8-benzamidochromen-4-one-2-carboxylic acids: Potent and selective agonists for the orphan G protein-coupled receptor GPR35

Article abstract of DOI:10.1021/jm400587g

8-Amido-chromen-4-one-2-carboxylic acid derivatives were identified as novel agonists at the G protein-coupled orphan receptor GPR35. They were characterized by a β-arrestin recruitment assay and optimized to obtain agonists with nanomolar potency for the

Full text of DOI:10.1021/jm400587g

Subscriber access provided by Penn State | University Libraries
Article
8-Benzamidochromen-4-one-2-carboxylic Acids - Potent and Selective
Agonists for the Orphan G Protein-Coupled Receptor GPR35
Mario Funke, Dominik Thimm, Anke C Schiedel, and Christa E Müller
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm400587g • Publication Date (Web): 28 May 2013
Downloaded from http://pubs.acs.org on June 13, 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 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
8
-Benzamidochromen-4-one-2-carboxylic Acids -
Potent and Selective Agonists for the Orphan G protein-
coupled Receptor GPR35
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
#
Mario Funke, Dominik Thimm, Anke C. Schiedel, and Christa E. Müller*
#
PharmaCenter Bonn, Pharmaceutical Institute, Pharmaceutical Chemistry I, University of Bonn, An der
Immenburg 4, D-53121 Bonn, Germany.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 56
KEYWORDS
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
agonist, β-arrestin, chromenone, GPR35, GPR35b, GPR55, G protein-coupled receptor, orphan G
protein-coupled receptors, species selectivity, structure-activity relationships
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
ABSTRACT
8
-Amido-chromen-4-one-2-carboxylic acid derivatives were identified as novel agonists at the
G protein-coupled orphan receptor GPR35. They were characterized by a β-arrestin recruitment assay,
and optimized to obtain agonists with nanomolar potency for the human GPR35. The compounds were
found to exhibit high selectivity versus the related GPR55. The most potent agonists were 6-bromo-8-
(4-methoxybenzamido)-4-oxo-4H-chromene-2-carboxylic acid (85, EC50 12.1 nM) and 6-bromo-8-(2-
chloro-4-methoxybenzamido)-4-oxo-4H-chromene-2-carboxylic acid (90, EC 11.1 nM), both of which
5
0
were >1700-fold selective versus GPR55. Most compounds were considerably less potent at rat and
mouse than at human GPR35. 6-Bromo-8-(2-methoxybenzamido)-4-oxo-4H-chromene-2-carboxylic
acid (87) was the only derivative that activated GPR35 of all three species at similar, low micromolar
concentration. Compounds 85 and 90 are the most potent agonists at the human GPR35 known to date
and might thus serve as powerful pharmacological tools to further elucidate the receptor’s
(patho)physiological role and its potential as a future drug target.
ACS Paragon Plus Environment

Page 3 of 56
Journal of Medicinal Chemistry
INTRODUCTION
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
G protein-coupled receptors (GPCRs) are mediators and modulators of a variety of
(patho)physiological processes. Endogenous agonists have been described for about 300 members of
this receptor class and many of them are considered as important or at least as promising, potential
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
1
pharmacotherapeutic targets. However, for nearly 100 (non-olfactory) GPCRs the endogenous agonist
1
remains unknown. Those are designated orphan receptors. One of the orphan GPCRs is the human
2
GPR35, which was first described in 1998. It was reported to be coupled to G and G proteins.
3-5
i/o
13
Later on a transcript variant was found that encoded for an identical protein but with 31 additional
6
amino acids at its N-terminus. That variant was designated GPR35b. For both, rat and mouse, only one
6
7, 8
isoform of GPR35 has been described so far.
There is only a moderate homology between the
orthologs with a sequence identity between human GPR35 and rodent orthologs of 72% and 71% for
9
, 10
mouse and rat GPR35, respectively, as it is observed for many other GPCRs.
One of the
1
1
phylogenetically closest relatives of the human GPR35 is GPR55.
GPR35 mRNA expression was found in the stomach and intestine of humans and rodents, with an
2
, 5, 6
increased expression in human gastric cancer cells.
Heterologous GPR35 expression was shown to
exhibit transforming activity in a mouse fibroblast cell line indicating that the receptor might play a role
6
in cell growth control. The human receptor was also found to be expressed in various immune cells,
5, 12,
1
3
while a further increase in receptor expression was detected in mast cells upon stimulation with IgE
1
antibodies. Following receptor activation mast cells showed reduced lipopolysaccharide-induced
3
5
TNFα secretion. For invariant natural killer T cells receptor activation caused a reduction of IL-4
12
secretion. Thus, GPR35 modulates immune responses. There is also evidence for the involvement of
GPR35 in pain perception: receptor mRNA expression was detected in murine neurons of the dorsal
1
4, 15
root ganglia, which are known to transmit nociceptive signals.
Antinociceptive effects were
14-16
observed after treating mice or rats with GPR35 agonists.
Furthermore, GPR35 was suggested to
1
7, 18
play a role in the regulation of blood pressure and in the pathogenesis of cardiovascular diseases.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 56
Although we are only at the very beginning of understanding the physiological role of GPR35, the
receptor appears to be a promising target for therapeutic intervention.
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
3
,
Figure 1. Selected GPR35 agonists with potencies at GPR35 orthologs determined in β-arrestin assays.
1
9-23
24-28
Potencies of compounds at other targets are displayed in gray.
AMPAR: α-Amino-3-hydroxy-5-
methylisoxazole-4-propionic acid receptor, NMDA: N-methyl-D-aspartic acid, GABA: γ-aminobutyric
acid, nAChR: nicotinic acetylcholine receptor, AHR: aryl hydrocarbon acid receptor, PDE:
phosphodiesterase, DNA pol.: DNA polymerase.
ACS Paragon Plus Environment

