Discovery of Balovaptan, a Vasopressin 1a Receptor Antagonist for the Treatment of Autism Spectrum Disorder

Article abstract of DOI:10.1021/acs.jmedchem.9b01478

We recently reported the discovery of a potent, selective, and brain-penetrant V1a receptor antagonist, which was not suitable for full development. Nevertheless, this compound was found to improve surrogates of social behavior in adults with autism spect

Full text of DOI:10.1021/acs.jmedchem.9b01478

pubs.acs.org/jmc
Drug Annotation
Discovery of Balovaptan, a Vasopressin 1a Receptor Antagonist for
the Treatment of Autism Spectrum Disorder
Patrick Schnider,* Caterina Bissantz, Andreas Bruns, Cosimo Dolente, Erwin Goetschi,
̈
Roland Jakob-Roetne, Basil Kunnecke, Thomas Mueggler, Wolfgang Muster, Neil Parrott,
Emmanuel Pinard, Hasane Ratni, Celine Risterucci, Mark Rogers-Evans, Markus von Kienlin,
and Christophe Grundschober
́
Cite This: https://dx.doi.org/10.1021/acs.jmedchem.9b01478
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: We recently reported the discovery of a potent, selective, and brain-
penetrant V1a receptor antagonist, which was not suitable for full development.
Nevertheless, this compound was found to improve surrogates of social behavior in
adults with autism spectrum disorder in an exploratory proof-of-mechanism study. Here we
describe scaffold hopping that gave rise to triazolobenzodiazepines with improved
pharmacokinetic properties. The key to balancing potency and selectivity while minimizing
P-gp mediated efflux was fine-tuning of hydrogen bond acceptor basicity. Ascertaining a
V1a antagonist specific brain activity pattern by pharmacological magnetic resonance
imaging in the rat played a seminal role in guiding optimization efforts, culminating in the
discovery of balovaptan (RG7314, RO5285119) 1. In a 12-week clinical phase 2 study in
adults with autism spectrum disorder balovaptan demonstrated improvements in Vineland-
II Adaptive Behavior Scales, a secondary end point comprising communication,
socialization, and daily living skills. Balovaptan entered phase 3 clinical development in
August 2018.
men with autism spectrum disorder, demonstrated improve-
ments in some exploratory biomarkers for social and
communication behaviors, including a statistically significant
increase in orienting preference to biological motion as
assessed by eye tracking, a significantly reduced ability to
detect lust, and a trend for reduced ability to detect fear.16
At the time, indole-3-carboxamides with spiro[1H-isobenzo-
furan-3,4′-piperidine] including 2 (Figure 1) stood out
particularly for their unprecedented selectivity against oxytocin
(OT) and vasopressin 2 (V2) receptors. While V2 selectivity is
important to avoid side effects on water retention by the
kidney,1 avoidance of OT receptor antagonism is important
since activation of the receptor is believed to have prosocial
effects.17 However, while 2 has a very good pharmacokinetic
profile after intravenous administration to humans with a half-
life of around 8 h,16 the in vivo total blood clearance in animals
is high, close to liver blood flow in the dog and exceeding liver
blood flow in rat and mouse. It is thus not feasible to achieve
sufficient exposures to establish robust safety margins in
nonclinical studies, rendering this compound unsuitable for full
INTRODUCTION
■
Arginine vasopressin (AVP) is a 9 amino acid peptide
activating G protein coupled receptors. Its functions include
the regulation of water retention through the kidney
vasopressin V2 receptor1 and blood pressure through vascular
smooth muscle V1a receptors.2 In the brain it is mainly
synthesized in the hypothalamic supraoptic and paraventricular
nuclei and acts as a neurotransmitter regulating the HPA axis,3
through the V1b receptor, and social and aggressive behavior
through activation of the V1a receptor in brain areas
implicated in social behavior.4 The V1a receptor brain
expression pattern has been shown to critically affect social
behavior in voles.5
In humans promoter polymorphisms in the V1a gene have
been associated with autism spectrum disorder,6−11 and the
V1a receptor is linked to social behavior. However, the exact
role of vasopressin in autism has not been fully elucidated yet.
Parker et al. showed that V1a receptor activation by intranasal
administration of vasopressin improved social behavior in a
small clinical study with 30 autistic children.12 In contrast,
intranasal administration of vasopressin impaired emotion
recognition13 and increased threat perception in healthy
volunteers.14 Recently we reported the discovery of the highly
potent and selective brain-penetrant hV1a antagonist
RO5028442 (RG7713), 2.15 A proof-of-mechanism study of
this compound, given as a single infusion over 2 h to 19 adult
Received: September 5, 2019
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
© XXXX American Chemical Society
J. Med. Chem. XXXX, XXX, XXX−XXX
A

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
exposed indole-N-substituent led to compounds with poor
brain exposure due to P-glycoprotein (P-gp) mediated efflux.
Since the identification of compounds from the indole-3-
carboxamide class of V1a receptor antagonists that were
simultaneously highly potent, selective, metabolically stable,
and devoid of P-gp-mediated efflux turned out to be elusive,15
we were prompted to explore a scaffold hop. To this end a 4-
aryl substituted 1,2,4-triazole looked most promising since
both 2 and the triazolobenzodiazepine PF-18456318 3 (Figure
1) feature a chlorophenyl group and a good H-bond acceptor
(an amide carbonyl oxygen in the case of 2 and a 1,2,4-triazole
nitrogen atom in 3) at approximately the same distance, which
suggested that these two classes of compounds might bind to
V1a receptors in a similar way.
4-Aryl substituted 1,2,4-triazole derivatives were for the first
time reported by researchers from Yamanouchi as potent and
selective V1a receptor antagonists.19 In a radioligand binding
assay “compound 19” 4 (Figure 1) was reported to have 1700-
fold selectivity for the human V1a vs the V2 receptor. In our
hands, the selectivity of this compound in functional calcium-
flux assays was found to be only 67-fold vs human V2 and 34-
fold vs OT receptors (Table 2). In contrast, the triazolo-
benzodiazepine 318 is more than 2400-fold selective against
both human V2 and OT receptors (Table 1). However, the
latter compound was also reported to have significant agonistic
activity at V2 receptors,18 which can lead to undesired
antidiuretic side effects.20 Moreover, 3 is a substrate of
human P-gp with an efflux ratio of 4 and hence limited brain
exposure. Nevertheless, the favorable in vitro as well as in vivo
pharmacokinetic profile of 3 in rat18 implied that the tricyclic
Figure 1. Known indole-3-carboxamide and 4-aryl-1,2,4-triazole V1a
antagonists and hybrid design structures 5 and 6 (RO5135117).
development as an oral treatment. All attempts to improve
metabolic stability in animals by modifications of the sterically
Table 1. In Vitro Potency, Selectivity, and Metabolic Stability Data of 3, 5, and Analogs of 5
a
b
pKb: negative logarithm of the equilibrium dissociation constant measured in a calcium flux functional assay. HLM CL: intrinsic clearance
c
measured in human liver microsomes. RLM CL: intrinsic clearance measured in rat liver microsomes.
B
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Table 2. In Vitro Potency and Selectivity Data of Compounds 4, 6, and Derivatives of 6
a
b
pKb: negative logarithm of the equilibrium dissociation constant measured in a calcium flux functional assay. LE: ligand efficiency.
triazolobenzodiazepine was an attractive scaffold, while the
conformationally much less restricted 4 has very low metabolic
stability in human as well as rat liver microsomes. On the other
hand, with a pKb of 8.0, 4 features much higher potency at the
rat V1a receptor compared to 3, which has a pKb of only 6.2.
We thus hypothesized that replacement of the N-(pyridin-2-
yl)piperidine “head-group” of this triazolobenzodiazepine
derivative by different groups such as a biphenyl or a
spiro[1H-isobenzofuran-3,4′-piperidine] group might give rise
to potent and selective V1a receptor antagonists retaining an
overall balanced pharmacokinetic profile suitable for an orally
bioavailable brain penetrating drug (Figure 1, hybrid structures
5 and 6 (RO5135117)).
Herein we describe our efforts toward this end which
culminated in the discovery of the 5,6-dihydro-4H-[1,2,4]-
triazolo[4,3-a][1,4]benzodiazepinebalovaptan 1. In vivo profil-
ing using pharmacological magnetic resonance imaging
(phMRI) in rats to record cerebral blood perfusion as a
proxy for spatiotemporal patterns of evoked neural activity
following drug administration21 played a seminal role in
guiding lead optimization efforts and the selection of 1 for
clinical development. In a double-blind, randomized, placebo-
controlled clinical phase 2 study in 223 adult men with autism
spectrum disorder (VANILLA; NCT01793441) 1 was well
tolerated and demonstrated dose-dependent and clinically
meaningful improvements in the Vineland-II Adaptive
Behavior Scales at oral doses of 4 mg and 10 mg daily, driven
primarily by changes in the socialization and communication
scores.22 A phase 2 trial in children and adolescents
(NCT02901431)23 and a phase 3 trial in adult men and
women (NCT03504917)24 are both ongoing at the time of
writing.
showed much higher clearance in rat liver microsomes than 3,
while metabolic turnover was low in human liver microsomes
for both compounds.
Variation of the spiropiperidine group of 5 led to
compounds 7, 8, 9, and 10, which had reduced potency and
even lower selectivity. We therefore hypothesized that
selectivity might be achievable in this class of V1a antagonists
with head-groups attached to the triazole scaffold via an sp3
carbon as in the highly selective PF-184563 3. Replacement of
the spiro[1H-isobenzofuran-3,4′-piperidine] nitrogen of 5 by
CH thus gave rise to the diastereomers 11 and 12 (Table 1).
Although the latter was found not only to have the highest
clearance of all compounds shown in Table 1 in rat liver
microsomes but also high clearance in human liver micro-
somes, it indeed showed clearly increased selectivity as
compared to 5.
The second hybrid designed between the triazolobenzodia-
zepine PF-184563 3 and the diphenyl substituted triazole 4,
compound 6 (RO5135117), retained potency as well.
