Discovery of G Protein-Biased Dopaminergics with a Pyrazolo[1,5-a]pyridine Substructure

Article abstract of DOI:10.1021/acs.jmedchem.6b01857

1,4-Disubstituted aromatic piperazines are privileged structural motifs recognized by aminergic G protein-coupled receptors. Connection of a lipophilic moiety to the arylpiperazine core by an appropriate linker represents a promising concept to increase binding affinity and to fine-tune functional properties. In particular, incorporation of a pyrazolo[1,5-a]pyridine heterocyclic appendage led to a series of high-affinity dopamine receptor partial agonists. Comprehensive pharmacological characterization involving BRET biosensors, binding studies, electrophysiology, and complementation-based assays revealed compounds favoring activation of G proteins (preferably G_o) over β-arrestin recruitment at dopamine D₂ receptors. The feasibility to design G protein-biased partial agonists as putative novel therapeutics was demonstrated for the representative 2-methoxyphenylpiperazine 16c, which unequivocally displayed antipsychotic activity in vivo. Moreover, combination of the pyrazolo[1,5-a]pyridine appendage with a 5-hydroxy-N-propyl-2-aminotetraline unit led to balanced or G protein-biased dopaminergic ligands depending on the stereochemistry of the headgroup, illustrating the complex structure-functional selectivity relationships at dopamine D₂ receptors.

Full text of DOI:10.1021/acs.jmedchem.6b01857

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Discovery of G Protein-Biased Dopaminergics
with a Pyrazolo[1,5-a]pyridine Substructure
Dorothee Möller, Ashutosh Banerjee, Taygun C. Uzuneser, Marika Skultety, Tobias Huth, Bianca Plouffe,
Harald Hübner, Christian Alzheimer, Kristina Friedland, Christian P. Müller, Michel Bouvier, and Peter Gmeiner
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01857 • Publication Date (Web): 01 Mar 2017
Downloaded from http://pubs.acs.org on March 1, 2017
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 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of G ProteinꢀBiased Dopaminergics with a
Pyrazolo[1,5ꢀa]pyridine Substructure
1
1
2
1
3
Dorothee Möller, Ashutosh Banerjee, Taygun C. Uzuneser, Marika Skultety, Tobias Huth,
4
1
3
5
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
Bianca Plouffe, Harald Hübner, Christian Alzheimer, Kristina Friedland, Christian P.
2
4
1
Müller, Michel Bouvier, Peter Gmeiner *
1
Department of Chemistry and Pharmacy, Medicinal Chemistry, FriedrichꢀAlexander
University ErlangenꢀNürnberg, Schuhstraße 19, 91052 Erlangen, Germany
2
Department of Psychiatry and Psychotherapy, FriedrichꢀAlexander University Erlangenꢀ
Nürnberg, Schwabachanlage 6, 91054 Erlangen, Germany
3
Institute of Physiology and Pathophysiology, FriedrichꢀAlexander University Erlangenꢀ
Nürnberg, Universitätsstraße 17, 91054 Erlangen, Germany
4
Institute for Research in Immunology and Cancer (IRIC), Department of Biochemistry and
Molecular Medicine, University of Montreal, Québec, Canada H3C 1J4
5
Department of Chemistry and Pharmacy, Molecular and Clinical Pharmacy, Friedrichꢀ
Alexander University ErlangenꢀNürnberg, Cauerstraße 4, 91058 Erlangen, Germany
KEYWORDS:
GPCR, dopamine receptor, functional selectivity, ligand bias, G protein, βꢀarrestin, GIRK
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 71
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
ABSTRACT
,4ꢀDisubstituted aromatic piperazines are privileged structural motifs recognized by
1
aminergic G proteinꢀcoupled receptors. Connection of a lipophilic moiety to the
arylpiperazine core by an appropriate linker represents a promising concept to increase
binding affinity and to fineꢀtune functional properties. In particular, incorporation of a
pyrazolo[1,5ꢀa]pyridine heterocyclic appendage led to a series of high affinity dopamine
receptor partial agonists. Comprehensive pharmacological characterization involving BRET
biosensors, binding studies, electrophysiology and complementationꢀbased assays revealed
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
compounds favoring activation of G proteins (preferably G
o
) over βꢀarrestin recruitment at
dopamine D receptors. The feasibility to design G proteinꢀbiased partial agonists as putative
2
novel therapeutics was demonstrated for the representative 2ꢀmethoxyphenylpiperazine 16c,
which unequivocally displayed antipsychotic activity in vivo. Moreover, combination of the
pyrazolo[1,5ꢀa]pyridine appendage with a 5ꢀhydroxyꢀNꢀpropylꢀ2ꢀaminotetraline unit led to
balanced or G proteinꢀbiased dopaminergic ligands depending on the stereochemistry of the
head group, illustrating the complex structureꢀfunctional selectivity relationships at dopamine
D receptors.
2
ACS Paragon Plus Environment

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
In recent years, a number of studies have identified 1,4ꢀdisubstituted aromatic
1
piperazines (1,4ꢀDAPs) as privileged scaffolds interacting with aminergic G proteinꢀcoupled
receptors (GPCRs), especially those belonging to the serotonergic, αꢀadrenergic 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
2
, 3
2 2
dopamine D receptor (D R) family. Consisting of an aromatic head group in vicinity to a
basic amine, they are able to recognize a threeꢀdimensional pattern of highly conserved amino
acids in the orthosteric binding pocket, a cavity formed within the helix bundle of the 7ꢀ
transmembrane receptor (7ꢀTM). 1,4ꢀDAPs are able to mimic the crucial interactions between
dopamine and the receptor and can therefore serve as surrogates for the endogenous agonist.
2
Although the core region of the aromatic piperazines is sufficient to allow binding to D R and
4
the closely related D R subtype, enlargement of the chemical structures by addition of a
3
flexible linker and a second, mostly lipophilic system (type I compounds, Figure 1) has been
2, 4, 5
found to promote enhanced affinity and subtype selectivity.
Figure 1. Chemical structure of type I compounds (1,4ꢀDAP) consisting of an aromatic core
structure (π ), a basic amine unit, a linker region and a second, mostly lipophilic system (π ).
1 2
The relevance of type I compounds is best illustrated by the growing number of
antipsychotic drugs with antagonist/partial agonist activity at D R and D R. Besides
2
3
6
7
8
aripiprazole (1a), cariprazine (1b) and brexiprazole (1c) have recently received FDAꢀ
approval for the treatment of schizophrenia and bipolar I manic/mixed episodes or major
9
depressive disorders, respectively (Figure 2).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 71
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
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
Figure 2. Chemical structures of approved antipsychotics drugs 1aꢀc and recently described
functionally selective dopamine receptor partial agonists (2aꢀc) bearing a 1,4ꢀDAP as
orthosteric head group.
Traditionally, different properties have been attributed to the three individual parts
(orthosteric head group, linker and lipophilic appendage) of 1,4ꢀDAPs. While the
arylpiperazine including its substituents has been considered responsible for the intrinsic
activity of a given ligand, the lipophilic appendage addressing a secondary binding pocket has
been thought to promote enhanced affinity with the linker simultaneously controlling subtypeꢀ
2
selectivity. More recently, a number of compounds with an enlarged 1,4ꢀDAP structure have
been described (e.g. compounds 2aꢀc, Figure 2), that are able to activate D
2
R in a functionally
This behavior, also described as ligand bias, is the capacity of a ligand
to direct the signaling of a GPCR to a preferential signal transducer while other signaling
3, 10ꢀ14
selective manner.
15, 16
pathways remain largely unaffected.
Interestingly, ligands preferably activating G
proteinꢀmediated signaling and those preferentially leading to a recruitment of βꢀarrestins
1
0ꢀ13
have been found to share high chemical similarity.
This behavior has been attributed to a
complex interplay between the orthosteric aromatic pharmacophores and the lipophilic
appendages connected by an appropriate spacer, enabling the ligands to adopt a bitopic
1
7ꢀ19
orthosteric/allosteric binding mode.
In 2014, we presented pyrazolo[1,5ꢀa]pyridine
heterocycles as valuable lipophilic moieties providing high D R/D R affinity, when attached
2
3
to a 2,3ꢀdichlorophenylpiperazine by a butoxyꢀlinker. The ligands including the oxime
ACS Paragon Plus Environment