Page 5 of 56
Journal of Medicinal Chemistry
The first GPR35 agonist described in the literature was the tryptophane metabolite kynurenic acid (1,
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
5
see Fig. 1), postulated to be the receptor’s endogenous agonist. However, its potency - at least at the
5
, 23
human ortholog - is relatively low with EC values typically in the high micromolar range.
It is not
50
clear, whether sufficiently high concentrations of 1 are reached under physiological conditions to
2
9-31
activate the receptor.
Among the agonists described so far are several other naturally occurring
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
4
compounds that might constitute endogenous agonists, including lysophosphatidic acid species, and
32
tyrosine metabolites. However, up to now GPR35 is still considered an orphan receptor, since the
physiological agonist has not been unambiguously identified yet.
Figure 2. Selected GPR35 antagonists.
During the past few years several synthetic GPR35 agonists have been identified, including the
8
, 13
antiallergic drug cromolyn (2) and zaprinast (3).
The latter has become the standard agonist for
GPR35. While cromolyn and zaprinast are full agonists, pamoic acid (4) was shown to be a partial
3
, 23
agonist.
The recently described and up to now most potent agonist at the human GPR35 is the
2
thiazolidinylidenemethylbenzoic acid derivative 5 (see Fig. 1). So far, only few potent GPR35
2
antagonists have been described, the most potent ones being CID 2745687 (6) and ML 145 (7) (see Fig.
2
3, 33
2).
There are several major drawbacks in investigating the receptor’s physiological role using
pharmacological tools. On the one hand, although many agonists have been described, only very few
show high potency with EC50 values in the lower nanomolar range. On the other hand, most known
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 56
agonists also bind to other targets, mostly with similar potency: e. g. kynurenic acid inhibits
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
AMPA/kainate receptors, NMDA receptors, GABA receptors, α7 nicotinic acetylcholine (α7 nAChR)
A
2
receptors and potently activates intracellular aryl hydrocarbon acid receptors (AHR), while zaprinast
4
34
inhibits cGMP-specific phosphodiesterases (PDEs), especially PDE5 and 6, and pamoic acid inhibits
2
6
DNA polymerase β. Another drawback is the fact that the known GPR35 antagonists are virtually
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
inactive at the rodent orthologs. Thus, there is a need of highly potent and selective GPR35 ligands.
1
In most published studies focusing on the identification of GPR35 receptor ligands large compound
libraries were screened. This led to the discovery of structurally diverse ligands. However, the number
of studies on structure-activity relationships (SARs) of GPR35 ligands is very limited. In the present
study, we intended to identify a suitable scaffold to develop highly potent and selective GPR35 agonists.
Therefore, we started by screening an in-house compound library (http://mueller-group.pharma.uni-
bonn.de/mueller-laboratory/compound-library). For this purpose we used a β-arrestin recruitment assay,
the most commonly used assay system to pharmacologically characterize GPR35 ligands. We identified
a chromen-4-one-2-carboxylic acid, which shared a sub-structure with cromolyn, as a GPR35 agonist
with moderate potency. Subsequently, we synthesized a series of mostly novel 8-substituted chromen-4-
one-2-carboxylic acid derivatives to study their structure-activity relationships as novel GPR35 agonists.
ACS Paragon Plus Environment

Page 7 of 56
Journal of Medicinal Chemistry
RESULTS AND DISCUSSION
Chemistry
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 large variety of chromen-4-one-2-carboxylic acids with different halogen atoms at the 6-position
and various alkyl- or arylamido residues at the 8-position were synthesized (Scheme 1). The compounds
were prepared starting from 8, or the corresponding 5-halogeno derivatives 9–11, respectively. Based on
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
5
a recently described synthesis of pranlukast derivatives, the acetophenones 8–11 were treated with
diethyloxalate and potassium tert-butoxide as a base in N,N-dimethylformamide (DMF) at 0–5 °C. The
resulting diketone intermediates were isolated after acidification, and subsequently reacted with
hydrochloric acid in ethanol under reflux yielding the chromen-4-one derivatives 12–15 in very good
3
6, 37
38, 39
yields (70–82%). Compounds 12
and 13
had previously been described in the literature using
4
0
different reaction conditions. The final product 16 was obtained in an excellent yield of 96% by
hydrolysis of its ethyl ester 12 using optimized mild basic conditions (potassium carbonate in a mixture
of tetrahydrofuran (THF), ethanol and water at room temperature), while strong basic conditions led to a
4
cleavage of the benzopyran-4-one ring. The nitro groups of intermediates 12–15 were reduced using
1
41
tin(II) chloride in diluted hydrochloric acid/ethanol yielding the corresponding amino derivatives 17,
3
8, 19, and 20 in excellent yields (80–90%). The final chromen-4-one-2-carboxylic acids 21 and 22
8
42
1
were formed by mild basic hydrolysis of 17 and 18, respectively, under the same conditions as described
above for 16.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 56
a
Scheme 1. Synthesis of chromen-4-one-2-carboxylic acids
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
a
t
Reagents and conditions: (a) (COOEt) , KO Bu, DMF, argon, 0–5 °C, 2–3 h; (b) concd. aq. HCl, ice-
2
water; (c) concd. aq. HCl, EtOH, reflux, overnight, yield 70–82%; (d) K CO , THF, EtOH, H O, rt,
2
3
2
overnight; aq. HCl (1 N), H O, yield 96%; (e) SnCl ꢀ2H O, aq. HCl (2 N), EtOH, 65–70 °C, 20–45 min,
2
2
2
yield 80–90%; (f) K CO , THF, EtOH, H O, rt, overnight; aq. HCl (1 N), H O, yield 96–99%; (g) acetic
2
3
2
2
2
anhydride, DIPEA, DCM, argon, rt, 2 d for 23; R COCl, DIPEA, DCM (THF), argon, rt, 1–4 d for 24–
b
6, yield 21–94%; (h) K CO , THF, EtOH, H O, rt, 2–24 h; aq. HCl (2 N), H O, yield 45–99%. For R
2
5
2
3
2
2
see Tables 1 and 2.
ACS Paragon Plus Environment

Page 9 of 56
Journal of Medicinal Chemistry
The amines 17–20 were reacted with various acid chlorides in the presence of N,N-
diisopropylethylamine (DIPEA) as a base in dichloromethane (DCM), or in a 6:1 mixture of DCM and
THF at room temperature (rt) yielding a series of ethyl 8-amido-5-halogeno-chromen-4-one-2-
carboxylates (Scheme 1). A large range of differently substituted benzamido residues were introduced at
the 8-position of the chromen-4-one ring system. The structures and yields of the synthesized
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
4
3
compounds are shown in Tables 1 and 2. Compound 23 was obtained by treatment of 17 with acetic
anhydride in the presence of DIPEA in DCM at rt after a reaction time of two days. In the last step, the
ethyl esters 23–56 were hydrolyzed yielding the desired chromen-4-one-2-carboxylic acids 57–90
(Scheme 1, see Tables 1 and 2 for structures and yields). Mild basic reaction conditions as described
above for the hydrolysis of compound 16 were applied.
Table 1. Yields of 8-amidochromen-4-one-2-carboxylic acids and their ethyl esters
1
R
2
R
a
a
compd
yield (%)
compd
yield (%)
3
R = CO Et
3
R = CO H
2
2
H
H
methyl
ethyl
23
24
82
80
57
58
77
82
H
25
85
59
90
H
H
H
H
cyclohexyl
benzyl
26
27
28
29
93
85
85
80
60
61
62
63
82
97
94
49
2-naphthyl
2-quinolinyl
a
Isolated yields.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 56
Table 2. Yields of substituted 8-benzamidochromen-4-one-2-carboxylic acids and their ethyl esters
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
a
a
substitution pattern
compd
yield (%)
compd
yield (%)
1
R
2
R
3
R = CO Et
3
R = CO H
ortho
H
meta
H
para
H
2
2
H
H
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
92
94
83
75
71
89
76
21
78
66
88
78
91
55
88
81
71
80
73
71
57
89
65
85
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
89
95
79
49
62
85
45
89
97
92
98
95
82
99
98
73
99
98
91
98
71
95
99
86
H
CH3
NO2
H
H
H
H
H
H
H
CH3
CF3
OCH3
Br
H
H
H
H
H
H
H
H
H
F
Cl
Cl
H
Cl
H
H
F
Cl
F
Cl
H
Cl
F
H
OCH3
H
Cl
Cl
Cl
Cl
Br
Br
Br
Br
Br
Br
Br
Br
H
H
H
CH3
H
H
H
NO2
OCH3
H
H
H
H
H
H
CH3
H
H
H
Cl
Cl
H
H
Cl
Cl
H
Cl
H
CN
OCH3
ethoxy
H
H
H
H
H
Br OCH3
H
ACS Paragon Plus Environment

Page 11 of 56
Journal of Medicinal Chemistry
Br
Br
Br
H
H
OCH3
OCH3
54
55
56
76
79
81
88
89
90
97
96
95
1
2
3
4
5
6
7
8
9
Cl
H
OCH3
a
Isolated yields.
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 final products 99 and 100 were obtained as follows. Treatment of 91 and 92, respectively, with
thionyl chloride in the presence of DMF as a catalyst in DCM at rt for 30 min afforded the
corresponding acid chlorides (Scheme 2). Coupling with amine 17, using the same conditions as
described for 23–56, yielded the amides 95 and 96 in good yields. In the last step, the ethyl ester groups
of 95 and 96 were hydrolyzed as described above to obtain the desired chromen-4-one-2-carboxylic
acids 99 and 100 in excellent yields of 99%. Acid 91 was synthesized starting from methyl 4-
hydroxybenzoate, which was converted with cyclopropylmethyl bromide by Williamson ether synthesis
4
potassium carbonate, DMF, 110–115 °C, overnight). Methyl 4-(cyclopropylmethoxy)benzoate
4
(
45
103) was obtained in a good yield (78%). The efficient hydrolysis of the ester, using a KOH solution
(
46
in methanol under microwave irradiation at 100 °C for 45 min provided the benzoic acid derivative 91
in an excellent yield of 98%. The acids 93 and 94 were converted to their acid chlorides as depicted in
Scheme 2, and subsequently reacted with amine 18 in the presence of DIPEA as a base in a 6:1 mixture
of DCM and THF at rt for two days yielding the amides 97 or 98. Good yields of 79–80% were
obtained. Ester hydrolysis afforded the target compounds 101 and 102 in yields of 68% and 98%,
respectively, using mild basic conditions.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 56
a
Scheme 2. Synthesis of 8-amidochromen-4-one-2-carboxylic acid derivatives
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
a
Reagents and conditions: (a) SOCl , cat. DMF, DCM, argon, rt, 30 min; (b) 17, DIPEA, DCM, argon,
2
rt, 24 h, yield 74–79% for 95–96; 18, DIPEA, DCM, THF, argon, rt, 2 d, yield 79–80% for 97–98; (c)
K CO , THF, EtOH, H O, rt, overnight; aq. HCl (2 N), H O, yield 68–99%.
2
3
2
2
Altogether, 41 final products (carboxylates) were synthesized, analyzed and tested, 39 of which are
new compounds not previously described in the literature. In addition, 37 of the prepared intermediate
esters were not previously described and therefore fully analyzed as well. The structures of the
1
13
synthesized compounds were confirmed by H and C NMR spectroscopy, in addition to HPLC
analysis coupled to electrospray ionization mass spectrometry (LC/ESI-MS), which was also used to
determine the compounds’ purity.
ACS Paragon Plus Environment

Page 13 of 56
Journal of Medicinal Chemistry
Hydrolytic stability
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
Hydrolytic stability tests were performed for the most active compound of the series, chromen-4-one-
2-carboxylic acid 85, at three different pH values. Therefore, three samples were prepared: (i) a 1.20
mM solution of 85 in 1:1 acetonitrile/water (pH 3.5) containing acetic acid (300 mM) and ammonium
acetate (2 mM), (ii) a 1.20 mM solution of 85 in 1:1 acetontrile/water (pH 5.5), and (iii) a 1.20 mM
solution of 85 in 1:1 acetontrile/water (pH 8.0) containing ammonia (1 M) and ammonium acetate (2
mM). All samples were analyzed by LC-MS directly, and after 90 min, 4 hours and 2–4 days,
respectively. In all cases, no degradation of 85 occurred. These results confirmed the hydrolytic stability
of chromen-4-one-2-carboxylic acids over a wide pH range from acidic to basic conditions, which can
be observed in the human body.
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
Pharmacological Evaluation
The Pathhunter® β-arrestin assay (DiscoverX, Fremont, CA) was employed, in which the receptors
were C-terminally tagged with a small enzyme fragment. This construct was inserted into a Chinese
hamster ovary (CHO) cell line stably expressing β-arrestin fused to an N-terminal deletion mutant of β-
galactosidase. Upon activation, receptors recruited β-arrestin fusion proteins, which allowed enzyme
complementation. Thus, subsequently measured enzyme activity correlated with agonist-induced β-
arrestin recruitment. The compounds were initially tested at a concentration of 10 µM for their agonistic
potency at the human GPR35. Effects were normalized to the signal induced by 30 µM zaprinast
corresponding to its maximal effect. If compounds showed effects >50%, concentration-response curves
were determined. Compounds that did not show agonistic activity were tested for their potency to
antagonize receptor activation by 5 µM zaprinast (~ EC ). A selection of compounds was additionally
8
0
tested at the human GPR35b. Compounds were further evaluated in β-arrestin assays at the mouse and
rat GPR35, and, in addition, at the human GPR55 to determine selectivity. Effects were normalized to
the signal induced by 30 and 10 µM zaprinast for mouse and rat GPR35, respectively, or by 1 µM
lysophospatidylinositol (LPI) for GPR55. The compounds were additionally examined for their potency
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 56
to inhibit GPR55-β-arrestin recruitment induced by 1 µM LPI. Results are summarized in Tables 3, 4, 5
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
and 6.
Structure-Activity Relationships
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A series of 41 mostly novel chromen-4-one-2-carboxylic acids and one intermediate ethyl ester
(compounds 16, 21, 22, 40, 57–90, and 99–102, see Tables 3 and 4) were evaluated at the human
GPR35. All compounds were shown to be agonists except for 21, 99 and 100, which were inactive.
Some compounds showed a somewhat increased or reduced efficacy compared to that of the standard
agonist zaprinast. However, there was no obvious correlation pattern between the compounds’ structures
and their efficacy, and most of the differences were not statistically significant. Thus, the majority of
compounds can be considered to be full agonists. The 8-nitro-substituted derivative 16 was the simplest
compound that showed weak agonistic properties (EC50 46300 nM). If the nitro group of 16 was
reduced to an amino function (in 21), the resulting compound was virtually inactive in both, agonist and
antagonist assays. This indicated that lipophilic residues and hydrogen bond-accepting groups might be
favorable at position 8. Due to this observation we chose an amide linker to introduce various larger 8-
substituents. The acetamide (57) showed a 2-fold increase in potency (EC 24100 nM) compared to the
5
0
nitro derivative 16. Introduction of longer and bulkier alkyl substituents further improved potency:
ethyl- (58, EC 13400 nM) and cyclohexyl- (60, EC 4200 nM) substituted compounds had 2- and 6-
5
0
50
fold increased potency, respectively, compared to 57. The introduction of an alkyl substituent with a
distal methyl ester (59, EC50 20200 nM), thus displaying increased polarity, appeared not to be
beneficial. In contrast, compounds with aromatic groups in that position, such as phenyl (64, EC 4910
5
0
nM), 2-naphthyl (62, EC 4670 nM) and 2-quinolinyl (63, EC 1340 nM) residues were equally or in
50 50
case of the latter even more potent than the cyclohexyl derivative 60. However, elongation of the linker
between the chromenone core and the distal aromatic group led to a somewhat decreased potency, as
observed for the benzyl derivative 61 (EC50 8960 nM). Bromo-substitution in the 6-position of the
ACS Paragon Plus Environment

Page 15 of 56
Journal of Medicinal Chemistry
chromenone scaffold appeared to have a strong effect on potency: 6-bromination of the inactive 8-amino
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
derivative 21 led to 22 with an EC value of 12600 nM (Table 3).
50
Table 3. Potency of 8-substituted chromen-4-one-2-carboxylic acid derivatives and standard agonists
at GPR35
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
human GPR35
mouse GPR35
rat GPR35
EC ± SEM
E
max
µM
a
a
a
EC ± SEM
(µM)
E
max
EC ± SEM
50
(µM)
E
max
50
50
1
2
R
compd
R
a
% Effect ±
SEM)
a
(% Effect ±
SEM)
a
(
(% Effect ±
SEM)
0.986 ±
2
, cromolyn
1.26 ± 0.17
1.96 ± 2.4
119
100
4.84 ± 0.60
102
100
104
100
0
.126
.0611 ±
.0061
0
3, zaprinast
1.60 ± 0.04
0
1
2
6
1
46.3 ± 3.2
(2 ± 4)
101
-
(4 ± 6)
(4 ± 2)
(18 ± 4)
(3 ± 1)
(12 ± 1)
-
-
-
-
-
(38 ± 3)
(40 ± 2)
(40 ± 4)
(31 ± 3)
(30 ± 5)
-
-
-
-
-
H
Br
H
22
57
58
12.6 ± 1.5
24.1 ± 3.3
13.4 ± 3.4
84
91
119
methyl
ethyl
H
5
9
H
20.2 ± 5.7
87
(15 ± 4)
-
6.65 ± 0.34
89
6
6
6
6
6
0
4
1
2
3
H
H
H
H
H
cyclohexyl
phenyl
4.20 ± 0.38
4.91 ± 0.65
8.96 ± 0.43
4.67 ± 1.00
118
107
107
125
81
(38 ± 5)
(13 ± 4)
-
-
1.20 ± 0.12
(39 ± 3)
(47 ± 5)
(46 ± 3)
(41 ± 6)
79
-
benzyl
(19 ± 4)
-
-
2-naphthyl
2.99 ± 0.51
3.59 ± 0.30
79
65
-
2-quinolinyl 1.34 ± 0.06
-
a
Screening was performed at a concentration of 10 µM. Effects were normalized to the signal induced
by 30 µM (human and mouse) and 10 µM zaprinast (rat), corresponding to a maximal response at the
respective receptor.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 56
Table 4. Potency of 8-benzamido-chromen-4-one-2-carboxylic acid derivatives at GPR35
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
1
R
2
R
human GPR35
mouse GPR35
rat GPR35
a
a
a
Emax
Emax
Emax
EC ±
SEM (µM)
EC ±
50
EC ±
50
50
compd
SEM (µM)
SEM (µM)
ortho meta
para
a
a
(% Effect ±
SEM)
a
(% Effect ±
SEM)
(% Effect
± SEM)
5
6
1
.44 ± 1.34
70
(18 ± 4)
-
-
2.75 ± 0.35
71
-
6
6
6
6
6
5
6
7
8
9
H
H
H
H
H
H
H
H
H
H
CH
3
H
H
.97 ± 0.30 112
.38 ± 0.14 104
(26 ± 3)
(43 ± 2)
(23 ± 2)
(32 ± 6)
(43 ± 3)
NO
H
2
-
-
CH
CF
3
1
5.5 ± 3.1
81
-
(38 ± 4)
-
H
3
0
.346 ±
.037
.804 ±
0.148
.382 ±
.056
.0170 ±
.0017
104 5.03 ± 0.67
66
2.16 ± 0.22
85
H
OCH
Br
3
0
0
122
106
113
131
123
80
(28 ± 1)
n. d.
-
-
-
-
-
-
-
(49 ± 6)
3.13 ± 0.28
(46 ± 3)
-
70
H
F
F
F
F
H
Cl
Cl
H
H
Cl
H
0
0
82
-
7
7
7
7
1
2
3
4
H
0
(24 ± 4)
(37 ± 4)
(30 ± 4)
(7 ± 2)
(31 ± 3)
Cl
0
0.117 ±
0.026
2.65 ± 0.36
1.76 ± 0.08
(1 ± 1)
80
78
-
Cl
H
Cl
0
.112 ±
.011
.112 ±
.024
0.430 ±
.063
H
OCH
3
0
0
40
0
115
2.84 ± 0.41
2.10 ± 0.21
65
92
7
5
Cl
H
H
H
H
0
1
1
.22 ± 0.17 102
(31 ± 5)
(18 ± 1)
(34 ± 3)
-
-
-
7
7
7
6
7
8
Cl
Cl
Cl
H
H
H
CH
3
.71 ± 0.16 130
3.14 ± 0.65 115
H
NO
2
0
.0168 ±
.0021
.303 ±
0.021
110
110
1.83 ± 0.27
2.75 ± 0.60
86
75
H
OCH
3
0
0
(40 ± 4)
-
79
Br
H
H
H
ACS Paragon Plus Environment

Page 17 of 56
Journal of Medicinal Chemistry
0
.842 ±
.165
.0251 ±
138
115
110
110
108
121
99
(27 ± 5)
(27 ± 4)
(12 ± 4)
(29 ± 2)
(41 ± 1)
(36 ± 1)
(36 ± 3)
4.95 ± 0.81
-
-
-
-
-
-
-
2.08 ± 0.33
(47 ± 1)
79
-
8
0
Br
Br
Br
Br
Br
Br
Br
H
H
Cl
H
H
H
H
CH
H
3
H
Cl
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
0
81
0.0019
0.0164 ±
(32 ± 3)
-
8
8
2
3
H
Cl
0
.0011
.0154 ±
.0030
.399 ±
0
3.25 ± 0.35
2.26 ± 0.21
1.40 ± 0.03
1.36 ± 0.16
1.10 ± 0.18
62
99
88
74
73
Cl
H
Cl
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
84
CN
OCH
0.060
.0121 ±
0
8
8
5
6
H
3
0
.0010
.634 ±
.076
0
H
ethoxy
H
0
1.66 ± 0.28
91
71
73
8
8
7
8
Br OCH
3
H
0.425 ±
0.075
102 6.40 ± 0.48
1.30 ± 0.16 106
Br
Br
H
H
OCH
OCH
3
3
0
.0599 ±
.0042
.0111 ±
107
(34 ± 4)
-
2.05 ± 0.55
86
8
9
0
0
120
-
(35 ± 2)
(13 ± 3)
-
-
4.17 ± 0.33
81
-
90
Br
H
Cl
H
H
H
OCH
3
0.0031
(10 ± 3)
cyclo-
propyl-
methoxy
prop-
(48 ± 1)
9
9
H
H
H
(4 ± 2)
-
(13 ± 4)
(27 ± 5)
-
-
3.61 ± 0.38
1.54 ± 0.21
80
68
1
1
00
01
oxy
0
.144 ±
.017
.0305 ±
.0016
115
0
0
117 4.98 ± 0.35
64
0.991 ±
0.064
94
102
0
a
Screening was performed at a concentration of 10 µM. Effects were normalized to the signal induced by 30 µM
human and mouse) and 10 µM zaprinast (rat), corresponding to a maximal response at the respective receptor.
(
The phenyl-substituted derivative 64 (8-benzamido-4-oxo-4H-chromene-2-carboxylic acid) was
selected for further modification due to its high potency, and because a large number of benzoic acid
derivatives are easily accessible. In addition, we introduced different halogen atoms in the 6-position
due to the initial observation that a bromine atom in that position considerably improved potency
(
compare 21 to 22). 8-Benzamidochromen-4-one-2-carboxylic acids with a variety of substituents in the
o-, p-, and m-position (mono- and disubstitutions) were investigated. We most thoroughly studied
substituents in the para-position, since some substituents in that position led to a considerable increase
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 56
in potency. The following rank order of potency among compounds with varying substituents in the
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
para-position was observed: methoxy (69, EC 346 nM) > bromo (70, EC 804 nM) > methyl (67,
5
0
50
EC 1380 nM) > hydrogen (64, EC 4910 nM) > trifluoromethyl (68, EC 15500 nM) > propoxy (99)
50
50
50
≈
cyclopropylmethoxy (100). meta-Substitution with a methyl (65, EC 5440 nM) or a nitro group (66,
50
EC50 6970 nM) did not increase potency compared to the unsubstituted benzamidochromenone
derivative 64 (EC50 4910 nM). Since para-methoxy substitution appeared most favorable so far (69,
EC 346 nM) we combined it with halogen substitution in the 6-position in order to investigate whether
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
50
a combination could further improve potency. After comparing 69 (6-H) with 74 (6-F, EC 112 nM) 78
5
0
(6-Cl, EC 16.8 nM), and 85 (6-Br, EC 12.1 nM) it became clear that any of the introduced halogen
50 50
atoms caused an increase in potency. However, the larger bromine substituent had the greatest impact,
followed by the chlorine and fluorine substituents. Although the difference in potencies between
bromine- and chlorine-substituted compounds was small, it was consistent (observable for 79, 6-Br,
EC 303 nM versus 75, 6-Cl, EC 430 nM; 85, 6-Br, EC 12.1 nM versus 78, 6-Cl, EC 16.8 nM; and
50
50
50
50
80, 6-Br, EC 842 nM versus 76, 6-Cl, EC 1200 nM). Interestingly, another group also found that the
5
0
50
introduction of a bromine atom as a substituent of a bicyclic aromatic core structure greatly increased
4
the compound’s potency at the human GPR35. Thus, there seems to be a lipophilic pocket, which is
7
filled by the bromine atom.
Since bromine was discovered to be the best 6-substituent, most further derivatives contained a
bromine atom at that position and only few 6-fluoro- and 6-chloro-substituted derivatives were
prepared. The p-methoxy substituent at the 8-benzamide residue was also found to the best
monosubstitution in the 6-bromo-substituted series (85, EC 12.1 nM). It was 2-fold more potent than
50
the p-chloro-substituted derivative 81 (EC50 25.1 nM). The rank order of potency for 6-bromo-
substituted compounds with varying substituents in the para-position was as follows: methoxy (85,
EC 12.1 nM) > chloro (81, EC 25.1 nM) > hydrogen (79, EC 303 nM) ≥ cyano (84, EC 399 nM)
50
50
50
50
>
ethoxy (86, EC 634 nM).
0
5
ACS Paragon Plus Environment

Page 19 of 56
Journal of Medicinal Chemistry
In the 6-chloro-substituted series, which behaved very similar to the 6-bromo-substituted series, it was
shown that a p-nitro function was not well tolerated, leading to a reduced potency for compound 77 with
an EC value of 1710 nM (compared to the 6-chloro-substituted benzamidochromenone 75, EC 430
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
5
0
50
nM).
A methoxy substitution in the o-position, investigated in the series of 6-bromo-substituted
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, was not well tolerated reducing potency by a factor of 6 compared to the unsubstituted
derivative (87, EC50 1660 nM and 79, EC50 303 nM). A methyl residue in the m-position was better
tolerated reducing potency only 2-fold (80, EC50 842 nM). A similar effect was observed in the 6-
chloro-substituted series: 3-fold reduction in potency by an m-methyl residue (76, EC50 1220 nM
compared to unsubstituted 75, EC 430 nM).
50
Furthermore disubstitution of the benzamide ring was investigated. In the 6-fluoro-substituted series
o,m- (71), o,p- (72) and m,p- (73) dichloro-substituted derivatives were investigated. The rank order of
potency was as follows: o,p-dichloro (EC 17.0 nM) >> m,p-dichloro (EC 117 nM) > o,m-dichloro
50
50
(EC50 382 nM). This result confirmed again the importance of a para-substituent such as a chlorine
atom (or a methoxy group). Therefore, in the 6-bromo-substituted series, we only investigated the o,p-
dichloro- (82) and the m,p-dichloro- (83) substituted derivatives. Here, both substitution patterns led to
potent GPR35 agonists with almost the same potency as the mono-methoxy-substituted compound 85
(82, EC 16.4 nM; 83, EC 15.4 nM; 85, EC 12.1 nM). This result demonstrated the interdependence
50 50 50
of the substituents in the 6- and the 8-position. Thus, SARs were not always parallel in the 6-fluoro- and
the 6-bromo-substituted series of compounds.
A m,p-dimethoxy substitution in the 6-bromo-substituted derivative 88 led to a large reduction in
activity compared to the p-mono-methoxy-substituted compound (88, EC 425 nM; 85, EC 12.1 nM).
5
0
50
The corresponding, less bulky m,p-methylendioxy-substitution derivative was better tolerated (89, EC50
9.9 nM). A compound in which an o-chloro substituent was combined with a p-methoxy substituent
did not show significantly improved potency, but was well tolerated (90, EC 11.1 nM; 85 EC 12.1
5
50
50
nM).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 56
Finally, the p-methoxybenzamide was replaced by corresponding nitrogen-containing
methoxypyridine rings. A nitrogen atom in the 3-position (102, EC 30.5 nM) was better tolerated than
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
5
0
in the 2-position (101, EC 144 nM), but both compounds showed reduced potency as compared to the
50
parent methoxybenzamido derivative.
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
Most, although not all of the GPR35 agonists described in literature are negatively charged under
physiological conditions. In order to address the question whether the negative charge of the chromen-4-
one-2-carboxylic acid derivatives contributed to their activity the intermediate compound 40, the ethyl
ester of the potent agonist 74, was also tested. Both compounds were equally potent (EC50 112 nM).
This is surprising because ethyl esters of other acidic GPR35 agonists were found to have a strongly
2
0, 47, 48
reduced potency.
Although 40 was equally potent, its efficacy was considerably reduced
compared to that of 74 (by 40%) and it thus appeared to be a partial agonist. One explanation might be a
cleavage of the ethyl ester during the assay, for example by esterases. However, the ester and the acid
clearly differed in their potency at the rat GPR35, the acid being active (EC 1.76 µM) while the ester
50
was inactive (no activation at 10 µM, see Table 4). Additional studies, including the preparation of
stable, non-acidic derivatives, will be performed in the future to address this issue.
All in all, we could show that the potency of the virtually inactive chromen-4-one-2-carboxylic acid 21
was greatly increased when a phenyl residue was attached via an amide linker in the 8-position of the
core structure (64, EC 4910 nM). The 6-bromo-substitution at the chromenone core (79, EC 303 nM)
5
0
50
and the introduction of a methoxy group in the para-position of the phenyl ring (69, EC50 346 nM)
further increased the compounds’ potency. Combination of both substitutions had additive effects on
potency and yielded an agonist with an EC50 value of 12.1 nM (85). Furthermore, we observed that
chlorine substituents at the phenyl ring of the 8-benzamido-chromen-4-one-2-carboxylic acid
derivatives also caused a significant increase in potency: the p-chloro-substituted 81 had an EC value
5
0
of 25.1 nM, while the potency of dichloro-substituted 82 and 83 was somewhat higher (EC values 16.4
0
5
ACS Paragon Plus Environment

Page 21 of 56
Journal of Medicinal Chemistry
and 15.4 nM, respectively; for concentration-response curves of the compounds mentioned above see
Fig. 3). Thus, we successively developed a set of novel, highly potent GPR35 agonists with 85 and 90
being the most potent agonists yet described for GPR35, at least when comparing data from β-arrestin
recruitment assays.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 56
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
O
O
1
2
1
20
00
80
R
R
R1
O
85
Br
Br
H
OCH3
H
1
OH
79
69
64
NH
O
OCH3
6
0
H
H
2
4
0
R
16
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
0
0
-
10
-9
-8
-7
-6
-5
-4
-3
10
10
10
10
10
10
10
10
[
compound], M
O
O
ortho
meta
para
120
00
R
R
R
Br
OH
O
90
Cl
H
H
Cl
H
H
H
OCH3
1
O
NH
R
83
82
81
Cl
80
Cl
H
Cl
Cl
H
60
40
79
H
20
0
-
10
-9
-8
-7
-6
-5
10
10
10
10
10
10
[compound], M
O
1
00
human GPR35
EC50 1660 nM
Br
OH
O
80
60
40
20
0
O
NH
O
rat GPR35
EC50 1100 nM
3
H CO
mouse GPR35
EC50 4950 nM
-8
-7
-6
-5
-4
10
10
10
10
10
[
87], M
ACS Paragon Plus Environment

