Discovery of highly selective brain-penetrant vasopressin 1a antagonists for the potential treatment of autism via a chemogenomic and scaffold hopping approach

Article abstract of DOI:10.1021/jm501745f

From a micromolar high throughput screening hit 7, the successful complementary application of a chemogenomic approach and of a scaffold hopping exercise rapidly led to a low single digit nanomolar human vasopressin 1a (hV1a) receptor antagonist 38. Initial optimization of the mouse V1a activities delivered suitable tool compounds which demonstrated a V1a mediated central in vivo effect. This novel series was further optimized through parallel synthesis with a focus on balancing lipophilicity to achieve robust aqueous solubility while avoiding P-gp mediated efflux. These efforts led to the discovery of the highly potent and selective brain-penetrant hV1a antagonist RO5028442 (8) suitable for human clinical studies in people with autism.

Full text of DOI:10.1021/jm501745f

Subscriber access provided by SELCUK UNIV
Article
Discovery of highly selective brain-penetrant vasopressin
a antagonists for the potential treatment of autism
1
via a chemogenomic and scaffold hopping approach
Hasane Ratni, Mark Rogers-Evans, Caterina Bissantz, Christophe Grundschober, Jean-
Luc Moreau, Franz Schuler, Holger Fischer, Ruben Alvarez-Sanchez, and Patrick Schnider
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 05 Feb 2015
Downloaded from http://pubs.acs.org on February 6, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of highly selective brainꢀpenetrant
vasopressin 1a antagonists for the potential treatment
of autism via a chemogenomic and scaffold hopping
approach
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Hasane Ratni*, Mark Rogers-Evans*, Caterina Bissantz*, Christophe Grundschober, Jean-Luc
Moreau, Franz Schuler, Holger Fischer, Ruben Alvarez Sanchez and Patrick Schnider
Pharmaceutical Research and Early Development, Roche Innovation Center Basel, F. Hoffmannꢀ
La Roche Ltd, Grenzacherstrasse 124, 4070 Basel, Switzerland
ABSTRACT
From a micromolar high throughput screening hit 7, the successful complementary application of
a chemogenomic approach and of a scaffold hopping exercise rapidly led to a low single digit
nanomolar human vasopressin 1a (hV1a) receptor antagonist 38. Initial optimization of the
mouse V1a activities delivered suitable tool compounds which demonstrated a V1a mediated
central in vivo effect. This novel series was further optimized through parallel synthesis with a
focus on balancing lipophilicity to achieve robust aqueous solubility while avoiding Pꢀgp
mediated efflux. These efforts led to the discovery of the highly potent and selective brainꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
penetrant hV1a antagonist RO5028442 (8) suitable for human clinical studies in people with
autism.
INTRODUCTION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Oxytocin and vasopressin are evolutionarily highly conserved 9ꢀamino acid cyclic
peptides, differing only by 2 amino acids. Both peptides are known to play an important role in
1
the regulation of social behavior in animals and humans. Vasopressin release in the rodent brain
is increased during stress, induced by social defeat or forced swim test, and causes a passive
2
coping behavior in the later. Peptidic vasopressin 1 (V1) receptor antagonists injected in the
2
amygdala reduce passive coping behavior, like antidepressant drugs, and reduce anxiety in the
3
elevated plus maze when injected in the septum. In human studies, using BOLD magnetic
resonance imaging during a face matching task, intranasal vasopressin administration has been
shown to modulate the activity of the cingulate cortex and to reduce the connectivity between
subgenual and supragenual cingulate. This work points to a potential neural network linking
vasopressin to the modulation of social behavior by changing a subgenual, supragenual cingulate
4
and amygdala negative feedback loop. This indicates that vasopressin is increasing the brain
response to socially threatening stimuli in humans, in line with the increased threat perception of
5
normal faces after intranasal vasopressin administration.
Three vasopressin G protein coupled receptors are known. V1a, V1b are expressed in rat
brain limbic areas, like hypothalamus, amygdala, septum and hippocampus. V1a receptors are
6
7
expressed in similar human and monkey brain regions with additional expression in cortical
areas in comparison to rats. In the periphery V1a is expressed in kidney, liver, platelets and
8
9
vascular smooth muscle and V1b in pituitary, heart and lung . V2 is mainly expressed in the
ACS Paragon Plus Environment

Page 3 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8a
kidney, mediating the antidiuretic effects of vasopressin. The oxytocin receptor is closely
1
0
1
related to the vasopressin receptors and mediates the reported proꢀsocial effects of oxytocin. In
addition, oxytocin has well known peripheral effects on uterine contraction during parturition
and the milk ejection reflex during lactation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A brain penetrant V1a antagonist may therefore have antidepressant and anxiolytic
properties, as well as proꢀsocial effects by modulating the social brain and may thus have
potential for the treatment of psychiatric disorders with social emotional dysfunction, including
autism, anxiety disorders and schizophrenia. It is important for such a compound not to block the
V2 and oxytocin receptors to avoid peripheral sideꢀeffects and not to counteract the proꢀsocial
effects of oxytocin.
Despite the therapeutic potential of modulating the vasopressin system in the brain, the
11
number of selective small molecule V1a receptor antagonists reported so far is limited.
12
Compounds which progressed to clinical trials included relcovaptan (SR49059, 1),
a
peripherally active molecule from SanofiꢀAventis which has shown initial positive results in the
1
3
14
treatment of Raynaud’s disease, dysmenorrhoea, and tocolysis, PFꢀ184536 (2) from Pfizer
targeted on the treatment of dysmenorrhea, and two CNS penetrating compounds from Azevan,
15
16
SRXꢀ246 (3) and SRXꢀ251 (4) (Figure 1). In a human fMRI study 3 blocked the effect of
intranasal vasopressin on the neural response to angry faces. Secondary analyses revealed
furthermore that 3 treatment was associated with significantly attenuated BOLD responses to
15c
angry faces in the right temporoparietal junction, precuneus, anterior cingulate, and putamen.
In 2014 a 12ꢀweek Phase II clinical trial with 3 for the treatment of Intermittent Explosive
1
7
Disorder was launched. At a preꢀclinical stage, Johnson and Johnson demonstrated in vivo
18
efficacy of 5 (JNJꢀ17308616) in a rat model of anxiety. More recently, MSD disclosed a novel
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
series of CNS penetrant V1a antagonists exemplified by 6 with a good affinity for the rat
1
9
receptor. Peripheral in vivo functional V1a antagonism was demonstrated by the reversal of
V1a mediated arginine vasopressin (AVP) induced blood pressure increases in conscious rats.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Relcovaptan
SR49059, 1
PF-184536, 2
SRX-246, 3
SRX-251, 4
JNJ-17308616, 5
6
HTS hit, 7
RO5028442, 8
Figure 1. Structure of compounds 1ꢀ6, HTS hit 7 and 8 (RO5028442)
To identify novel starting points for the discovery of an orally bioavailable, CNS
penetrant and selective V1a antagonist ligand, we performed a high throughput screening (HTS)
campaign of our library using a functional FLIPR assay. We screened around seven hundred
thousand compounds at a single concentration of 10ꢀM and obtained a hit rate of 1.48%.
Binding affinities of the confirmed hits were determined on both human V1a (hV1a) and human
V1b (hV1b) with a Scintillation Proximity Assay (SPA). All hits identified displayed a much
ACS Paragon Plus Environment

