Antagonists for the orphan G-protein-coupled receptor GPR55 based on a coumarin scaffold

Article abstract of DOI:10.1021/jm4005175

The orphan G-protein-coupled receptor GPR55, which is activated by 1-lysophosphatidylinositol and interacts with cannabinoid (CB) receptor ligands, has been proposed as a new potential drug target for the treatment of diabetes, Parkinson's disease, neuropathic pain, and cancer. We applied β-arrestin assays to identify 3-substituted coumarins as a novel class of antagonists and performed an extensive structure-activity relationship study for GPR55. Selectivity versus the related receptors CB₁, CB₂, and GPR18 was assessed. Among the 7-unsubstituted coumarins selective, competitive GPR55 antagonists were identified, such as 3-(2-hydroxybenzyl)-5-isopropyl-8- methyl-2H-chromen-2-one (12, PSB-SB-489, IC₅₀ = 1.77 μM, pA ₂ = 0.547 μM). Derivatives with long alkyl chains in position 7 were potent, possibly allosteric GPR55 antagonists which showed ancillary CB receptor affinity. 7-(1,1-Dimethyloctyl)-5-hydroxy-3-(2-hydroxybenzyl)-2H- chromen-2-one (69, PSB-SB-487, IC₅₀ = 0.113 μM, K_B = 0.561 μM) and 7-(1,1-dimethylheptyl)-5-hydroxy-3-(2-hydroxybenzyl)-2H- chromen-2-one (67, PSB-SB-1203, IC₅₀ = 0.261 μM) were the most potent GPR55 antagonists of the present series.

Full text of DOI:10.1021/jm4005175

Subscriber access provided by Otterbein University
Article
Antagonists for the orphan G protein-coupled
receptor GPR55 based on a coumarin scaffold
Viktor Rempel, Nicole Volz, Franziska Gläser, Martin Nieger, Stefan Bräse, and Christa E Müller
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm4005175 • Publication Date (Web): 16 May 2013
Downloaded from http://pubs.acs.org on May 21, 2013
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 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Antagonists for the orphan G proteinꢀcoupled
receptor GPR55 based on a coumarin scaffold
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
Viktor Rempel, # Nicole Volz, § Franziska Gläser, § Martin Nieger, & Stefan Bräse, § and
Christa E. Müller#,*
#PharmaCenter Bonn, Pharmaceutical Institute, Pharmaceutical Chemistry I, An der Immenburg
§
4, Dꢀ53121 Bonn, Germany; Institute of Organic Chemistry, University of Karlsruhe, Karlsruhe
Institute of Technology (KIT), FritzꢀHaberꢀWeg 6, Dꢀ76131 Karlsruhe, Germany; &Laboratory of
Inorganic Chemistry, Department of Chemistry, P.O. Box 55 (A.I.Virtasen aukio 1), FINꢀ00014
University of Helsinki, Finland.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
1
2
3
4
Abstract
5
6
7
8
9
The orphan G proteinꢀcoupled receptor GPR55, which is activated by 1ꢀlysophosphatidylinositol
and interacts with cannabinoid (CB) receptor ligands, has been proposed as a new potential drug
target for the treatment of diabetes, Parkinson’s disease, neuropathic pain, and cancer. We
applied βꢀarrestin assays to identify 3ꢀsubstituted coumarins as a novel class of antagonists, and
performed an extensive structureꢀactivity relationship study for GPR55. Selectivity versus the
related receptors CB1, CB2 and GPR18 was assessed. Among the 7ꢀunsubstituted coumarins
selective, competitive GPR55 antagonists were identified, such as 3ꢀ(2ꢀhydroxybenzyl)ꢀ5ꢀ
isopropylꢀ8ꢀmethylꢀ2Hꢀchromenꢀ2ꢀone (12, PSBꢀSBꢀ489, IC50 1.77 ꢁM, pA2 0.547 ꢁM).
Derivatives with long alkyl chains in position 7 were potent, possibly allosteric GPR55
antagonists which showed ancillary CB receptor affinity. 7ꢀ(1,1ꢀDimethyloctyl)ꢀ5ꢀhydroxyꢀ3ꢀ(2ꢀ
hydroxybenzyl)ꢀ2Hꢀchromenꢀ2ꢀone (69, PSBꢀSBꢀ487, IC50 0.113 ꢁM, KB 0.561 ꢁM) and 7ꢀ(1,1ꢀ
dimethylheptyl)ꢀ5ꢀhydroxyꢀ3ꢀ(2ꢀhydroxybenzyl)ꢀ2Hꢀchromenꢀ2ꢀone (67, PSBꢀSBꢀ1203, IC50
0.261 ꢁM) were the most potent GPR55 antagonists of the present series.
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
Keywords
GPR18, GPR55, orphan GPCR, cannabinoid receptors, coumarin, antagonist,
ꢀ9ꢀtetrahydrocannabinol
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
1
2
3
4
Introduction
5
6
7
8
The various physiological effects of natural and synthetic cannabinoids cannot be fully
explained by their interaction with the two known cannabinoid (CB) receptor subtypes, CB1 and
CB2.1ꢀ3 An intensive search for additional CB receptor subtypes resulted in the identification of
the orphan receptor GPR55 as a promising candidate for a third CB receptor subtype.4, 5 Like the
CB receptors, GPR55 belongs to the rhodopsinꢀlike seven transmembrane G proteinꢀcoupled
receptor (GPCR) superfamily.6 Despite its low amino acid identity with CB1 (13.5%) and CB2
(14.4%) several cannabinoids were found to interact with GPR55, including the CB1ꢀselective
antagonist/inverse agonist rimonabant (1), the nonꢀselective CB1/CB2 full agonist CP55,940 (2),
and the partial CB agonist ꢂ9ꢀtetrahydrocannabinol (ꢂ9ꢀTHC, 3).3, 7ꢀ9
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
Figure 1. Structures of cannabinoids with GPR55 activity (1ꢀ3) and the endogenous GPR55
agonist 4.