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substitutedꢀpyrazolo[1,5ꢀa]pyridine 2a (Figure 2) displayed partial agonist properties for G
2
protein activation, while they were inhibitors for the recruitment of βꢀarrestinꢀ2 at the D R,
10
thus being high affinity biased agonists. This finding was supported by a very recent report
on G proteinꢀbiased agonists with a very similar structural architecture replacing the
pyrazolo[1,5ꢀa]pyridine moiety by a bioisosteric benzothiazole unit (compound 2c 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
1
2
analogs). Extending our previous investigations, we aimed at exploring the different effects
of modifications in linker region as well as variation of the orthosteric recognition element.
Thus, we replaced the 2,3ꢀdichlorophenylpiperazine unit by a 2ꢀmethoxyphenylpiperazine
group. Moreover, an exchange of the oxymethylene part of the linker by an aminocarbonyl
group was investigated. To complement the phenylpiperazineꢀtype head group by a classical
dopamine bioisostere, we planned the incorporation of both enantiomeric forms of 5ꢀhydroxyꢀ
Nꢀpropylꢀ2ꢀaminotetraline, a conformationally restricted scaffold known for its agonist
20
properties at D
2
ꢀlike receptors. Based on this comprehensive set of compounds, differential
studies of signaling profiles and SAR studies should allow to identify molecular determinants
of functional selectivity. Guided by the hypothesis that in vitro profiles similar to those of the
approved drug 1a should predict in vivo antipsychotic activity, we performed a proof of
concept study in rats. Results indicate the feasibility of pyrazolo[1,5ꢀa]pyridines as useful
scaffolds for the development of functionally selective CNS drugs.
RESULTS
1
a, an approved antipsychotic drug has been previously described as partial D
2
R
agonist, preferentially activating G proteinꢀmediated pathways with either little or no efficacy
10, 12, 13, 21
for the recruitment of βꢀarrestinꢀ2.
Employing our recently described affinityꢀ
R partial
generating pyrazolo[1,5ꢀa]pyridine moieties, we aimed at developing new, biased D
2
agonists with putative antipsychotic activity. The heterocyclic moieties were combined with
different orthosteric dopamine surrogates using either butoxyꢀ or butyramidoꢀderived spacers,
since ligand bias at D R results from a complex interplay between orthosteric
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 71
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
pharmacophores, linker regions and liphophilic moieties addressing a secondary binding
3
, 10ꢀ14, 19
pocket.
5ꢀButyramido pyrazolo[1,5ꢀa]pyridines. In a first series of derivatives, the highly
flexible butoxyꢀlinker connecting the orthosteric head group with the heterocyclic moiety was
replaced by an amideꢀmoiety, thereby ideally constraining the ligands into a bioactive
conformation. The respective 5ꢀbutyramidoꢀpyrazolo[1,5ꢀa]pyridines were synthesized by a
combination of two independently prepared building blocks (Scheme 1). For the synthesis of
the orthosteric pharmacophore and linker region, two commercially available
phenylpiperazines (2,3ꢀdichloroꢀ and 2ꢀmethoxyphenylpiperazine) were subjected to
nucleophilic displacement reactions with methyl 4ꢀbromobutyrate, leading to the methyl
carboxylates 3a and 3b. Subsequent ester hydrolysis with sodium hydroxide in
methanol/water resulted in formation of the sodium carboxylates 4a and 4b, which were
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
22
isolated in excellent yield by precipitation. Carboxylates 4a and 4b were activated using
HATU and coupled to 5ꢀaminopyrazolo[1,5ꢀa]pyridine (5), which was synthesized in analogy
23, 24
to previously described protocols.
While the resulting amides 6a and 6b represent target
compounds on their own, previous studies suggested that the pyrazolo[1,5ꢀa]pyridine scaffold
10, 23, 25
can be further modified in position 3 to enhance affinity for D
2
ꢀlike receptors.
Thus, the
azaindoles 6a and 6b were reacted under Vilsmeyer conditions to yield the carbaldehydes 7a
and 7b. Subsequent conversion into oxime derivatives was realized by refluxing with
2
6
hydroxylamine hydrochloride in ethanol/water. The oximes 8a and 8b were isolated as an
inseparable 1:1 mixture of two isomers which were identified as (E) and sꢀcis/sꢀtrans isomers
by diagnostic NOESY experiments. Finally, the oximes were reacted with acetic acid
27
28, 29
anhydride under reflux conditions or cyanuric chloride in DMF at room temperature
to
afford the carbonitriles 9a and 9b.
ACS Paragon Plus Environment

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
†
Scheme 1. Synthesis of pyrazolo[1,5-a]pyridine-derived dopaminergics containing an amide
linker.
H
R2
X
N
NH
N
N
a
c
N
N
O
N
O
N
N
H2N
N
N
R1
R1
R1
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
6
a: R = 2,3-diCl, R = H
5
1 2
R = 2,3-diCl
R = 2-OCH3
3a: R = 2,3-diCl, X = OCH
1 3
1
6b: R = 2-OCH , R = H
1 3 2
d
3b: R = 2-OCH X = OCH3
1
b
1 3,
7a: R = 2,3-diCl, R = CHO
1 2
-
+
4a: R = 2,3-diCl, X = O Na
1
7
b: R = 2-OCH R = CHO
1
3,
2
4b: R = 2-OCH X = O Na+
-
e
1
3,
8a: R = 2,3-diCl, R = CHNOH
1 2
8
b: R = 2-OCH , R = CHNOH
1
3
2
f or g
9a: R = 2,3-diCl, R = CN
1 2
9
b: R = 2-OCH R = CN
1
3,
2
†
Reagents and conditions: (a) methyl 4ꢀbromobutyrate, TEA, DMF, rt, 24 h (88–98%); (b) 1 M NaOH, MeOH,
65 °C, 6 h (93–94%); (c) 5ꢀaminopyrazolo[1,5ꢀa]pyridine, HATU, DIPEA, DMF/CH Cl , 24h, rt (68–95%); (d)
POCl , DMF, 0 °C to rt, 1 h (58–59%); (e) hydroxylamine ⋅ HCl, EtOH/H O, reflux, 2 h (77–83%); (f) for 9a
TCT, DMF, rt, 5 h (75%); (g) for 9b Ac O, reflux, 6 h (12%).
2
2
3
2
2
Prior to functional characterizations, binding affinities of the newly synthesized
compounds 6a,b–9a,b were determined employing radioligand displacement studies with cell
membranes expressing dopamine receptors or related receptors of interest (D
1
R, D2LR, D2SR,
D
3
R, D R, 5ꢀHT1AR, 5ꢀHT2AR, α R, Table 1). For the entire set of test compounds,
4
1
competition experiments revealed high binding affinities in the low nanomolar to
subnanomolar range for the family of D ꢀlike receptors. Among the unmodified pyrazolo[1,5ꢀ
2
a]pyridines 6a,b and their formylꢀsubstituted congeners 7a,b, affinities were found to depend
on the orthosteric phenyliperazine. Test compounds bearing the 2,3ꢀdichlorophenylpiperazine
moiety showed 5ꢀfold higher affinity for D2LR, D2SR and D
residue was beneficial for D R affinity, leading to subnanomolar
for the carbaldehyde 7a (0.30, 0.41 and 1.7 for D R, D R and D R). Further modifications
3
R. Introduction of the formyl
2
K
i
values at D2LR and D2S
R
2
L
2S
3
of the substituents only slightly influenced binding affinities for D2LR, D2SR, preserving the
low nanomolarꢀ to subnanomolar binding characteristics (0.50ꢀ1.11 nM for 8a,b 9a,b at
D
2LR/D2SR). Hence, formylꢀ, oximeꢀ and cyanoꢀsubstituted compounds displayed binding
affinities for D R/D R that are highly similar to those of the antipsychotic drug 1a (Ki 0.54
2
L
2S
and 0.45 nM for D2LR and D2SR). Independent from the orthosteric phenylpiperazine and the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 71
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
substitution pattern of the pyrazolo[1,5ꢀa]pyridine appendage, target compounds displayed
only small differences in their affinity for D
nM). At the closely related 5ꢀHT2AR subtype the 2,3ꢀdichlorophenylpiperazines 6aꢀ9a
showed low nanomolar affinity ( 2.6ꢀ6.3 nM), whereas affinities of the 2ꢀ
methoxyphenylpiperazines 6bꢀ9b were roughly 50ꢀfold lower ( 120ꢀ380 nM). While the
compound family displayed week binding affinity (210ꢀ2900 nM) and hence selectivity over
R, all 5ꢀbutyramidopyrazolo[1,5ꢀa]pyridines also had substantial affinity (1.6ꢀ10 nM) for
R.
4 i i
R (K 8.1ꢀ34 nM) and for 5ꢀHT1AR (K 21ꢀ40
K
i
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
K
i
D
1
α
1
Table 1. Binding affinities and functional properties of 5ꢀbutyramidopyrazolo[1,5ꢀa]pyridines
a,bꢀ9a,b.
6
†
35
K
i
[nM]
[ S]GTPγS
1
,4ꢀ
DAP
,3ꢀ
p5ꢀ
HT1A HT2A
h5ꢀ
hD2S/GαoA
3
hD /GαoA
‡
§
§
§
§
§§
R
2
hD
1
hD2L hD2S hD
3
hD
4
pα1
†
†
‡‡
†††
‡‡‡
†††
‡‡‡
EC50
Emax
EC50
Emax
2
diCl
310 0.53 0.45
± 140 ± 0.24 ± 0.06 ± 1.2 ± 16
240 8.1 7.8 5.6 14
± 0.0 ± 3.5 ± 2.1 ± 1.2 ± 3.1
700 46 36 97 14
840 ± 29 ± 24 ± 20 ± 5.1 ± 2.8 ± 0.0 ± 1.1
210 0.30 0.41 1.7 30 21 6.3 4.9
± 74 ± 0.12 ± 0.19 ± 0.56 ± 7.6 ± 2.8 ± 4.2 ± 0.56
200 1.4 1.5 70 34 29 300 4.6
± 770 ± 0.43 ± 0.81 ± 35 ± 16 ± 18 ± 49 ± 1.0
2900 0.57 0.50 2.2 31 27 2.6 10
± 1900 ± 0.31 ± 0.14 ± 0.80 ± 14 ± 9.9 ± 1.0 ± 7.0
00 0.97 0.52 13 8.1 22 120 1.6
560 ± 0.62 ± 0.05 ± 4.6 ± 1.9 ± 7.1 ± 7.1 ± 1.1
250 0.91 0.52 2.3 32 40 5.6 8.5
± 100 ± 0.58 ± 0.11 ± 0.08 ± 6.1 ± 4.2 ± 1.7 ± 2.2
860 1.1 0.55 28 17 16 380 2.0
± 630 ± 0.73 ± 0.17 ± 12 ± 4.1 ± 1.4 ± 21 ± 0.14
2.3
81
77
3.7
17
67
± 2
52
± 3
1
a
ꢀ
39
20
± 22 ± 1.1 ± 0.71
23 4.9 7.5
± 12 ± 2.1 ± 1.6
31 310 5.3
2,3ꢀ
diCl
50
± 2
65
± 4
6a
H
H
240
630
13
45
66
37
79
3.8
38
62
42
1
±
38
± 3
45
± 3
6
b 2ꢀMeO
2
,3ꢀ
43
± 2
83
± 3
7a
CHO
diCl
1
41
± 2
53
± 4
7
8
b 2ꢀMeO CHO
150
600
470
870
84
2,3ꢀ
diCl
47
± 3
74
± 3
a
oxime
9
37
± 3
58
± 3
8b 2ꢀMeO oxime
±
2
,3ꢀ
50
± 4
83
± 4
9
a
CN
diCl
32
± 2
57
± 4
9
b 2ꢀMeO CN
†
Data represent mean ± S.D. of two to six independent experiments, each performed in triplicate. Radioligands:
‡
‡
3
§ 3
†† 3
‡‡ 3
§§ 3
†††
[
H]SCH23990, [ H]spiperone, [ H]WAY100635, [ H]ketanserin, [ H]prazosin.
Potency given in nM.
‡‡
Efficacy normalized to the reference agonist quinpirole (100%) and buffer (0%). Data represent mean ±
S.E.M. from the pooled curve of four to seven individual experiments, each performed in triplicate.
To determine functional properties of the test compounds at the Gα ꢀcoupled D R
i/o
2S
3
5
and D
3
R, we performed [ S]GTPγS incorporation assays with membranes from transiently
ACS Paragon Plus Environment