Page 5 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
greater affinity for the V1a rather than the V1b receptor. The binding selectivity versus human
V2 and oxytocin receptors (hOTR) were also determined. Compound 7 which had binding
affinities of 2830 nM and 12000 nM for hV1a and hV1b receptors, respectively, particularly
caught our attention due to its high chemical tractability. An efficient combination of
chemogenomic and parallel library synthesis methods with a focus on optimization of in vitro
affinity at the human V1a receptor as well as physicochemical and DMPK properties quickly led
to the discovery of 8 suitable for human clinical studies (Figure 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Chemistry.
The synthesis of the derivatives described in Table 2 was performed in three steps
2
0
starting from the commercially available 6ꢀchloroꢀ1Hꢀindole 9 (Scheme 1). Regioselective
acetylation using trifluoroacetic anhydride in DMF yielded exclusively the substituted 3ꢀ
2
1
trifluoroacetylindole derivative 10. This reaction was readily performed on a 50ꢀgram scale
with 83% isolated yield. We found that this procedure could be applied to numerous electron rich
or poor Nꢀunsubstituted indoles, whereas azaꢀindoles and Nꢀsubstituted indoles could only be
acylated when activated with electron donating substituents. Upon treatment with an aqueous
solution of sodium hydroxide at reflux, the corresponding substituted indole 3ꢀcarboxylic acid 11
was obtained. This versatile key intermediate was ultimately prepared on a hundredꢀgram scale.
Finally, derivatives 12ꢀ23 were obtained in high yields by standard amide coupling upon reaction
of 11 with a range of amines (called ‘head groups’ herein after), commercial or readily prepared,
such as spiropiperidines, piperidines or piperazines in the presence of HOBt and EDC.
Alternatively, the carboxylic acid moiety was converted into an acid chloride prior to reaction
with the amines to give 12ꢀ23.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 1. Synthesis of compounds described in Table 2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
c, d
or e
9
10
11
12-23
a
Reagents and conditions: (a) TFAA, DMF, RT, 1h; (b) aq. NaOH, 70 ºC, 16h; (c) oxalyl
chloride, DMF, THF; (d) (R) NH, Et N, RT; (e) (R) NH, HOBt, EDC, CH Cl , RT.
2
3
2
2
2
Subsequent elaboration of Nꢀsubstituted indole was easily performed starting from 12 to
2
0b
provide compounds described in Table 3 (Scheme 2). Compounds 24 and 25 were obtained by
indole Nꢀalkylation with the corresponding commercially available 2ꢀchloroacetamides.
Alternatively, sequential Nꢀalkylation with ethylꢀ2ꢀbromoacetate followed by hydrolysis of the
ester gave 26, which underwent amide coupling with N',N'ꢀdimethylethaneꢀ1,2ꢀdiamine to form
27.
a
Scheme 2. Synthesis of compounds described in Table 3
a
b, c
d
2
4, 25
12
26
27
ACS Paragon Plus Environment

Page 7 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: (a) NaH, R R NCOCH Cl, DMF (b) NaH, DMF, ethylꢀ2ꢀ
1
2
2
bromoacetate, RT; (c) NaOH, H O, EtOH, RT; (d) R R NH, HOBt, EDC, CH Cl , RT.
2
1
2
2
2
The synthesis of the derivatives described in Table 4 started as well from the Nꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0b
unsubstituted indole 12 (Scheme 3).
A straightforward Nꢀalkylation with 2ꢀchloroꢀN,Nꢀ
dimethylꢀethanamine led to 8, which progressed to clinical studies. This compound was prepared
in a total of only four highꢀyielding steps from the commercially available 6ꢀchloro indole. The
2
2
other derivatives were prepared via an alkylation with the corresponding mesylate electrophile
followed by a reductive amination with aqueous formaldehyde.
a
Scheme 3. Synthesis of compounds described in Table 4
a
b, c
d
8
12
28
29
e, c
d
3
0
31
a
Reagents and conditions: (a) NaH, 2ꢀchloroꢀN,N-dimethylꢀethanamine.HCl, DMF; (b) NaH,
tertꢀbutyl (2S)ꢀ2ꢀ(methylsulfonyloxymethyl)pyrrolidineꢀ1ꢀcarboxylate, DMF; (c) TFA, CH Cl ,
RT; (d) aq. HCHO, AcOH, NaBH CN, MeOH; (e) Cs CO , tertꢀbutyl 4ꢀ
2
2
3
2
3
methylsulfonyloxypiperidineꢀ1ꢀcarboxylate, DMF.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Lead optimization.
Chemogenomics allows discovering new ligands for a protein based on a comparison of
its binding site with binding sites of other proteins, taking advantage of the idea that proteins
with similar binding sites should bind similar ligands. This is a complementary approach to the
HTS, which allowed us to virtually screen and identify compounds that were initially not part of
an HTS screen. In our search for new antagonists of the V1a receptor we thus clustered all
human class A GPCRs based on their binding site similarity with the goal to find another GPCR
with a highly similar binding site for which ligands are already known. According to this
hypothesis these ligands should have a high probability to also bind to the V1a receptor.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To create the dataset for the chemogenomics approach, all nonꢀolfactory, human class A
2
3
GPCR receptor sequences were retrieved from the UniProt database. Fragments and sequences
were discarded for which not exactly seven transmembrane domains were detected, leading to a
24
total of 298 sequences. We aligned the sequences by a previously published method, and
2
5
extracted the 35 residues that form the transmembrane pocket (Ligand binding Pocket Vector).
We then encoded all residues by descriptors. As descriptors we used the ZZꢀscales.
26
Additionally, each residue was described by a set of 5 pharmacophoric descriptors (hydrophobic,
aromatic, Hꢀbond donor and/or acceptor, positively ionizable, negatively ionizable). Thus each
GPCR was described by a fingerprint of 350 descriptors. Principal Component Analysis was
carried out using the SIMCA+10.5 software (Umetrics AB, Umea, Sweden).
Historically, GPCRs are clustered into families based on their natural ligands. GPCRs
binding the same natural ligand should have similar binding sites. We thus expect that in our
bindingꢀsite based clustering, receptors of the same families form clusters (this concept has been
validated inꢀhouse). This matrix was applied to the search for new V1a antagonists. Figure 2
ACS Paragon Plus Environment