Surprisingly, those ligands display opposite effects at the GPR55 as compared to CB receptors:
the CB1 antagonist 1 was found to activate GPR559, while the CB agonists 29 and 37 display
antagonistic effects at GPR55. Thus, GPR55 has been referred to as an “antiꢀcannabinoid”
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
1
2
3
4
5
6
7
8
9
receptor.10 GPR55 and CB2 receptors have recently been shown to interact and crossꢀtalk with
each other on the level of their signaling pathways, either synergistically or contrarily.11, 12 In
addition, GPR55 and CB1 receptors were shown to form receptor heteromers thereby modulating
each others function.13
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
Nevertheless the classification of GPR55 as a CB receptor is highly controversial. The
pharmacology of further cannabinoids, especially that of the endocannabinoids anandamide
(arachidonoylethanolamide) and 2ꢀarachidonoylglycerol at GPR55, is not clear.2, 8, 14 Ryberg et
al. reported an activation of GPR55 by anandamide and 2ꢀarachidonoylglycerol, while several
other research groups could not detect agonistic effects by these eicosanoids.3, 9, 15 The only
unambiguously confirmed endogenous agonist for GPR55 so far appears to be
1ꢀlysophosphatidylinositol (LPI), a lipid that does not interact with CB receptors.1, 3, 8 Okuno et
al. reported that the biological activity of arachidonic acidꢀcontaining LPI species was markedly
higher compared to species containing other fatty acids, indicating that 2ꢀarachidonoylꢀLPI (4)
might be the true endogenous agonist for GPR55.16, 17 Another potential CB receptor subtype
might be the orphan GPR18, since some cannabinoids, such as ꢂ9ꢀTHC and anandamide, were
recently shown to interact with that receptor, thereby further emphasizing the complex and not
yet fully understood pharmacology of cannabinoids.18ꢀ20
The GPR55 is a peculiar GPCR since it has not been found to couple to classical G proteins
(Gi, Gs, Gq), but only to G12,13ꢀproteins thereby activating rhoA (ras homolog gene family
member A) and ROCK (Rhoꢀassociated protein kinase), which may in turn induce phospholipase
Cꢀmediated inositol 1,4,5ꢀtriphosphate formation and subsequent calcium release from
11
intracellular stores.4,
In addition, GPR55ꢀmediated activation of the ERKꢀcascade was
confirmed in several studies.5, 7, 21
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mRNA for the receptor is highly expressed in the brain (highest levels in caudate nucleus,
nucleus accumbens, putamen and striatum), on cells and organs of the immune system
(lymphocytes, spleen), and in stomach and intestine.4 High GPR55 expression has also been
found on many cancer cells, and GPR55 expression appears to correlate with the proliferation
rate and aggressiveness of those cells.4, 22 Furthermore, inflammatory and neuropathic pain,
vasorelaxation and angiogenesis, and bone resorption may be mediated by GPR55.1, 4 Very
recently the LPI/GPR55 system has been found to be positively associated with obesity in
humans.23 Thus, GPR55 is of interest as a novel potential drug target, e.g., for the treatment of
diabetes, Parkinson’s disease, neuropathic pain, and cancer.3, 4, 22
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
Most of the published pharmacological studies investigating the physiological role of GPR55
were performed with antagonists known to interact with further receptors, such as GPR18 and
CB receptors. The lacking selectivity of the utilized compounds considerably limits their
suitability as pharmacological tools.3, 4, 19, 20, 24
In order to be able to conduct unequivocal in vitro and in vivo studies, and to validate GPR55
as a drug target, appropriate, selective tool compounds are urgently needed to further explore the
(patho)physiological roles of GPR55.4 Recently, the first synthetic GPR55 agonists and
antagonists have been identified by the screening of compound libraries, but they have been
neither fully characterized nor further optimized.8, 25 Thus, the goal of the present study was to
identify antagonists for GPR55 and to perform extensive studies on their SARs.
As many of the commercially available cannabinoids did not only activate CB1 and CB2
receptors, but were shown to additionally interact with GPR18 and GPR55, our approach was to
investigate a series of recently published coumarinꢀbased CB receptor ligands (see Figure 2) as
well as inactive analogs of the same series for their potential interaction with both orphan
receptors, GPR18 and GPR55. After obtaining several hits for GPR55, but not for GPR18, we
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
1
2
3
4
5
6
7
8
9
extended the series via optimizing the compounds for interaction with GPR55. This approach led
us to identify and develop potent and selective GPR55 antagonists and to carry out an extensive
SAR analysis for this pharmacologically attractive receptor.
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
Structural considerations
We recently reported on the development of coumarin derivatives as novel ligands for CB
receptors.26, 27 A comparison of the structural features of the 7ꢀalkylꢀ3ꢀbenzylcoumarins 5ꢀ7 with
recently described GPR55 antagonists isolated from the plant Cannabis sativa, ꢂ9ꢀ
tetrahydrocannabivarin (8), an analog of 3, and cannabidivarin (9), a cannabidiol analog, is
depicted in Figure 2.7 Compounds 8 and 9 were shown to exhibit antagonistic activity at GPR55:
they reduced the maximum effect of LPIꢀmediated ERK1/2 phosphorylation in human embyonic
kidney (HEK) 293 cells stably expressing the human GPR55 by 50% (8) and 56% (9) at a
concentration of 1 ꢁM.7
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 2. Structural comparison of recently identified cannabinoid receptor antagonists (5) and
agonists (6, 7) with the GPR55 antagonists ꢂ9ꢀtetrahydrocannabivarin (8) and cannabidivarin
(9).7, 27
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
It becomes evident that the lipophilic sideꢀchain of the GPR55 antagonists 8 and 9 is two
methylene units shorter as compared to those in the structurally related, coumarinꢀbased CB
receptor ligands 5ꢀ7. Since shorter sideꢀchains appeared to be favourable for GPR55 receptor
interaction, we decided to include coumarin derivatives with no or with only a short lipophilic
substituent in position 7 in our studies, most of which had not shown high affinity for CB
receptors in previous investigations.7, 26 Based on the obtained test results at GPR55 we
additionally synthesized a series of new compounds.