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
transfected HEK293T cells. The ability of the ligands to stimulate nucleotide exchange was
measured using GαoA as representative G protein subunit, which is known to be activated by
30ꢀ33
2 3
both, D R and D R.
For both subtypes the antipsychotic drug 1a elicited submaximal
R, respectively), which is
responses (67 % and 52 % of the quinpirole effect at D2SR and D
3
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
good agreement with its partial agonist nature (Table 1, Supplementary Figure 1a). Similar
potencies of 39 nM at D R and 20 nM at D R were observed in the two test systems,
2
S
3
indicating little preference for one receptor subtype over the other. In contrast, pyrazolo[1,5ꢀ
a]pyridines 6a,b, 7b, 8a,b and 9a,b showed a preferential activation of GαoA via D R over
2SR (Supplementary Figure 1b,c and 1eꢀi) with up to 150ꢀfold higher potencies (EC50 3.8ꢀ79
3
D
3
nM and 84ꢀ870 nM at D R and D2SR, respectively) as well as higher efficacies (Emax 45ꢀ83%
at D R vs. 32ꢀ50 % at D R). Only the 2,3ꢀdichlorphenylꢀpiperazine 7a showed a slightly
3
2S
higher potency for the activation of D2SR (13 nM vs. 37 nM at D
3
R, Supplementary Figure
R (EC50 3.8
1d). The closely related oxime 8a, however, was the most potent activator of D
3
nM, Emax 83 %, Supplementary Figure 1f). Additionally, compounds 6a,bꢀ9a,b were
examined for their capacity to promote recruitment of βꢀarrestinꢀ2 to D2SR. Similar to the
reference antipsychotic 1a, none of the pyrazolo[1,5ꢀa]pyridines with amideꢀsubstructure
induced a significant effect (Supplementary Figure 2a) when they were tested at a
concentration of 10 µM in the PathHunter assay. This antagonism stands in contrast to the
35
partial agonist properties observed in the [ S]GTPγS binding experiments, identifying 1,4ꢀ
DAPs with butyramidoꢀspacer and a pyrazolo[1,5ꢀa]pyridine appendage as G proteinꢀbiased
ligands at D R. Compared to the functional properties of chemical analogs comprising a
2
2
butoxyꢀspacer (e.g. the antipsychotic 1a), inferior D R partial agonist properties (reduced
potencies, efficacies, or both) were observed for 6a,bꢀ9a,b. In spite of their subnanomolar
binding affinities reduction of the conformational flexibility by introduction of the amideꢀ
spacer did not prove successful in constraining the ligands in a bioactive conformation that
fosters receptor activation.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 71
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
Modification of the orthosteric recognition element. Complimentary to the
modifications of the spacer region and the lipophilic appendage, we explored the effects of
different aromatic head groups, which serve as primary recognition element for aminergic
2
receptors. While previous studies have widely used 2,3ꢀdichlorophenylpiperazines as
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
10, 13
dopamine surrogate,
we employed closely related 2ꢀmethoxyphenylpiperazines but also
more divergent 5ꢀhydroxyꢀNꢀpropylꢀ2ꢀaminotetralines, rigidized dopamine scaffolds known
20
2
for their D Rꢀagonist properties.
†
Scheme 2. Synthesis of pyrazolo[1,5ꢀa]pyridineꢀderived dopaminergics containing a butoxyꢀ
linker.
†
Reagents and conditions: (a) HBr 48%, reflux, 16 h; (b) KI, DIPEA or TEA, ACN, 85°C, 6–16 h, (19–89%);
(
c) LiAlH
EtOH/H O, reflux, 2 h (85%); (f) TCT, DMF, rt, 7 h (54%). (g) POCl
hydroxylamine ⋅ HCl, EtOH/H O, reflux, 2 h (35–89%); (i) Ac O, reflux, 6 h (79%).
4
, Et
2
O, 0 °C to rt, 1 h (74%), (d) MnO
2
, CH
2
Cl
2
, rt, 14 h (97%); (e) hydroxylamine hydrochloride,
2
3
, DMF, 0 °C to rt, 1 h (79–98%); (h)
2
2
Synthesis of target compounds was accomplished starting from either commercially
available 2ꢀmethoxyphenylpiperazine or (R)/(S)ꢀ5ꢀmethoxyꢀNꢀpropylꢀ2ꢀaminotetraline, which
ACS Paragon Plus Environment

Page 11 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was transformed into the phenol (R)/(S)ꢀ11 by refluxing in hydrobromic acid. Subsequent
coupling to the pyrazolo[1,5ꢀa]pyridine appendage was conducted by nucleophilic
10
displacement reaction with 5ꢀ or 6ꢀ(4ꢀbromo)butoxypyrazolo[1,5ꢀa]pyridines 10a or 10b,
yielding target compounds 12a,b and (R)/(S)ꢀ13a,b, respectively. In the same way, 2ꢀ
substituted pyrazolo[1,5ꢀa]pyridine 12c was obtained starting from 2ꢀ
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
0
methoxyphenylpiperazine and the 2ꢀethoxyꢀcarbonylpyrazolo[1,5ꢀa]pyridine derivative 10c .
The ethylꢀcarboxylate 12c was transformed into the primary alcohol 12d using a solution of
4
LiAlH in diethyl ether. Reoxidation with manganese(IV) oxide afforded the aromatic
aldehyde 12e in excellent yield. The aldehyde could be converted to the aldehydoxime 12f by
treatment with hydroxylamine hydrochloride in aqueous ethanol. Subsequent reaction with
1
0
cyanuric chloride in DMF gave the carbonitrile 12g. Since addition of polar substituents into
position 3 of the pyrazolo[1,5ꢀa]pyridine moiety has previously been shown to have
10, 23
beneficial effects on dopamine D
2
Rꢀlike affinity,
dopaminergics 12a,b and (S)ꢀ13a,b
10, 23, 25
were further functionalized in analogy to previously described procedures.
of formyl residues was realized by Vilsymeyerꢀtype formylation affording target compounds
4a,b and (S)ꢀ15a,b in good yields. The following conversion to the corresponding oximes by
Introduction
1
refluxing with an aqueous solution of hydroxylamine hydrochloride in ethanol led to a
mixture of (E)ꢀs-cis/s-trans configured isomers 16aꢀc, (S)ꢀ17a,b which could only be
separated in the case of the 5ꢀbutoxyꢀphenylpiperazines 16b,c. For the latter, a final
conversion into the corresponding 3ꢀcyanoꢀsubstituted pyrazolo[1,5ꢀa]pyridine 18 was
performed by refluxing in acetic acid anhydride.
Binding affinities. Affinities of target compounds 12aꢀg to 18 for dopamine D
1
R,
D2LR, D2SR, D
3
R and D R and related 5ꢀHT1AR, 5ꢀHT2AR and α R were determined in
4
1
competition binding experiments. Our initial observations were directed towards the
identification of the optimal attachment point of the butoxyꢀlinker and the positioning of
potential functional groups at the pyrazolo[1,5ꢀa]pyridine appendage. When the two
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 71
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
unsubstituted pyrazolo[1,5ꢀa]pyridines 12a and 12b with the 2ꢀmethoxyphenylpiperazine
head group were compared, the 6ꢀbutoxy analog 12a showed superior binding properties
2 i
towards all D ꢀlike receptor subtypes (K 0.69ꢀ5.0 nM compared to 13ꢀ18 nM for 12b).
Introduction of an ethyl carboxylateꢀ, hydroxymethylꢀ or formylꢀsubstituent into position 2 of
the 5ꢀbutoxypyrazolo[1,5ꢀa]pyridine moiety led to an approximately two fold improvement of
D R, D R and D R affinities for test compounds 12cꢀe (Ki 5.1ꢀ8.1 nM, 5.5ꢀ7.4 nM, 7.4ꢀ15
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
L
2S
3
nM at D2LR, D2SR and D
Interestingly, no significant differences in the D
observed. Conversion into the oxime 12f also slightly improved D
3.3, 1.9, 3.2 and 7.8 nM at D2LR, D2SR, D R and D R), indicating a beneficial effect of
positionꢀ2 substituents on D /D receptor affinity. In contrast, introduction of the linear and
3
R, respectively) compared to the unsubstituted precursor 12b.
R affinity of test compounds 12bꢀe were
R, D R and D R affinity
4
2
3
4
(K
i
3
4
2
3
small cyanoꢀgroup (12g) did not significantly alter the binding affinity for D
2 3
R/D R compared
to 12b. However, none of the position 2 modifications could achieve the subnanomolar D
2
R
binding affinity of the unsubstituted 6ꢀbutoxypyrazolo[1,5ꢀa]pyridine 12a. In comparison to
modifications at position 2 of the pyrazolo[1,5ꢀa]pyridine moiety, we investigated the effects
of similar substituents in position 3 of the aromatic heterocycle, since this position had
10, 23
previously been identified as suitable attachment point for different functional groups.
Interestingly, the introduction of these substituents had opposite effects within the group of 2ꢀ
methoxyphenylpiperazines. While the introduction of a formyl residue led to an improvement
of the D
2 2
R affinity for the 5ꢀbutoxyꢀderivative 14b, D R affinity was reduced for the 6ꢀbutoxy
analog 14a (Ki 1.8 and 2.3 nM at D R and D R for 14b vs. 7.8 and 7.7 nM for 14a). The
2S
2L
same trend was also observed for the oximeꢀsubstituted analogs 16a (6ꢀbutoxy,
for D2LR/D2SR) compared to 16b/c. Especially, the (E) and s-cis configured isomer 16c
showed ultraꢀhigh affinity for D2LR and D2SR ( 0.16 and 0.094 nM). Subꢀnanomolar binding
affinity at D2LR /D2SR was also observed for the 3ꢀcyanoꢀderivative 18 ( 0.25 and 0.16 nM).
i
K 4.2 and 1.7
K
i
K
i
In contrast to the set of 2ꢀmethoxyphenylpiperazines, substitution patterns of the
ACS Paragon Plus Environment