However, the selectivity against V2 and OT receptors was
even lower than that of 4 (Table 2). Yet the finding that 6 is
not a substrate of P-gp warranted further profiling of this
compound. Indeed it was found that it had an acceptable
pharmacokinetic profile in rats, with a clearance of 50−60% of
liver blood flow and a Vss of 4−6 L/kg, which resulted in a half-
life of approximately 1.5 h after iv administration. Absorption
was complete following oral administration, confirming the
drug-like potential of this class of compounds. On the basis of
extrapolation of unbound plasma concentrations and a Kb of
12 6 nM at the rat V1a receptor, it was predicted that more
than 90% brain V1a receptor occupancy should be achieved
after iv injection of a single dose of 30 mg/kg in rats, and since
a selectivity screen in a 50-target panel showed no off-target
activities (see Supporting Information, Table 2), we selected
this compound as a tool to probe the effect of central V1a
receptor antagonism on brain circuitry in a preclinical phMRI
study. In line with the prediction of the efficacious dose 6
indeed showed a strong modulation of brain circuitry in a
RESULTS AND DISCUSSION
■
Discovery of Balovaptan (1). The hybrid structure 5 was
indeed found to be a potent V1a receptor antagonist (Table 1).
However, the selectivity against V2 and OT receptors was an
order of magnitude lower than that of 3. Furthermore, 5
C
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 2. Evolution of the exploration of the effect of the polarity of R on P-gp mediated efflux. For detailed structures see Figure 3.
dose-dependent manner when tested in rat using arterial spin-
labeling (ASL)25 based phMRI as discussed in more detail
below.
thus plausible that log D is a useful predictor of the likelihood
of P-gp mediated efflux, the data set described herein illustrates
that more subtle effects such as steric accessability and
efficiency of polar interactions with P-gp within the membrane
may play an important role as discussed below.
Encouraged by the in vivo profile of 6 and the preliminary
evidence suggesting that selectivity may be achievable with
head-groups attached to the triazole core via an sp3 carbon, we
envisaged saturation of the first phenyl ring of 6. This gave rise
to two compounds, the racemate 13 and 14, with significantly
increased potency at the human V1a receptor and excellent
ligand efficiency but, disappointingly, no improved selectivity
versus V2 and OT receptors (Table 2). Replacement of the
phenyl ring of 14 by 2-pyridyl, however, led to a dramatic
surge in selectivity, while V1a potency increased only slightly.
Yet, like its close analog PF-184563 3, 15 is a substrate of
human P-gp with a comparable efflux ratio of 4.4, while 14 is
devoid of efflux. This result suggested that an H-bond acceptor
in this part of the ligand is crucial for selectivity. At the same
time the increase in polarity renders 15 a substrate of P-gp.
Indeed the degree of hydrogen bond basicity was previously
described as an important determinant of P-gp mediated
efflux.26−28
The evolution of our efforts to explore replacements of the
2-pyridyl group of 15 (Table 2) in search of a compound that
is devoid of P-gp mediated efflux is outlined in Figure 2. A set
of 27 analogs differing only in the substituent in the 4-position
of the cyclohexane ring were synthesized and tested in
transfected LLC-PK1 cells expressing human P-gp. Linear
regression analysis of a plot of the logarithm of the efflux ratio
against log D, measured at pH 7.4, gave model 1 (Figure 3). In
spite of a modest r2 of 0.272, the efflux ratios of 22 of the 27
compounds were predicted with an error of less than 2-fold
using this simplistic model.
Well in accordance with model 1, substitution of the
pyridine ring of our lead structure 15 by fluoride, chloride, or
methyl gave rise to compounds 16−21 with increased log D
values between 2.5 and 3.2, all of which showed decreased
efflux, while the pyridazine and pyrimidine analogs 22 and 23,
which are substantially more polar by 1.5 and 1.3 log units,
respectively, showed stronger efflux. Unexpectedly, however,
the dimethyl substituted pyrimidine 24, which is 1 order of
magnitude more lipophilic (log D = 2.0) than the
unsubstituted pyrimidine derivative 23, was found to be a
very good substrate of P-gp. Interestingly, pyrazine (compound
25) conferred substantial polarity without effecting efflux. This
may be explicable by a reduced interaction with P-gp due to
the modest hydrogen bond basicity of pyrazine (pKHB = 1.22)
as compared to, for example, pyridine (pKHB = 1.86).30
A 5-methyl-1,2,4-oxadiazole ring (compound 26) was found
to be equipolar to pyrazine, giving even weaker P-gp mediated
efflux. Again this is understandable by the low predicted H-
bond basicity of the two nitrogen atoms (pKHB = 0.5 and
0.4).31 Replacing the oxadiazole N4 by CH raised log D by 0.6
units; yet the slightly increased efflux of the resulting isoxazole
27 as compared to oxadiazole 26 may be rationalized by the
moderately higher H-bond basicity (pKHB = 0.8 31) of the
isoxazole nitrogen. Addition of a second methyl group in
position 4 of the isoxazole led to an increase in log D by 0.4
units as expected (compound 28); similarly as for the
pyrimidine−4,6-dimethylpyrimidine pair, methylation again
led to a significant increase in the P-gp mediated efflux ratio
from 2.4 to 3.8. It is not clear whether this is due to an increase
in H-bond basicity by the introduction of the modestly
electron-donating methyl group or rather a conformational
effect. Conversely, the isomeric 5-ethylisoxazole 29, which has
almost the same log D but should have about the same H-bond
basicity as the 5-methylisoxazole 27, showed similarly low
log 10 (efflux ratio) = 1.07 − (0.25 log D)
(1)
Lipophilicity is a composite of a volume and a polarity term.
The latter is determined by dipolarizability, H-bond donor
acidity, and H-bond acceptor basicity.29 Volume and
descriptors of polarity were also reported to be the main
parameters predicting P-gp substrate specificity.27 While it is
D
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 3. log D (pH 7.4) vs P-gp mediated efflux ratio (ER). Solid line: linear regression line. Dotted lines: within 2-fold of linear regression line.
Green dotted line: ER = 3. Red dotted line: ER = 10.
efflux as the latter. Increasing electron density of the nitrogen
and hence H-bond basicity by dearomatization gave the
substantially polar 4,5-dihydroisoxazole 30, which had a log D
of 1 and hence an efflux ratio of 6.1.
Reducing the electron density of the isoxazole by fusion with
a phenyl ring gave the benzisoxazole 31, with a log D above 3
and no efflux (efflux ratio = 1.4). The two regioisomeric
isoxazolopyridine analogs 32 and 33 were more polar by up to
approximately 1 order of magnitude and showed slightly
increased yet still modest efflux in accordance with model 1. It
is worth noting that 34, the isothiazole analog of 32, was found
to be a distinctly better substrate of P-gp in spite of its higher
lipophilicity, presumably due to the lower electronegativity of
sulfur as compared to oxygen, resulting in increased H-bond
basicity of the adjacent nitrogen. The same trend was seen with
the benzisothiazole 35. A much higher efflux ratio of 10.9 than
expected based on the log D of 2.4 was measured for the
benzotriazole 36, suggesting that this ring system is
substantially more H-bond basic than benzisothiazole.
In search of further weakly H-bond basic isosteres of bicyclic
aromatic groups, such as (aza)benzisoxazole, we prepared the
phenyl, pyridyl, and pyrazinyl ethers 37, 1, and 38. All three
were devoid of P-gp mediated efflux, consistent with the very
weak H-bond basicity of the oxygen atom of aromatic
E
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Table 3. Scope of the N-Substituent of Phenyl Cyclohexyl Ether Derivatives: In Vitro Potency, Selectivity, and log D
a
b
pKb: negative logarithm of the equilibrium dissociation constant measured in a calcium flux functional assay. LE: ligand efficiency.
ethers.32,30 In sharp contrast, the 2-aminopyridine 39, resulting
from replacement of the ether oxygen of 1 by NH, was found
to be a very good substrate of P-gp with an efflux ratio of 14.
This huge difference can be attributed to the H-bond basicity
of the pyridine nitrogen, which is increased by at least 1 order
of magnitude,30 as well as the additional H-bond donor; P-gp
mediated efflux has been described as being particularly
sensitive to the number of H-bond donors.33 To further
explore the scope of ethers, we prepared the cyclobutyl and
cyclopentyl ethers 40 and 41. By comparison with alkyl aryl
ethers, dialkyl ethers are significantly better H-bond accept-
ors.32,34 In line with this, 40 and 41 had slightly increased
efflux ratios compared to the aryl ethers, yet well predicted by
their log D using model 1.
After learning that even subtly increasing the polarity of the
substituent R in the 4-position of the cyclohexane ring could
confer P-gp mediated efflux, we chose to use the compound
with the lowest efflux ratio, the phenyl cyclohexyl ether 37, a
highly potent V1a antagonist with good selectivity against hV2
and hOT receptors, as a template to explore the scope of the
N-substituent. As summarized in Table 3 potency at the
human V1a receptor was retained with a number of
replacements of the N-methyl group, with the exception of
hydrogen (42) and BOC (48). Beyond this the secondary
amine 42 turned out to be a substrate of P-gp with an efflux
ratio of 4.3. It should be pointed out that model 1 would
predict an efflux ratio of 2.5 based on its log D value of 2.7,
which highlights that this model based on a single parameter is
certainly too simplistic to universally predict P-gp mediated
efflux and that other factors such as hydrogen bonding need to
be taken into account. For example in the case of 42 the
increased efflux can be attributed to the introduction of the H-
bond donor.33
Selectivity decreased with bulkier N-substituents which
tended to increase lipophilicity (43, 44, and 48−50), whereas
derivatives carrying substituents effecting a decrease in log D
such as hydroxyethyl (45), acetyl (46), and N,N-dimethyla-
minoacetyl (47) showed enhanced selectivity, albeit at the cost
of a higher molecular weight and hence decreased ligand
efficiency (Table 3). Since we had shown that introduction of
functionality with enhanced H-bonding capacity would
increase the risk of P-gp mediated efflux, N-methyl substitution
appeared as overall most favorable.
All but one of the N-methyl substituted cylcohexyl-
triazolobenzodiazepine derivatives discussed above, for which
P-gp data had been generated, were very potent antagonists of
the human V1a receptor with Kb values below 3 nM (the full
table of results is provided as Supporting Information). Among
the nonsubstrates of P-gp (efflux ratio of <2) the 2-pyridyl
ether 1, which had a Kb of 0.39 nM at the human V1a receptor,
clearly stood out for its outstanding functional selectivity
against human V2 (3888-fold) and OT (2735-fold) receptors
(Figure 4). In a diverse 120-target selectivity screen, 1 showed
no significant off-target activities except for the human 5HT2B
receptor (see Supporting Information Table 3). In a functional
assay 1 was found to be an antagonist of the 5HT2B receptor
with a Kb of 366
97 nM (n = 3; 938-fold functional
selectivity over hV1a).