Page 9 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
shows a 3D representation of the matrix around the V1a receptor, and Table 1 shows the distance
of the V1a receptor to its closest neighbors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
V1b
V2
Oxytocin
Figure 2. Different colors of the markers represent different GPCR families. The principal
components PC1, PC2 and PC3 represent different properties of the binding sites. The largest
weight in PC1 have the properties of the residues in the positions 2.58, 6.51, 7.40 and 3.32
according to the Weinstein nomenclature. The largest weight in PC2 have properties in position
5
.47, 7.43 and 5.39. Properties of residues 5.47 and 5.39 have a major weight in PC3. However,
overall each PC is a mixture of properties of all residues.
Table 1. Nearest neighbor list (top 10 neighbors) of the human V1A receptor.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
GPCR name (family)
GP173 (orphan, SREB)
Distance
0.715
0.899
0.901
0.924
0.930
0.937
0.950
0.957
1.012
1.041
MC5R (melanocortin)
Q8NEN2 (orphan, SREB)
V1BR (vasopressinꢀlike)
V2R (vasopressinꢀlike)
MSHR (melanocortin)
OXYR (vasopressinꢀlike)
MC3R (melanocortin)
NK1R (tachykinin)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
GPR21 (orphan)
Additionally, two orphan GPCRs are located closely to the vasopressin family (GP173
and Q8NEN2 of the SREB family). From this nearest neighbor list, nonꢀpeptidic small molecule
ligands are mainly known for the NK1 receptor, a member of the family of mammalian
tachykinin receptors, and to a smaller extent for the MSHR receptor. Due to this fact and the inꢀ
house availability of different NK1 antagonist series, we decided to select a diverse set of 31
NK1 antagonists for biological testing which were not part of the library we used in the HTS.
Nine compounds proved to possess hV1a affinity better than 10000 nM, two of them showed an
affinity better than 1000 nM, one of them, 32 had an 81 nM affinity. As this novel hit 32
originated from a NK1 program, we will in course of its optimization for hV1a affinity, monitor
closely its side activity on both the NK receptors and the closely associated melanocortin family.
Interestingly, the most potent compound differs from our HTS hit 7 only by the substitution of
the 4ꢀbenzylꢀ4ꢀhydroxypiperidine with an isomeric spiro analog, leading to a 30ꢀfold increase in
in vitro hV1a affinity and a marked increase in ligand efficiency from 0.26 to 0.33 (compound
32, Figure 3). While 7 showed weak affinity for the hV1b receptor, no affinity could be
measured for 32. All compounds tested during the course of optimization of this lead were tested
inactive at the hV1b receptor.
ACS Paragon Plus Environment

Page 11 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HTS hit, 7
hV1a Ki: 2830 nM
hV1b Ki: 12000 nM
32
hV1a Ki: 81 nM
hV1b Ki: inactive
Figure 3. Identification of a novel ’head group‘ via chemogenomics.
In parallel to our chemogenomic approach, we explored the scope of our HTS hit by
replacing the 2ꢀethoxynaphtaleneꢀ1ꢀcarboxylic acid part in search of further enhanced in vitro
affinity (Figure 4). Removal of the ethoxy group (33) led to a moderate loss of affinity, which
can be rationalized by the minor impact of this substituent on the torsional angle between the
naphthalene moiety and the amide plane. We next examined the importance of a fused biaryl
system by synthesizing a small subset of substituted phenyl derivatives such as 34. However, a
complete loss of affinity was encountered, highlighting the importance of the second aromatic
ring. Finally, replacement of the naphthalene moiety with an Nꢀbenzylꢀ3ꢀindole fragment led to a
tenfold hV1a affinity improvement (e.g. 35).
HTS hit, 7
hV1a Ki: 2830 nM
33
34
hV1a Ki: 6620 nM
hV1a Ki: Inactive
3
5
36
37
hV1a Ki: 4990 nM
hV1a Ki: 295 nM
hV1a Ki: 3900 nM
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 4. Naphthalene scaffold replacement
From the initial hit, the exchange of the head group by a spiropiperidine or the
naphthalene replacement by a substituted indole moiety led to a thirtyꢀ and tenꢀfold in vitro
affinity gain, respectively. Gratifyingly, the combination of these two fragments proved to be
synergistic, resulting in the low nanomolar V1a antagonist lead 38. Starting from an HTS hit
with a micromolar affinity, we were thus able to increase the in vitro affinity by more than three
log units by preparing only two dozen derivatives (Figure 5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
2
hV1a Ki: 81 nM
HTS hit, 7
hV1a Ki: 2830 nM
38
hV1a Ki: 2 nM
3
5
hV1a Ki: 295 nM
Figure 5. From the micromolar HTS hit 7 to the single digit nanomolar V1a antagonist 38
ACS Paragon Plus Environment

Page 13 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Although a high affinity on hV1a was achieved with 38, this compound suffered from
poor aqueous solubility (< 1 ꢀg/mL at pH 6.5) and excessive lipophilicity (kow_cLogP = 5.7,
LogD not measurable). Comparison of this compound with the original HTS hit and its close
analog 33 (Figure 4) containing only a bare naphthalene moiety suggested that neither the benzyl
nor the methyl substituent on the indole were critical (Figure 7). Although the in vitro affinity of
the unsubstituted indole derivative 39 was reduced, the ligand efficiency (LE) and the lipophilic
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
7
ligand efficiency (LLE), efficient metrices for lead optimization,
as well as the
physicochemical properties were markedly improved. Further fine tuning of the indole core
restored the in vitro affinity upon introduction of a chlorine atom in the 6ꢀposition, with
compound 12 displaying low nanomolar hV1a affinity and further increased ligand efficiency,
high selectivity versus V2 and oxytocin receptors (> 6750 and 1881ꢀfold, respectively) and an
improved but not yet satisfactory physicoꢀchemical profile (LogD and solubility) (Figure 6).
3
8
39
12
hV1a Ki: 2 nM
hV1a Kb: 3.6 nM
mV1a Ki: 210 nM
hV2 Ki: 9308 nM
hOT Ki: Inactive
kow_cLogP: 5.7
LogD: >3
hV1a Ki: 32 nM
mV1a Ki: 6620 nM
kow_cLogP: 2.9
LogD: > 3
Solubility: 17 mg/L
LE: 0.41
hV1a Ki: 4 nM
hV1a Kb: 2.9 nM
mV1a Ki: 1012 nM
hV2 Ki: > 27000 nM
hOT Ki: 7524 nM
kow_cLogP: 3.7
LogD: 3.7
LLE: 4.6
Solubility: <1 mg/L
LE: 0.36
Solubility: 4 mg/L
LE: 0.44
LLE: 3.1
LLE: 4.8
Figure 6. Optimization of binding and lipophilic ligand efficiency
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
However, the large species difference between hV1a and mV1a affinity for 12 precluded
any in vivo behavioral assessment in mice. In search of derivatives with increased mouse V1a
affinity, a wide range of analogs with an alternative head group was therefore prepared (Table
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2).
Table 2: Head group evaluation
N
O
R
N
Compd
12
4
13
5
14
15
39
16
17
16
hV1a Ki (nM)
hV1a Kb (nM)
mV1a Ki (nM)
9
5
5
2.9
1012
1.8
1018
4.2
7.8
72
2980
4816
Inactive
Inactive
R
Compd
18
27
19
24
20
32
38
21
8
22
23
23
118
193
1972
hV1a Ki (nM)
hV1a Kb (nM)
mV1a Ki (nM)
20
11
14
43
5890
5886
Inactive
2395
1491
Ki: Binding affinity measured in a radioligand competition binding assay. Kb: Apparent affinity
measured in a functional antagonist calciumꢀflux assay
ACS Paragon Plus Environment