Page 13 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pyrazolo[1,5ꢀa]pyridine appendage had only marginal influence on the affinity of compounds
within the series of 5ꢀaminotetraline head groups. Compared to the unsubstituted parent
compounds (S)ꢀ13a/b, introduction of formyl or oxime moieties 3 did not alter D
2
R affinities
R were
more than twoꢀfold. In general, low nanomolar to subnanomolar affinities for D R/D
2
3
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
observed within the entire compound family. More importantly, the absolute stereochemistry
of the 5ꢀhydroxyꢀNꢀpropylꢀ2ꢀaminotetraline was found to be responsible for the superꢀ
affinity, when the (S)ꢀisomers 13a/b were identified as eutomers with a 45ꢀ to 190ꢀfold higher
D
2
R and 15ꢀfold higher D
.71ꢀ0.80 nM at D R and D
3a/b). All compounds showed weak binding to the D
3
R affinity compared to the (R)ꢀenantiomers (
R for (S)ꢀ13a/b compared to 39ꢀ120 nM and 11ꢀ13 nM for (R)ꢀ
R subtype ( 210ꢀ8600 nM).
i
K 0.63ꢀ2.1 nM and
0
2
3
1
1
K
i
Noteworthy, the α adrenoceptor affinity was found to be strongly correlated to the nature of
1
the orthosteric head group. While all target compounds comprising
methoxyphenylpiperazine moiety had binding affinities in the subnanomolar or low
nanomolar range ( 0.32ꢀ7.2 nM), the α R affinity of ligands with a 5ꢀhydroxyꢀNꢀpropylꢀ2ꢀ
aminotetraline unit was at least 100ꢀfold lower ( 77ꢀ360 nM). Although not as pronounced,
a similar trend was observed at serotonergic 5ꢀHT R, when aminotetralines displayed 2ꢀto
a
2ꢀ
K
i
1
K
i
2
A
i
10ꢀfold lower affinities than 2ꢀmethoxyphenylpiperazines (K 880ꢀ2300 nM and 78ꢀ1500 nM).
All compounds also displayed high to moderate affinity for 5ꢀHT1AR (
i
K 0.087ꢀ25 nM).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 71
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
Table 2. Binding affinities of pyrazolo[1,5ꢀa]pyridines 12aꢀg to 18.
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
K
i
[nM]
hD
‡
§
§
§
§
††
‡‡
§§
R
1
R
2
hD
1
hD2L
hD2S
hD
3
4
p5ꢀHT1A h5ꢀHT2A
pα1
1
±
100
1.8
0.74
5.0
± 2.0 ± 0.032
20 13
± 1.4 ± 0.71
0.69
3.3
± 0.6
3.0
± 1.4
1.8
± 0.67
0.68
± 0.41
1.0
± 0.40
0.41
± 0.13
0.60
± 0.28
5.3
± 2.5
25
± 8.5
7.7
± 0.78
9.1
± 1.3
0.59
± 0.3
1.1
210
1.3
12a
12b
H
H
6ꢀA
5ꢀA
5ꢀA
260 ± 0.63 ± 0.16
± 13 ± 0.0
1100 3.4
± 420 ± 0.92
78 2.2
± 12 ± 0.14
250 1.6
± 28 ± 0.50
180 1.8
± 42 ± 0.28
94 7.2
± 29 ± 7.0
410 3.9
± 210 ± 1.2
2000 360
± 1000 ± 130
2800
18
± 2.8
8.1
± 5.4
6.0
± 2.7
5.1
± 2.0
3.3
± 2.5
12
± 4.6
0.63
14
± 2.1
7.4
± 1.8
5.8
± 3.7
5.5
± 3.0
1.9
± 1.0
12
± 3.5
0.75
±
850
4
10
64
20
7.4
± 2.2
15
11
± 2.6
21
12c
2ꢀCOOEt
±
9
1
2
2d 2ꢀCH OH 5ꢀA
±
540
70
300
± 6.7
7.4
± 7.0
19
4
1
2e
2f
2g
S)ꢀ13a
2ꢀCHO
2ꢀoxime
2ꢀCN
H
5ꢀA
5ꢀA
5ꢀA
6ꢀB
6ꢀB
5ꢀB
5ꢀB
6ꢀA
5ꢀA
6ꢀB
5ꢀB
6ꢀA
5ꢀA
5ꢀA
6ꢀB
±
± 1.1
3.2
± 3.1
7.8
4
40
85
80
1
±
± 1.9
12
± 4.5
5.2
2
1
±
2100
7.1
± 5.6
0.71
± 2.5
1.5
(
± 570 ± 0.24 ± 0.12 ± 0.41
± 1.1
54
± 24
1.2
8600
± 4500
800
120
± 39
2.1
48
± 23
0.70
11
± 5.7
0.80
1100
± 320
1200
160
± 14
96
(
R)ꢀ13a
S)ꢀ13b
H
(
H
± 430 ± 0.29 ± 0.28 ± 0.27 ± 0.28
± 70 ± 5.7
3700
± 3500
1400
96
± 30
7.8
± 5.1
1.8
39
± 17
7.7
± 2.8
2.3
13
± 1.7
17
84
± 39
4.3
1600
± 71
1500
± 160 ± 0.77
350 1.1
± 85 ± 0.23
2300
± 280
2200
± 350
100
± 28 ± 0.40
91 0.70
± 13 ± 0.01
100 0.62
± 23 ± 0.18
310
± 71
1.3
(
R)ꢀ13b
H
1
4a
4b
S)ꢀ15a
S)ꢀ15b
6a
3ꢀCHO
3ꢀCHO
CHO
CHO
3ꢀoxime
± 71
± 14 ± 0.92
430
12
± 5.0
1.6
7.1
± 2.5
7.9
1
±
1
240 ± 0.67
400 2.4
± 1.4
0.94
± 0.29
6.6
± 3.3
1.6
± 0.71
0.087
± 0.054
6.2
± 0.21
6.6
120
± 14
150
± 42
1.2
(
±
1
350 ± 0.89 ± 0.19 ± 0.87
300 2.3 0.90 1.1
± 4.5
2.2
(
±
350 ± 0.19 ± 0.21 ± 0.71 ± 0.71
1
±
000
680
710
4.2
± 2.5 ± 0.34 ± 0.78 ± 0.96
1.2 0.72 4.5 1.5
± 2.2 ± 0.28
3.3 0.73
± 1.2 ± 0.12
1.7
2.3
4.7
1
3
ꢀoxime
E)ꢀs-trans
ꢀoxime
E)-s-cis
16b
16c
(
± 28 ± 0.27 ± 0.32
860 0.16 0.094
± 78 ± 0.02 ± 0.027
00 2.5 0.84
160 ± 1.1 ± 0.28 ± 0.18
3
(
± 2.1
5.0
± 1.9
6
0.41
6.5
880
± 360
77
± 26
(S)ꢀ17a 3ꢀoxime
±
± 4.8
ACS Paragon Plus Environment

