A receptor antagonist. Part 1. Synthesis and SAR
development of urea derivatives. Bioorg. Med. Chem. Lett. 30, 127416.
́
́
̈
́
́
́
́
(14) Bozo, E., Baska, F., Lovei, K., Szanto, G., Domany-Kovacs, K.,
́
́
́
Kurko, D., Szondine Kordas, K., Szokoli, T., and Bata, I. (2020) New
V1A receptor antagonist. Part 2. Identification and optimization of
triazolobenzazepines. Bioorg. Med. Chem. Lett. 30, 127417.
(15) Bryan, M. C., Falsey, J. R., Frohn, M., Reichelt, A., Yao, G.,
Bartberger, M. D., Bailis, J. M., Zalameda, L., Miguel, T. S., Doherty,
E. M., and Allen, J. G. (2013) N-substituted azaindoles as potent
inhibitors of Cdc7 kinase. Bioorg. Med. Chem. Lett. 23, 2056−2060.
(16) Kawai, M., Sakurada, I., Morita, A., Iwamuro, Y., Ando, K.,
Omura, H., Sakakibara, S., Masuda, T., Koike, H., Honma, T., Hattori,
K., Takashima, T., Mizuno, K., Mizutani, M., and Kawamura, M.
(2007) Structure-activity relationship study of novel NR2B-selective
antagonists with arylamides to avoid reactive metabolites formation.
Bioorg. Med. Chem. Lett. 17, 5537−5542.
(17) Hopkins, A. L., Keseru, G. M., Leeson, P. D., Rees, D. C., and
̈
Reynolds, C. H. (2014) The role of ligand efficiency metrics in drug
discovery. Nat. Rev. Drug Discovery 13, 105−121.
(18) Latorraca, N. R., Venkatakrishnan, A. J., and Dror, R. O. (2017)
GPCR Dynamics: Structures in Motion. Chem. Rev. 117, 139−155.
́
(19) Schorscher-Petcu, A., Sotocinal, S., Ciura, S., Dupre, A., Ritchie,
J., Sorge, R. E., Crawley, J. N., Hu, S.-B., Nishimori, K., Young, L. J.,
Tribollet, E., Quirion, R., and Mogil, J. S. (2010) Oxytocin-Induced
Analgesia and Scratching Are Mediated by the Vasopressin-1A
Receptor in the Mouse. J. Neurosci. 30, 8274−8284.
(20) WO2004074291 A1 patent for PF-184563 (Example 102) and
WO2010/060836 A1 patent for balovaptan (Example 69).
I
https://dx.doi.org/10.1021/acschemneuro.0c00486
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX