Identification of novel GPR81 agonist lead series for target biology evaluation

Article abstract of DOI:10.1016/j.bmcl.2020.126953

GPR81 is a novel drug target that is implicated in the control of glucose and lipid metabolism. The lack of potent GPR81 modulators suitable for in vivo studies has limited the pharmacological characterization of this lactate sensing receptor. We performed a high throughput screen (HTS) and identified a GPR81 agonist chemical series containing a central acyl urea scaffold linker. During SAR exploration two additional new series were evolved, one containing cyclic acyl urea bioisosteres and another a central amide bond. These three series provide different selectivity and physicochemical properties suitable for in-vivo studies.

Full text of DOI:10.1016/j.bmcl.2020.126953

Bioorganic&MedicinalChemistryLetters30(2020)126953
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of novel GPR81 agonist lead series for target biology
evaluation
⁎
⁎
Öjvind Davidssona, , Kristina Nilssona, , Jonas Brånalta, Terese Anderssona,f, Kristina Berggrena,b
,
Yantao Chena, Ola Fjellströma,c, Henrik Gradéna, Linda Gustafssona,h, Nils-Olov Hermanssond,
Frank Jansene, Petra Johannessona,g, Bengt Ohlssona,i, Christian Tyrchana,b, Annika Wellnera,b
Eric Wellnera,j, Maria Ölwegård-Halvarssona
,
a CVRM Medicinal Chemistry, Research and Early Development, Cardiovascular, Renal and Metabolism, BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden
b RIA Medicinal Chemistry, Research and Early Development, Respiratory, Inflammation and Autoimmune, BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden
c Research and Early Development, Cardiovascular, Renal and Metabolism, BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden
d Discovery Biology, Discovery Sciences, R&D, AstraZeneca, Gothenburg, Sweden
e Mechanistic Biology & Profiling, Discovery Sciences, R&D, AstraZeneca, Gothenburg, Sweden
f Early Product Development, Pharmaceutical Sciences, BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden
g Global Patient Safety CVRM, Chief Medical Office, AstraZeneca R&D, Gothenburg, Sweden
A R T I C L E I N F O
A B S T R A C T
Keywords:
GPCR
GPR81 is a novel drug target that is implicated in the control of glucose and lipid metabolism. The lack of potent
GPR81 modulators suitable for in vivo studies has limited the pharmacological characterization of this lactate
sensing receptor. We performed a high throughput screen (HTS) and identified a GPR81 agonist chemical series
containing a central acyl urea scaffold linker. During SAR exploration two additional new series were evolved,
one containing cyclic acyl urea bioisosteres and another a central amide bond. These three series provide dif-
ferent selectivity and physicochemical properties suitable for in-vivo studies.
GPR81
GPR109A
GHSR1a
Lipolysis
HTS
GPR81 (HCA1) is a member of the hydroxyl-carboxylic acid re-
ceptor family of G-protein coupled receptors, and lactate is its en-
dogenous ligand.1 It is highly expressed in adipose tissue and is also
found in kidney, skeletal muscle and liver.2 GPR81 displays about 52%
sequence identity to its closest homologue, the nicotinic acid receptor
GPR109A (HCA2).3 Both GPR81 and GPR109A are predominantly ex-
pressed in adipose tissue, couple to Gi, and have been shown to mediate
antilipolytic effects.4 GPR81 and GPR109A agonists have the potential
to lower free fatty acids (FFA) in plasma through antilipolytic actions
and thereby decrease ectopic lipid accumulation in liver and skeletal
muscle to improve insulin sensitivity in insulin resistant and type 2
diabetic patients.5 However, the failure of GPR109A agonists to achieve
durable free fatty acid lowering and improved glucose control in the
clinic, in addition to the side effect of flushing, motivate GPR81 to be
carefully explored as an alternative drug target.6
To date, only a few reports on GPR81 agonists have been pub-
lished.7 The GPR81 agonists, including lactate, hydroxybenzoic acids,
acrylamidocyclohexene-carboxylic acids and aminothiazoles have all
shown acute FFA lowering in vivo. Only the latter of these has been
reported as a GPR81 agonist with sub-micromolar potency. There are
only two studies yet published showing effects in vivo after repeated
dosing of a GPR81 agonist.8 Thus, there is still a need for potent and
selective GPR81 agonist or antagonist tool compounds to further study
the pharmacology of the GPR81 receptor in vivo.
In the present study, we describe the identification and SAR ex-
ploration of novel GPR81 agonist lead series. Several compounds were
tested in vivo and revealed that GPR81 agonists can durably lower FFA
and improve glucose control.8b This report focuses on the medicinal
chemistry evolution of three novel GPR81 lead series, Acyl Urea-,
Constrained Analogue- and Amide-Series.
⁎ Corresponding authors.
E-mail addresses: ojvind.davidsson@astrazeneca.com (Ö. Davidsson), kristina.a.nilsson@astrazeneca.com (K. Nilsson).
h Current address: Randstad Life Sciences, Gothenburg, Sweden.
i Current address: Emeriti Bio AB, AZ Bioventure Hub, c/o AstraZeneca, Gothenburg, Sweden.
j Current address: R&D Department, Mölnlycke Health Care AB, Gothenburg, Sweden.
https://doi.org/10.1016/j.bmcl.2020.126953
Received 15 November 2019; Accepted 1 January 2020
Availableonline07January2020
0960-894X/©2020ElsevierLtd.Allrightsreserved.

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
Table 1
Potency and physicochemical data for HTS-hits 1 and 2.
O
S
O
Cl
O
O
N
Cl
O
O
N
N
S
N
H
N
H
S
N
H
N
H
N
N
O
O
N
O
N
2
N
1
Compd
hGPR81
hGPR109A
EC50 (µM)
hGHS-R1a
IC50 (µM)a
Sol. (µM)
LogD7.4
LLEb
a
a
EC50 (µM)
1
2
0.023
0.074
0.123
2.16
0.008
0.096
0.2
4
3.0
2.9
4.6
4.2
a
b
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50-LogD7.4
.
Results and discussion
Acyl urea series
A high throughput screening (HTS) of the AstraZeneca proprietary
compound collection was conducted. The primary assay in the
screening cascade was a functional cell-based assay using a Chinese
hamster ovary cell line overexpressing GPR81. GPR81 agonism was
assessed through measurement of inhibition of forskolin stimulated
cAMP production.8b Screening of compounds started at single con-
centration (10 μM) and the actives were retested and screened in a
counter assay using the parental cell line. The functional activity was
followed up in concentration-response (CR) cAMP assay. Subsequently,
compounds were screened in an orthogonal CR assay measuring GPR81
agonism through accumulation of GTP bound Gi protein ([35S]GTPγS
binding assay.9 An acyl urea containing cluster originating from an in-
house growth hormone secretagogue receptor type 1a (GHS-R1a) in-
verse agonist project was identified as essentially the only viable series
confirmed in the orthogonal GTPγS assay.10 The acyl ureas 1 and 2
(Table 1) showed high potency and full agonism in both GPR81 assays
(GTPγS data not shown). However, the selectivity towards GHS-R1a
and GPR109A was low.
The initial SAR exploration (Table 2), focused on improving the
hGPR81 potency and LLE, was directed towards the substituents on the
phenyl ring at the left-hand side (LHS) of the molecule. In the R1-po-
sition the chloro substituent was found to provide the best fit with re-
spect to balanced size and electronic properties for combined hGPR81
potency, efficacy and LLE.
Mapping the importance of the substituent in the R2-position
showed that the electron donating group ethoxy gave the best hGPR81
potency and LLE, see compound 1 (Table 3). Removing or replacing the
ethoxy substituent either with a smaller group like methyl or a more
electronegative atom as fluoro improved LogD but lowered the potency
and LLE drastically. Also the efficacy was decreased.
Exploring the R3-position, summarized in Table 4, showed that
saturated rings such as morpholines 2 and 9 and pyrrolidine 10 as well
as heterocycles such as the 1-pyrazole 11 and 1-pyrrole 12 gave com-
pounds with high GPR81 potency, efficacy and LLE. In comparison, the
unsubstituted compound 14 showed a drop in potency and LLE.