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
±
800
2.2
0.87
0.59
2.0
3.0
± 1.5
1.0
1600
± 640 ± 7.1
110 0.32
± 44 ± 0.04
140
(
S)ꢀ17b 3ꢀoxime
3ꢀCN
5ꢀB
5ꢀA
500 ± 0.52 ± 0.20 ± 0.03 ± 0.28
210
0.25
0.16
2.9
±2.4
1.3
± 1.2
18
±
57 ± 0.12 ± 0.02
±0.0
†
Data represent mean ± S.D. of two to fifteen independent experiments, each performed in triplicate.
Radioligands: [ H]SCH23990, [ H]spiperone, [ H]WAY100635, [ H]ketanserin, [ H]prazosin.
‡
3
§ 3
†† 3
‡‡ 3
§§ 3
Activation of G proteins. The influence of the different attachment points of the 2ꢀ
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
methoxyphenylpiperazine via the butoxyꢀspacer (position 5 or 6 of the heterocycle) and
varying polar substituents of the pyrazolo[1,5ꢀa]pyridine moiety on D2SR and D
3
R activation
35
was investigated in [ S]GTPγS binding experiments (Table 3). In addition to studying D R
2
S
3
and D R mediated nucleotide exchange with GαoA, we also investigated the D2SR promoted
3
4
activation of the widely expressed Gαi2 subunit, since D
2
R is known to promiscuously
3
1
couple to the Gαi/o family. In all three test systems, 2ꢀmethoxyphenylpiperazines were found
to activate the receptors with submaximal efficacy, indicating partial agonists profiles similar
to the antipsychotic 1a. However, differences in the relative potencies and efficacies were
detected in a careful data analysis. Compared to 1a, the two unsubstituted pyrazolo[1,5ꢀ
a]pyridines 12a and 12b showed lower efficacies for the activation of D2SR. Whereas 12b
displayed potencies similar to 1a, 12a with the butoxyꢀspacer in position 6 of the heterocycle
was about tenꢀfold more potent for the activation of D2SR. At the D
were less pronounced but 1a was less effective than the two target compounds
Supplementary Figures 1a, 3a,b). Introduction of a formyl residue into position 3 of the
3
R differences in potencies
(
pyrazolo[1,5ꢀa]pyridine heterocycle preserved the activation properties in the case of the 6ꢀ
butoxy derivative 14a (EC50 5.9, 150 nM, and 19 nM, Emax 38 %, 25 % and 71 % at
D2SR/GαoA, D2SR/Gαi2 and D
3
R/GαoA). The same modification led to an increase in potency
and intrinsic activity for D R/Gα activation for the 5ꢀbutoxy analog 14b (EC 17 nM, E
2
S
oA
50
max
58 %), compared to its precursor 12b (Supplementary Figure 3a,b and 3j,k). Interestingly, the
efficacy was reduced by a factor of two for D2SR/Gαi2, leading to a substantial preference for
GαoA over Gαi2 activation at D2SR for the 3ꢀcarbaldehyde 14b. In agreement with their high
binding affinities, the corresponding oximes 16aꢀc also showed highly potent receptor
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 71
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
activation characteristics. Oxime 16a with a 6ꢀbutoxy spacer showed a relative preference for
D
3
R/GαoA activation (Supplementary Figure 3m), while the different stereoisomers with the
ꢀbutoxyꢀspacer 16b,c preferentially activated GαoA proteins via D2SR (EC50 3.4 and 0.74 nM
for D2SR/GαoA vs. 41 nM and 18 nM for D R/GαoA Supplementary Figure 3n,o).
5
3
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
Table 3. Functional properties of pyrazolo[1,5ꢀa]pyridines 12aꢀg to 14a,b and 16aꢀc and 18
35
compared to the reference 1a, determined by [ S]GTPγS binding.
35
†
[
S]GTPγS binding
D2SGαoA
D2SGαi2
3
D GαoA
‡
§
max
‡
§
max
‡
§
R
1
R
2
pEC50
EC50
E
pEC50
EC50
E
pEC50
EC50
E
max
1a
ꢀ
ꢀ
7.41 ± 0.09 39
67 ±2 6.80 ± 0.20 160 29 ± 3 7.71 ± 0.17 20 52 ± 3
1
2a
12b
H
H
6ꢀA 8.39 ± 0.15 4.1 30 ± 1 7.64 ± 0.21 23 25 ± 2 7.54 ± 0.08 29 62 ± 2
5ꢀA 7.38 ± 0.05 42 48 ± 1 6.73 ± 0.12 190 31 ± 2 7.25 ± 0.16 56 79 ± 5
1
2c 2ꢀCOOEt 5ꢀA 6.72 ± 0.15 190 37 ± 3 6.16 ± 0.15 690 28 ± 3 7.30 ± 0.15 50 63 ± 3
2d 2ꢀCH OH 5ꢀA 7.14 ± 0.15 73 47 ± 3 6.02 ± 0.13 960 39 ± 3 7.57 ± 0.19 27 65 ± 5
1
2
12e
2ꢀCHO 5ꢀA 7.06 ± 0.12 88 46 ± 2 5.52 ± 0.19 2900 31 ± 6 8.02 ± 0.20 9.7 54 ± 4
2ꢀoxime 5ꢀA 6.40 ± 0.12 400 35 ± 2 5.97 ± 0.17 1100 31 ± 4 7.16 ± 0.17 70 52 ± 3
1
2f
2g
S)ꢀ13a
1
2ꢀCN
H
5ꢀA 5.85 ± 0.10 1400 42 ± 3 6.00 ± 0.21 980 29 ± 4 7.51 ± 0.22 31 48 ± 4
6ꢀB 8.69 ± 0.05 2.0 95 ± 2 7.59 ± 0.06 26 92 ± 2 9.32 ± 0.12 0.48 92 ± 4
5ꢀB 8.31 ± 0.04 4.9 102 ± 2 7.26 ± 0.06 55 98 ± 2 9.37 ± 0.10 0.43 104 ± 3
(
(
S)ꢀ13b
H
14a
3ꢀCHO 6ꢀA 8.23 ± 0.09 5.9 38 ± 1 6.81 ± 0.26 150 25 ± 3 7.73 ± 0.07 19 71 ± 2
3ꢀCHO 5ꢀA 7.76 ± 0.11 17 58 ± 3 7.80 ± 0.34 16 17 ± 2 8.11 ± 0.16 7.7 54 ± 3
3ꢀoxime 6ꢀA 8.54 ± 0.06 2.9 57 ± 1 9.36 ± 0.46 0.44 14 ± 2 9.49 ± 0.09 0.32 58 ± 2
14b
16a
3ꢀoxime
E)-s-trans
16b
5ꢀA 8.47 ± 0.09 3.4 47 ± 1 6.56 ± 0.17 270 25 ± 2 7.39 ± 0.14 41 74 ± 4
(
3
ꢀoxime
16c
18
5ꢀA 9.13 ± 0.09 0.74 50 ± 1 6.93 ± 0.15 120 21 ± 2 7.74 ± 0.10 18 76 ± 3
5ꢀA 7.89 ± 0.12 13 47 ± 2 6.60 ± 0.27 250 22 ± 3 7.72 ± 0.19 19 75 ± 6
(
E)-s-cis
3ꢀCN
†
Data represent mean ± S.E.M. from the pooled curve of four to ten independent experiments, each performed in
‡
§
triplicate. Potency given in nM. Efficacy normalized to the reference agonist quinpirole (100%) and buffer
0%).
(
ACS Paragon Plus Environment

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In contrast to the beneficial effects of substituents in position 3 of the pyrazolo[1,5ꢀ
2
a]pyridine heterocyle, the same functional groups did not have a positive impact on D R
activation when they were attached to position 2 (Supplementary Figure 3cꢀg). On the
contrary, compounds 12cꢀg display lower potencies for promoting nucleotide exchange in
membrane preparations from HEK293T cells expressing D2SR together with GαoA or Gαi2
(EC 73ꢀ1400 nM and E 35ꢀ47% for Gα ; EC 690ꢀ2900 nM and E 28ꢀ39% for Gαi2)
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
5
0
max
oA
50
max
compared to the unsubstituted congener 12b (42 nM, 48% for GαoA and 190 nM, 31% for
Gαi2). At the same time, introduction of substituents had only a marginal influence on the
3
potency observed for the activation of D R (EC50 9.7ꢀ70 nM vs. 56 nM), while efficacies
(
E
max 52ꢀ65 %) were slightly decreased relative to 12b (Emax 79 %). Compared to the partial
agonist profiles observed for 2ꢀmethoxyphenylpiperazines, two representative pyrazolo[1,5ꢀ
a]pyridines with an aminotetraline head group ((S)ꢀ13a,b) displayed full agonist
characteristics for the three investigated systems (D2SR + GαoA, D2SR + Gαi2, and D
GαoA, Supplementary Figure 3h,i). Both compounds showed similar low to subnanomolar
potencies and the same rank order with the highest potency for the activation of D
3
R +
3
R/GαoA
(
EC50 0.48 and 0.43 nM) followed by D R/Gα (EC50 2.0 and 4.9 nM) and finally
2S oA
D
2SR/Gαi2 (EC50 26 and 55 nM).
Further functional characterization. Although measurements of nucleotide
35
exchange in form of [ S]GTPγS binding experiments represent a direct method to assess
receptorꢀmediated activation of G proteins without the amplification of a second messenger
35, 36
pathways, they also have their limitations.
For instance, the different Gα isoforms vary in
the nucleotide exchange rates, leading to diverse but generally low signal to background
35, 37
ratios, which in turn may require experiments to be carried out under different conditions.
Very recently, it has been demonstrated that assay kinetics also play an important role when
38
ligand bias or functional selectivity is examined, as it may lead to an observational bias. In
order to overcome these issues, we employed recently developed bioluminescence resonance
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 71
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
3
9ꢀ42
energy transfer (BRET) test system for the activation of G proteins.
Based on resonance
and hence
energy transfer between Renilla Luciferase (RLucII) taggedꢀGα and GFP10ꢀGγ
2
the separation of Gβγ from Gα subunits, these highly sensitive systems allow a rapid
investigation of receptorꢀmediated G protein activation under highly similar conditions for a
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
9ꢀ42
broad variety of G proteins and different receptors.
a
b
c
d
e
f
Figure 3. Functional properties of pyrazolo[1,5ꢀa]pyridines 16c, (S)ꢀ13b and (R)ꢀ13b in comparison to the full
agonist quinpirole and the reference antipsychotic 1a at D2SR. Activation of GαoA (a,c) and Gαi2 (b,d) was
determined either by [ S]GTPγS binding (a,b) or by BRET changes occurring upon separation of RLucIIꢀGαi/o
from GFP10ꢀGγ (c,d). Ligandꢀstimulated recruitment of βꢀarrestinꢀ2 was determined employing the PathHunter
assay (e) or an approach based on bystander BRET (f). Data wase normalized relative to the response of
quinpirole (100 %) and vehicle (0 %) and represent mean ± S.E.M. from three to ten independent experiments,
each performed at least in duplicate.
3
5
2
A series of BRET titration experiments was performed to confirm coupling of the
different Gαi/o biosensors with D2SR, D2LR and D
3
R. D2SR and D2LR gave robust BRET
ACS Paragon Plus Environment

Page 19 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
changes upon stimulation with the endogenous agonist dopamine for both Gαi1ꢀ3 and GαoA/B
(Supplementary Figure 4aꢀj), which coincides with earlier reports on promiscuous coupling of
31
D
2
R to the entire family Gαi/o proteins. Robust BRET changes upon dopamine treatment
30
indicated a preferential interaction with GαoA/B for D
3
R, while only small BRET changes
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
3
were observed with Gαi1ꢀ3, suggesting a minor interaction of D R with these subunits
(Supplementary Figure 4kꢀo). From these titrations, optimal transfection conditions were
derived for the individual biosensors to obtain doseꢀresponse curves for a subset of the newly
synthesized ligands and the five Gαi/o subunits in coexpression with D2SR (Table 4,
o 3
Supplementary Figure 5aꢀp) and D2LR and the two Gα isoforms for D R (Supplementary
Table 1). In all investigated systems, the two reference agonists quinpirole and dopamine
showed a highly similar behavior (Supplementary Figure 5a,b). In general, only small
differences were observed among the three Gα
found to be slightly better (2ꢀ5 fold) for the activation of the Gα
dopamine was a little less effective in activating the Gα pathways (Emax 87ꢀ92 %), which is
i
or two Gα
o
isoforms although potencies were
o
proteins. However,
o
35
in good agreement with the results from [ S]GTPγS binding experiments. Results obtained
with D R showed no significant differences to those obtained with the short isoform D R.
2L
2S
All compounds sharing an (S)ꢀaminotetraline moiety were found to be highly potent agonists
EC50 0.11ꢀ0.96 nM) for the D2SRꢀmediated activation of the different Gαi/o isoforms. In
(
agreement with the reduced binding affinities, enantiomers (R)ꢀ13a/b were 10ꢀ to 20ꢀfold less
potent (EC50 1.5ꢀ9.8 nM), while retaining full agonist efficacies (Supplementary Figure 5eꢀ
h,k,l,o,p). On the contrary, all ligands bearing a 2ꢀmethoxyphenylpiperazine pharmacophore
shared partial agonist properties for the activation of all Gαi/o proteins. Although efficacies
3
5
were generally found to be higher when compared to the results from [ S]GTPγS binding, the
preferential activation of Gα over Gα was preserved in the BRET readings. Both, efficacies
and potencies were found to be higher for the activation of GαoA/B, (Emax 77ꢀ86 % vs. 39ꢀ62 %
and EC 1.5ꢀ4.3 nM vs. 2.1ꢀ13.1 nM for Gα vs. Gαi1ꢀ3) indicating a functionally selective
o
i
5
0
oA/B
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 71
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
D R activation mode for the investigated 1,4ꢀDAPs. Importantly, when compared to the
2
approved antipsychotic 1a, 1,4ꢀDAPs comprising a pyrazolo[1,5ꢀa]pyridine appendage, were
found to show similar activation profiles with higher overall potency (Figure 3aꢀd).
Recruitment of βꢀarrestins. In addition to the investigations of canonical G protein
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
signaling, we determined the ligands’ ability to induce βꢀarrestinꢀ2 recruitment upon binding
10, 23
to D R employing the PathHunter assay, as previously described.
Experiments
2
S
unambiguously showed the impact of the orthosteric head group on the intrinsic activity
Table 4, Figure 3e,f). Similar to the reference antipsychotic 1a, the investigated 2ꢀ
(
methoxyphenylpiperazines displayed antagonist, or at most weak partial agonist properties
with a maximum effect less than 10 % of the reference agonist quinpirole at D2SR (Table 4,
Supplementary Figure 2b). In contrast, all compounds sharing a (S)ꢀ5ꢀhydroxyꢀ2ꢀ
aminotetraline head group were found to be potent agonists. Observed maximum effects (Emax
93ꢀ105 %) were comparable to the reference agonist quinpirole or dopamine and in agreement
with their superior binding affinities and potencies to activate G proteins, ligands (S)ꢀ13a,b,
(S)ꢀ15a,b and (S)ꢀ17a,b were found to be up to 120ꢀfold more potent (EC50 1.3ꢀ3.0 nM) than
the endogenous agonist. Similar to the above described results, addition of functional groups
to the pyrazolo[1,5ꢀa]pyridine appendage did not significantly influence βꢀarrestinꢀ2
recruitment for ligands with an 5ꢀhydroxyꢀ2ꢀaminotetraline head group. Importantly,
inversion of the stereogenic center into its (R)ꢀconfiguration converted the potent agonists into
partial agonists with only moderate potencies (Emax 49ꢀ51%, EC50 161ꢀ218 nM). Since the
same compounds were found to fully activate the panel of Gα proteins, the inversion of the
i/o
stereogenic center converted the two balanced agonists (S)ꢀ13a,b into Gꢀproteinꢀbiased
ligands ((R)ꢀ13a,b). The apparent preference for G protein mediatedꢀsignaling of (R)ꢀ13a,b
was further confirmed by a quantitative analysis of ligand bias using the operational model of
4
3ꢀ45
agonism.
As representative pathways, we selected the activation of GαoA as G protein
readout and compared it to the recruitment of βꢀarrestinꢀ2 determined in the PathHunter assay.
ACS Paragon Plus Environment

Page 21 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
While (S)ꢀ13a/b showed only small differences from the reference ligand quinpirole with a
slight preference for the arrestin signaling (ꢁꢁlog(τ/K
Table 2) the respective (R)ꢀenantiomers displayed an almost 10ꢀfold preference for the
canonical signaling pathway (ꢁꢁlog(τ/K ) = 0.96 and 0.88, bias factors 9.09 and 7.65 for (R)ꢀ
3a and (R)ꢀ13b respectively). Interestingly, the quantification also a revealed a preferential
A
) = ꢀ0.25 and ꢀ0.46, Supplementary
A
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
activation of βꢀarrestinꢀ2 signaling the oximeꢀsubstituted analogs (S)ꢀ17a,b (ꢁꢁlog(τ/K ) = ꢀ
A
0.65 and ꢀ0.58) which was statistically significant but not obvious from the visual analysis.
Since the operational model is not feasible for the analysis of very weak partial agonists or
46
antagonists, ligand bias between GαoA activation and βꢀarrestinꢀ2 recruitment could not be
quantified for the family of 1,4ꢀDAPs.
Considering the importance of cell background, kinetics and experimental conditions
3
8, 47
in the context of putative ligand bias,
we intended to complement our βꢀarrestin
recruitment studies by an orthogonal BRET assay. Thus, we employed a recently described
approach, which is based on enhanced bystander BRET between RLucIIꢀβꢀarrestinꢀ1 or βꢀ
arrestinꢀ2 and the CAAXꢀmotif of KRas Cꢀterminally fused to Renilla reniformis green
4
8
fluorescent protein (rGFP). This system allowed to asses G protein activation and βꢀarrestin
recruitment on the same time scale (within minutes) and in the same cell background
(HEK293). We also coexpressed G proteinꢀcoupled receptor kinase 2 (GRK2), as it has been
1
2
2
shown to enhance ligandꢀinduced βꢀarrestin recruitment, especially for D R partial agonists.
Using this setup, highly similar results were obtained for pyrazolo[1,5ꢀa]pyridineꢀderived test
compounds. Indeed, the propensity to induce βꢀarrestin recruitment at D R (Table 4) or D R
2
S
2L
(Supplementary Table 1) was confirmed to depend on the orthosteric head group. While (S)ꢀ
aminotetralines acted as full agonists (Emax 83ꢀ100 %, EC50 10.5ꢀ17.7 nM), their (R)ꢀ
enantiomers showed moderate partial agonist characteristics (Emax 26ꢀ31 %, EC50 48ꢀ98 nM)
for the two βꢀarrestin isoforms. Again no significant βꢀarrestin recruitment was observed for
ligands containing a 2ꢀmethoxyphenylpiperazine pharmacophore. These antagonist properties
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 71
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
were further confirmed by inhibition experiments with both, the PathHunter and BRETꢀbased
assays, when the ligands were able to inhibit the effect of 0.1 µM or 1 µM quinpirole,
respectively (Supplementary Figure 6aꢀd).
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
ACS Paragon Plus Environment