F
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 4. Selectivity of cylcohexyltriazolobenzodiazepine derivatives with P-gp data against hV2 and hOT receptors. Fold selectivity was based on
functional Kb values. Green dotted lines represent 1000-fold selectivity against hOT (vertical line) and hV2 (horizontal line).
Figure 5. Dose-dependency of the overall response strength plotted as root-mean-square of normalized perfusion difference to vehicle across 45
ROIs. Values are given as group mean change SEM at iv doses of 1, 3, 10, and 30 mg/kg for RO5135117 (6) and 3, 10, 30, 60, and 90 mg/kg for
RO5285119 (1).
In in vivo pharmacokinetic studies in rats, dogs, and
cynomolgus monkeys 1 showed a plasma clearance which
was high (close to liver blood flow) in rats but moderate or low
in dogs and monkeys. Distribution volume was between 0.5 L/
kg and 1.3 L/kg in the preclinical species, indicating moderate
extravascular distribution, and oral bioavailability was moder-
ate (40% in rats) to high (130% in monkeys) indicative of
good oral absorption. For the prediction of human
pharmacokinetics the biggest uncertainty was associated with
clearance. Hepatic metabolism was believed to be the major
route and was qualitatively similar across species. However,
quantitative clearance scaling from in vitro data generated in
hepatocytes was inconsistent in preclinical species, being
reasonable in rat and dog but showing a large overprediction in
G
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Figure 6. Activation maps showing brain (de)activation for RO5135117 (6) and RO5285119 (1) at the iv dose of 30 mg/kg. Individual columns
show ROI template and group perfusion images as difference to vehicle [%] for four selected anteriorities relative to Bregma (see Experimental
Section for a complete list and abbreviations of ROIs/compartments).
Figure 7. Perfusion profiles of RO5135117 (6) and RO5285119 (1). Values are given as difference of normalized perfusion vs vehicle [%] change
SEM at a dose of 30 mg/kg: *p < 0.05; ROI-wise nested ANOVA, group [study] with FDR control at 1%. Plotted are only those ROIs with a
significant response to at least one of the nine compound/dose conditions shown in Figure 5 (see Experimental Section for a complete list and
abbreviations of ROIs/compartments).
monkey. Scaling from human hepatocytes predicted a
clearance in the range of 4−6 mL min−1 kg−1, but given the
overprediction seen in monkey, a clearance as low as 1 mL
min−1 kg−1 was considered possible. Overall, when taking the
predicted human pharmacokinetics in combination with the
reasonably high free fraction in plasma of 13% and the high
H
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
a
Scheme 1. Synthesis of Triazolobenzodiazepine V1a Antagonists
a
Reagents and conditions: (a) SOCl2, Et3N, CH2Cl2, 24−32 °C, 3 h, 99%; (b) glycine ethyl ester hydrochloride, Et3N, ethanol, reflux, 8 h, 99%
crude; (c) (BOC)2O, DMAP, CH2Cl2, 0 °C to rt, 18 h, 51%; (d) H2 (1 bar), ZnBr2, Pd/C, EtOAc, rt, 48 h, 95%; (e) KO-t-Bu, THF, 5 °C to rt, 3
h, 88%; (f) Lawesson’s reagent, THF, reflux, 3 h, 86%; (g) CHONHNH2, 1,4-dioxane, 90 °C, 18 h, 60%; (h) NBS, THF, reflux, 62%; (i)
substituted piperidine 55 (10 equiv), 120 °C; (j) 1.25 M HCl in MeOH, 50 °C, 1 h; (k) aldehyde/ketone, Et3N, MeOH, rt to reflux; NaBH3CN, 0
°C to rt, or aldehyde/ketone, AcOH, MeOH, NaBH(OAc)3, rt; (l) hydrazide 56, n-BuOH, reflux, 16−72 h; (m) Et3N, THF, filtration; carboxylic
acid, EDC, 50 °C, 18 h; (n) R4-Cl, Et3N, CH2Cl2, rt, 18−72 h.
potency at the human V1a receptor, it was expected to be
possible to achieve high levels of target engagement at a
reasonable daily dose. Thus, even with a clearance of 6 mL
min−1 kg−1, which is at the higher end of the expected range, a
daily dose of 10 mg was expected to achieve >70% receptor
occupancy over 12 h at steady state when considering the Kb of
0.39 nM and a free fraction in plasma of 13%. Simulations are
provided in the Supporting Information.
phMRI. To investigate the brain activity patterns elicited by
our selective V1a antagonists, we measured region-specific
alterations of perfusion as a proxy of neural activity changes
using ASL-based phMRI35 in stress-hyperresponsive rats
(Fischer, F344). Acute intravenous administration of 6 and 1
resulted in consistent and dose-dependent changes of the
neural activity pattern as compared to treatment with vehicle.
Figure 5 shows the dose-dependency of the overall response
strength measured as root-mean-square (de)activation across
the regions of interest (ROIs).
At the dose of 30 mg/kg, at which the rat brain V1a receptor
occupancy of 6 and 1 is estimated to be clearly greater than
90% (see Supporting Information), both compounds elicited a
highly similar activity pattern comprising several cortical and
subcortical brain regions. This regional change in neural
activity for both compounds is depicted in Figure 6 showing
brain perfusion maps in coronal orientation at selected
anteriorities relative to Bregma that comprise prefrontal,
entorhinal, and sensory cortices as well as regions of the
basal ganglia, limbic, and brain stem nuclei.
The region-specific changes are plotted in Figure 7 as
changes in normalized perfusion for 32 ROIs with significant
response to at least one compound/dose condition. 1, which
showed a higher number of (de)activated areas than 6 at the
selected dose, elicited a pronounced decrease in perfusion in
the medial prefrontal, motor, and visual cortex on the order of
−20% compared to vehicle; only in the entorhinal piriform
cortex was an increase in perfusion observed. In contrast, areas
of the midbrain (VTA, SN, Hyp), raphe nuclei, and amygdala
showed pronounced perfusion increase upon acute dosing with
1. Only moderate perfusion changes were observed in caudate
putamen (CPU), hippocampus (dHpc), and midbrain tectum
(SC, IC). Overall, the regional distribution of decreased and
increased normalized perfusion was very similar for 6, albeit at
the selected dose the decrease in perfusion vs vehicle in cortical
areas was generally lower. The most pronounced effect for 6
was the increase in perfusion in VTA, SN, raphe, and
hypothalamic nuclei.
In phMRI the measured perfusion changes can reflect either
direct local receptor engagement or regulation by intercon-
nected upstream regions. Since V1a receptor expression is
widely distributed in several subcortical and cortical brain
regions, interpretation of the regional activity changes as being
due to either local or circuitry engagement is not
straightforward. Clearly, several areas showing V1a antago-
nist-induced brain activity changes possess a dense V1a
receptor expression in particular in the hypothalamic and
raphe nuclei, accumbens shell, hippocampus, VTA, and
piriform cortex.36 The observed perfusion changes in these
brain structures hence can reflect local receptor engagement. In
areas with only minor V1a receptor expression, such as the
amygdala or PFC, the pronounced perfusion changes are
rather induced through ascending projections from areas with
high V1a receptor expression.
Taken together, we observed V1a receptor mediated effects
in brain regions that are part of the social decision-making
I
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
network which combines brain regions of the mesolimbic
reward system with those of the social behavior network.37 Of
particular interest is the pronounced decrease in perfusion
elicited by V1a antagonists in the PFC, a key node controlling
social behavior. It recently has been demonstrated that
increasing excitation of the PFC leads to a robust decrease
in sociability in both rats and mice.38
Following these encouraging results, which led to a
Breakthrough Therapy Designation by the U.S. FDA in
January 2018,40 balovaptan 1 entered phase 3 clinical
development in adults in August 2018,24 and a phase 2 study
in children and adolescents is also currently ongoing.23
EXPERIMENTAL SECTION
■
General Information. This section contains synthesis details for
key intermediates and general procedures for the preparation of the
compounds discussed in the text. The specific syntheses of balovaptan
1 as well as compounds 5 and 6 are shown in this section; details for
the syntheses of all other compounds discussed are in the Supporting
Information. All materials for the reactions were purchased from
commercial sources and used without further purification. LC−MS
data were recorded on Waters UPLC/MS systems equipped with an
Agilent Zorbax Eclipse Plus C18 column (2.1 mm × 30 mm, 1.8 μm,
part no. 959731-902), a CTC PAL autosampler, and a Waters SQD
single quadrupole mass spectrometer using electrospray ionization
(ESI) modes (positive and/or negative). Chromatographic purifica-
tions were carried out using silica gel cartridges of 4−120 g (average
particle size: 35−70 μm) on TELEDYNE ISCO CombiFlash Systems
or by preparative RP-HPLC using a Gemini NX 3u C18 50 mm × 4.6
mm column with water/acetonitrile as eluent. LC−MS high-
resolution spectra were recorded with an Agilent LC system
consisting of Agilent 1290 high-pressure gradient system, a CTC
PAL autosampler, and an Agilent 6520 QTOF. Ionization was
performed in Agilents multimode source. The mass spectrometer was
run in “2 GHz extended dynamic range” mode, resulting in a
resolution of about 10 000 at m/z = 922. Mass accuracy was ensured
by internal drift correction. LC−MS high-resolution data for all
compounds are provided in the Supporting Information. The mean
UVA/B purity of all compounds was ≥95% with three exceptions: 30,
CHEMISTRY
■
All compounds described were prepared from the common
thiolactam intermediate 51 as described in Scheme 1.
Chlorination of 5-chloro-2-nitrobenzyl alcohol 52 with thionyl
chloride was followed by N-alkylation with glycine ethyl
ester,N-BOC protection, and selective hydrogenation of the
nitro group using zinc bromide modulated palladium on
carbon39 to give 53. The thiolactam 51 was obtained by
cylization of 53 using potassium tert-butoxide and treatment of
the resulting lactam with Lawesson’s reagent.