Syntheses
Syntheses of 3ꢀmethylꢀ and 3ꢀbenzylꢀsubstituted coumarins with modifications in various
positions were recently described.26ꢀ30 We established a straightforward procedure to obtain
coumarins in a oneꢀpot synthesis from appropriately substituted salicylaldehydes and α,βꢀ
unsaturated aldehydes.26ꢀ28 According to this procedure (see Experimental and Supporting
Information) further derivatives were obtained with the goal to optimize the compounds for
interaction with GPR55. The products were purified by flash chromatography. The structures
were confirmed by 1H and 13C NMR spectra, and in some cases by EIꢀHRMS spectra; purity was
confirmed by elemental analysis (for details see Experimental Section and Supporting
Information). For one final product, 14, a crystal structure was obtained (see Supporting
Information).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
1
2
3
4
5
6
Biological Evaluation
7
8
9
Activities at the human GPR55 were investigated in βꢀarrestin translocation assays using
Chinese hamster ovary (CHO) cells stably expressing the receptor.3, 9, 21 To monitor βꢀarrestin
recruitment to the activated receptors, the βꢀgalactosidase enzyme fragment complementation
technology was applied (βꢀarrestin PathHunter™ assay, DiscoverX, Fremont, CA, USA). In these
cells, the βꢀarrestin molecule is fused to an inactive deletion mutant of βꢀgalactosidase, which
functions as an enzyme acceptor (EA). The investigated receptor is extended by a small lowꢀ
affinity fragment derived from the sequence deleted in the EA.31 Activation of the receptor by a
ligand will cause recruitment of βꢀarrestin molecules to the receptor, resulting in a
complementation of the two fragments to form a functional enzyme, which is then able to cleave
an added detection reagent resulting in light emission.31 The compounds were initially screened
for agonistic and antagonistic activity at the receptor at a concentration of 10 ꢁM. Agonistic
effects of test compounds were compared to the effect of the GPR55 agonist LPI at a
concentration of 1 ꢁM (set at 100%). IC50 values were determined for compounds showing an
inhibition of >60% of the LPI signal at 1 ꢁM. Testing of the compounds at the human GPR18
was carried out by the same assay technology. For receptor stimulation, the GPR18 agonist 3 was
used at a concentration of 10 ꢁM. The affinities of the 3ꢀbenzylcoumarin derivatives at human
CB receptors were determined in radioligand binding studies as previously described.26, 27 As a
CB receptor radioligand [3H](ꢀ)ꢀcisꢀ3ꢀ[2ꢀhydroxyꢀ4ꢀ(1,1ꢀdimethylheptyl)phenyl]ꢀtransꢀ4ꢀ(3ꢀ
hydroxypropyl)cyclohexanol (2) was used. As a source for human CB1 and CB2 receptors,
membrane preparations of CHO cells stably expressing the respective receptor subtype were
utilized. Rat CB1 receptors were obtained from rat brains as previously described.26 Initially the
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
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds were screened at a concentration of 10 ꢁM. In cases where inhibition of radioligand
binding was about 50% or more, full concentrationꢀinhibition curves were determined in order to
calculate Ki values. Functional properties at CB receptors were investigated in cAMP assays
using CHO cells stably expressing the respective human CB receptor subtype as previously
described.27 Effects of test compounds (1 ꢁM) on forskolinꢀstimulated cAMP levels were
determined relative to the maximal effect observed with the full agonist 2.
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
Structureꢀactivity relationships
To validate the applied GPR55 test system, we initially investigated some standard ligands. 3
had been reported to be an agonist,32 an antagonist,7 or to display no activity at GPR55.9 In our
experiments measuring GPR55ꢀinduced βꢀarrestin recruitment 3 behaved as a moderately potent
antagonist at GPR55 (IC50 14.2 ꢁM, see Table 1). We had previously reported that 2 behaved as a
relatively potent GPR55 antagonist displaying an IC50 value in the low micromolar range (1.61
ꢁM).33 This is in agreement with results published by other groups.8, 9 Furthermore, we recently
confirmed that 1 acts as a GPR55 agonist33 (see Table 1) in agreement with literature data.8, 21
None of the coumarins investigated in the present study showed agonistic activity at GPR55.
However, many coumarin derivatives were found to inhibit LPIꢀinduced GPR55 activation (see
Table 1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
1
2
3
4
Table 1. Potencies of coumarin derivatives at GPR55a
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
Compd
R3
R5
R6
R7
R8
βꢀarrestin assay
human GPR55
IC50 ± SEM (ꢁM)
Standard agonists and antagonists
1.00 ± 0.25
(agonist, EC50)
2.0133
4
1
2
3
(agonist, EC50)
1.6133
14.2 ± 5.4
Coumarin derivatives I: with small 7ꢀsubstituents
methyl
benzyl
isopropyl H
isopropyl H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
methyl
methyl
methyl
methyl
isopropyl
H
3.45 ± 0.36
5.33 ± 1.10
1.77 ± 0.23
7.14 ± 2.66
5.70 ± 1.62
> 10 (46%)b
> 10 (40%)b
> 10 (28%)b
> 10 (13%)b
~ 10 (54%)b
> 10 (45%)b
> 10 (16%)b
> 10 (7%)b
> 10 (5%)b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2ꢀhydroxyꢀ isopropyl H
benzyl
2ꢀmethoxyꢀ methyl
benzyl
2ꢀmethoxyꢀ methyl
benzyl
H
H
benzyl
methyl
benzyl
methyl
benzyl
methyl
methyl
methyl
benzyl
H
H
H
H
H
H
H
H
H
I
I
H
H
H
methoxy
H
methoxy
H
Cl
Br
H
Br
methoxy
methoxy
methoxy
H
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
see above for structure
see above for structure
2.81 ± 1.16
> 10 (28%)b
> 10 (43%)b
> 10 (6%)b
> 10 (2%)b
9.38 ± 0.58
~ 10 (54%)b
3.99 ± 0.75
> 10 (45%)b
6.74 ± 2.04
> 10 (42%)b
0.981 ± 0.140
12.8 ± 3.2
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
benzyl
benzyl
benzyl
benzyl
methoxy
hydroxy
methoxy
methoxy
H
H
H
H
H
H
H
H
H
H
H
H
H
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
H
Br
H
H
3ꢀmethoxyꢀ methoxy
benzyl
3ꢀmethoxyꢀ methoxy
benzyl
2ꢀmethylꢀ hydroxy
benzyl
H
H
Br
H
methyl
methyl
methyl
H
benzyl
methoxy
H
4ꢀmethylꢀ hydroxy
benzyl
2ꢀmethoxyꢀ methyl
benzyl
H
methoxy methyl
methoxy methyl
hydroxy methyl
methoxy methyl
hydroxy methyl
methoxy methyl
hydroxy methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
H
benzyl
methyl
methyl
methyl
methyl
benzyl
9.32 ± 1.05
13.5 ± 4.3
4ꢀfluoroꢀ
benzyl
4ꢀfluoroꢀ
benzyl
10.3 ± 0.7
2ꢀmethylꢀ methyl
benzyl
2ꢀmethylꢀ methyl
benzyl
3ꢀmethoxyꢀ methoxy
benzyl
3ꢀmethoxyꢀ methoxy
benzyl
7.69 ± 1.71
5.16 ± 0.73
> 10 (13%)b
> 10 (18%)b
~ 10 (53%)b
> 10 (25%)b
H
H