Page 23 of 71
Journal of Medicinal Chemistry
1
2
3
Table 4. Functional properties of selected pyrazolo[1,5ꢀa]pyridines compared to the reference 1a, for the
activation of the five different Gα subunits D R and βꢀarrestin recruitment, determined by BRET or the
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
i/o
2S
PathHunter assay.
†
D
2SR activation
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
BRET
PathHunter
βꢀarr2
Gαi1
Gαi2
Gαi3
GαoA
GαoB
βꢀarr1
βꢀarr2
‡
§
‡
§
‡
§
‡
§
‡
§
‡
§
‡
§
‡
§
EC50
E
max EC50
E
max EC50
E
max EC50
E
max EC50
E
max EC50
E
max EC50
E
max
EC50
E
max
3.0 100
0.5
5.1 100
± 1 ± 1.3
5.9 100
± 1 ± 0.7
1.2 100
1.2 100 295 100 151 100
±1 ± 100 ± 1 ± 10 ± 1
64 100
quinpirole
dopamine
±
±
± 1 ± 0.10
± 1 ± 0.1
± 8
± 1
2.8 100
3.9 102
± 6 ± 0.6
4.5
± 3 ± 0.2
95 0.67
± 3 ± 0.17
92 0.78
± 3 ± 0.15
87 226 115 163 113
±5 ± 53
156
92
0.6
± 7 ± 35
± 5
± 36
± 4
14.9
7.8
70 19.4
± 2 ± 9.7
71 17.5
± 2 ± 7.5
62
8.7
81
4.8
± 6 ± 1.9
82
±5
††
††
††
††
1a
n.d.
n.d.
< 15 n.d.
< 10
n.d.
<10
±
± 1 ± 4.2
2
0.2
.1
53
4.4
± 5 ± 0.5
47
5.0
± 6 ± 0.2
39
1.5
81 1.9
±2 ± 0.2
81
±3
††
††
12a
< 10 n.d.
< 10
n.d.
< 10
±
± 5 ± 0.10
0.30
0.02
95 0.61 102 0.81 102 0.13
± 1 ± 0.05 ± 1 ± 0.05 ± 2 ± 0.01
98 0.11
± 1 ± 0.01
99 13.7
± 2 ± 2.8
95 15.9 100
± 2 ± 1.3
3.0
93
(
S)ꢀ13a
±
±
± 4
± 0.4
± 5
5.5
93 8.7 93 9.6 83 1.7
± 1 ± 0.8 ± 2 ± 0.7 ± 2 ± 0.4
98 2.2 ± 100
± 3 0.4 ± 2 ± 40
98
31 61
± 4 ± 20
26
± 1
218
51
(
R)ꢀ13a
S)ꢀ13b
±
0.2
± 33
± 3
0.30
0.01
97 0.65 103 0.80 101 0.12
99 0.12 102 14.5 100 10.5
± 1 ± 0.01
99
± 5
1.3
97
(
± 1 ± 0.03
± 3 ± 0.18
± 1 ± 0.01
± 1 ± 3.0
± 4 ± 1.0
31 54
± 0.1
± 5
5
0.7
.2 101
8.4
± 3 ± 0.8
96 9.8
± 1 ± 1.2
88 1.5
95
1.7 100
± 1 ± 0.1
48
27
± 3
161
49
(
R)ꢀ13b
±
± 2 ± 0.3
47 2.6
± 2 ± 3.0
± 4 ± 22
± 22
± 3
4
0.2
.2
61
7.7
± 1 ± 1.2
54 10.1
± 3 ± 0.78
84
2.6
± 1 ± 0.2
82
± 2
††
††
††
1
4a
n.d.
< 10 n.d.
< 10
n.d.
< 10
±
± 3 ± 0.6
0
0.04
.31
97 0.54 100 0.72
± 1 ± 0.18 ± 1 ± 0.12
99 0.13
± 1 ± 0.02
99 0.12 100 17.7
± 1 ± 0.02 ± 1 ± 5.5
95 13.4
± 6 ± 1.2
96
± 2
3.0
97
(
S)ꢀ15a
S)ꢀ15b
±
±
± 0.9
± 4
0.39 103 0.71 106 0.74 102 0.11
99 0.11 100 12.5
± 1 ± 0.01 ± 1 ± 1.6
99 12.4
± 3 ± 1.5
91
± 6
2.9
96
(
0.05
± 1 ± 0.12
± 1 ± 0.02
± 2 ± 0.01
± 0.8
± 4
6
0.7
.8
59 10.7
± 5 ± 0.3
53 13.1
± 5 ± 2.8
45 2.9
82
4.3
±2 ± 0.4
86
± 2
††
††
††
††
††
1
6a
n.d.
n.d.
n.d.
< 10 n.d.
< 10
< 10
< 10
n.d.
< 10
< 10
< 10
±
± 4 ± 0.8
43 1.8
3
0.9
.7
59
5.7
± 6 ± 0.2
56
5.7
± 8 ± 0.8
86 2.1
84
± 1
††
††
1
6b
< 10 n.d.
n.d.
±
±
± 5 ± 0.4
±1 ± 0.2
77 2.0
5.0
62
5.8
± 3 ± 2.2
59 6.2
± 2 ± 2.4
50 2.5
± 2 ± 0.7
82
± 4
††
††
1
6c
< 15 n.d.
n.d.
2.2
± 5 ± 0.7
0
0.09
.39
99 0.60 103 0.94
± 4 ± 0.02 ± 3 ± 0.12
99 0.19
± 2 ± 0.02
99 0.16 103 13.1
± 1 ± 0.01 ± 1 ± 1.3
92 12.1
± 6 ± 1.3
96
± 5
1.5
99
(
S)ꢀ17a
S)ꢀ17b
±
±
± 0.2
± 4
0.49
99 0.67 101 0.96
± 1 ± 0.08 ±1 ± 0.14
98 0.21 101 0.17 101 11.9
±1 ± 0.04 ±2 ± 0.03 ± 1 ± 1.6
83 13.2
±2 ± 2.2
92
±2
1.7 105
± 0.4 ± 6
(
0.08
†
§
‡
Data represent mean ± S.E.M. from three to six independent experiments, each performed in duplicates. EC50 given in nM.
††
Emax relative to the effect of vehicle (0 %) and the saturating effect of quinpirole (100 %). Not determined.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 71
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
Activation of inwardly rectifying potassium channels. In addition to the
measurements of nucleotide exchange or resonance energy transfer at the G protein level, we
examined the intrinsic activity of representative partial agonists further downstream at the
level of G proteinꢀcoupled, inwardly rectifying potassium (GIRK) channels. GIRKꢀmediated
signaling is involved in a number of important physiological processes, including nociception,
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
49
rewardꢀrelated behavior, cognition, mood and the regulation of the heart rate. GIRK
+
50
activation increases outward K ꢀcurrents, leads to cellular hyperpolarization and is therefore
51
known as key inhibitory mediator for the control of electrical excitability. Recently, a
genomeꢀwide association study revealed a link between the KCNJ3, the gene encoding
52,
GIRK1, and the susceptibility for schizophrenia and bipolar disorders in Asian populations.
5
3
Furthermore, it has been demonstrated that lithium exerts its efficacy in bipolar disorders
5
4
2 3
by a dual regulation of GIRKs. Both, D R and D R have been shown to couple to GIRK
55, 56
channels in vitro.
Interestingly, typical antipsychotics including haloperidol act as
R mediated GIRK activation, while newer
generation antipsychotics including 1a have been described as antagonists or partial
antagonists or even inverse agonists for D
2
57, 58
agonists.
To compare the properties of the high affinity, high potency pyrazolo[1,5ꢀ
a]pyridine 16c with the marketed drug 1a and the endogenous agonist dopamine, we
59
performed voltage clamp experiments in analogy to a previously described procedure. When
HEK293T cells expressing either D2LR or D R together with GIRK1/2, the predominant
3
60
GIRK heteromer in the CNS, were stimulated with the endogenous agonist dopamine,
+
robust K ꢀcurrents were observed (Figure 4a). Currentꢀvoltage diagrams revealed typical
inwardly rectifying profiles (Supplementary Figure 7a,b) and currents were efficiently
2
+
2 3
blocked by application of Ba or the D R/D Rꢀantagonist haloperidol (Figure 4a,
Supplementary Figure 7c,d). Currents were evoked in a Gαi/oꢀspecific manner, as overnight
incubation with pertussis toxin completely abolished the dopamineꢀmediated effects
(Supplementary Figure 7e,f). Although not to the same extent as dopamine, phenylpiperazines
ACS Paragon Plus Environment