Condensation of 51 with formylhydrazine and bromination
using N-bromosuccinimide gave the bromotriazole 54.
Nucleophilic aromatic substitution with an excess of a
substituted piperidine 55 at elevated temperature, N-
deprotection with HCl in methanol, and reductive N-
methylation with paraformaldehyde and sodium cyanoborohy-
dride gave compounds 5 and 7−10.
Triazolobenzodiazepine V1a antagonists 1, 6, 11, 12, and
14−50 with head-groups linked to the triazole core via a
carbon atom were prepared by cylocondensation of the
thiolactam 51 with a hydrazide 56. After cleavage of the
BOC protecting group under acidic conditions modification of
the N-substitution was effected either by reductive alkylation,
amide formation using an acid and a coupling reagent such as
1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC), or by
treatment with an acid, sulfonyl, or sulfamoyl chloride in the
presence of a tertiary amine base.
1
94.0%; 31, 94.3%; 36, 94.3%. H NMR spectra were recorded on a
Bruker Avance 300 MHz or an Avance III 600 MHz spectrometer, in
CDCl3 (δ = 7.26 ppm) or DMSO-d6 (δ = 2.50 ppm) with all chemical
shifts reported downfield from tetramethylsilane internal standard (δ
1
in ppm). H resonances are reported to the nearest 0.01 ppm. NMR
abbreviations are as follows: s, singlet; d, doublet; t, triplet; q, quartet;
quint, quintet; spt, septet; m, multiplet; br s, broad singlet. Coupling
constants (J) are given in hertz (Hz).
General Procedure A: Synthesis of 4-Aryloxycyclohexane
Carboxylic Acid Hydrazides. (a) Synthesis of 4-Aryloxycyclo-
hexane Carboxylic Acid Esters. To a solution of triphenylphos-
phine (1.2 equiv) in dry tetrahydrofuran (0.1 M) is added diethyl
azodicarboxylate (1.2 equiv) at 0 °C. After 20 min a phenol derivative
(1.2 equiv) and a solution of a 4-hydroxycyclohexanecarboxylic acid
ester in tetrahydrofuran (1−3 M) are added consecutively at 5 °C.
After completion of addition the cooling bath is removed, and the
reaction mixture is allowed to warm to room temperature and stirred
for 3−18 h. The solvent is evaporated, and the residue is dissolved in
ethyl acetate. The ethyl acetate solution is washed with one to two
portions of 1 M aqueous sodium hydroxide solution. The aqueous
layer is extracted with one to two portions of ethyl acetate. The
combined organic layers are dried over anhydrous sodium sulfate and
concentrated in vacuo. Purification by flash chromatography gives a 4-
aryloxycyclohexane carboxylic acid ester. The main product obtained
results from inversion of the absolute configuration in the 4-position.
(b) Synthesis of 4-Aryloxycyclohexane Carboxylic Acid
Hydrazides. A mixture of a 4-aryloxycyclohexane carboxylic acid
ester (1 equiv) and a hydrazine hydrate (5 equiv) is heated at 120 °C
for 5 h. After cooling to room temperature the reaction mixture is
partitioned between dichloromethane and water. The layers are
separated, and the organic layer is washed with water and brine. The
organic layer is dried over anhydrous sodium sulfate and concentrated
in vacuo to give the crude hydrazide, which is used in the next step
without further purification.
CONCLUSION
■
To date, no centrally acting V1a receptor antagonist has
received marketing authorization. Herein we describe the
discovery of balovaptan 1, a highly potent and selective brain
penetrating V1a antagonist. The key to balancing potency and
selectivity while minimizing P-gp mediated efflux at the
blood−brain barrier was careful fine-tuning of hydrogen
bond acceptor basicity. In the absence of a validated
pharmacodynamic in vivo model for central V1a receptor
antagonists, assessment of a V1a antagonist specific brain
activity pattern by pharmacological magnetic resonance
imaging in rats played a seminal role in strengthening
confidence in the potential of the class of triazolo-
benzodiazepines for optimization toward a candidate drug.
These effects were well in line with calculated brain V1a
receptor engagement in rats based on in vitro potency, thus
laying a solid foundation to the application of the same
methodology to predict an efficacious dose in humans.
Consistent with these early predictions balovaptan 1 was
well tolerated and showed dose-dependent and clinically
meaningful improvements in the socialization and communi-
cation domains of the Vineland-II Adaptive Behavior Scale at a
dose of 4 or 10 mg/day in secondary analyses from a 12-week
placebo-controlled clinical phase 2 study in adult men with
moderate/severe autism spectrum disorder.22
General Procedure B: Synthesis of N-Methylated
Triazolobenzodiazepines. (a) Cyclocondensation. A mixture of
a hydrazide derivative (1−1.5 equiv) and 7-chloro-2-thioxo-1,2,3,5-
J
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
tetrahydrobenzo[e][1,4]diazepine-4-carboxylic acid tert-butyl ester 51
(1 equiv) in n-butanol (0.1−0.2 M) is heated at reflux for 16−72 h.
After cooling to room temperature, the solvent is evaporated, and the
residue is purified by flash chromatography to give a triazole
derivative. In some cases partial cleavage of the N-BOC group is
observed under these conditions.
(b) Cleavage of the N-BOC Group. A solution of an N-BOC
protected triazolobenzodiazepine (1 equiv) in 1.25 M methanolic or
1.5 M ethanolic hydrogen chloride solution (10−20 equiv HCl) is
heated at 50 °C for 15−60 min. After cooling to room temperature,
the reaction mixture is concentrated in vacuo to give a secondary
amine derivative as hydrochloride salt. The free base can be obtained
by partitioning the hydrochloride salt between 1 M aqueous sodium
hydroxide solution and an organic solvent such as ethyl acetate or
dichloromethane. The layers are separated, and the aqueous layer is
extracted with two portions of the organic solvent. The combined
organic layers are dried over anhydrous sodium sulfate, filtered, and
concentrated in vacuo to give the free base.
(c) Reductive N-Methylation. A mixture of an N-unsubstituted
triazolobenzodiazepine as free base or as hydrochloride salt (1 equiv,
0.1−0.2 M), triethylamine (1 equiv when the hydrochloride salt is
used), and paraformaldehyde (8 equiv) in methanol is heated at reflux
for 2−6 h. After cooling to 0 °C sodium cyanoborohydride (2−3
equiv) is added. The reaction mixture is stirred for 3−16 h at room
temperature and quenched with 1 M aqueous sodium hydroxide
solution. The aqueous layer is extracted with ethyl acetate. The
combined organic layers are dried over anhydrous sodium sulfate and
concentrated in vacuo. Flash chromatography gives an N-methyl
derivative.
trans-8-Chloro-5-methyl-1-[4-(pyridin-2-yloxy)cyclohexyl]-
5,6-dihydro-4H-2,3,5,10b-tetraazabenzoazulene (1). The title
compound was obtained in 65% yield as white solid according to the
general procedure B from trans-4-(pyridin-2-yloxy)cyclohexane-
carboxylic acid hydrazide, which had been prepared according to
the general procedure A from cis-4-hydroxycyclohexanecarboxylic acid
methyl ester and 2-hydroxypyridine. 1H NMR (600 MHz, CDCl3): δ
8.13 (ddd, 1H, J = 5.0, 1.9, 0.6 Hz), 7.51−7.56 (m, 3H), 7.29−7.32
(m, 1H), 6.83 (ddd, 1H, J = 7.1, 5.1, 0.9 Hz), 6.66 (d, 1H, J = 8.4
Hz), 5.10 (tt, 1H, J = 10.7, 4.2 Hz), 3.11−3.99 (m, 4H), 2.83−2.95
(m, 1H), 2.48 (s, 3H), 2.23−2.39 (m, 2H), 1.85−2.21 (m, 2H), 1.78
(m, 2H), 1.41−1.58 (m, 2H). 13C NMR (151 MHz, CDCl3): δ 163.1,
146.9, 138.6, 132.7, 124.0, 116.5, 111.4, 72.1, 55.9, 47.7, 47.6, 43.8,
33.8, 31.4, 29.3.
2H), 3.85 (s, 2H), 3.82 (s, 2H), 3.30−3.45 (m, 4H), 1.92−2.11 (m,
2H), 1.72−1.89 (m, 2H). MS m/e: 408 (M + H+).
(c) 8-Chloro-5-methyl-1-(1′H,3H-spiro[2-benzofuran-1,4′-
piperidin]-1′-yl)-5,6-dihydro-4H-[1,2,4]triazolo[4,3-a][1,4]-
benzodiazepine (5). The title compound was obtained as a light
yellow solid (0.043 g, 58%) from 1′-(8-chloro-5,6-dihydro-4H-
benzo[f ][1,2,4]triazolo[4,3-a][1,4]diazepin-1-yl)-3H-spiro-
[isobenzofuran-1,4′-piperidine] according to step c of the general
1
procedure B. H NMR (600 MHz, CDCl3): δ 7.73 (d, 1H, J = 8.6
Hz), 7.50 (br d, 2H, J = 8.5 Hz), 7.27−7.30 (m, 2H), 7.22 (dd, 1H, J
= 5.0, 3.4 Hz), 7.18 (dd, 1H, J = 5.2, 3.5 Hz), 5.06 (s, 2H), 3.60−3.76
(m, 2H), 3.44−3.59 (m, 2H), 3.31−3.42 (m, 4H), 2.54 (br s, 3H),
2.01 (ddd, 2H, J = 13.5, 10.3, 7.1 Hz), 1.73 (br d, 2H, J = 12.5 Hz).
13C NMR (151 MHz, CDCl3): δ 157.0, 145.0, 139.0, 133.9, 131.5,
127.8, 127. 8, 127.4, 127.3, 123.3, 121.2, 120.7, 99.1, 84.1, 56.2, 70.9,
48.4, 45.9, 44.1, 35.5.