Page 23 of 56
Journal of Medicinal Chemistry
Figure 3. β-Arrestin recruitment assays at GPR35. (A, B) Concentration-response curves of selected,
successively optimized compounds at the human GPR35; EC values, 85: 12.1 nM, 79: 303 nM, 69:
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
50
346 nM, 64: 4910 nM, 87: 46300 nM, 90: 11.1 nM, 83: 15.4 nM, 87: 16.4 nM, 81: 25.1 nM. (C)
Concentration-response curves of compound 87 at GPR35 orthologs.
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
Activity on the isotypic human GPR35b
In humans a second receptor isotype was identified, designated GPR35b. This receptor protein has an
6
extended N-terminus (by 31 amino acids) but is otherwise identical to the human GPR35. A series of
compounds were tested at the human GPR35b. The results are summarized in Table 5. The potency for
almost every compound at GPR35b was not significantly different from that at the shorter GPR35
receptor isotype. This is in agreement with published results by other groups, who had also not found
2
3, 49
any difference in the two isotypes’ pharmacological properties.
The physiological role of GPR35b is
still unclear. However, it has previously been hypothesized to be a precursor of GPR35 that might be
6
activated” upon cleavage yielding GPR35. The N-terminus of GPR35b might also be cleaved off to
“
release GPR35 and a signaling peptide. Such signaling peptides are frequently found in class B GPCRs,
5
0
but have also been suggested to exist in class A GPCRs. Interestingly, GPR35b has only been detected
on the mRNA level, but never on the protein level. Possibly, the GPR35b mRNA could partially be
translated into the shorter isoform, because the start codon of the GPR35 is included in the sequence and
might alternatively be used (a process known as leaky scanning). For future studies, it would be of
interest to show whether the GPR35b exists on the protein level, and to address the question whether
and under which conditions its N-terminus is cleaved off.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 56
Table 5. Potency of selected derivatives at GPR35b
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
human GPR35b
a
compd
EC ± SEM (µM)
Emax
5
0
a
% Effect ± SEM)
b
(
ns
2.24 ± 0.47
100
115
113
104
110
136
113
125
111
130
116
109
136
101
118
129
3
ns
0.472 ± 0.057
69
70
71
72
73
74
78
80
81
82
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
ns
0.678 ± 0.100
0
.873 ± 0.100*
ns
0
0
.0205 ± 0.0010
ns
0
.106 ± 0.004
.143 ± 0.028
ns
0
ns
.0286 ± 0.0041
ns
0
.699 ± 0.061
ns
ns
ns
ns
0
0
0
0
.0405 ± 0.0092
.0346 ± 0.0093
.0272 ± 0.0054
.0221 ± 0.0036
83
85
88
89
90
ns
0
.628 ± 0.047
ns
0
.128± 0.025
ns
0
.0147± 0.0014
a
b ns
Effects were normalized to the effect induced by 30 µM zaprinast. Results of a two-tailed t test: data
not significantly different from those at the human GPR35, * p > 0.05.
Species selectivity
To investigate species differences of the chromenone derivatives, we additionally determined their
potency to induce β-arrestin recruitment at mouse and rat GPR35. The results are displayed in Tables 3
and 4. Most compounds appeared to be agonists at the rodent receptors as well. However, they were
ACS Paragon Plus Environment

Page 25 of 56
Journal of Medicinal Chemistry
mostly far less potent than at the human GPR35. Moreover, if concentration-response curves were
determined, observed maximal effects for many compounds were below that of zaprinast. Thus, most of
these compounds are partial agonists at the rodent GPR35.
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
At the mouse receptor, compounds with long and bulky substituents introduced via an amide linker in
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
the 8-position (e. g. 2-naphthyl in 62, EC 2990 nM, 2-quinolinyl in 63, EC 3590 nM) were much
5
0
50
more potent than simpler derivatives like the 8-amino-substituted 21. This is in agreement with the
results obtained for the human receptor. However, the increase in potency could only be found for
compounds with bicyclic substituents or (very few) substituted phenyl derivatives including p-
methoxyphenyl (69, EC 5030 nM), o-methoxyphenyl (87, combined with 6-bromo-substitution, EC
50
5
0
4950 nM), and m,p-dimethoxyphenyl (88, combined with 6-bromo-substitution, EC50 6400 nM).
However, the simple benzamido-chromenone 64, which we used as a lead structure for developing
agonists for the human receptor, was nearly inactive at the mouse GPR35. Another major difference in
the SARs between mouse and human is the fact that halogen substituents in the 6-position of the
chromenone core caused a reduction in potency at the mouse receptor, which becomes evident by
comparison of the 6-unsubstituted p-methoxybenzamido derivative 69 (EC50 5030 nM) with the
corresponding halogenated analogs 74 (6-F), 78 (6-Cl), and 85 (6-Br) (EC >10000 nM). Furthermore,
5
0
chloro-substitution in different positions of the phenyl ring of the 8-benzamido residue did not lead to
potent agonists. In agreement with results at the human receptor, substitution of the phenyl ring of 64
with a p-methoxy residue (69, EC 5030 nM) increased potency at the mouse receptor. Exchange of the
5
0
phenyl ring of 85 for a 3-pyridyl residue (102, EC 4980 nM), but not for a 2-pyridyl ring (101, EC
50
50
>10000 nM), led to an increase in potency.
At the rat GPR35 chromenone derivatives with small substituents in the 8-position like the 8-amino
derivative 21 displayed low potency. Similar as for the human receptor, introduction of a cyclohexyl
moiety at the 8-position of the core structure connected by an amide linker increased potency (60, EC50
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 56
1
200 nM). However, this effect could not be observed for compounds with aromatic residues in the
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
same position (phenyl in 64, benzyl in 61, 2-naphthyl in 62 or 2-quinolinyl in 63, all with EC >10000
5
0
nM) in contrast to the human GPR35, and in contrast to the mouse receptor regarding the derivatives
with bicyclic residues (62, 63). We next examined derivatives of 8-benzamido-4-oxo-4H-chromene-2-
carboxylic acid (64) with various substituents at the phenyl ring. The p-methoxy substitution (69, EC50
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
160 nM) caused an increase in potency, which appeared to be a general effect observed for all
orthologs. We also observed that bioisosteric replacement of the phenyl ring in 85 (EC 1400 nM) by a
5
0
2-pyridyl residue (102, EC 991 nM), or a 3-pyridyl residue (101, EC 1540 nM) was tolerated. These
50
50
observations are consistent with those for the mouse receptor. However, there are a few SARs unique
for the rat GPR35: halogen substituents in the 6-position of the chromenone core structure barely
affected the compounds’ potency at the rat receptor, observable e. g. for the p-methoxybenzamido
derivatives: 69 (6-H, EC 2160 nM), 74 (6-F, EC 1760 nM), 78 (6-Cl, EC 1830 nM), and 85 (6-Br,
5
0
50
50
EC 1400 nM). Comparison of the compounds with chloro substituents (71, o,m-dichloro (6-F), EC
50
50
3130 nM; 72, o,p-dichloro (6-F), EC >10000 nM; 73, m,p-dichloro (6-F), EC 2650 nM; 81, p-chloro
5
0
50
(6-Br), EC >10000 nM; 82, o,p-dichloro (6-Br), EC >100000 nM, and 83, m,p-dichloro (6-Br), EC
50 50 50
3
250 nM) showed that mainly a chlorine atom in the meta-position increased potency. Contrary to the
human receptor, the negative charge appeared to be important: the ethyl ester 40 was inactive in contrast
to the corresponding acid 74 (EC 1760 nM). Interestingly, apart from 88 (m,p-dimethoxy-substituted,
50
EC 1300 nM), 77 (with a nitro group in the para-position, EC 3140 nM) and 84 (cyano-substituted
50
50
in the para-position, EC 2260 nM) the efficacies of the agonists appeared to be somewhat lower than
50
that of the standard compound zaprinast.
We showed that most compounds are highly species-selective for the human GPR35, e. g. 85 (>100-fold
compared to rat GPR35, >1000-fold compared to mouse GPR35). However, we could also identify
several compounds that activated two of the three investigated orthologs in a nearly equipotent manner:
(i) human and rat GPR35: 60 (EC , human: 4200 nM, rat: 1200 nM), 65 (EC , human: 5440 nM, rat:
50 50
ACS Paragon Plus Environment