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
a and 16c stimulated GIRK1/2 activation when they were applied at a concentration of 1 µM
to cells expressing D2LR and GIRK1/2. Again, the effect was inhibited by application of the
2
+
pore blocker Ba . At lower concentration (0.1 µM) only 16c but not 1a induced a significant
effect (Supplementary Figure 7g) pointing towards a higher potency of the pyrazolo[1,5ꢀ
a]pyridine.
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
Figure 4. Ligandꢀinduced activation of GIRK1/2 channels. When HEK293T cells expressing D R and
2
L
+
GIRK1/2 cells were stimulated with 10 µM dopamine (a) 1 µM 1a (b) or 1 µM 16c (c), a robust K current is
+
induced at the holding potential of ꢀ70 mV in high K external solution. Currents are completely inhibited by
2
subsequent addition of a 2 mM BaCl into the bath solution. Quantification of D2LR activation from ramp
currents shows a partial response is elicited by 1a and 16c compared to the endogenous agonist dopamine (d). In
HEK293T cells expressing D R and GIRK1/2 only dopamine, but not 1 µM solutions of 1a or 16c evoke a
3
+
significant K current (e). Data represent mean ± S.E.M. from 9 to 30 individual cells, VEH = vehicle, ** = p <
0.01, *** p < 0.001, oneꢀway ANOVA.
+
In the absence of receptors, neither dopamine nor 1a or 16c elicited an increase in K ꢀ
currents (Supplementary Figure 7h). In cells expressing the D R, no effect was observed for
3
the partial agonists (Figure 4e), indicating a different capacity of the ligands to stimulate
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 71
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
GIRK1/2 by means of D R or D R. Moreover, differences in GIRK1/2 activation may reflect
2L
3
3
the distinctions observed in signaling signatures for D2LR and D R observed by BRET. While
D
2LR is capable to equally engage Gα and Gα , D R shows a strong preference for the Gα
i
o
3
o
subtypes and only minor coupling to Gαi1ꢀ3 (Supplementary Figure 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
In vivo antipsychotic activity. In order to examine if G proteinꢀbiased D
2
R/D
3
R
partial agonists comprising a pyrazolo[1,5ꢀa]pyridine appendage possess antipsychotic
activity in vivo, we investigated the representative test compound 16c in comparison to 1a,
regarding their effects on schizophreniaꢀrelated behaviors in rats. As readouts, we examined
locomotion and anxiety, which have been used extensively in animal models of
61ꢀ63
schizophrenia.
Additionally, we investigated light induced activity (LIA), a nonꢀ
64, 65
aversively motivated sensorimotor processing measure,
to assess the intactness of the
64, 66
sensorimotor gating system.
In order to induce schizophreniaꢀlike alterations in animals,
we used an amphetamine (AMPH)ꢀsensitization regimen that has been developed by Pelegꢀ
62
Raibstein et al. and shown to effectively induce these disruptions. Thus, after a six day
sensitization with escalating doses of AMPH, antipsychotic 1a, test compound 16c (both at a
dose of 1.5mg/kg/day) or vehicle were continuously administered via an osmotic Alzet mini
pump over the course of seven days. LIA was measured on day five after mini pump
implantation and behavioral testing in the open field (OF) environment was performed on
treatment day seven before (baseline) and after an acute AMPH (1.5 mg/kg) challenge. Preꢀ
planned comparisons revealed a decrement in the baseline locomotor activity in all treatment
groups relative to the SAL/VEH control group (Figure 5a,b, Supplementary Results).
Compared to this baseline activity, we found increased horizontal and vertical (rearing)
locomotor responses to the low dose amphetamineꢀchallenge in AMPHꢀsensitized animals,
67
which has been frequently reported (for a review, see ref. ). These elevations were reversed
by continuous treatment with both 1a and 16c. 16c treatment was arguably more effective
than 1a treatment in inhibiting AMPHꢀinduced hyperꢀlocomotor activity (Figure 5d,g,e,h).
ACS Paragon Plus Environment

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
However, baseline activity was influenced by the treatments. Indeed, 16c treatment inhibited
the baseline locomotor activity as opposed to no alterations observed with 1a (Figure 5a,b).
6
8
Central activity in the OF test is a useful parameter to estimate the anxiety level of animals,
69
and several comorbid anxiety disorders have been associated with AMPH abusers 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
7
0, 71
schizophrenia patients.
AMPHꢀpretreated animals showed elevated anxiety, as indicated
by a decreased time spent in the center of the OF (Figure 5c). Treatment with 1a or 16c did
not alter baseline anxiety levels. However, 16c but not 1a, at least partially reversed the
effects of an acute amphetamine challenge (Figute 5f,i). Finally, AMPHꢀsensitization caused
disruptions in sensorimotor gating, which was measured by lightꢀinduced activity test. These
disruptions were partially improved by both 1a and 16c treatments (Supplementary Figure
8a,b). The results show that 16c treatment effectively reverses amphetamineꢀinduced hyperꢀ
locomotor activity in rats, but unlike 1a, 16c treatment exacerbates the baseline locomotionꢀ
related attenuations induced by AMPHꢀpretreatment. Extrapyramidal deficits are common
72, 73
side effects of many antipsychotic drugs, including both typical and atypical ones.
Altogether, these results indicate that 16c treatment can effectively improve schizophreniaꢀ
like disruptions in vivo, but as a side effect, may reduce general locomotor activity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 71
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
b
c
d
e
f
L
o
c
o
m
o
ontion
g
AUC
Bas
B
e
a
l
s
inel
e
in
L
e
o
Lo
c
c
o
o
m
mo
o
ti
t
o
i
n
on
L
o
c
o
moti
AUC
50
35
s
S
a
A
l
L/V
/V
/1
/16c
ꢀ
E
V
H
EH
100
80
60
40
20
0
Amphetamine
AMPH
***
a
A
m
M
p
P
h
H
ꢀ
V
E
E
H
H
30
40
a
A
m
M
p
P
h
H
ꢀ
A
a
RI
25
a
A
m
M
p
P
h
H
ꢀ
M
S
#
30
*
*
2
0
5
#
1
2
0
0
0
*
**
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
10
5
1
0
5
10
15
i ni n
20
T
T
i
i
m
m
e
e
[
(
m
m
]
)
h
AUC
AUC
Bas
B
e
asli
e
n
li
e
ne
R
R
e
e
a
ar
r
inin
g
g
Re
Amphetamine
a
R
ri
e
ngring
a
60
80
70
60
sa
S
A
l
L/V
/V
/1
/1
ꢀ
E
V
H
EH
250
200
AMPH
am
A
M
p
P
h
H
ꢀ
V
E
E
H
H
50
***
am
A
M
p
P
h
H
ꢀ
A
a
RI
am
A
M
p
P
h
H
ꢀ
M
6
S
c
40
#
#
50
150
*
*
30
40
#
##
#
100
30
20
*
**
20
50
10
10
0
0
0
5
10
15
[ m( mi ni n
20
T
T
i
i
m
m
e
e
]
)
i
Baseline Center Time
Center Time
Center Time
AUC
AUC
Baseline Center Time
120
100
sa
S
A
l
L/V
/V
/1
/1
ꢀ
E
V
H
EH
250
200
150
100
50
Amphetamine
AMPH
a
A
m
M
p
Ph
H
ꢀ
VE
E
H
H
100
a
A
m
M
p
P
h
H
ꢀ
A
a
RI
80
60
40
20
0
a
A
m
M
p
P
h
H
ꢀ
M
6
S
c
8
0
0
6
*
*
*
40
*
*
20
0
0
5
10
15
i ni n
20
T
T
i
i
m
m
e
e
[
(
m
m
]
)
Figure 5. The effects of 16c compared to 1a on schizophreniaꢀlike behavior in rats. Animals received i.p. injections of
amphetamine (AMPH) with a sensitization regimen or saline (SAL). Subsequently, the animals were continuously treated with 1a
1.5 mg/kg/day), 16c (1.5 mg/kg/day), or vehicle (VEH) for 7 days. Baseline activity: Baseline horizontal locomotion (a), rearing
b) and center duration (c) in the open field on treatment day 7 are shown for SAL/VEH (n = 11), AMPH/VEH (n = 12),
AMPH/1a (n = 12), and AMPH/16c (n = 9) groups as a 20 min total activity. AMPHꢀinduced activity: The effects of a challenge
AMPH injection (1.5 mg/kg) on horizontal locomotion (d), rearing (e), and center duration (f) are shown in 5 min intervals, and as
area under the curve (AUC) total activity (g, h, i respectively). AMPHꢀinduced behaviors are calculated as ∆ baseline (vs. last 5
min of baseline). Values are shown as mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001 compared to SAL/VEH. # p < 0.05, ##
(
(
p
<
0.01,
###
p
<
0.001
compared
to
AMPH/VEH.
ACS Paragon Plus Environment

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
DISCUSSION AND CONCLUSIONS
The combination of 1,4ꢀdisubstituted phenylpiperazines with pyrazolo[1,5ꢀa]pyridine
heterocyles by an appropriate linker led to a series of ligands with subnanomolar affinity for
3
5
2
the dopamine D R. Intensive pharmacological investigations employing classical [ S]GTPγS
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
binding experiments but also highly sensitive BRET biosensors revealed partial agonist
properties for the activation of Gαi/o proteins at D R, D R and D R. Receptor subtypeꢀ
2L
2S
3
selectivity as well as potencies and efficacies could be modulated by the substitution pattern
and the attachment of the linker subunit to the pyrazolo[1,5ꢀa]pyridine appendage. While
substituents in position 2 of the heterocycle or attachment of the linker via an aminocarbonyl
moiety led to relatively weak efficacies and/or potencies, oxymethylene spacers and the
addition of polar functional groups to position 3 of the heterocycle were highly beneficial for
the intrinsic activity of 1,4ꢀDAPꢀbased dopaminergics. Interestingly, the family of
phenylpiperazines preferentially activated the Gα
potencies when compared to Gα activation. At the same time, the compounds bearing either a
,3ꢀdichloroꢀ or 2ꢀmethoxyphenylpiperazine moiety were inhibitors of βꢀarrestin recruitment
o
subtypes, with both, higher efficacies and
i
2
at D R and D R, indicating substantial G protein bias.
2L
2S
When the substituted phenylpiperazine unit was replaced by a 5ꢀhydroxyꢀNꢀpropylꢀ2ꢀ
aminotetraline group, superꢀaffinity dopamine receptor agonists were discovered. Here,
receptor affinities and potencies were only slightly influenced by substitutions at the
pyrazolo[1,5ꢀa]pyridine appendage, but strongly dependent on the absolute configuration of
the stereogenic center. Compounds in the (S)ꢀconfiguration were found to be full agonists for
3
the activation of all signaling pathways at D2SR, D2LR and D R. In contrast, (R)ꢀ13a,b
activated G proteins to the full extent, but were only partial agonists for the recruitment of βꢀ
arrestins at D2SR and D2LR. Although these ligands might not serve as antipsychotics due to
their high efficacy for G proteinꢀmediated signaling, they present an interesting starting point
for further SAR for understanding the molecular determinants of functional selectivity at the
ACS Paragon Plus Environment