1-Biphenyl-4-yl-8-chloro-5-methyl-5,6-dihydro-4H-
2,3,5,10b-tetraazabenzoazulene (6). The title compound was
obtained as a white solid according to the general procedure B from 4-
1
biphenylcarboxylic acid hydrazide. H NMR (600 MHz, CDCl3): δ
7.60−7.67 (m, 4H), 7.55−7.60 (m, 3H), 7.45−7.49 (m, 2H), 7.37−
7.42 (m, 1H), 7.31 (dd, 1H, J = 8.5, 2.4 Hz), 6.94 (d, 1H, J = 8.6 Hz),
3.78 (br s, 2H), 3.62 (br s, 2H), 2.57 (s, 3H). 13C NMR (151 MHz,
CDCl3): δ 152.7, 143.1, 139.9, 134.5, 133.4, 131.5, 129.3, 129.0,
129.0, 128.1, 127.5, 127.1, 125.7, 125.1, 56.3, 48.2, 44.4.
7-Chloro-2-thioxo-1,2,3,5-tetrahydrobenzo[e][1,4]-
diazepine-4-carboxylic Acid tert-Butyl Ester (51). (a) 4-Chloro-
2-chloromethyl-1-nitrobenzene. To a solution of 5-chloro-2-
nitrobenzyl alcohol 52 (80 g, 0.42 mol, 1 equiv) and triethylamine
(64 mL, 0.46 mol, 1.1 equiv) in dichloromethane (840 mL) was
added dropwise thionyl chloride (34 mL, 0.46 mol, 1.1 equiv) during
a period of 30 min while the internal temperature was kept below 32
°C by cooling with a water bath. The reaction mixture was stirred for
3 h. The solvent was evaporated, and the residue was triturated in
warm tert-butyl methyl ether (970 mL). The ammonium salts were
removed by filtration and the filtrate was concentrated in vacuo to
give the title compound (85 g, 99%) as a brown oil which was used in
1
the next step without purification. H NMR (300 MHz, CDCl3): δ
8.06 (d, 1H, J = 8.88 Hz), 7.73 (d, 1H, J = 2.22 Hz), 7.42−7.55 (m,
1H), 4.97 (s, 2H). MS m/e: 205 (M+).
(b) (5-Chloro-2-nitrobenzylamino)acetic Acid Ethyl Ester. A
mixture of 4-chloro-2-chloromethyl-1-nitrobenzene (85 g, 0.41 mol, 1
equiv), glycine ethyl ester hydrochloride (70 g, 0.50 mol, 1.2 equiv),
and triethylamine (121.4 mL, 0.8665 mol, 2.1 equiv) in ethanol (1000
mL) was heated at reflux for 8 h. The solvent was evaporated, and the
residue was triturated in warm tert-butyl methyl ether. The
ammonium salts were removed by filtration, and the filtrate was
concentrated in vacuo to give the title compound (111 g, 99%) as an
amorphous brown solid which was used in the next step without
purification. MS m/e: 273 (M + H+).
8-Chloro-5-methyl-1-(1′H,3H-spiro[2-benzofuran-1,4′-pi-
peridin]-1′-yl)-5,6-dihydro-4H-[1,2,4]triazolo[4,3-a][1,4]-
benzodiazepine (5). General Procedure C. (a) tert-Butyl-8-
chloro-1-(3H-spiro[isobenzofuran-1,4′-piperidin]-1′-yl)-4H-
benzo[f ][1,2,4]triazolo[4,3-a][1,4]diazepine-5(6H)-carboxy-
late. A mixture of 1-bromo-8-chloro-4H,6H-2,3,5,10b-tetraazabenzo-
[e]azulene-5-carboxylic acid tert-butyl ester 54 (0.15 g, 0.375 mmol, 1
equiv) and 3H-spiro[isobenzofuran-1,4′-piperidine] (0.71 g, 3.75
mmol, 1 equiv) was heated at 120 °C and stirred for 5 h. The reaction
mixture was partitioned between ethyl acetate (50 mL) and saturated
ammonium chloride solution. The aqueous layer was extracted with
three 50 mL portions of ethyl acetate. The combined organic layers
were washed with one 30 mL portion of brine, dried over Na2SO4,
filtered, and concentrated. Trituration from ethyl acetate/n-heptane
(1:1, 6 mL) as eluent gave the title compound (0.112 g, 59%) as a
(c) [tert-Butoxycarbonyl-(5-chloro-2-nitrobenzyl)amino]-
acetic Acid Ethyl Ester.
A solution of (5-chloro-2-
nitrobenzylamino)acetic acid ethyl ester (110 g, 0.403 mol, 1
equiv), di-tert-butyl dicarbonate (180 g, 0.807 mol, 2.0 equiv), and
4-N,N-dimethylaminopyridine (2.51 g, 0.0202 mol, 0.05 equiv) in
dichloromethane (1200 mL) was stirred for 2 h at 0 °C and further 16
h at room temperature. The solvent was evaporated, and the crude
product was purified by flash chromatography with a cyclohexane/
ethyl acetate mixture as eluent to give the title compound (76.4 g,
1
white solid. H NMR (300 MHz, CDCl3): δ 7.68−7.87 (m, 1H),
7.46−7.63 (m, 2H), 7.28−7.32 (m, 2H), 7.14−7.15 (m, 1H), 7.12−
7.14 (m, 1H), 5.06 (s, 2H), 4.20−4.82 (m, 4H), 3.30−3.52 (m, 4H),
1.94−2.15 (m, 2H), 1.67−1.79 (m, 2H), 1.50 (s, 9H). MS m/e: 509
(M + H+).
1
51%) as a light yellow viscous oil. H NMR (600 MHz, CDCl3): δ
7.95−8.06 (m, 1H), 7.58−7.63 (m, 1H), 7.37−7.43 (m, 1H), 4.87 (s,
1H), 4.81 (s, 1H), 4.18−4.24 (m, 2H), 4.04 (s, 1H), 3.93 (s, 1H),
1.48 (s, 4.7 H), 1.37 (s, 4.3 H), 1.24−1.32 (m, 3H). MS m/e: 373 (M
+ H+).
(d) [(2-Amino-5-chlorobenzyl)-tert-butoxycarbonylamino]-
acetic Acid Ethyl Ester (53). To a solution of [tert-butoxycarbon-
yl-(5-chloro-2-nitrobenzyl)amino]acetic acid ethyl ester (69.0 g, 0.186
mol, 1 equiv) in ethyl acetate (1200 mL) was added zinc bromide
(8.5 g, 0.037 mol, 0.20 equiv). The reaction mixture was purged with
argon after 15 min. After addition of the palladium catalyst (10% on
(b) 1′-(8-Chloro-5,6-dihydro-4H-benzo[f ][1,2,4]triazolo[4,3-
a][1,4]diazepin-1-yl)-3H-spiro[isobenzofuran-1,4′-piperidine]
Hydrochloride. The title compound was obtained as a light brown
solid (0.070 g, 86%) from tert-butyl 8-chloro-1-(3H-spiro-
[isobenzofuran-1,4′-piperidin]-1′-yl)-4H-benzo[f ][1,2,4]triazolo[4,3-
a][1,4]diazepine-5(6H)-carboxylate according to step b of the general
1
procedure B. H NMR (300 MHz, CDCl3): δ 7.65−7.79 (m, 1H),
7.44−7.56 (m, 2H), 7.27−7.31 (m, 2H), 7.12−7.24 (m, 3H), 5.06 (s,
K
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
activated charcoal, 7.9 g, 0.0074 mol, 0.040 equiv) the mixture was
hydrogenated at ambient pressure during a period of ∼48 h until ∼13
L of hydrogen gas had been consumed. The catalyst was removed by
filtration, and the filtrate was washed with two portions of saturated
aqueous sodium bicarbonate solution and brine, each. The organic
layer was dried over anhydrous sodium sulfate and concentrated in
vacuo to give the title compound (60.6 g, 95.5%) as yellow waxy solid.
1H NMR (600 MHz, CDCl3): δ 7.05 (dd, 1H, J = 8.4, 2.2 Hz), 6.92−
glutamate, 200 μg/mL geneticin (and 100 μg/mL hygromycin only
for hV2-Gqs5) at 37 °C in a 10% CO2 incubator at 95% humidity.
Cells were plated for 24 h at 50 000 cells/well in clear-bottomed 96-
well plates and were dye loaded for 60 min with 2 μM Fluo-4-AM in
assay buffer. After cell washing, the plate was loaded on a fluorometric
imaging plate reader (FLIPR), compound dilution series added to the
cells, and the calcium signal recorded for 5 min in order to detect
agonist activity. After 20 min of incubation with compound, a
concentration of natural agonist (oxytocin or vasopressin depending
on the receptor) giving 80% of the maximum signal was added to the
plate and the calcium signal recorded for 5 min in order to detect
antagonist activity of the test compound.
6.96 (m, 1H), 6.59 (br d, 1H, J = 8.4 Hz), 4.40 (s, 2H), 4.15 (q, J =
7.1 Hz, 2H), 3.78 (br s, 2H), 1.45 (br s, 9H), 1.19−1.29 (m, 3H). MS
m/e: 343 (M + H+).
(e) 7-Chloro-2-oxo-1,2,3,5-tetrahydrobenzo[1,4]diazepine-
4-carboxylic Acid tert-Butyl Ester. To a solution of [(2-amino-
5-chlorobenzyl)-tert-butoxycarbonylamino]acetic acid ethyl ester 53
(60 g, 0.18 mol, 1 equiv) in tetrahydrofuran (600 mL) was added
potassium tert-butoxide (22 g, 0.19 mol, 1.05 equiv) in small portions
at 5 °C under cooling on an ice−water batch. After addition was
completed the cooling bath was removed and reaction mixture was
stirred for 3 h at room temperature followed by addition of water
(400 mL), saturated aqueous ammonium chloride solution (280 mL),
and ethyl acetate (800 mL). After 10 min the precipitate was collected
by filtration. The organic layer was separated from the filtrate, dried
over anhydrous sodium sulfate, and concentrated in vacuo. The
residue was combined with the precipitate, which had previously been
collected by filtration, and crystallized from hot ethyl acetate to give
the title compound (46 g, 88%) as a white solid. 1H NMR (600 MHz,
DMSO-d6): δ 10.03−10.24 (m, 1H), 7.23−7.39 (m, 2H), 7.10 (br d,
1H, J = 8.36 Hz), 4.36−4.55 (m, 2H), 4.15−4.34 (m, 2H), 1.19−1.45
(m, 9H). MS m/e: 295 (M − H−).