H
H
hydroxymethyl
bromomethyl
methoxy
H
benzyl
H
H
methyl
H
methoxy
H
Coumarin derivatives II: 7ꢀpentylꢀsubstitution
4ꢀmethoxyꢀ methoxy
3,5ꢀ
dimethylꢀ
benzyl
2ꢀmethoxyꢀ methoxy
benzyl
H
pentyl
H
> 10 (44%)b
46
H
H
pentyl
pentyl
H
H
6.35 ± 2.66
> 10 (37%)b
47
48
2ꢀhydroxyꢀ hydroxy
benzyl
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
1
2
3
4
5
6
7
8
9
3ꢀmethoxyꢀ methoxy
benzyl
3ꢀhydroxyꢀ hydroxy
benzyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
pentyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
3.23 ± 0.31
10.6 ± 4.9
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
benzyl
methoxy
> 10 (36%)b
> 10 (0%)b
benzyl
hydroxy
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
2ꢀmethylꢀ methoxy
benzyl
2ꢀmethylꢀ hydroxy
benzyl
3ꢀmethylꢀ methoxy
benzyl
3ꢀmethylꢀ hydroxy
benzyl
5.08 ± 1.05
> 10 (19%)b
> 10 (27%)b
> 10 (19%)b
9.00 ± 2.44
> 10 (0%)b
2ꢀchloroꢀ
benzyl
2ꢀchloroꢀ
benzyl
3ꢀchloroꢀ
benzyl
3ꢀchloroꢀ
benzyl
4ꢀchloroꢀ
benzyl
4ꢀfluoroꢀ
benzyl
4ꢀbromoꢀ
benzyl
methoxy
hydroxy
methoxy
hydroxy
methoxy
methoxy
methoxy
> 10 ꢁM (36%)b
> 10 (30%)b
3.29 ± 1.30
~ 10 (57%)b
3.76 ± 1.46
Coumarin derivatives III: long, branched 7ꢀsubstituent
benzyl
benzyl
methoxy
hydroxy
H
H
H
H
H
H
H
H
H
H
1,1ꢀdimethylheptyl
1,1ꢀdimethylheptyl
1,1ꢀdimethylheptyl
1,1ꢀdimethylheptyl
1,1ꢀdimethyloctyl
1,1ꢀdimethyloctyl
1ꢀbutylcylopentyl
1ꢀbutylcylopentyl
1ꢀbutylcylcohexyl
1ꢀbutylcylcohexyl
H
H
H
H
H
H
H
H
H
H
~ 10 (51%)b
0.358 ± 0.089
> 10 (25%)b
0.261 ± 0.181
> 10 (31%)b
0.113 ± 0.020
> 10 (30%)b
0.854 ± 0.454
> 10 (44%)b
0.961 ± 0.431
64
65
66
67
68
69
70
71
72
73
2ꢀmethoxyꢀ methoxy
benzyl
2ꢀhydroxyꢀ hydroxy
benzyl
2ꢀmethoxyꢀ methoxy
benzyl
2ꢀhydroxyꢀ hydroxy
benzyl
2ꢀmethoxyꢀ methoxy
benzyl
2ꢀhydroxyꢀ hydroxy
benzyl
2ꢀmethoxyꢀ methoxy
benzyl
2ꢀhydroxyꢀ hydroxy
benzyl
aAll data result from three independent experiments, performed in duplicates
bPercent inhibition of LPI (1 ꢁM)ꢀinduced βꢀarrestin recruitment by test compounds (10 ꢁM)
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
The investigated coumarin derivatives can be divided into three groups according to the
residues in position seven of the coumarin scaffold (R7): the first group consists of coumarins
with small residues in that position (H, Br, methyl, hydroxymethyl, bromomethyl, methoxy, 10ꢀ
45). The second group contains coumarins with a pentyl residue (46ꢀ63), and the third group
comprises compounds with branched aliphatic residues in position 7 (64ꢀ73). The nature of the 7ꢀ
substituent had a big influence on the activity of the coumarin derivatives at GPR55. Even more
so, it had been found to be crucial for affinity towards CB receptors.27 Compounds with no or
small moieties in position 7 of the coumarin (coumarin derivatives type I) had been shown to
generally exhibit no or only low affinity for CB receptors (see Table 3 and references 8, 25). In
contrast, all compounds of this group, that possessed at the same time an alkyl residue like
methyl or isopropyl at the 8ꢀposition, were identified as antagonists at GPR55 (Table 1).
Regarding the 3ꢀsubstituent, different residues from a small methyl (10) to a larger, aromatic 2ꢀ
methoxybenzyl (13) were tolerated (e.g. compounds 10, 13, 11, 12). At the 5ꢀposition a
substituent appeared to be required since all 5ꢀunsubstituted derivatives were inactive. Methyl,
isopropyl, and methoxy residues were well tolerated at C5. Isopropyl substitution at that position
conferred GPR55 selectivity versus CB receptors (see below).
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
The 6ꢀposition remained unsubstituted in most cases, but substituents like hydroxy (41) and
methoxy (35) were tolerated when combined with small substituents in other positions. Hydroxy
at the 6ꢀposition appeared to be slightly, but consistently superior to methoxy substitution
(compare 36/37, 38/39, 40/41). An annelated benzene ring in the 5,6ꢀposition increased potency
(24).
In position 7 a substituent was not required, but small residues such as bromo and methyl were
accepted and could even somewhat improve potency (compare 30/31, 29/33). Larger, lipophilic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
1
2
3
4
5
6
7
8
9
residues in the 7ꢀposition were also tolerated and led to an increase in affinity in some cases (see
below, coumarin subgroups II and III). Interdependence of substituents in different positions was
observed. In case of the benzocoumarins (24, 25) only the 3ꢀbenzylꢀsubstituted derivative 24 was
active, while the 3ꢀmethyl derivative 25 did not display any significant GPR55ꢀinhibitory
activity.
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
The second subgroup of investigated coumarins consisted of derivatives with a pentyl moiety
in the 7ꢀposition. The presence of a lipophilic residue in that position had been shown to be
essential for good affinity at CB receptors, thus all compounds of this subgroup displayed high
affinity for both CB receptor subtypes (see Table 3).27 Coumarins of this subgroup possessing a
methoxy group in the 5ꢀposition were shown to exhibit higher CB receptor affinity than their
hydroxy analogues.27 Interestingly, the same was true for the potency of the compounds at human
GPR55: all compounds with a methoxy moiety in that position showed a somewhat higher
antagonistic potency at GPR55 than their hydroxylꢀsubstituted analogs (see Table 3). In the 7ꢀ
pentyl series (coumarins II) benzyl residues with various substituents were present in the 3ꢀ
position. Substitution of the benzyl residue was required to achieve good inhibitory activity at
GPR55: 3ꢀmonosubstitution in the orthoꢀ, metaꢀ, or paraꢀposition led to compounds with IC50
values in the low micromolar range (e.g. oꢀmethoxy (47), oꢀmethyl (53), mꢀmethoxy (49), pꢀCl
(61), pꢀBr (63)).
The compounds of the second subgroup displayed a comparable profile as the plantꢀderived
cannabinoid prototype 3 and its synthetic derivative 2, binding to both CB receptor subtypes and
acting as antagonists with moderate potency at GPR55. The difference of the presented series in
comparison with 3 or 2 was that the identified GPR55 antagonists showed a complex functional
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
behavior at CB receptors. For example, GPR55 antagonists that displayed at the same time
CB1/CB2ꢀagonistic (47, 49), CB1/CB2ꢀantagonistic (50, 63) or CB1ꢀantagonistic/CB2ꢀagonistic
activity (53, 61) were identified (for intrinsic activities of selected compounds see Supporting
Information and reference27).