Page 15 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
While high hV1a affinity was achieved with a large diversity of head groups, no
improvement of the affinity for neither the human nor the mouse V1a receptor over the spiro[1Hꢀ
isobenzofuranꢀ3,4'ꢀpiperidine] derivative 12 was achieved. We therefore conserved this head
group in our next round of optimization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Interestingly, it was reported that 5 has a high human (5 nM) and much weaker mouse
1
8
and rat V1a receptor affinity (428 and 216 nM, respectively). However, this species difference
for binding affinity was smaller compared to 12. Furthermore, the modest rodent V1a affinity for
1
8
5 was sufficient to demonstrate efficacy in a behavior paradigm of anxiety. A V1a homology
model based on the Xꢀray structure of the bovine rhodopsine receptor was constructed which
allowed us to perform ligandꢀbased alignments within the binding pocket between 12 and 5
2
8
(
Figure 7). The first model was designed by a previously published method. Consequently 5
was manually docked into the binding pocket such that it binds in a low energy conformation.
The resulting ligandꢀreceptor complex was consequently minimized. Finally, 12 was docked into
the optimized binding pocket in a low energy conformation as well. The resulting ligand
alignment clearly indicated the presence of a binding pocket filled by the 2ꢀ(N,N-
dimethylamino)ethyl amide substituent of 5 but not with 12. The model suggested that the indole
nitrogen of 12 is perfectly oriented for the introduction of a substituent to efficiently occupy this
pocket. We also tried to rationalized the species affinity difference with this model. However, we
could not link the observed differences to one of the specific residue exchanges between the
human and mouse receptor sequence.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Q333
Q311
E54
A334
D112
A335
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F307
F308
T221
Q108
S338
Q131
N340
Q104
W304
I224
Figure 7. Alignment of 39 and 5 guided by their predicted docking poses to the V1a receptor
homology model. Carbon atoms of 39 are depicted in cyan, carbon atoms of 5 in green and those
of the receptor in grey. Oxygen atoms are depicted in red, nitrogen atoms in blue.
The introduction of polar acetamide substituents was readily achieved by indole N-
alkylation. Gratifyingly, this did not only lead to the anticipated increase in solubility but also to
a marked improvement of the mouse V1a binding affinity to as low as 14 nM, while the high
affinity at the human receptor was retained (Table 3). However, these compounds turned out to
be strong substrates of both human and mouse Pꢀglycoprotein (Pꢀgp) . Since all showed high
passive permeability, Pꢀgp mediated efflux likely accounts for the poor brain penetration
observed in mice.
ACS Paragon Plus Environment

Page 17 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3: Evaluation of polar amide substituents
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
HN
N
R
H
Compd
hV1a Ki (nM)
hV1a Kb (nM)
12
4
2.9
24
0.5
0.1
25
1
0.2
27
2
0.7
mV1a Ki (nM)
1012
> 6750
1881
4
weak
n.d.
n.d.
n.d.
14
41
177
Binding selectivity vs hV2
Binding selectivity vs hOTR
Solubility (ꢀg/mL)
hPꢀgp transport
Brain / plasma
> 60000
9160
56
strong
0.05
1.2
11850
5690
193
n.d.
0.07
1.5
> 15000
2214
> 650
strong
0.1
Vss (l/kg)
Cl (mL/min/kg)
4.9
14
40
39
Ki: Binding affinity measured in a radioligand competition binding assay. Kb: Apparent affinity
measured in a functional antagonist calciumꢀflux assay. Brain/Plasma concentration ratio in
mouse @ 10 mg/kg PO. Volume of distribution (Vss) and plasma clearance (Cl) in mouse @ 2
mg/kg i.v.
Central AVP administration stimulates V1a, V1b and oxytocin receptors and induces
2
9
30
scratching behavior in mice. In V1a knockout mice AVP does not induce scratching. In order
to probe the potential of our new class of V1a receptor antagonists to antagonize brain V1a
receptors, 25 was administered to mice i.c.v. prior to the treatment with AVP. Gratifyingly 25
was found to doseꢀdependently suppress scratching. Based on its superior pharmacokinetic
profile due to the lower clearance and higher volume of distribution compared to 24 and 25 we
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
consequently selected 27 for peripheral i.p. administration. In spite of its poor brain penetration,
27 also showed a doseꢀdependent suppression of AVP induced scratching (Figure 8).
1
00
75
50
25
0
1
00
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
e
s
n
o
p
s
75
e
s
n
o
p
s
*
e
r
5
0
x
*
e
r
a
m
f
x
* *
a
o
25
0
*
m
f
%
o
%
veh
0.1
0.3
1
νg
veh
0.3
1
3
10
30 mg/kg
25; (i.c.v. administration)
2
7; (i.p. administration)
Figure 8. Inhibition AVPꢀinduced scratching by 25 and 27
Encouraged by these promising pharmacodynamic results, we were seeking derivatives
which are devoid of Pꢀgp mediated efflux while retaining a high human V1a affinity, selectivity
and solubility (Table 4). HꢀBonding capacity has been described to be a key parameter
determining Pꢀgp mediated efflux. We thus envisioned to replace the aminoalkyl acetamide N-
indole substituents containing two strong Hꢀbond acceptors by aminoalkyl residues comprising
only one Hꢀbond acceptor. Derivatives with secondary amine substituents were found to be
potent hV1a receptor antagonists, although a drop of affinity at the mV1a receptor was often
observed. Unfortunately, as exemplified by pyrrolidine derivative 28, they were also found to be
strong substrates of both human and mouse Pꢀgp resulting in poor brain penetration. In contrast,
Nꢀalkylation to give rise to the corresponding less strongly Hꢀbonding tertiary amines resulted in
derivatives which are devoid of Pꢀgp mediated efflux and consequently showed excellent brain
penetration. Compounds 29, 31 and 8 have excellent binding and functional affinity on hV1a,
moderate mouse affinity and excellent selectivity versus hV2 and hOT receptors. All compounds
ACS Paragon Plus Environment

Page 19 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
showed high solubility. The N,Nꢀdimethylaminoethyl derivative 8 was found to be highly
3
1
selective against a panel of 89 targets. Finally, 8 was identified as a suitable compound for
clinical studies. The entry into human enabling and clinical studies in people with autism with
this compound will be reported elsewhere in due course.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4: Evaluation of polar amine substituents
R
Compd
hV1a Ki (nM)
28
1.4
29
1
0.5
13
6071
1975
140
weak
2
31
4
3.3
8
1
2.6
hV1a Kb (nM)
1.4
mV1a Ki (nM)
125
11
39
Binding selectivity vs hV2
Binding selectivity vs hOTR
Solubility (mg/L)
hPꢀgp transport
Brain / plasma
> 19286
1514
108
strong
0.05
6.8
> 6750
1196
12
weak
1
> 30000
9891
205
weak
1.4
Vss (l/kg)
2.6
67
7.2
52
5.8
91
Cl (mL/min/kg)
32
Ki: Binding affinity measured in a radioligand competition binding assay. Kb: Apparent affinity
measured in a functional antagonist calciumꢀflux assay. Brain/Plasma concentration ratio in
mouse @ 10 mg/kg PO. Volume of distribution (Vss) and plasma clearance (Cl) in mouse @ 2
mg/kg i.v.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SUMMARY AND CONCLUSIONS
A brain penetrant V1a antagonist may have antidepressant and anxiolytic properties, as
well as proꢀsocial effects by modulating the social brain. It is important for such a compound not
to block the V2 and oxytocin receptors to avoid peripheral sideꢀeffects and counteracting the proꢀ
social effects of oxytocin. In our search for an orally bioavailable potential drug fulfilling these
criteria, we performed a high throughput screening (HTS) campaign of our library.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Among the few weakly active HTS hits, 7 (hV1a Ki =2830 nM) was considered the most
promising. An efficient chemogenomic approach allowed us to identify a novel head group,
where the substitution of the 4ꢀbenzylpiperidinꢀ4ꢀol with the isomeric spiropiperidine led to a 30ꢀ
fold increase in in vitro affinity as with 32. Concomitantly, the replacement of the southern part
of the hit, the naphthalene moiety by a substituted indole, led to 35 with a 10ꢀfold in vitro affinity
improvement. The combination of these two fragments proved to be synergistic resulting in the
singleꢀdigit nanomolar hV1a antagonist 38. The affinity of the HTS hit was thus improved by
more than three log units after preparing only two dozen derivatives.
V1a antagonists have often been reported to suffer from a large discrepancy between
human and rodent affinity. This was also the case for our novel class of indoleꢀ3ꢀcarboxamide
V1a antagonists. Efforts to improve mouse V1a affinity while keeping an overall balanced
profile met with only partial success. Nevertheless, we were able to demonstrate in vivo central
target engagement with tool compounds (25 and 27). No compounds from this class were further
tested in any in vivo model of autism.
Encouraged by these promising pharmacodynamic results, we were seeking derivatives
which are devoid of Pꢀgp mediated efflux while retaining high affinity on the human V1a
ACS Paragon Plus Environment