The calcium signal increase due to agonist activity of the
compounds was fitted to the equation with variable slope Y =
Bottom + (Top − Bottom)/(1 + 10((log EC −X)·Hillslope)) where Y is the
50
% normalized fluorescence, Bottom is the minimum Y, Top is the
maximum Y, EC50 is the concentration yielding 50% of the maximum
induced fluorescence, X is the logarithm of the concentration of the
test compound, and Hillslope is the Hill coefficient.
The calcium signal reduction due to the antagonist activity of the
compounds was fitted to a single site competition equation with
variable slope and formula Y = Bottom + (Top − Bottom)/(1 +
10((log IC −X)Hillslope)), where Y is the % normalized fluorescence,
50
Bottom is the minimum Y, Top is the maximum Y, IC50 is the
concentration inhibiting 50% of the agonist-induced fluorescence, X is
the logarithm of the concentration of the competing compound, and
Hillslope the Hill coefficient. All compounds were tested at least 3
times in duplicate.
P-gp. Human P-glycoprotein (P-gp, MDR1) mediated efflux ratios
were determined as previously described.41
(f) 7-Chloro-2-thioxo-1,2,3,5-tetrahydrobenzo[e][1,4]-
diazepine-4-carboxylic Acid tert-Butyl Ester (51). A mixture of
7-chloro-2-oxo-1,2,3,5-tetrahydrobenzo[1,4]diazepine-4-carboxylic
acid tert-butyl ester (41.1 g, 0.139 mol. 1 equiv) and 2,4-bis(4-
methoxyphenyl)-1,3,2,4-dithiadiphosphetane-2,4-disulfide (31.5 g,
0.0763 mol, 0.55 equiv) in tetrahydrofuran (1100 mL) was heated
at reflux for 3 h. The solvent was evaporated, and the residue was
triturated in tert-butyl methyl ether. The precipitate was removed by
filtration, and the filtrate was concentrated to dryness. The residue
was crystallized from hot ethanol to give the title compound (37.5 g,
86.4%) as a light yellow solid. 1H NMR (600 MHz, CDCl3): δ 9.34−
9.87 (m, 1H), 7.28−7.45 (m, 2H), 6.97 (d, 1H, J = 8.46 Hz), 4.28−
4.91 (m, 4H), 1.33−1.55 (m, 9H). MS m/e: 311 (M − H−).
1-Bromo-8-chloro-4H,6H-2,3,5,10b-tetraazabenzo[e]-
azulene-5-carboxylic Acid tert-Butyl Ester (54). (a) 8-Chloro-
4H,6H-2,3,5,10b-tetraazabenzoazulene-5-carboxylic Acid tert-
Butyl Ester. A mixture of 7-chloro-2-thioxo-1,2,3,5-tetrahydrobenzo-
[e][1,4]diazepine-4-carboxylic acid tert-butyl ester 51 (5.7 g, 18.3
mmol, 1 equiv) and formyl hydrazine (5.5 g, 91.4 mmol, 5 equiv) in
1,4-dioxane (113 mL) was heated at 90 °C and stirred for 18 h. The
solvent was evaporated. The residue was partitioned between
dichloromethane (150 mL) and water (200 mL). The aqueous
layer was extracted with two 75 mL portions of dichloromethane. The
combined organic layers were dried over Na2SO4, filtered, and
concentrated. Purification by column chromatography with n-
heptane/ethyl acetate as eluent gave the title compound (3.5 g,
60%) as a white solid.
(b) 1-Bromo-8-chloro-4H,6H-2,3,5,10b-tetraazabenzo[e]-
azulene-5-carboxylic Acid tert-Butyl Ester (54). A solution of
8-chloro-4H,6H-2,3,5,10b-tetraazabenzoazulene-5-carboxylic acid tert-
butyl ester (1.3 g, 4.05 mmol, 1 equiv) and NBS (0.79 g, 4.47 mmol,
1.1 equiv) in THF (25 mL) was heated at reflux and stirred for 4 h.
Purification by column chromatography with n-heptane/ethyl acetate
as eluent gave the title compound (1.0 g, 62%) as an off-white solid.
In Vitro Pharmacology. Stable Cell Culture and Calcium
Flux Assay Using Fluorescent Imaging. CHO cells were stably
transfected with an expression plasmid encoding one of the receptor
of interest and neomycin resistance gene. The cell line expressing hV2
was additionally cotransfected with a plasmid coding for Gqs5 to
redirect the cAMP signal to calcium. The cells were grown in F-12 K,
containing 10% fetal bovine serum, 1% penicillin−streptomycin, 1% L-
log D. Water/octanol distribution coefficients or effective lip-
ophilicities (log D) were determined as previously described.42
phMRI. Animals and Animal Preparation. All procedures were
conducted in strict adherence to the Swiss federal regulations on
animal protection and the rules of the Association for Assessment and
Accreditation of Laboratory Animal Care International (AAALAC)
and with the explicit approval of the local veterinary authorities.
Animals were group-housed (four per cage) in a temperature- and
light-controlled environment (12 h light/dark cycle) with ad libitum
access to food and water. For MRI, male Fischer (F344) rats (RCC,
Switzerland) weighing 300−350 g were anesthetized with isoflurane
in carrier gas composed of oxygen and air (1:5) supplied to the
spontaneously breathing animals initially in an inhalation box (3.5%
isoflurane) and later using a face mask. During MRI isoflurane
concentration was adjusted to be between 1.8% and 2.4% to maintain
stable respiration rates of 50−60 breaths per minute. Rats were
positioned in a Perspex cradle with the head immobilized in a
stereotaxic holder, and body temperature was maintained at 37 °C via
a feedback-regulated electric heating blanket. Breathing rate and
concentrations of exhaled oxygen and CO2 were monitored
continuously with a capnometer (Datex, Helsinki, Finland) and
recorded with a PowerLab data acquisition system (ADInstruments,
Spechbach, Germany).
Magnetic Resonance Imaging. MRI was performed essentially
according to previously published procedures.35,43 In more detail, data
were acquired on a 4.7 T/40 cm Bruker Biospec horizontal-bore
small-animal scanner (Bruker BioSpin, Ettlingen, Germany), equipped
with a 72 mm bird-cage resonator for excitation and an actively
decoupled quadrature surface receiver coil (Rapid Biomedical,
Rimpar, Germany) for head imaging. On scout images, the most
rostral extension of the corpus callosum was used as a landmark for
selecting eight coronal image planes at −10.0, −7.8, −5.3, −2.9, −1.6,
−0.3, +1.0, and +2.3 mm from Bregma.44 All subsequent images were
acquired from these planes, with a field of view of 4 cm × 4 cm and a
slice thickness of 1 mm. The first imaging volume was a set of RARE
T2-weighted anatomical images (TR/TEeff = 1800 ms/39 ms, RARE
factor 8, matrix 256 × 256).45 Next, a T1-weighted image series
required to quantitatively calibrate perfusion readouts was obtained
using an inversion−recovery snapshot FLASH sequence with 8
inversion times (TR/TE = 3400 ms/1.4 ms, matrix 128 × 64).46
L
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
Finally, perfusion-weighted images were acquired using continuous
arterial spin labeling (CASL)47 with single-slice centered RARE
readout (TR/TE 3750 ms/5.7 ms, RARE-factor 32, matrix 128 × 64,
2.5 s labeling pulse, 0.4 s postlabeling delay). Appropriate frequencies
for the tagged and control images were chosen so as to minimize
magnetization transfer contrast and off-resonance effects. For the
assessment of compound effect three consecutive volumes of CASL
perfusion images were acquired over an acquisition time of 4 min per
volume followed by the compound bolus injection and a series of 9
additional volumes to track the time course of perfusion changes.
Compounds 6 and 1 were formulated in saline containing Tween80
(0.3%) and 2-hydroxypropyl-β-cyclodextrin (30%), respectively, and
during phMRI administered intravenously via a tail vein catheter as a
bolus of 3 mL/kg volume. The following doses were tested: 6 at 1, 3,
10, and 30 mg/kg; 1 at 3, 10, 30, 60, and 90 mg/kg.
Image Processing. Images were processed and analyzed using in-
house developed software written in IDL (RSI, Boulder, CO, USA)
and MATLAB (The MathWorks Inc., Natick, MA, USA). The
anatomical volume of each individual animal was co-registered to an
in-house established rat-brain template using the open-source
software SPM5 (Wellcome Trust Centre for Neuroimaging, London,
U.K.). Spatial normalization comprised a 12-parameter affine and a
nonlinear transform, which were then applied identically to all
pharmacological images of the same subject. The template was aligned
and annotated with an in-house digital atlas delineating 45 regions of
interest (ROIs) adapted from the Paxinos & Watson rat-brain atlas.44
T1 maps were calculated on a voxel-wise basis by fitting a three-
parameter exponential to the image intensities across the eight
inversion times.48 These T1 maps were then combined with the
related CASL images to obtain quantitative absolute perfusion maps
as described elsewhere.49,35 In order to account for possible systemic
changes affecting global brain perfusion and to eliminate part of the
interindividual variability, perfusion maps of each individual were
normalized slicewise to the brain mean value, which was set to 100%.
The temporal evolution of perfusion was quantified by fitting a three-
parameter γ-variate function to the time series of each voxel, yielding
the amplitude of the compound-induced perfusion changes as the
parameter of interest. Amplitudes were averaged ROI-wise and pooled
across hemispheres for bilateral ROIs.
ROI Labels and Abbreviations. The full list of 45 ROIs is given
below. The name of each ROI is followed by its abbreviation (used for
example in Figure 7) and, if applicable, the numeric label that is used
in Figure 6, both in parentheses: cingulate cortex (Cg, 1); prelimbic
and infralimbic cortex (PrL & IL, 2a and 2b); dorsal peduncular
cortex (DP, 3); orbitofrontal cortex (OFC, 4); somatosensory cortex
(SSC, 5/6); primary motor cortex (M1, 7); visual cortex (VC);
entorhinal cortex (Ent); piriform cortex (Pir, 10); ectorhinal cortex
(Ect, 11); perirhinal cortex (PRh, 12); insular cortex (InsC, 13);
caudate putamen (CPU, 14); nucleus accumbens core and shell
(AcbC & AcbSh, 15); ventral pallidum (VP, 16); bed nucleus of the
stria terminalis (BNST, 17); ventral tegmental area (VTA, 18);
substantia nigra (SN, 19); locus coeruleus (LC); thalamus, lateral and
mediodorsal & habenula and other ventral nuclei (lTh and mdTh &
Hb & vmThX, 20); amygdala, basolateral and lateral & central and
basomedial & medial and cortical (BLA & LA & CeA & BMA & MeA
& CoA, 21); dorsal hippocampus (dHpc, 22); posterior dorsal
hippocampus (pdHpc, 23); subiculum (Sb); ventral hippocampus
(vHpc, 26); septal region (Spt, 27); dorsal and ventral periaqueductal
gray (dPAG and vPAG); dorsal and median raphe nuclei (DRN &
MRN); superior and inferior colliculi (SC & IC); hypothalamus,
lateral and median & paraventricular (LH & MH & PVH, 32).