The 5-1H-pyrazole compound 13, a regioisomer of 11, was designed
to investigate possible formation of a favourable intramolecular hy-
drogen bond between the pyrazole-NH and the oxygen of the ethoxy
Table 2
Table 3
SAR exploration of R1-position.
SAR exploration of R2-position.
O
O
R1
O
O
N
N
S
Cl
O
O
N
S
N
H
N
H
S
N
N
H
N
H
O
N
R2
N
O
N
N
a
a
Compd
R1
hGPR81 EC50 (µM)
LogD7.4
LLEb
Compd
R2
hGPR81 EC50 (µM)
LogD7.4
LLEb
2
3
4
5
Cl
0.074
2.9
2.9
3.3
3.0
4.2
1
6
7
8
EtO
Me
F
0.023
3.0
2.5
2.4
2.3
4.6
2.7
3.0
2.9
Me
CF3
H
> 3.7
< 2.5
2.6
6.1 (70%)c
3.6 (70%)c
7.0 (74%)c
1.4
0.74 (47%)c
3.1
H
a
b
a
b
Mean values from a minimum of two experiments.
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50
-
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50-
LogD7.4
.
LogD7.4
.
c
c
Efficacy in brackets.
Efficacy in brackets.
2

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
Table 4
Table 5
SAR exploration of R3-position.
SAR exploration of the right-hand side (RHS).
Cl
O
O
Cl
O
O
N
RHS
N
N
H
N
H
S
N
H
N
H
O
N
O
N
O
R3
a
Compd
R3
hGPR81 EC50 (µM)
0.074
LogD7.4
2.9
LLEb
4.2
Compd RHS
hGPR81 EC50
(µM)a
LogD7.4 LLEb
2
N
O
2
0.074
2.9
4.2
N
N
9
0.032
3.3
4.2
S
N
O
N
N
15
0.299
4.0
2.5
10
11
12
13
0.013
0.019
0.022
0.36
4.0
3.8
4.0
3.9
3.9
3.9
3.6
2.5
N
S
S
16
17
0.103
0.260
4.2
4.4
2.8
2.2
N
N
N
N
S
18
19
0.769
0.120
3.2
4.2
2.9
2.7
N
N
HN
S
S
N
14
H
0.56
3.4
2.9
a
b
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50
LogD7.4
-
.
20
0.719
2.3
3.8
O
N
substituent. However, this compound resulted in a 20-fold drop in po-
tency compared to 11 which is also reflected in the LLE.
N
N
N
S
The initial SAR exploration was, besides the LHS, also focused on
the right-hand side (RHS) of the molecule (Table 5). Larger variations
such as substituted bicyclic-systems and 5-membered heterocycles were
investigated.
21
22
0.895
0.166
1.7
2.2
4.3
4.6
O
N
OH
Matched pairs to the HTS hit 2 with different RHS:s were all found
to give decent to high hGPR81 potency as well as high efficacy but with
varying LLE:s. The unsubstituted benzothiazole 15 and the phenyl-
substituted thiazole 16 both kept decent potencies although they
dropped in LLE due to the increased lipophilicity compared to 2.
Removing nitrogen in the benzothiazole 15 to give the benzothiophene
17 resulted in a compound with similar potency and LLE, indicating
that the thiazole-nitrogen is not essential for the binding to the hGPR81
receptor. Reduction in size of the RHS was tolerated as illustrated by
the methyl-thiazole in compound 18. Furthermore, the 6-membered
ring of the benzothiazole in 15 was saturated and found to retain
hGPR81 potency but still with low LLE, exemplified in 19.11 Interest-
ingly, LogD and LLE were significantly improved by incorporating an
acylated nitrogen into the saturated ring as illustrated by compounds
20 and 21. The most striking example of improving LLE among the
modifications performed is the sulphone linked N-methylpiperidine on
the benzothiazole, exemplified in 22, a compound almost equipotent to
2 and with somewhat higher LLE.
S
O
O
S
S
N
a
b
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50
LogD7.4.