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
Coumarins of series III with branched aliphatic chains in the 7ꢀposition, i.e. 1ꢀbutylcyclopentyl,
1ꢀbutylcyclohexyl, 1,1ꢀdimethylheptyl, or 1,1ꢀdimethyloctyl residues, displayed different
structureꢀactivity relationships at GPR55 than their pentylꢀsubstituted analogs (series II). We had
previously observed that in this series III 5ꢀhydroxylꢀsubstituted coumarin derivatives were more
potent at CB receptors than their methoxyꢀsubstituted analogs –in contrast to the 7ꢀpentyl series
II (see Table 3).27 Similar SARs were observed at GPR55: while compounds with a methoxy
group in the 5ꢀposition showed only weak inhibtory effects on LPIꢀmediated βꢀarrestin
recruitment, all of the respective hydroxy analogues fully inhibited LPIꢀinduced effects in initial
screening assays at 10 ꢁM concentration (for results see Table 1). IC50 values of all further
investigated coumarin derivatives from series III were in the submicromolar concentration range.
The rank order of potency for the 7ꢀsubstituent was as follows: 1,1ꢀdimethyloctyl (69) > 1,1,ꢀ
dimethylheptyl (67) > 1ꢀbutylcyclopentyl (71) ≥ 1ꢀbutylcyclohexyl (73). The most potent GPR55
antagonist of the present series was 69 (7ꢀ(1,1ꢀdimethyloctyl)ꢀ5ꢀhydroxyꢀ3ꢀ[(2ꢀ
hydroxyphenyl)methyl]chromenꢀ2ꢀone) with an IC50 value of 0.113 ꢁM. In the 3ꢀposition, benzyl
as well as 3ꢀhydroxybenzyl was well tolerated.
Mechanism of GPR55 inhibition
The pharmacology of cannabinoids at GPR55 appeared to be complex, and published data are
in some cases inconsistent. For instance, in some studies 1 and 3 were found to behave as GPR55
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
1
2
3
4
agonists, while in others antagonistic activities were reported for those compounds.3, 7, 24, 32
A
5
6
7
8
recently published study indicated that some cannabinoids may bind to a binding site distinct
from the orthosteric LPI site and thus may represent allosteric rather than orthosteric ligands.7
Due to the lack of an appropriate radioligand and/or mutagenesis studies these findings have yet
to be definitely confirmed.7 However, an allosteric binding site would explain the inconsistent
data concerning the intrinisic activities of some cannabinoids, as allosteric modulation is known
to be highly probeꢀ, systemꢀ and assayꢀdependent.34, 35
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
Our data indicate that 7ꢀpentylcoumarins and coumarins with a branched, long alkyl chain in
position 7 might also interact with an allosteric binding site of GPR55, while compounds without
a lipophilic moiety in that position are most likely to be orthosteric antagonists, indicating that
the residue at position R7 might be crucial for the compounds’ binding mode at GPR55. Figure 3
shows examplary curves for one compound of each subgroup. Curves of further compounds (35,
47, 71) are displayed in Supporting Information. In addition to measuring concentrationꢀ
dependent inhibition of LPIꢀinduced βꢀarrestin translocation by antagonists at a single
concentration of agonist (1 ꢁM of LPI), we also determined full concentrationꢀresponse curves
for LPI in the presence and absence of the most potent antagonists. As far as the potency and
solubility of the coumarin derivatives allowed it, we tested 3ꢀ4 different antagonist
concentrations. To quantify the antagonistic potency, pA2 values were calculated by using
equiactive agonist concentrations in the absence and presence of antagonist. In case of potentially
allosteric modulation the pA2 value was determined using equiactive agonist concentrations at a
level of 30% of the maximal response of the depressed concentrationꢀresponse curve.36
Coumarins with no or only a small substituent in position 7 like 12 and 35 (series I) led to a
parallel rightward shift of the agonist curve without affecting its efficacy (Figure 3A and Figure
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
S2A) thereby indicating competitive antagonism. The corresponding agonistꢀinhibition curve
using 1 ꢁM LPI and a range of antagonist concentrations was also consistent with binding to the
orthosteric ligand binding site since the LPI signal was completely inhibited (Fig. 3B and Figure
S2B); the curves even went below zero indicating inverse agonism. It appears that GPR55 is
already tonically activated to some extent in the employed recombinant CHO cell line, and
coumarins 12 and 35 were able to not only block the signal induced by the added GPR55 agonist
LPI, but to additionally block the basal activity of the receptor measured in the test system. For
the most potent compounds of this subgroup, 35 and 12, IC50 values of 0.981 ꢁM and 1.77 ꢁM
were determined. In addition, a KBꢀvalue of 1.87 ꢁM was determined for 35 by applying the
Schild regression (see Supporting Information, Figure S3).36 For 12 the calculated pA2 value was
0.547 ꢁM. Thus, the potency values obtained for the compounds by different methods were
consistent. While 35 showed also some affinity for CB receptors (Ki CB1 3.66 ꢁM, CB2 0.338
ꢁM), probably due to its substituent (methyl) in the 7ꢀposition, 12 was inactive at CB receptors
and thus selective for GPR55 (see Table 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
The curves determined for 49 belonging to the 7ꢀpentylꢀsubstituted coumarin series II might
indicate an allosteric binding mode of 49, as the antagonist (10 ꢁM) led to a rightward shift of the
agonist curve (pA2 11.1 ± 1.9 ꢁM) and at the same time appeared to reduce the maximal effect of
LPI (Figure 3C). A KB value of 16.5 ꢁM was estimated by linear transformation of the data
according to the method of Gaddum (see Supporting Information, Figure S4).36 The assumption
of an allosteric binding mode is further supported by the agonistꢀinhibition curve displayed in
Figure 3D. It can be seen that 49 is not able to fully inhibit the effect of LPI, as the inhibition
curve does not reach basal levels. This phenomenon can be observed for partial agonists and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
5
6
7
8
9
allosteric inhibitors. Since the compounds showed no agonistic effects in screening assays, a
partial agonism can be excluded. Thus allosteric modulation might be an explanation for the
observed effects. Compound 49 turned out to be the most potent compound of this subgroup II at
human GPR55, exhibiting an IC50 value of 3.23 ꢁM at the human GPR55. However, it was even
more potent at CB receptor, where it showed Ki values in the low nanomolar range (see Table 3
and Figure 3D).