Statistical Analyses. The data presented here are compiled from
5 studies comprising 12 active dose groups (among them 3 pairs with
identical treatment in 2 studies, respectively) and 5 vehicle groups,
i.e., negative control groups (1 per study), totaling 130 animals. The
statistical model was an ANOVA with “dose group” as the between-
subject factor of interest, which was nested within the factor “animal
batch” (i.e., study) to absorb potential nuisance variability caused by
general differences among batches of animals. The analysis included
10 post hoc contrasts of interest, comprising each of the 9
compound/dose combinations vs vehicle, plus an interaction contrast
of the two compounds vs each other at 30 mg/kg. Readouts were
least-squares estimates of group differences along with their standard
errors, plus t test statistics. This analysis was conducted independently
for each of the 45 ROIs, and the false discovery rate (FDR) across
ROIs and contrasts (i.e., 450 single tests) was controlled at 1% using
the Benjamini−Hochberg procedure.50
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jmedchem.9b01478.
Syntheses of compounds 7−50; high-resolution mass
spectrometry and UVA and UVB mean purity data for
compounds 1 and 3−50; results for RO5135117
(compound 6) from 50-target CEREP selectivity screen;
results for balovaptan (compound 1) from 120-target
CEREP selectivity screen; in vitro pharmacology data for
compounds 1 and 3−50; log D and P-gp efflux ratios for
compounds 1 and 3−50; prediction of free brain
concentrations of compounds 1 and 6 in rodent and
human; calculation of receptor occupancy; prediction of
pharmacokinetics and dose in human for balovaptan
(compound 1) (PDF)
Molecular formula strings and some data (CSV)
AUTHOR INFORMATION
■
Corresponding Author
Patrick Schnider − F. Hoffmann-La Roche Ltd., Basel,
Switzerland; orcid.org/0000-0002-5102-8818;
Email: patrick.schnider@roche.com
Other Authors
Caterina Bissantz − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Andreas Bruns − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Cosimo Dolente − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Erwin Goetschi − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Roland Jakob-Roetne − F. Hoffmann-La Roche Ltd.,
Basel, Switzerland
Basil Kunnecke − F. Hoffmann-La Roche Ltd., Basel,
̈
Switzerland
Thomas Mueggler − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Wolfgang Muster − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Neil Parrott − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Emmanuel Pinard − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Hasane Ratni − F. Hoffmann-La Roche Ltd., Basel,
Switzerland; orcid.org/0000-0002-8151-8295
́
Celine Risterucci − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Mark Rogers-Evans − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
M
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
vasopressin receptor 1A (AVPR1A) polymorphisms in autism. Mol.
Psychiatry 2002, 7, 503−507.
Markus von Kienlin − F. Hoffmann-La Roche Ltd., Basel,
Switzerland
Christophe Grundschober − F. Hoffmann-La Roche
Ltd., Basel, Switzerland
(7) Wassink, T. H.; Piven, J.; Vieland, V. J.; Pietila, J.; Goedken, R.
J.; Folstein, S. E.; Sheffield, V. C. Examination of AVPR1a as an
autism susceptibility gene. Mol. Psychiatry 2004, 9, 968−972.
(8) Yirmiya, N.; Rosenberg, C.; Levi, S.; Salomon, S.; Shulman, C.;
Nemanov, L.; Dina, C.; Ebstein, R. P. Association between the
arginine vasopressin 1a receptor (AVPR1a) gene and autism in a
family-based study: mediation by socialization skills. Mol. Psychiatry
2006, 11, 488−494.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jmedchem.9b01478
Notes
The authors declare the following competing financial
interest(s): All authors are employees or former employees
of F. Hoffmann-La Roche Ltd.
(9) Yang, S. Y.; Cho, S. C.; Yoo, H. J.; Cho, I. H.; Park, M.; Kim, B.
N.; Kim, J. W.; Shin, M. S.; Park, T. W.; Son, J. W.; Chung, U. S.;
Kim, H. W.; Yang, Y. H.; Kang, J. O.; Kim, S. A. Association study
between single nucleotide polymorphisms in promoter region of
AVPR1A and Korean autism spectrum disorders. Neurosci. Lett. 2010,
479, 197−200.
(10) Tansey, K. E.; Hill, M. J.; Cochrane, L. E.; Gill, M.; Anney, R.
J.; Gallagher, L. Functionality of promoter microsatellites of arginine
vasopressin receptor 1A (AVPR1A): implications for autism. Mol.
Autism 2011, 2, 3.
(11) Yang, S. Y.; Kim, S. A.; Hur, G. M.; Park, M.; Park, J. E.; Yoo,
H. J. Replicative genetic association study between functional
polymorphisms in AVPR1A and social behavior scales of autism
spectrum disorder in the Korean population. Mol. Autism 2017, 8, 44.
(12) Parker, K. J.; Oztan, O.; Libove, R. A.; Mohsin, N.; Karhson, D.
S.; Sumiyoshi, R. D.; Summers, J. E.; Hinman, K. E.; Motonaga, K. S.;
Phillips, J. M.; Carson, D. S.; Fung, L. K.; Garner, J. P.; Hardan, A. Y.
A randomized placebo-controlled pilot trial shows that intranasal
vasopressin improves social deficits in children with autism. Sci.
Transl. Med. 2019, 11, No. eaau7356.
ACKNOWLEDGMENTS
■
We gratefully acknowledge Jens Barthold, Serge Burner, Rafael
Flu
̈
ckiger, Sonja Keppeler, Simone Moser, Katja Ostmann, and
Christian Rink for the synthesis of some of the compounds
described, Andy Stucki, Muriel Schmitt, Gabrielle Py, and
Alain Rudler for the in vitro affinity and potency assessments,
Christian Bartelmus for generating high-resolution mass
spectrometry data, Caroline Wyss Gramberg for recording
NMR spectra, and Sebastien Debilly, Stephanie Schoeppen-
thau, Ciril Waelti, and Thomas Bielser for MRI measurements.
Furthermore, we extend our gratitude to all teams within the
Pharmaceutical Research and Early Development Department
of F. Hoffmann-La Roche AG who supported this program
culminating in the discovery and development of balovaptan
for the treatment of autism spectrum disorder. Funding was
provided by F. Hoffmann-La Roche Ltd. for this work and
third-party editing assistance, which was provided by Dinakar
Sambandan, Ph.D., of Envision Pharma Group, and Nick Fitch,
Ph.D., CMPP, of ArticulateScience (London, U.K.).
(13) Uzefovsky, F.; Shalev, I.; Israel, S.; Knafo, A.; Ebstein, R. P.
Vasopressin selectively impairs emotion recognition in men.
Psychoneuroendocrinology 2012, 37, 576−580.
(14) Thompson, R.; Gupta, S.; Miller, K.; Mills, S.; Orr, S. The
effects of vasopressin on human facial responses related to social
communication. Psychoneuroendocrinology 2004, 29, 35−48.
(15) Ratni, H.; Rogers-Evans, M.; Bissantz, C.; Grundschober, C.;
Moreau, J. L.; Schuler, F.; Fischer, H.; Alvarez Sanchez, R.; Schnider,
P. Discovery of highly selective brain-penetrant vasopressin 1a
antagonists for the potential treatment of autism via a chemogenomic
and scaffold hopping approach. J. Med. Chem. 2015, 58, 2275−2289.
(16) Umbricht, D.; Del Valle Rubido, M.; Hollander, E.;
McCracken, J. T.; Shic, F.; Scahill, L.; Noeldeke, J.; Boak, L.;
Khwaja, O.; Squassante, L.; Grundschober, C.; Kletzl, H.; Fontoura,
P. A single dose, randomized, controlled proof-of-mechanism study of
a novel vasopressin 1a receptor antagonist (RG7713) in high-
functioning adults with autism spectrum disorder. Neuropsychophar-
macology 2017, 42, 1914−1923.
ABBREVIATIONS USED
■
ANOVA, analysis of variance; AVP, arginine vasopressin; FDR,
false discovery rate; V1a, vasopressin receptor 1a; V2,
vasopressin receptor 2; OT, oxytocin; P-gp, P-glycoprotein;
ROI, region of interest; HPA, hypothalamic−pituitary−
adrenal; LLC-PK1 cells, porcine kidney epithelial cells; H-
bond, hydrogen bond; phMRI, pharmacological magnetic
resonance imaging; ASL, arterial spin-labeling
REFERENCES
■
(17) Gulliver, D.; Werry, E.; Reekie, T. A.; Katte, T. A.; Jorgensen,
W.; Kassiou, M. Targeting the oxytocin system: new pharmacother-
apeutic approaches. Trends Pharmacol. Sci. 2019, 40, 22−37.
(18) 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.; Russell, R.;
Spark, D.; Stobie, A.; Underwood, T. J.; Ward, R.; Wheeler, S.
Discovery of PF-184563, a potent and selective V1a antagonist for the
treatment of dysmenorrhoea. The influence of compound flexibility
on microsomal stability. Bioorg. Med. Chem. Lett. 2011, 21, 5684−
5687.
(19) Kakefuda, A.; Suzuki, T.; Tobe, T.; Tsukada, J.; Tahara, A.;
Sakamoto, S.; Tsukamoto, S. Synthesis and pharmacological
evaluation of 5-(4-biphenyl)-3-methyl-4-phenyl-1,2,4-triazole deriva-
tives as a novel class of selective antagonists for the human
vasopressin V(1A) receptor. J. Med. Chem. 2002, 45, 2589−2598.