-
larger acceptance for size and lipophilicity was observed for R3 and the
RHS (Tables 4 and 5), indicating that these positions could be suitable
for further modifications of physicochemical properties.
The second part of SAR building was focused on modifying or re-
placing the acyl urea linker. Besides hGPR81 potency and LLE, se-
lectivity towards hGHS-R1a and hGPR109A in addition to solubility
were investigated.
Summarizing the first part of SAR exploration of the acyl urea series
focusing on hGPR81 potency and LLE, modifications in the LHS showed
tight SAR for R1 and R2 substituents. Specific groups were needed to
achieve highly potent and efficacious compounds (Table 2 and 3). A
Several variations in the acyl urea linker between the LHS and the
RHS were synthesized (Table 6). Methylation of the acyl urea, ex-
emplified in compound 23, increased the solubility more than 100-fold
3

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
Table 6
Constrained analogue series
SAR exploration of the linker (L) between the left-hand side (LHS) and the right-
hand side (RHS).
The SAR discussed above shows that the linker in the acyl urea
compounds seems to direct the LHS and to some extent the RHS into
favorable locations for interaction with the hGPR81 receptor.
Therefore, a certain element of conformational constrained would be
favorable.
Cl
L
N
S
O
N
O
Conformational analysis of the acyl urea motif substructure
(Fig. 1A) combined with a search in Cambridge Structural Database
(CSD)12 revealed a conservative conformation enforced by an in-
tramolecular hydrogen bond (Fig. 1B). Subsequently, the 3D con-
formations generated for compound 2 (Fig. 1C)13 showed conforma-
tional similarities with the CSD search. Based on these results we
hypothesized that intra- and intermolecular hydrogen bonds might be
an integral part of the bioactive conformation and receptor interaction,
respectively. To test this hypothesis further, we designed con-
formationally locked acyl ureas considering the prerequisite to have
one hydrogen bond donor (HBD) and one hydrogen bond acceptor
(HBA) at correct positions from previous findings (Fig. 1D).
N
N
Compd
L
hGPR81
hGHS-R1a
IC50 (µM)
LogD7.4 Sol. (µM)
a
EC50 (µM)a
0.074
2
0.096
2.9
2.5
3.2
4
O
O
O
O
O
N
N
H
H
23
24
25
> 16
> 33
> 33
> 50b
> 4.97b
492
188
Initially this resulted in the preparation of a pyridone linker shown
in compound 26 and a pyrimidone linker 27, to replace the acyl urea
and ensuring that the identified important element is available for
hydrogen bond interaction with the hGPR81 receptor (Table 7).
For the pyridone linker in 26, the hGPR81 potency increased > 10-
fold compared with the matched pair acyl urea 15 (Table 5). The
hGPR81 selectivity over hGHS-R1a went from no selectivity for the acyl
urea 15 (hGHS-R1a IC50 0.378 μM) to excellent 100-fold selectivity for
the pyridone compound 26. Furthermore, the hGPR81 selectivity over
hGPR109A shows a 29-fold separation for the pyridone linker com-
pound 26, which seems to be no improvement over the matched pair
acyl urea 15 (> 25-fold, hGPR109A IC50 > 7.5 μM). Instead the dif-
ference is more accentuated for the pyrimidone analogue, exemplified
in 27, where the hGPR109A activity is very weak (> 45-fold).
Furthermore, introducing the RHS from the acyl urea 21 (Table 5)
to the pyridone scaffold linker yielded the matched pair 28 (Table 7). A
highly potent compound, 40-fold more potent than the acyl urea, with a
good LLE of 5.3. The selectivity towards hGPR109A and hGHS-R1a is
decent to good, 41- and 82-fold, respectively. All these properties make
28 a very interesting compound for further evaluation. In addition,
using the RHS from the acyl urea 22 (Table 5) forming the pyridone
compound 30 resulted in one of the most potent compounds (IC50
0.005 µM) overall. The two pyridones 28 and 30 are 30–40-fold more
potent than the matched pair acyl ureas 21 and 22, respectively. Since
these pyridones show a logD increase of only 10-fold, the potency in-
crease is not only driven by lipophilicity, but also from an entropy effect
due to imposed constraints in the linker. Furthermore, the pyrimidone
compound 31, a matched pair to the pyridone 30, gave a less potent
compound, 50-fold decrease in hGPR81 potency. The lipophilicity of
the pyrimidone compound 31 was also reduced significantly (logD 1.6
for 31 vs 3.4 for 30) which is reflected in similar LLE:s, showing that the
potency is lipophilicity driven.