Page 21 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
receptor, selectivity and solubility. These efforts ultimately led to the discovery of 8 which was
32
identified as suitable for clinical studies in people with autism. The entry into human enabling
and clinical studies in autistic subjects with this compound will be reported elsewhere in due
course.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Compound Synthesis and Characterization. Chemistry.
Reactions were carried out under argon atmosphere. Unless otherwise mentioned, all reagents
and chemicals were obtained from commercial suppliers and used without further purification.
All reactions were followed by TLC (TLC plates F254, Merck) or LCMS (liquid
chromatographyꢀmass spectrometry) analysis. The purity of final compounds as measured by
HPLC was at least above 95%. Flash column chromatography was carried out either using
cartridges packed with silica gel (Isolute Columns, Telos Flash Columns) or on glass columns on
silica gel 60 (32ꢀ60 mesh, 60Å). LC high resolution spectra were recorded with a Agilent LCꢀ
system consisting of Agilent 1290 high pressure system, a CTC PAL auto sampler and a Agilent
6520 QTOF. The separation was achieved on a Zorbax Eclipse Plus C18 1,7 µm 2.1*50mm
column at 55°C; A=0.01% formic acid in Water; B= 0.01% formic acid in acetonitrile at flow 1
mL/min. gradient: 0 min 5%B, 0.3 min 5%B, 4.5 min 99 %B 5 min 99%B. The NMR spectra
were measured on a Bruker 600 MHz machine in a 5 mm TCI cryoprobe at 298 K. TMS was
used for referencing.
[6-Chloro-1-[2-(dimethylamino)ethyl]indol-3-yl]-spiro[1H-isobenzofuran-3,4'-
piperidine]-1'-yl-methanone (8). To a solution of (6ꢀchloroꢀ1Hꢀindolꢀ3ꢀyl)ꢀspiro[1Hꢀ
isobenzofuranꢀ3,4'ꢀpiperidine]ꢀ1'ꢀylꢀmethanone 12 (50 mg, 0.136 mmol) in DMF (4 mL) cooled
at 0 °C, was added sodium hydride dispersion (18 mg, ~55% in oil, 0.41 mmol) and after 0.5h, 2ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chloroꢀN,N-dimethylꢀethanamine hydrochloride (39 mg, 0.34 mmol). After 12h at RT, aqueous
NH Cl (2 mL) was added and the reaction mixture was concentrated under vacuum. The residue
4
was partitioned between ethyl acetate and water. The organic portion was dried over Na SO ,
2
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
filtered and concentrated. The crude product was purified by column chromatography, eluting
1
with ethyl acetate to produce 8 (51 mg, 84%) as a white solid. H NMR (600 MHz,
CHLOROFORMꢀd) δ 7.68 (d, J=8.56 Hz, 1H), 7.55 (s, 1H), 7.36 (d, J=1.61 Hz, 1H), 7.27ꢀ7.32
(
m, 2H), 7.21ꢀ7.25 (m, 1H), 7.18 (dd, J=1.81, 8.46 Hz, 1H), 7.11ꢀ7.14 (m, 1H), 5.11 (s, 2H),
4.41 (br s, 2H), 4.18 (t, J=6.90 Hz, 2H), 3.46 (br s, 2H), 2.71 (t, J=6.85 Hz, 2H), 2.30 (s, 6H),
.92 (br s, 2H), 1.79 (br d, J=12.89 Hz, 2H); LCꢀHRMS: m/z = 438.1969 [(M+H)+ calculated
for C H ClN O = 438.1943; Diff = 2.6 mD].
1
25 28
3
2
1-(6-Chloro-1H-indol-3-yl)-2,2,2-trifluoro-ethanone (10). To a solution of 6ꢀchloroꢀ
Hꢀindole 9 (15.0 g, 98.9 mmol) in DMF (150 mL) at 0 °C, was added drop wise trifluoroacetic
1
anhydride (15.8 mL, 114 mmol). After 2 h, another portion of trifluoroacetic anhydride (15.8
mL, 114 mmol) was added, and stirring was continued for an additional 0.5 h. An aqueous
saturated sodium carbonate solution (500 mL) was added to the reaction mixture and the product
was extracted with three portions (250 mL × 3) of ethyl acetate. The combined organic layers
were
dried
over
sodium
sulfate
and
concentrated
in
vacuo.
The residue was triturated in tert.ꢀbutyl methyl ether and a filtration produced 10 (20.4 g (82%)
as a white solid. The mother liquor was concentrated in vacuo and triturated in tert.ꢀbutyl methyl
1
ether followed by filtration gave another portion of 10 (1.2 g, 5%). H NMR (600 MHz,
CHLOROFORMꢀd) δ 8.77ꢀ9.01 (m, 1H), 8.33 (d, J=8.56 Hz, 1H), 8.07 (qd, J=1.65, 3.32 Hz,
1H), 7.49 (d, J=1.51 Hz, 1H), 7.36 (dd, J=1.81, 8.46 Hz, 1H); LCꢀHRMS: m/z = 245.9960 [(Mꢀ
H)ꢀ calculated for C H ClF NO = 245.9939; Diff = 2.1 mD].
1
0
5
3
ACS Paragon Plus Environment