(20) Pouzet, B.; Serradeil-Le Gal, C.; Bouby, N.; Maffrand, J. P.; Le
Fur, G.; Bankir, L. Selective blockade of vasopressin V2 receptors
reveals significant V2-mediated water reabsorption in Brattleboro rats
with diabetes insipidus. Nephrol., Dial., Transplant. 2001, 16, 725−
734.
(1) Birnbaumer, M. Vasopressin receptors. Trends Endocrinol. Metab.
2000, 11, 406−410.
(2) Koshimizu, T. A.; Nasa, Y.; Tanoue, A.; Oikawa, R.; Kawahara,
Y.; Kiyono, Y.; Adachi, T.; Tanaka, T.; Kuwaki, T.; Mori, T.; Takeo,
S.; Okamura, H.; Tsujimoto, G. V1a vasopressin receptors maintain
normal blood pressure by regulating circulating blood volume and
baroreflex sensitivity. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 7807−
7812.
(3) Engelmann, M.; Landgraf, R.; Wotjak, C. T. The hypothalamic-
neurohypophysial system regulates the hypothalamic-pituitary-adrenal
axis under stress: an old concept revisited. Front. Neuroendocrinol.
2004, 25, 132−149.
(4) Insel, T. R. The challenge of translation in social neuroscience: a
review of oxytocin, vasopressin, and affiliative behavior. Neuron 2010,
65, 768−779.
(5) Hammock, E. A.; Young, L. J. Oxytocin, vasopressin and pair
bonding: implications for autism. Philos. Trans. R. Soc., B 2006, 361,
2187−2198.
(6) Kim, S. J.; Young, L. J.; Gonen, D.; Veenstra-VanderWeele, J.;
Courchesne, R.; Courchesne, E.; Lord, C.; Leventhal, B. L.; Cook, E.
H., Jr.; Insel, T. R. Transmission disequilibrium testing of arginine
N
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Drug Annotation
(21) Jenkins, B. G. Pharmacologic magnetic resonance imaging
(phMRI): imaging drug action in the brain. NeuroImage 2012, 62,
1072−1085.
substituted nitroarenes, alkenes, benzyl ethers, and aromatic ketones.
Synthesis 2003, 2003, 1657−1660.
(40) FDA. Roche’s Autism Drug Balovaptan Granted Breakthrough
Therapy Designation. https://www.fdanews.com/articles/185437-
roches-autism-drug-balovaptan-granted-breakthrough-therapy-
designation (accessed Nov 26, 2019).
(41) Poirier, A.; Weetall, M.; Heinig, K.; Bucheli, F.; Schoenlein, K.;
Alsenz, J.; Bassett, S.; Ullah, M.; Senn, C.; Ratni, H.; Naryshkin, N.;
Paushkin, S.; Mueller, L. Risdiplam distributes and increases SMN
protein in both the central nervous system and peripheral organs.
Pharmacol. Res. Perspect. 2018, 6, No. e00447.
(22) Bolognani, F.; Del Valle Rubido, M.; Squassante, L.; Wandel,
C.; Derks, M.; Murtagh, L.; Sevigny, J.; Khwaja, O.; Umbricht, D.;
Fontoura, P. A phase 2 clinical trial of a vasopressin V1a receptor
antagonist shows improved adaptive behaviors in men with autism
spectrum disorder. Sci. Transl. Med. 2019, 11, No. eaat7838.
(23) ClinicalTrials.gov. A Study to Investigate the Efficacy and
Safety of RO5285119 in Participants With Autism Spectrum Disorder
(ASD). https://www.clinicaltrials.gov/ct2/show/NCT02901431 (ac-
cessed Nov 26, 2019).
(42) Schnider, P.; Dolente, C.; Stalder, H.; Martin, R. E.;
Reinmu
H.; Alker, A. M.; Mu
̈
ller, V.; Marty, R.; Wyss Gramberg, C.; Wagner, B.; Fischer,
(24) ClinicalTrials.gov. A Study of Balovaptan in Adults With
Autism Spectrum Disorder With a 2-Year Open-Label Extension.
https://www.clinicaltrials.gov/ct2/show/NCT03504917 (accessed
Nov 26, 2019).
(25) Detre, J. A.; Leigh, J. S.; Williams, D. S.; Koretsky, A. P.
Perfusion imaging. Magn. Reson. Med. 1992, 23, 37−45.
(26) Seelig, A.; Landwojtowicz, E. Structure-activity relationship of
P-glycoprotein substrates and modifiers. Eur. J. Pharm. Sci. 2000, 12,
31−40.
(27) Didziapetris, R.; Japertas, P.; Avdeef, A.; Petrauskas, A.
Classification analysis of P-glycoprotein substrate specificity. J. Drug
Target. 2003, 11, 391−406.
̈
ller, K. Modulation of pharmacologically relevant
properties of piperidine derivatives by functional groups in an
equatorial or axial β-position to the amino group. ChemBioChem
[Online early access]. DOI: 10.1002/cbic.201900474. Published
Online: Sep 6, 2019. https://onlinelibrary.wiley.com/doi/abs/10.
1002/cbic.201900474 (accessed Nov 26, 2019).
(43) Bruns, A.; Mueggler, T.; Kunnecke, B.; Risterucci, C.; Prinssen,
E. P.; Wettstein, J. G.; von Kienlin, M. “Domain gauges”: A reference
system for multivariate profiling of brain fMRI activation patterns
induced by psychoactive drugs in rats. NeuroImage 2015, 112, 70−85.
(44) Paxinos, G., Watson, C., Eds. The Rat Brain in Stereotaxic
Coordinates; Academic Press: San Diego, CA, 1998.
(28) Fish, P. V.; Barta, N. S.; Gray, D. L.; Ryckmans, T.; Stobie, A.;
Wakenhut, F.; Whitlock, G. A. Derivatives of (3S)-N-(biphenyl-2-
ylmethyl)pyrrolidin-3-amine as selective noradrenaline reuptake
inhibitors: Reducing P-gp mediated efflux by modulation of H-bond
acceptor capacity. Bioorg. Med. Chem. Lett. 2008, 18, 4355−4359.
(29) El Tayar, N.; Testa, B.; Carrupt, P. A. Polar intermolecular
interactions encoded in partition coefficients: an indirect estimation
of hydrogen-bond parameters of polyfunctional solutes. J. Phys. Chem.
1992, 96, 1455−1459.
(45) Hennig, J.; Nauerth, A.; Friedburg, H. RARE imaging: a fast
imaging method for clinical MR. Magn. Reson. Med. 1986, 3, 823−
833.
(46) Haase, A.; Frahm, J.; Matthaei, D.; Hanicke, W.; Merboldt, K.
D. FLASHimaging - rapid NMRimaging using low flip-angle pulses. J.
Magn. Reson. 1986, 67, 258−266.
(47) Williams, D. S.; Detre, J. A.; Leigh, J. S.; Koretsky, A. P.
Magnetic resonance imaging of perfusion using spin inversion of
arterial water. Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 212−216.
(48) Deichmann, R.; Hahn, D.; Haase, A. Fast T1 mapping on a
whole-body scanner. Magn. Reson. Med. 1999, 42, 206−209.
(49) Alsop, D. C.; Detre, J. A. Reduced transit-time sensitivity in
noninvasive magnetic resonance imaging of human cerebral blood
flow. J. Cereb. Blood Flow Metab. 1996, 16, 1236−1249.
(50) Benjamini, Y.; Hochberg, Y. Controlling the false discovery
rate: a practical and powerful approach to multiple testing. J. R. Statist.
Soc. B 1995, 57, 289−300.
(30) Berthelot, M.; Laurence, C.; Safar, M.; Besseau, F. Hydrogen-
bond basicity pKHB scale of six-membered aromatic N-heterocycles.
J. Chem. Soc., Perkin Trans. 2 1998, 283−290.
(31) Kenny, P. W.; Montanari, C. A.; Prokopczyk, I. M.; Ribeiro, J.
F.; Sartori, G. R. Hydrogen bond basicity prediction for medicinal
chemistry design. J. Med. Chem. 2016, 59, 4278−4288.
(32) Abraham, M. H.; Duce, P. P.; Prior, D. V.; Barratt, D. G.;
Morris, J. J.; Taylor, P. J. Hydrogen bonding. Part 9. Solute proton
donor and proton acceptor scales for use in drug design. J. Chem. Soc.,
Perkin Trans. 2 1989, 1355−1375.
(33) Hitchcock, S. A. Structural modifications that alter the P-
glycoprotein efflux properties of compounds. J. Med. Chem. 2012, 55,
4877−4895.
(34) Berthelot, M.; Besseau, F.; Laurence, C. The hydrogen-bond
basicity pKHB scale of peroxides and ethers. Eur. J. Org. Chem. 1998,
1998, 925−931.
(35) Bruns, A.; Kunnecke, B.; Risterucci, C.; Moreau, J. L.; von
Kienlin, M. Validation of cerebral blood perfusion imaging as a
modality for quantitative pharmacological MRI in rats. Magn. Reson.
Med. 2009, 61, 1451−1458.
(36) Smith, C. J.; Poehlmann, M. L.; Li, S.; Ratnaseelan, A. M.;
Bredewold, R.; Veenema, A. H. Age and sex differences in oxytocin
and vasopressin V1a receptor binding densities in the rat brain: focus
on the social decision-making network. Brain Struct. Funct. 2017, 222,
981−1006.
(37) O’Connell, L. A.; Hofmann, H. A. The vertebrate mesolimbic
reward system and social behavior network: a comparative synthesis. J.
Comp. Neurol. 2011, 519, 3599−3639.
(38) Benekareddy, M.; Stachniak, T. J.; Bruns, A.; Knoflach, F.; von
Kienlin, M.; Kunnecke, B.; Ghosh, A. Identification of a cortico-
habenular circuit regulating socially directed behavior. Biol. Psychiatry
2018, 83, 607−617.
(39) Wu, G.; Huang, M.; Richards, M.; Poirier, M.; Wen, X.; Draper,
R. W. Novel ZnX2-modulated Pd/C and Pt/C catalysts for
chemoselective hydrogenation and hydrogenolysis of halogen-
O
https://dx.doi.org/10.1021/acs.jmedchem.9b01478
J. Med. Chem. XXXX, XXX, XXX−XXX