N
H
N
N
H
N
H
O
> 50b
3.5
20
N
H
N
H
a
b
Mean values from a minimum of two experiments.
Value from one experiment.
RHS
H
RHS
O
O
O
N
X
O
RHS
N
H
O
LHS
N
H
LHS
N
H
LHS
N
H
X = CH or N
A
B
D
C
Fig. 1. A) Substructure motif of acyl urea used in Cambridge Structural
Database, CSD. B) Suggested constrained conformation of acyl urea. C) 3D
conformation analysis of 2. D) Pyridone and pyrimidone scaffold linkers, cyclic
acyl urea bioisosteres.
Finally, selective methylation of either the nitrogen or oxygen of the
pyridone completely abolished hGPR81 potency (data not shown) all in
concordance with the previously described SAR for the N-methylated
matched pair acyl urea 23 (Table 6).
compared to 2 but was accompanied by a complete loss of potency for
hGPR81 and hGHS-R1a. Removal of a carbonyl oxygen, as in compound
24, improved solubility up to 50-fold but again the activity of hGPR81
and hGHS-R1a was lost. Furthermore, to shorten the acyl urea linker to
a urea linker, as in 25, also resulted in an inactive compound, and only
slightly increased the solubility (5-fold) compared to 2.
To summarize the constrained analogues, the findings support the
importance of conformational restraint to stabilize the bioactive con-
formation needed for hGPR81 potency, in combination with the minimal
pharmacophore consisting of a hydrogen bond donor (HBD) and a hy-
drogen bond acceptor (HBA). Interestingly, the selectivity towards
hGPR109A and hGHS-R1a showed more variations depending on the RHS
used compared to the acyl ureas. In general, the difference in activity was
100-fold in favour of hGPR81 vs hGHS-R1a, which may indicate a se-
paration in SAR between hGPR81 and hGHS-R1a for this class of com-
pounds, whereas the selectivity between hGPR81 and hGPR109A showed
large variations from 6-fold for 31 to > 45-fold for 27.
It was hypothesized that the acyl urea linker may be part of an in-
termolecular hydrogen bond network and/or forming an in-
tramolecular hydrogen bond and thereby influences the solubility and
potency. The results from the two first modifications above, as in 23
and 24, are consistent with this hypothesis. In both cases the solubility
was significantly increased indicating that a crystal-packing and/or an
intramolecular constrained conformation may have been disrupted. In
addition, the potency was lost for both compounds.
4

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
Table 7
SAR exploration of constrained analogue compounds.
RHS
O
X
Cl
N
H
O
N
O
a
a
Compd
X
RHS
hGPR81 EC50 (µM)a
0.021
hGPR109A EC50(µM)
0.61
hGHS-R1a IC50(µM)
1.3
LLEb
LogD7.4
> 3.8
Sol. (μM)
< 0.7
26
CH
< 3.9
N
S
S
27
28
29
30
N
0.49
> 22
0.91
1.9
5.2
2.4
5.3
4.5
4.9
3.9
2.4
2.6
3.4
< 0.31
7
N
CH
CH
CH
0.022
0.075
0.005
1.8c
O
N
N
N
N
OH
S
> 50
0.46c
29
O
N
O
S
0.037
< 0.5
O
S
O
O
S
N
31
N
0.24
1.6
> 18
5.0
1.6
61
O
S
N
S
N
a
b
c
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50-LogD7.4
.
Value from one experiment.
Amide series
(Table 6) and the methylated pyridone compound 29 (Table 7) (data
not shown), supporting the importance of a minimal pharmacophore
consisting of an HBD and HBA.