Page 23 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
-Chloro-1H-indole-3-carboxylic acid (11). A solution of 1ꢀ(6ꢀchloroꢀ1Hꢀindolꢀ3ꢀyl)ꢀ
2
,2,2ꢀtrifluoroꢀethanone (10, 21.6 g, 87.2 mmol) in an aqueous solution of potassium hydroxide
(
4M, 110 mL) was heated at reflux for 2 h. The reaction mixture was cooled to 0 °C and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
neutralized (final pH of 5) by addition of a concentrated aqueous hydrochloric acid solution (37
mL). The resulting precipitate was collected by filtration, washed with water and dried to afford
1
1
1 (16.4 g, 96%) as a white solid. H NMR (600 MHz, DMSOꢀd ) δ 11.93 (br s, 2H), 8.05 (s,
6
1H), 8.00 (d, J=8.56 Hz, 1H), 7.53 (d, J=1.91 Hz, 1H), 7.19 (dd, J=1.91, 8.46 Hz, 1H); LCꢀ
HRMS: m/z = 194.0024 [(MꢀH)ꢀ calculated for C H ClNO = 194.0014; Diff = 1 mD].
9
6
2
(6-Chloro-1H-indol-3-yl)-spiro[1H-isobenzofuran-3,4'-piperidine]-1'-yl-methanone
(12). General Procedure A. To a solution of 6ꢀchloroꢀ1Hꢀindoleꢀ3ꢀcarboxylic acid (11, 5 g, 25.6
mmol) in THF (90 mL) was added DMF (39.6 ꢁl, 0.55 mmol). The mixture was cooled to 0 °C
and oxalyl dichloride (2.13 mL, 24.3 mmol) was added drop wise. The temperature was raised to
RT and stirring was pursued for 3 h, before the mixture was cooled to 5 °C. A solution of
3
3
spiro[1Hꢀisobenzofuranꢀ3,4'ꢀpiperidine] (4.99 g, 25.6 mmol) in CH Cl (20.0 mL) and
2
2
triethylamine (7.1 mL, 51.1 mmol) were added. After 0.5 h at RT, the reaction was complete and
H O (5 mL) was added. The reaction mixture was concentrated under vacuum and the residue
2
taken up in ethyl acetate and washed with water. The organic phase was dried over sodium
sulfate, filtered, concentrated before the crude product was purified by flash chromatography
1
eluting with 40% ethyl acetate in heptane to afford 12 (8.2 g, 87%) as a white solid. H NMR
(600 MHz, CHLOROFORMꢀd) δ 9.10 (br s, 1H), 7.67 (d, J=8.56 Hz, 1H), 7.43 (d, J=2.72 Hz,
1
H), 7.33 (d, J=1.61 Hz, 1H), 7.27ꢀ7.31 (m, 2H), 7.21ꢀ7.25 (m, 1H), 7.18 (dd, J=1.86, 8.51 Hz,
H), 7.10ꢀ7.15 (m, 1H), 5.11 (s, 2H), 4.43 (br d, J=11.99 Hz, 2H), 3.48 (br s, 2H), 1.92 (br s,
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
H), 1.81 (br s, 2H); LCꢀHRMS: m/z = 367.1225 [(M+H)+ calculated for C H ClN O =
2
1
19
2
2
3
67.1208; Diff = 1.7 mD].
6-Chloro-1H-indol-3-yl)-spiro[5H-furo[3,4-b]pyridine-7,4'-piperidine]-1'-yl-
methanone (13). General procedure A, with 11 (0.03 g, 0.15 mmol) and spiro[5Hꢀfuro[3,4ꢀ
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
20a, 34
b]pyridineꢀ7,4'ꢀpiperidine]
(0.032 g, 0.17 mmol), was used to produce 13 (0.039 g, 69%) as
1
a white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 8.99 (br s, 1H), 8.46ꢀ8.55 (m, 1H),
7
.67 (d, J=8.56 Hz, 1H), 7.58 (dd, J=1.31, 7.56 Hz, 1H), 7.43 (d, J=2.72 Hz, 1H), 7.35 (d,
J=1.41 Hz, 1H), 7.21 (dd, J=4.94, 7.66 Hz, 1H), 7.18 (dd, J=1.81, 8.56 Hz, 1H), 5.11 (s, 2H),
.40 (br s, 2H), 3.50 (br t, J=12.24 Hz, 2H), 2.14 (dt, J=4.63, 13.00 Hz, 2H), 1.77 (br d, J=12.79
Hz, 2H); LCꢀHRMS: m/z = 368.118 [(M+H)+ calculated for C H ClN O = 368.1160; Diff = 2
4
20 18
3
2
mD].
(6-Chloro-1H-indol-3-yl)-spiro[1H-furo[3,4-c]pyridine-3,4'-piperidine]-1'-yl-
methanone (14). General procedure A, with 11 (0.03 g, 0.15 mmol) and spiro[1Hꢀfuro[3,4ꢀ
c]pyridineꢀ3,4'ꢀpiperidine] (0.032 g, 0.17 mmol), was used to produce 14 (0.038 g, 67%) as a
1
white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 9.08 (br s, 1H), 8.54 (d, J=5.04 Hz, 1H),
8
.45 (d, J=0.91 Hz, 1H), 7.66 (d, J=8.46 Hz, 1H), 7.47 (d, J=2.72 Hz, 1H), 7.35 (d, J=1.81 Hz,
H), 7.22 (dd, J=0.96, 4.99 Hz, 1H), 7.19 (dd, J=1.81, 8.56 Hz, 1H), 5.11 (s, 2H), 4.41 (br s,
1
2H), 3.49 (br s, 2H), 1.97 (br s, 2H), 1.84 (br d, J=12.79 Hz, 2H); LCꢀHRMS: m/z = 368.1181
(M+H)+ calculated for C H ClN O = 368.1160; Diff = 2.1 mD].
[
20
18
3
2
(6-Chloro-1H-indol-3-yl)-spiro[7H-furo[3,4-b]pyridine-5,4'-piperidine]-1'-yl-
methanone (15). General procedure A, with 11 (0.03 g, 0.15 mmol) and spiro[7Hꢀfuro[3,4ꢀ
20a
b]pyridineꢀ5,4'ꢀpiperidine] (0.032 g, 0.17 mmol), was used to produce 15 (0.054 g, 95%) as a
ACS Paragon Plus Environment