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
All coumarins with branched aliphatic residues in position 7 (series III) that were identified as
GPR55 antagonists completely inhibited the LPIꢀmediated effect at a high concentration of 10
ꢁM (Table 1). Full concentrationꢀresponse curves for LPI in the absence and in the presence of
various coumarin concentrations showed that increasing concentrations of 69 led to a rightward
shift and at the same time a depression of the maximal effect of the agonistꢀresponse curve
(Figure 3E). A KB value of 0.561 ꢁM was estimated by application of the method of Gaddum,
which was in good accordance to the determined pA2 value of 0.483 ꢁM (see Supporting
Information, Figure S5).36 Due to a long incubation time of 3.5 h for the antagonist (1 h of
preincubation) and 2.5 h for the agonist in the applied assay system, the observed results are
probably not due to different dissociation kinetics of both ligands. Rather they may indicate an
allosteric mode of GPR55 inhibition. An allosteric mode of action had previously been proposed
for other CB ligands, such as 1 and 3.7 However to clearly confirm an allosteric binding mode of
the investigated compounds, dissociation kinetic studies would have to be performed, which are
currently not feasible due to the lack of an appropriate radioꢀ or fluorescentꢀlabelled ligand.7
Compound 69 was the most potent compound of this subgroup at human GPR55, exhibiting an
IC50 value of 0.113 ꢁM (Figure 3F).
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
Coumarin series I
7
8
9
(A)
(B)
100
50000
LPI
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
LPI + 12 10 ꢀM
LPI + 12 3 ꢀM
LPI + 12 1ꢀM
40000
30000
20000
10000
0
75
50
25
0
10-7
10-6
10-5
10-4
-7
-6
-5
-4
-10000
-25
[12], M
[LPI], M
Coumarin series II
(D)
100
(C)
80000
70000
60000
50000
40000
30000
20000
10000
0
LPI
LPI + 49 10 ꢀM
75
50
25
0
10-8
10-7
10-6
[49], M
10-5
10-4
10-7
10-6
10-5
10-4
[LPI], M
Coumarin series III
(F)
(E)
100
70000
60000
50000
40000
30000
20000
10000
0
LPI
LPI + 69 10 ꢀM
LPI + 69 3 ꢀM
LPI + 69 1ꢀM
LPI + 69 0,3 ꢀM
75
50
25
0
10-7
10-6
[LPI], M
10-5
10-4
10-9
10-8
10-7
[69], M
10-6
10-5
-10000
Figure 3. Functional properties of selected coumarin derivatives in βꢀarrestin recruitment
assays performed with CHO cells stably expressing the human GPR55. Compounds with no
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
1
2
3
4
5
6
7
8
9
lipophilic moiety in position 7 (A, 12 subgoup I) had no impact on Emax of LPI and led to a
rightwardshift of the agonist curve (see Table 2). A pA2ꢀvalue of 0.547 ± 0.119 ꢁM could be
determined. Coumarins with a pentyl residue in position 7 (C, 49 subgroup II) exhibited
antagonistic properties (pA2: 11.1 ± 1.9 ꢁM) and a moderate reduction of maximal LPIꢀresonse,
while compounds with branched aliphatic residues (E, 69 subgroup III) displayed considerable
concentrationꢀdependent inhibition of LPI’s maximal reponse (see Table 2). A pA2 value of 0.483
± 0.198 ꢁM could be determined for 69. Concentrationꢀdependent inhibtion of LPI (1 ꢁM)ꢀ
response by test compounds revealed IC50 values of (B) 1.77 ± 0.23 ꢁM for 12, (D) 3.23 ± 0.31
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
ꢁM for 49 and (F) 0.113 ± 0.020 ꢁM for 69. Data points represent means ± SEMs of three
independent experiments, performed in duplicates. For additional curves (compounds 35 (group
I), 47 (group II) and 71 (groupIII)) see Supporting Information (Figure S2, S6, and S7).
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
Table 2. Potencies and maximal effect of LPI in the presence and absence of selected coumarin
7
8
9
derivatives at GPR55a
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
Compd
LPI
%Emax
EC50 ± SEM
of LPI (ꢁM)
Coumarin series I
100
89 (ns)b
87 (ns)b
ndc
2.47 ± 0.89
10.6 ± 5.3
LPI + 12
(1 ꢁM)
LPI + 12
(3 ꢁM)
LPI + 12
(10 ꢁM)
17.3 ± 6.8
ndc
Coumarin series II
2.57 ± 0.89
4.17 ± 1.47
Coumarin series III
2.25 ± 1.12
3.29 ± 1.28
3.99 ± 1.61
0.878 ± 0.386
ndc
LPI
100
LPI + 49
(10 ꢁM)
72 (*)d
LPI
100
LPI + 69
(0.3 ꢁM)
LPI + 69
(1 ꢁM)
LPI + 69
(3 ꢁM)
93 (ns)b
52 (*)d
23 (*)d
19 (*)d
LPI + 69
(10 ꢁM)
aall data result from three independent experiments, performed in duplicates
bns = not significantly different from Emax=100%
cnd = not determinable
d* = significantly different from Emax=100% (Pꢀvalues < 0.05, paired tꢀtest)
Selectivity for GPR55
In order to further investigate the selectivity of the compounds they were additionally
investigated for an interaction with the orphan receptor GPR18, a putative member of the
cannabinoid receptor family.18, 19 3 was used as an agonist. In our hands 3 showed an EC50 value
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
1
2
3
4
5
of 4.61 ꢁM, which is consonant with literature values (0.96 ꢁM).19 1 and 2 behaved as
moderately potent GPR18 antagonists (IC50 1, 10.1 ꢁM; 2, 5.99 ꢁM).
6
7
8
9
14, 36, 39 (subgroup I), as well as 61 (subgroup II), and 65, 67 and 69 (subgroup III), showed
weak antagonistic properties at GPR18 (IC50 values ~ 10 ꢁM), while all other compounds failed
to display either agonistic or antagonistic activity at GPR18 at a test concentration of 10 ꢁM (see
Supporting Information and Table 3). These results indicate that coumarins with short, as well as
bulky residues in position 7 may exhibit weak antagonistic activity at the human GPR18. In
addition, the results showed that the compounds identified as GPR55 antagonists in the present
study are selective versus the putative third CB receptors subtype GPR18. We used exactly the
same assay for both receptors, determining agonistꢀinduced βꢀarrestin translocation by fragment
complementation methodology measuring luminescence. This type of assays is very specific, and
artefacts are rare.37 In addition the lack of potency of the coumarin derivatives at GPR18 further
proves that the coumarins identified to be GPR55 antagonists exert their effects by direct
interaction with the GPR55 receptor protein.
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
As discussed above, coumarins of series I possessing no substituents in the R7ꢀposition showed
generally no or only low CB receptor affinity, while they exhibited remarkable inhibitory effects
at GPR55. In particular, coumarins with an additional isopropyl residue at the 5ꢀposition
displayed high GPR55 selectivity versus CB receptors (10, 11, 12). Introduction of a methyl
residue in position 7 increased affinity for CB receptors, particularly for the CB2 receptor
subtype. 5ꢀMethoxyꢀcoumarins of series II, and 5ꢀhydroxyꢀcoumarins of series III displayed
antagonistic effects at GPR55, but showed also high affinity for CB receptors, acting as CB
receptor agonists, partial agonists, or antagonists, respectively (see Supporting Information).