In a final SAR exploration, the acyl urea linker was replaced by an
amide bond giving rise to a new group of compounds with different
properties compared to the acyl ureas and their constrained analogues.
The amide compounds were initially synthesized with different LHS
and RHS identified in the acyl urea series with the aim to improve
hGPR81 potency and selectivity against GPR109A and hGHS-R1a, in
addition to increasing solubility (Table 8).
Turning to the LHS, substitution in the morpholine ring was in-
vestigated. Mono- and di-substitution next to the morpholine-oxygen
were performed using 35 as matched pair. Monomethylated compound,
as in 36, and di-methylated, as in the trans isomers 37 (S,S/R,R), gave
slightly higher hGPR81 potency (~3-fold) compared to 35, but no
improvement in hGPR109A and hGHS-R1a selectivity. However, in-
troducing two methyl groups cis to each other, as in 38, gave a 10-fold
more potent hGPR81 compound, in addition to an excellent hGHS-R1a
selectivity (500-fold), compared to the matched pair 35 (80-fold). No
difference in hGPR109A selectivity was achieved though. In addition,
the solubility was good (95 µM) although less than 35 (323 µM).
Notably, 38 shows the same potency and LLE as the HTS hit 2 which is
an unselective compound towards hGHS-R1a. This makes 38 an inter-
esting compound for further evaluation.
One of the first synthesized compounds, 32, matched pair to the
acyl urea 2 and the urea 25 (Table 6), showed less potency (5-fold) than
2 but was still active in contrast to the inactive urea 25. Another RHS
identified in the acyl ureas and their constrained analogues was used in
the amide 33. Again, the potency was less compared to the matched
pair acyl urea 22 (Table 5) and constrained analogues, 30 and 31
(Table 7). The solubility was low for all of the compounds (33, 22 and
30, < =1 µM) with exception of the constrained analogue 31 (61 µM).
However, to introduce a carbon linker at the RHS between the
sulphonyl- and N-methylpiperazine group, as in compound 35, did not
improve potency and selectivity but increased the solubility > 200-fold
vs the amides 32 and 33, since a larger degree of rotational freedom
was provided.
To summarize the amide series, the amide linked compounds are in
general ~2-fold less active towards hGPR81 than the original acyl urea
series. Selectivity towards hGPR109A is similar (~10-fold) or some-
times less than the acyl ureas, irrespectively of LHS or RHS used, in-
dicating a common SAR between hGPR81 and hGPR109A also for the
amide linked series. In contrast to acyl ureas, amides show excellent
hGHS-R1a selectivity (> 70-fold, Table 8), again irrespectively of the
RHS or LHS used. This may indicate a separation in SAR between
Selective methylation of the amide nitrogen in 32, completely
abolished hGPR81 potency (data not shown). This is also in agreement
with the result from the N-methylated matched pair acyl urea 23
5

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
Table 8
SAR exploration of amide compounds.
X
O
N
Cl
S
N
H
O
R3
a
a
Compd
R3
X
hGPR81 EC50 (µM)a
0.41
hGPR109A EC50 (µM)
> 11
hGHS-R1a IC50 (µM)
> 50
LLEb
3.6
LogD7.4
2.8
Sol. (µM)
1.5
32
N
N
N
O
N
33
34
O
S
0.74
0.11
> 9
1.6
> 50
> 30
3.6
4.1
2.5
2.9
1
5
O
O
N
N
O
N
O
N
O
S
35
36
0.60
0.16
> 6.3
1.4
> 50
21c
4.2
4.4
2.0
2.4
323
470
O
O
N
N
N
N
O
N
O
S
O
c
O
S
37 (S,S/R,R)
0.35/0.27
3.2/1.4
(10/14)
3.4/3.5
3.0/3.0
155/231
O
O
N
O
N
N
N
N
38
39
0.054
1.9
0.66
12
27c
4.6
1.4
2.7
4.3
95
O
S
N
O
> 50
0.09
H
N
O
a
b
c
Mean values from a minimum of two experiments.
Ligand lipophilicity efficiency (LLE), defined as GPR81 potency pIC50-LogD7.4
.