Page 25 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 8.96 (br s, 1H), 8.51 (dd, J=1.41, 4.94
Hz, 1H), 7.67 (d, J=8.56 Hz, 1H), 7.46 (d, J=2.72 Hz, 1H), 7.44 (dd, J=1.41, 7.66 Hz, 1H), 7.36
(
d, J=1.61 Hz, 1H), 7.17ꢀ7.22 (m, 2H), 5.09 (s, 2H), 4.15ꢀ4.68 (m, 2H), 3.41ꢀ3.64 (m, 2H), 1.71ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
.07 (m, 4H); LCꢀHRMS: m/z = 368.1179 [(M+H)+ calculated for C H ClN O = 368.1160;
20 18 3 2
Diff = 1.9 mD].
1
'-(6-Chloro-1H-indole-3-carbonyl)spiro[isobenzofuran-3,4'-piperidine]-1-one (16).
General procedure A, with 11 (0.150 g, 0.77 mmol) and spiro[isobenzofuranꢀ3,4'ꢀpiperidine]ꢀ1ꢀ
1
one (0.156 g, 0.77 mmol), was used to produce 16 (0.289 g, 99%) as a white solid. H NMR (600
MHz, CHLOROFORMꢀd) δ 8.70 (br s, 1H), 7.92 (d, J=7.56 Hz, 1H), 7.66ꢀ7.73 (m, 2H), 7.53ꢀ
7
.59 (m, 2H), 7.39ꢀ7.43 (m, 2H), 7.23 (dd, J=1.76, 8.41 Hz, 1H), 4.52 (br s, 2H), 3.35ꢀ3.84 (m,
1H), 3.35ꢀ3.84 (m, 1H), 2.17 (br s, 2H), 1.77 (br d, J=13.40 Hz, 2H); LCꢀHRMS: m/z =
81.1001 [(M+H)+ calculated for C H ClN O = 381.1000; Diff = 0.1 mD].
3
2
1
17
2
3
(
6-Chloro-1H-indol-3-yl)-spiro[2H-benzofuran-3,4'-piperidine]-1'-yl-methanone
17). General Procedure B. To a solution of 6ꢀchloroꢀ1Hꢀindoleꢀ3ꢀcarboxylic acid (11, 0.227 g,
.16 mmol) in CH Cl (15 mL) was added HOBt (0.186 g, 1.50 mmol), EDC (0.263 g, 1.50
(
1
2
2
mmol) and Et N (0.19 mL, 1.50 mmol). The reaction mixture was stirred at 30 °C for 1 h before
3
spiro[2Hꢀbenzofuranꢀ3,4'ꢀpiperidine] (0.200 g, 1.056 mmol) was added. After 24 h, the mixture
was poured into a separating funnel and washed successively with a saturated aqueous
ammonium chloride solution, sodium bicarbonate solution and then brine. The organic phase was
dried over sodium sulfate, filtered, concentrated before the crude product was purified by flash
chromatography eluting with 50% ethyl acetate in heptane to afford 17 (0.191 g, 42%) as a white
1
solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 8.49 (br s, 1H), 7.66 (d, J=8.56 Hz, 1H), 7.54
(d, J=2.62 Hz, 1H), 7.42 (d, J=1.71 Hz, 1H), 7.21 (dd, J=1.76, 8.51 Hz, 1H), 7.15ꢀ7.18 (m, 1H),
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
.14ꢀ7.19 (m, 1H), 6.91 (t, J=7.10 Hz, 1H), 6.83 (d, J=7.66 Hz, 1H), 4.48 (s, 2H), 4.37 (br s,
H), 3.19 (br t, J=12.09 Hz, 2H), 1.89ꢀ2.01 (m, 2H), 1.82 (br d, J=13.50 Hz, 2H); LCꢀHRMS:
2
m/z = 365.1086 [(MꢀH)ꢀ calculated for C H ClN O = 365.1062; Diff = 2.4 mD].
2
1
19
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(6-Chloro-1H-indol-3-yl)-spiro[indoline-3,4'-piperidine]-1'-yl-methanone (18). To a
solution of 6ꢀchloroꢀ1Hꢀindoleꢀ3ꢀcarboxylic acid 11 (0.02 g, 0.10 mmol) and spiro[indolineꢀ3,4'ꢀ
piperidine] (0.019 g, 0.10 mmol) in DMF (1 mL) at RT was added HATU (0.038 g, 0.10 mmol)
and Hunig’s base (0.034 mL, 0.2 mmol). After 2h, the reaction was complete and purification by
1
preparative HPLC gave 18 (0.012 g, 35%) as a white solid. H NMR (600 MHz,
CHLOROFORMꢀd) δ 8.77 (br s, 1H), 7.66 (d, J=8.56 Hz, 1H), 7.47 (d, J=2.62 Hz, 1H), 7.37 (d,
J=1.71 Hz, 1H), 7.19 (dd, J=1.86, 8.51 Hz, 1H), 7.10 (s, 1H), 7.05ꢀ7.09 (m, 1H), 6.77 (dt,
J=0.86, 7.43 Hz, 1H), 6.67 (d, J=7.66 Hz, 1H), 5.30 (s, 1H), 4.36 (br s, 2H), 3.56 (s, 2H), 3.20
(br t, J=11.74 Hz, 2H), 1.89 (br d, J=10.28 Hz, 2H), 1.76ꢀ1.85 (m, 2H); LCꢀHRMS: m/z =
3
66.1378 [(M+H)+ calculated for C H ClN O = 366.1368; Diff = 1 mD].
21 20
3
(6-Chloro-1H-indol-3-yl)-(2,2-dioxospiro[1H-2-benzothiophene-3,4'-piperidine]-1'-
yl)methanone (19). General procedure A, with 11 (0.300 g, 1.53 mmol) and spiro[1Hꢀ2ꢀ
35
benzothiopheneꢀ3,4'ꢀpiperidine] 2,2ꢀdioxide (0.365 g, 1.53 mmol), was used to produce 19
1
(0.230 g, 36%) as a white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 8.49 (br s, 1H), 7.68
(d, J=8.66 Hz, 1H), 7.58 (d, J=2.72 Hz, 1H), 7.42 (d, J=1.81 Hz, 1H), 7.41 (br d, J=8.36 Hz,
1H), 7.36 (dt, J=1.21, 7.56 Hz, 1H), 7.29 (d, J=8.26 Hz, 1H), 7.25 (d, J=8.06 Hz, 1H), 7.22 (dd,
J=1.81, 8.56 Hz, 1H), 4.39 (s, 2H), 4.08ꢀ4.59 (m, 2H), 3.54ꢀ3.85 (m, 2H), 2.44 (br d, J=12.79
Hz, 2H), 2.09 (br d, J=10.38 Hz, 2H); LCꢀHRMS: m/z = 415.0890 [(M+H)+ calculated for
C H ClN O S = 415.0878; Diff = 1.2 mD].
21
19
2
3
ACS Paragon Plus Environment