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Activation of GPR55 and CB receptors is reported to result in contrary physiological effects.
For instance, CB1 and CB2 receptor agonists have been shown to suppress neuropathic
nociception in several models, whereas GPR55 receptor antagonists are considered as potential
therapeutics for inflammatory and neuropathic pain.4, 10, 38, 39 Moreover, CB1 receptor agonists
have been reported to inhibit cancer cell proliferation and migration, while similar effects were
reported for GPR55 antagonists.22, 24, 40 It is tempting to speculate that ligands targeting both
receptors, by acting as CB receptor agonists and GPR55 receptor antagonists at the same time,
may exhibit additive or even synergistic effects, compared to drugs targeting only one of the
respective receptors. Dual receptor ligands featuring both, GPR55ꢀantagonistic and CBꢀagonistic
activity, like 47, 49 and 65 (GPR55 antagonists, CB1 and CB2 agonists) and 67, 69, 71 and 73
(GPR55 antagonists, CB2 agonists) may be interesting pharmacological tools to further explore a
therapeutic dual target strategy.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
1
2
3
4
Table 3. Potencies and affinities of selected coumarin derivatives at GPR55, GPR18, CB1 and
5
6
CB2 receptorsa
7
8
Compd
human GPR55
human GPR18
rat/human CB1
human CB2
9
βꢀarrestin recruitment assay
IC50 ± SEM (ꢁM)
Radioligand binding assays
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
vs. [3H]CP55,940
Ki ± SEM (ꢁM)
Standard CB receptor agonists and antagonists
2.0133
Agonist (EC50)
1.6133
10.1 ± 1.3
0.012633
0.0012833
0.90033
0.0014233
1
2
3
5.99 ± 1.88
14.2 ± 5.4
4.61 ± 0.50
0.00388 ± 0.00091
0.0716 ± 0.0024
Agonist (EC50)
Coumarin derivatives I: with small 7ꢀsubstituents
3.45 ± 0.36
5.33 ± 1.10
1.77 ± 0.23
7.14 ± 2.66
5.70 ± 1.62
3.99 ± 0.75
6.74 ± 2.04
0.981 ± 0.140
12.8 ± 3.2
> 10 (11%)c
> 10 (12%)c
> 10 (32%)c
> 10 (0%)c
11.3 ± 2.0
> 10 (26%)d
> 10 (24%)d
> 10 (31%)d
> 10 (16%)d
> 10 (29%)d
> 10 (38%)d
≥ 10 (45%)d
3.66 ± 0.51
≥ 10 (42%)d
> 10 (21%)d
~ 10 (49%)d
> 10 (34%)d
7.86 ± 2.52
≥ 10 (45%)d
>> 10 (0%)d
> 10 (24%)d
> 10 (24%)d
6.83 ± 1.16
0.795 ± 0.254
2.4527
10
11
12
13
14
31
33
35
36
37
38
39
40
41
> 10 (31%)c
> 10 (24%)c
> 10 (27%)c
≥ 10 (46%)c
> 10 (16%)c
> 10 (32%)c
≥ 10 (47%)c
> 10 (33%)c
> 10 (38%)c
≥ 10 (45%)d
0.338 ± 0.146
1.77 ± 1.00
3.42 ± 0.90
1.48 ± 0.33
0.444 ± 0.007
0.660 ± 0.171
1.56 ± 0.78
9.32 ± 1.05
13.5 ± 4.3
10.3 ± 0.7
7.69 ± 1.71
5.16 ± 0.73
Coumarin derivatives II: 7ꢀpentylꢀsubstitution
6.35 ± 2.66
3.23 ± 0.31
> 10 (20%)c
> 10 (33%)c
0.032227
0.0453 27
0.049227
0.14327
47
49
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10.6 ± 4.9
5.08 ± 1.05
> 10 (19%)b
9.00 ± 2.44
> 10 (0%)b
3.29 ± 1.30
3.76 ± 1.46
> 10 (27%)c
> 10 (34%)c
> 10 (10%)c
> 10 (25%)c
> 10 (0%)c
~ 10 (50%)c
> 10 (30%)c
16.227
5.1527
0.40527
50
53
54
57
58
61
63
0.022227
1.6427
3.5727
0.032727
19.127
0.18527
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
1.9427
0.613 ± 0.131
10.2 ± 3.4
0.578 ± 0.167
0.578 ± 0.136
Coumarin derivatives III: long, branched 7ꢀsubstituent
~ 10 (51%)b
0.358 ± 0.089
> 10 (25%)b
0.261 ± 0.181
> 10 (31%)b
0.113 ± 0.020
> 10 (30%)b
0.854 ± 0.454
> 10 (44%)b
0.961 ± 0.431
> 10 (31%)c
8.10 ± 0.58
> 10 (26%)c
15.9 ± 4.9
1.4327
2.6327
4.1227
0.46527
64
65
66
67
68
69
70
71
72
73
1.0227
3.0127
0.24427
~ 10 (51 %)d
1.1727
0.21027
> 10 (33%)c
12.5 ± 2.9
~ 10 (47 %)d
0.29227
> 10 (25%)c
≤ 10 (57%)c
> 10 (33%)c
≤ 10 (59%)c
0.59827
1.5827
1.1427
0.067627
≥ 10 (45 %)d
0.049127
≥ 10 (42 %)d
4.8927
aall data result from three independent experiments, performed in duplicates.
b% inhibition of LPI (1 ꢁM)ꢀinduced βꢀarrestin recruitment by test compounds in a concentration of 10
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢁM
c% inhibition of 3 (10 ꢁM)ꢀinduced βꢀarrestin recruitment by test compounds in a concentration of 10 ꢁM
d% inhibition of radioligand binding at 10 ꢁM
Conclusions
We successfully applied a βꢀarrestin recruitment assay to identify and characterize novel
GPR55 antagonists with a coumarin scaffold and to analyze their SARs. Our results showed that
the moiety in position 7 of the coumarin scaffold is crucial for interaction with GPR55, and CB
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
1
2
3
4
5
6
7
8
9
receptors as well. Compounds with no lipophilic residue in that position possessed in general no
or only moderate affinity for CB receptors, and no affinity for the putative third CB receptor
subtype GPR18, while they acted as selective, competitive GPR55 antagonists, in particular when
combined with a methyl moiety in position 8 of the coumarin scaffold. Thus 8ꢀmethylcoumarins
represent an appropriate lead structure for the development of more potent and selective GPR55
antagonists in the future. The most potent compound of this subset of compounds, 6ꢀmethoxyꢀ3ꢀ
(2ꢀmethoxybenzyl)ꢀ5,7,8ꢀtrimethylꢀ2Hꢀchromenꢀ2ꢀone (35, PSBꢀSBꢀ258), exhibited an IC50
value of 0.981 ꢁM at GPR55, which was in accordance with the determined KB value of 1.87
ꢁM. 3ꢀ(2ꢀHydroxybenzyl)ꢀ5ꢀisopropylꢀ8ꢀmethylꢀ2Hꢀchromenꢀ2ꢀone (12, PSBꢀSBꢀ489, IC50 1.77
ꢁM, pA2 0.547 ꢁM) and 5ꢀisopropylꢀ3,8ꢀdimethylꢀ2Hꢀchromenꢀ2ꢀone (10, PSBꢀSBꢀ115, IC50
3.45 ꢁM) are further potent GPR55 antagonist of this subset, possessing selectivity for GPR55.