Page 27 of 56
Journal of Medicinal Chemistry
2
750 nM), 76 (EC , human: 1220 nM, rat: 2100 nM), 77 (EC , human: 1710 nM, rat: 3140 nM), and
50 50
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
8
0 (EC , human: 842 nM, rat: 2080 nM); (ii) human and mouse GPR35: 62 (EC , human: 4670 nM,
50 50
mouse: 2990 nM) and 63 (EC , human: 1340 nM, mouse: 3590 nM); (iii) rat and mouse GPR35: 69
5
0
(EC , rat: 2160 nM, mouse: 5030 nM). Compound 87 is the only compound that was shown to activate
50
all three investigated orthologs equipotently with EC50 values in the low micromolar range (human:
660 nM, rat: 1100 nM, mouse: 4950 nM, see Figure 3C).
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
1
Receptor selectivity
In order to assess the selectivity of the new GPR35 agonists versus the closely related human orphan
receptor GPR55, we tested the compounds’ potency to induce β-arrestin recruitment by GPR55. We
found that none of the chromenone derivates activated GPR55 at a high concentration of 10 µM. Thus,
the investigated compounds are selective GPR35 agonists versus GPR55. Moreover, these results
showed that the effects observed for GPR35 cannot represent assay artifacts, since the same assay
system (Pathhunter® β-arrestin assay) was used for assaying the receptors, employing CHO cell lines
recombinantly expressing either GPR35 or GPR55. Subsequently, the compounds were evaluated as
potential GPR55 antagonists. Several compounds were found to inhibit LPI-induced β-arrestin
recruitment with moderate potency. For the most potent antagonists full concentration-inhibition curves
were determined. We identified several moderately potent GPR55 antagonists with IC50 values in the
lower micromolar range. Thus, we discovered some novel, however non-selective GPR55 antagonists.
Interestingly, 85 and 90 were not only among the most potent GPR55 antagonists, but were also the
most potent GPR35 agonists. However, both compounds were still more than 1700-fold selective for
GPR35.
Table 6. Potency of chromen-4-one-2-carboxylic acid derivatives at human GPR55
compd
human GPR55
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 56
IC ± SEM (µM)
5
0
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
EC ± SEM (µM)
50
a
% Effect ± SEM)
(% Inhibition ±
b
SEM)
(
1
6
(-6 ± 2)
(-11 ± 4)
(4 ± 2)
(-1 ± 5)
(1 ± 7)
(15 ± 4)
(-4 ± 7)
(17 ± 7)
(-11 ± 3)
(6 ± 5)
(-6 ± 4)
(3 ± 2)
(10 ± 6)
(12 ± 7)
(6 ± 6)
(1 ± 3)
(0 ± 6)
(8 ± 7)
(5 ± 4)
(-8 ± 5)
(3 ± 0)
(7 ± 3)
(20 ± 4)
(4 ± 4)
(8 ± 2)
(-2 ± 3)
(23 ± 7)
(-8 ± 4)
(32 ± 8)
(19 ± 2)
(17 ± 3)
(26 ± 1)
(-41 ± 9)
(16 ± 4)
(31 ± 8)
(16 ± 6)
(26 ± 10)
29.1 ± 17.3
(31 ± 3)
(24 ± 2)
(25 ± 3)
(20 ± 3)
(16 ± 8)
(5 ± 14)
(-3 ± 3)
(9 ± 4)
21
22
40
57
58
59
60
61
62
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
8
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(40 ± 8)
(18 ± 9)
(54 ± 5)
(34 ± 7)
(23 ± 1)
(40 ± 6)
(14 ± 9)
(47 ± 6)
(21 ± 6)
(46 ± 6)
ACS Paragon Plus Environment

Page 29 of 56
Journal of Medicinal Chemistry
8
8
8
8
8
8
8
8
8
9
9
1
2
3
4
5
6
7
8
9
0
9
(12 ± 4)
(10 ± 9)
(16 ± 5)
(0 ± 2)
(40 ± 5)
1
2
3
4
5
6
7
8
9
(43 ± 5)
(47 ± 7)
(38 ± 10)
21.7 ± 6.2
(39 ± 7)
(0 ± 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
(-1 ± 7)
(9 ± 5)
(44 ± 2)
(-12 ± 1)
(9 ± 3)
(45 ± 5)
(46 ± 4)
(10 ± 2)
(-5 ± 7)
(10 ± 1)
(4 ± 4)
21.7 ± 1.9
(24 ± 6)
1
1
1
00
01
02
(21 ± 9)
8.92 ± 0.64
6.20 ± 0.36
(9 ± 6)
a
Screening was performed at a concentration of 10 µM. Effects were normalized to the signal induced
b
by 1 µM LPI. Inhibition of β-arrestin recruitment induced by 1 µM LPI.
Physicochemical properties of selected compounds
Several physicochemical parameters can be calculated to evaluate the druglikeness of a given
compound, including the partition coefficient (clogP), the polar surface area (PSA), the ligand efficiency
51-53
(
LE) and the ligand-lipophilicity efficiency (LLE or LipE).
We calculated these parameters for
selected 8-benzamidochromen-4-one-2-carboxylic acids and compared them to the previously described
most potent agonist 5 (see Table 7). The clogP value was in all cases lower than 5, which is consistent
5
4, 55
with Lipinski’s rule of 5 for druglikeness.
Compounds 69, 88 and 102 exhibit a logP value between
2 and 3, which is often considered optimal for perorally administered drugs. The PSA is defined as the
sum over all polar atoms, mainly oxygen and nitrogen including the hydrogen atoms attached to them.
This parameter is used to assess the compound’s ability to permeate cells. For the most potent
ACS Paragon Plus Environment