Page 27 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
'-(6-Chloro-1H-indole-3-carbonyl)spiro[1H-3,1-benzoxazine-4,4'-piperidine]-2-one
(
20). General procedure B, with 11 (0.024 g, 0.122 mmol) and spiro[1Hꢀ3,1ꢀbenzoxazineꢀ4,4'ꢀ
piperidine]ꢀ2ꢀone (0.027 g, 0.122 mmol), was used to produce 20 (0.025 g, 51%) as a white
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 8.48 (br s, 1H), 7.66 (d, J=8.56 Hz, 1H), 7.57
(d, J=2.72 Hz, 1H), 7.43 (d, J=1.71 Hz, 1H), 7.27ꢀ7.30 (m, 1H), 7.21 (dd, J=1.81, 8.56 Hz, 1H),
7
.15 (br d, J=6.25 Hz, 2H), 7.09ꢀ7.12 (m, 1H), 6.78 (d, J=7.96 Hz, 1H), 4.09ꢀ4.80 (m, 2H), 3.66
br s, 2H), 2.17 (br d, J=13.00 Hz, 2H), 2.05 (br s, 2H); LCꢀHRMS: m/z = 394.0983 [(MꢀH)ꢀ
calculated for C H ClN O = 394.0964; Diff = 1.9 mD].
(
2
1
18
3
3
3
-[1-(6-Chloro-1H-indole-3-carbonyl)-4-piperidyl]-1H-benzimidazol-2-one
(21).
General procedure B, with 11 (0.045 g, 0.23 mmol) and 3ꢀ(4ꢀpiperidyl)ꢀ1Hꢀbenzimidazolꢀ2ꢀone
1
(0.050 g, 0.23 mmol), was used to produce 21 (0.040 g, 44%) as a white solid. H NMR (600
MHz, DMSOꢀd ) δ 11.73 (br d, J=1.61 Hz, 1H), 10.88 (s, 1H), 7.83 (d, J=2.72 Hz, 1H), 7.75 (d,
6
J=8.56 Hz, 1H), 7.52 (d, J=1.81 Hz, 1H), 7.29 (d, J=7.15 Hz, 1H), 7.16 (dd, J=1.91, 8.56 Hz,
1
H), 6.96ꢀ7.07 (m, 3H), 4.29ꢀ4.60 (m, 3H), 3.13 (br t, J=11.08 Hz, 2H), 2.33 (dq, J=4.28, 12.61
Hz, 2H), 1.78 (br d, J=10.17 Hz, 2H); LCꢀHRMS: m/z = 395.1270 [(M+H)+ calculated for
C H ClN O = 395.1269; Diff = 0.1 mD].
21
19
4
2
(6-Chloro-1H-indol-3-yl)-(4-phenyl-1-piperidyl)methanone (22). General procedure
B, with 11 (20 mg, 0.101 mmol) and 4ꢀphenylpiperidine (16 mg, 0.101 mmol), was used to
1
produce 22 (4.9 mg, 14%) as a white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 9.23 (br
s, 1H), 7.65 (d, J=8.56 Hz, 1H), 7.34ꢀ7.36 (m, 1H), 7.31ꢀ7.34 (m, 1H), 7.30ꢀ7.33 (m, 2H), 7.21ꢀ
7
1
3
.25 (m, 3H), 7.17 (dd, J=1.81, 8.56 Hz, 1H), 4.55 (br s, 2H), 3.07 (br s, 2H), 2.81 (tt, J=3.63,
2.14 Hz, 1H), 1.93 (br d, J=11.89 Hz, 2H), 1.74 (br d, J=11.08 Hz, 2H); LCꢀHRMS: m/z =
39.1278 [(M+H)+ calculated for C H ClN O = 339.1259; Diff = 1.9 mD].
2
0
19
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
6-Chloro-1H-indol-3-yl)-[4-(2-chlorophenyl)piperazin-1-yl]methanone (23). General
procedure B, with 11 (36 mg, 0.184 mmol) and 1ꢀ(2ꢀchlorophenyl)piperazine (36 mg, 0.184
1
mmol), was used to produce 23 (24 mg, 35%) as a white solid. H NMR (600 MHz,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CHLOROFORMꢀd) δ 8.56 (br s, 1H), 7.67 (d, J=8.56 Hz, 1H), 7.53 (d, J=2.72 Hz, 1H), 7.41 (d,
J=1.81 Hz, 1H), 7.38 (dd, J=1.51, 7.96 Hz, 1H), 7.22ꢀ7.26 (m, 1H), 7.20 (dd, J=1.81, 8.56 Hz,
1
H), 7.01ꢀ7.06 (m, 1H), 6.99ꢀ7.03 (m, 1H), 3.91 (br s, 4H), 3.09 (br s, 4H); LCꢀHRMS: m/z =
74.0831 [(M+H)+ calculated for C H Cl N O = 374.0821; Diff = 1 mD].
3
1
9
17
2
3
2-[6-Chloro-3-(spiro[1H-isobenzofuran-3,4'-piperidine]-1'-carbonyl)indol-1-yl]-N,N-
dimethyl-acetamide (24). To a solution of (6ꢀchloroꢀ1Hꢀindolꢀ3ꢀyl)ꢀspiro[1Hꢀisobenzofuranꢀ
,4'ꢀpiperidine]ꢀ1'ꢀylꢀmethanone 12 (60 mg, 0.16 mmol) in DMF (4 mL) cooled at 0 °C, was
3
added sodium hydride dispersion (8 mg, ~55% in oil, 0.17 mmol) and after 0.5h, 2ꢀchloroꢀN,Nꢀ
dimethylꢀacetamide (21 mg, 0.17 mmol). After 2h at RT, aqueous NH Cl (2 mL) was added and
4
the reaction mixture was concentrated under vacuum. The residue was partitioned between ethyl
acetate and water. The organic portion was dried over Na SO , filtered and concentrated. The
2
4
crude product was purified by column chromatography, eluting with ethyl acetate / heptane to
1
produce 24 (13 mg, 18%) as a white solid. H NMR (600 MHz, CHLOROFORMꢀd) δ 7.68 (d,
J=8.56 Hz, 1H), 7.44 (s, 1H), 7.27ꢀ7.31 (m, 2H), 7.24 (d, J=1.61 Hz, 1H), 7.21ꢀ7.23 (m, 1H),
7
.18 (dd, J=1.76, 8.51 Hz, 1H), 7.09ꢀ7.14 (m, 1H), 5.10 (s, 2H), 4.88 (s, 2H), 4.17ꢀ4.61 (m, 2H),
3.49 (s, 2H), 3.13 (s, 3H), 3.02 (s, 3H), 1.92 (br s, 2H), 1.79 (br d, J=12.69 Hz, 2H); LCꢀHRMS:
m/z = 452.1752 [(M+H)+ calculated for C H ClN O = 452.1735; Diff = 1.7 mD].
2
5
26
3
3
2
-[6-Chloro-3-(spiro[1H-isobenzofuran-3,4'-piperidine]-1'-carbonyl)indol-1-yl]-1-(4-
methylpiperazin-1-yl)ethanone (25). To a solution of (6ꢀchloroꢀ1Hꢀindolꢀ3ꢀyl)ꢀspiro[1Hꢀ
isobenzofuranꢀ3,4'ꢀpiperidine]ꢀ1'ꢀylꢀmethanone 12 (60 mg, 0.16 mmol) in DMF (4 mL) cooled at
ACS Paragon Plus Environment

Page 29 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
0
°C, was added sodium hydride dispersion (8 mg, ~55% in oil, 0.17 mmol) and after 0.5h, 2ꢀ
chloroꢀ1ꢀ(4ꢀmethylpiperazinꢀ1ꢀyl)ethanone (30 mg, 0.17 mmol). After 2h at RT, aqueous NH Cl
4
(2 mL) was added and the reaction mixture was concentrated under vacuum. The residue was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
partitioned between ethyl acetate and water. The organic portion was dried over Na SO , filtered
2
4
and concentrated. The crude product was purified by column chromatography, eluting with ethyl
1
acetate / heptane to produce 25 (20 mg, 24%) as a white solid. H NMR (600 MHz,
CHLOROFORMꢀd) δ 7.68 (d, J=8.56 Hz, 1H), 7.43 (s, 1H), 7.27ꢀ7.31 (m, 2H), 7.24 (d, J=1.71
Hz, 1H), 7.21ꢀ7.24 (m, 1H), 7.19 (dd, J=1.76, 8.51 Hz, 1H), 7.09ꢀ7.14 (m, 1H), 5.10 (s, 2H),
4.88 (s, 2H), 4.43 (br s, 2H), 3.69 (br s, 2H), 3.58 (br s, 2H), 3.46 (br s, 2H), 2.49 (br d, J=9.17
Hz, 4H), 2.36 (s, 3H), 1.92 (br s, 2H), 1.79 (br d, J=12.39 Hz, 2H); LCꢀHRMS: m/z = 507.2170
[(M+H)+ calculated for C H ClN O = 507.2157; Diff = 1.3 mD].
28
31
4
3
2
-[6-Chloro-3-(spiro[1H-isobenzofuran-3,4'-piperidine]-1'-carbonyl)indol-1-
yl]acetic acid (26). Step 1. To a solution of (6ꢀchloroꢀ1Hꢀindolꢀ3ꢀyl)ꢀspiro[1Hꢀisobenzofuranꢀ
,4'ꢀpiperidine]ꢀ1'ꢀylꢀmethanone 12 (0.50 g, 1.36 mmol) in DMF (10 mL) cooled at 0 °C, was
3
added sodium hydride dispersion (69 mg, ~50% in oil, 1.43 mmol) and after 0.5h, ethyl 2ꢀ
bromoacetate (0.16 mL, 1.43 mmol). After 2h at RT, aqueous NH Cl (2 mL) was added and the
4
reaction mixture was concentrated under vacuum. The residue was partitioned between ethyl
acetate and water. The organic portion was dried over Na SO , filtered and concentrated. The
2
4
crude product was purified by column chromatography, eluting with ethyl acetate / heptane to
produce ethyl 2ꢀ[6ꢀchloroꢀ3ꢀ(spiro[1Hꢀisobenzofuranꢀ3,4'ꢀpiperidine]ꢀ1'ꢀcarbonyl)indolꢀ1ꢀ
yl]acetate (0.56 g, 91%) as a light yellow oil. Step 2. This intermediate was dissolved in EtOH
12 mL) and an aqueous solution of NaOH (1M, 2 mL) was added. The reaction mixture was
(
stirred overnight, before being acidified with HCl (final pH ~3). The resulting solid was
ACS Paragon Plus Environment