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
Introduction of a pentyl residue in position 7 of the coumarin scaffold resulted in a dramatically
increased affinity at CB receptors, and a possible shift from orthosteric to allosteric antagonism at
GPR55. 5ꢀMethoxyꢀ3ꢀ(3ꢀmethoxybenzyl)ꢀ7ꢀpentylꢀ2Hꢀchromenꢀ2ꢀone (49, PSBꢀSBꢀ435) was
the most potent antagonist at GPR55 of this subgroup II (IC50 3.23 ꢁM, pA2 11.1 ꢁM),
possessing at the same time high agonistic potency at both CB receptor subtypes (EC50 CB1:
0.430 ꢁM; CB2: 0.092 ꢁM, see Supporting Information).
Increase in length and bulk of the alkyl moiety (subgroup III) further increased the antagonistic
potency of the compounds at GPR55. 7ꢀ(1,1ꢀDimethyl)alkylꢀ5ꢀhydroxycoumarins were shown to
be very potent antagonists at GPR55 and at the same time agonists at the CB2 receptor (see
Supporting Information). 7ꢀ(1,1ꢀDimethyloctyl)ꢀ5ꢀhydroxyꢀ3ꢀ(2ꢀhydroxybenzyl)ꢀ2Hꢀchromenꢀ2ꢀ
one (69, PSBꢀSBꢀ487, IC50 GPR55 0.113 ꢁM, KB 0.561 ꢁM) and 7ꢀ(1,1ꢀdimethylheptyl)ꢀ5ꢀ
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydroxyꢀ3ꢀ(2ꢀhydroxybenzyl)ꢀ2Hꢀchromenꢀ2ꢀone (67, PSBꢀSBꢀ1203, IC50 GPR55 0.261 ꢁM)
were found to be the most potent GPR55 antagonists of this subgroup. They may represent
interesting tools to study a therapeutic dual target strategy, activating CB2 receptors and blocking
GPR55 at the same time.
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
Our findings provide a basis for the development of potent GPR55 antagonists with selectivity
versus GPR18 and CB receptors, as well as dual GPR55/CB receptor ligands. The coumarins are
readily accessible and can easily be modified by a straightforward synthetic procedure. This
extensive SAR study of GPR55 ligands along with their detailed characterization may contribute
to further elucidating the physiological role of this orphan GPCR, and to explore its potential as a
future drug target.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
1
2
3
4
Experimental Section
5
6
7
8
All commercially available reagents were obtained from various producers and used without
further purification. Purity of all tested compounds was ≥95% unless otherwise noted as
confirmed by elemental analysis and HPLCꢀMS measurements. NMR spectra were recorded on a
Bruker AM 400 (100 MHz), a Bruker Avance 300 (300 MHz) or a Bruker Avance 400
(400 MHz). Deuterated CDCl3 was used as a solvent unless otherwise noted. The chemical shifts
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
1
13
of the remaining protons of the solvent were used as internal standard: H, 7.26 ppm; C, 77.0
ppm. All chemical shifts (δ) were expressed in ppm. Coupling constants (J) are given in Hertz
(Hz). The reactions were monitored by thin layer chromatography (TLC) using aluminum sheets
with silica gel 60 F254 (Merck). Elemental analyses were performed in the Institute of Organic
Chemistry, University of Karlsruhe with the device Elementar vario MICRO. Mass spectra were
collected on an Finnigan MAT 95 and IR spectra on a Bruker IFS 88.
Syntheses
General procedures for the preparation of coumarin derivatives
Under an atmosphere of argon, 1.00 equiv. of substituted salicylaldehyde, 1.20 equiv. of
potassium carbonate, 2.50 equiv. of cinnamaldehyde and 1.20 equiv. of 1,3ꢀdimethylimidazolium
dimethylphosphate were suspended in a pressure tube in toluene (7 mL/mmol salicylaldehyde).
The reaction mixture was stirred at 200 watt, heated to 110 °C and 7 bar in the microwave for
50 min. It was cooled to room temperature and quenched by addition of water. The product was
extracted with ethyl acetate and the combined organic phases were dried over Na2SO4, the
solvent removed under low pressure and the crude product purified by flash column
chromatography.
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
1
2
3
4
General procedures for the deprotection of coumarin derivatives
5
6
Under an atmosphere of argon, 5.00 equiv. of bortribromide (1
M in dichloromethane) was
7
8
9
added drop wise to a solution of 1.00 equiv. of coumarin in dichloromethane (10 mL/mmol
coumarin) at – 78 °C. The reaction mixture was stirred for 30 min. at this temperature and was
then stirred for another 24 h at room temperature. The reaction was quenched at 0 °C with
saturated NaHCO3ꢀsolution, extracted with dichloromethane and washed with aqua dest. and
brine. The combined organic phases were dried over Na2SO4, the solvent removed under low
pressure and the product purified by flash column chromatography.
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
Crystal Structure Determination of 14
The singleꢀcrystal Xꢀray diffraction study was carried out on a BrukerꢀNonius KappaꢀCCD
diffractometer at 123(2) K using MoKα radiation (λ = 0.71073 Å). Direct Methods
(SHELXSꢀ97)41 were used for structure solution and refinement was carried out using
SHELXLꢀ9741 (fullꢀmatrix leastꢀsquares on F2). Non hydrogen atoms were refined
anisotropically, hydrogen atoms were localized by difference electron density determination
and refined using a riding model. A semiꢀempirical absorption correction was apllied.
14: yellow crystals, C22H24O3, M = 336.41, crystal size 0.50 x 0.35 x 0.25 mm, monoclinic,
space group P21/c (No. 14), a = 7.5689(6) Å, b = 17.1760(18) Å, c = 14.2218(4) Å, ß =
100.855(4)°, V = 1815.(2) Å3, Z = 4, ρ(calc) = 1.213 Mg mꢀ3, F(000) = 720, ꢁ = 0.081 mmꢀ1,
17360 reflections (2θmax = 55°), 4136 unique [Rint = 0.031], 228 parameters, R1 (for 3432 I >
2σ(I)) = 0.039, wR2 (all data) = 0.102, S = 1.02, largest diff. peak and hole 0.278 and ꢀ0.233
e Åꢀ3.
Crystallographic data (excluding structure factors) for the structure reported in this work
ACS Paragon Plus Environment