Value from one experiment.
hGPR81 and hGHS-R1a for this class of compounds, which is also the
case with the constrained analogues. The R3 position of the LHS seems
to be important for the hGPR81 potency. The stereochemistry of dis-
ubstituted morpholine shows large difference in potency (> 5-fold)
between cis- and trans-configuration. A larger structural acceptance for
RHS was observed and the solubility showed high variability depending
on the RHS-group used, indicating that the RHS part could be suitable
for further modifications of physicochemical properties.
The overlapping SAR amongst the three scaffold linkers above is not
transferrable to the amide central linker (core 1). The non-overlapping
SAR is explained both by the smaller size of the amide and different
vector orientation towards the RHS substituents (assuming LHS over-
lapping), which denies similar RHS 3D overlap.
Summary and conclusion
A high throughput screen was performed with the aim to identify
GPR81 agonists. Only one chemical series was judged suitable to pursue
further, suggesting that GPR81 is not a highly druggable target and that
identification of suitable tool compounds for this receptor is not trivial.
Extensive characterization, SAR evaluation and optimization of this
acyl urea containing series was undertaken, and led to several findings.
First, SAR and conformational modelling suggested that the acyl ureas
form an internal hydrogen bond in their GPR81 bioactive conformation
and consistent with this, the pyridone and the pyrimidone scaffold
linkers were identified as bioisosteres of the conformationally con-
strained acyl urea. Secondly, replacing the acyl urea linker with an
amide bond gave GPR81 active compounds but with different
Acyl urea, constrained analogue and amide series
The general hGPR81 potency SAR trends for the central scaffold
linkers are summarized in Fig. 2. The acyl urea (core 2), the pyridine-2-
one (core 3) and the pyrimidine-6-one (core 4) follow a similar SAR
trend, although pyridine-2-one (core 3) display higher potency. This
general SAR trend can be explained by molecular superposition mod-
elling which suggests that all these three central groups provide similar
vectors towards the left- and right-hand sides of the molecules pro-
viding good overall molecular overlays. Thus, pyridine-2-one and pyr-
imidine-6-one act as bioisosteres of the acyl urea.
6

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
R1
O
R1
N
O
O
O
R1
R1
R2
N
H
N
H
R2
N
H
R2
N
H
O
R2
N
H
1
3
4
2
CORE
Fig. 2. hGPR81 potency SAR with respect to central scaffold linkers (cores 1–4), based on a matched pair analysis. Points connected with lines represent compounds
with identical LHS and RHS substituents and only their cores being different, as indicated in the figure.
selectivity and physicochemical properties than the acyl ureas and the
constrained analogues. Compounds from all series were further tested
in a rat antilipolysis in vitro assay and showed clear correlations be-
tween human and rat GPR81 cAMP and rat antilipolysis in vitro.
Furthermore, compounds were tested in vivo and clear correlation was
also shown between rat antilipolysis in vitro and rat FFA lowering in vivo
(see supplementary material in reference 8b, Figure S1, A, B and C.8b We
have demonstrated that conformational analysis contributed to the
evolution of non-selective HTS hits to highly potent and selective
GPR81 compounds.
vitreous lactate concentration. NMR Biomed. 1994;7:263–268. https://doi.org/10.
1002/nbm.1940070603
(f) Ellmerer M, Schaupp L, Sendlhofer G, et al. Lactate metabolism of
subcutaneous adipose tissue studied by open flow microperfusion. J Clin.
Endocrinol. Metab. 1998;83:4394–4401. https://doi.org/10.1210/jcem.83.12.
5303
(g) Langemann H, Alessandri B, Mendelowitsch A, Feuerstein T, Landolt H, Gratzl
O. Extracellular levels of glucose and lactate measured by quantitative
microdialysis in the human brain. Neurol Res. 2001;23:531–536. https://doi.org/
10.1179/016164101101198785
(h) MacLean DA, Bangsbo J, Saltin B. Muscle interstitial glucose and lactate levels
during dynamic exercise in humans determined by microdialysis. J Appl Physiol.
1999;87:1483–1490. https://doi.org/10.1152/jappl.1999.87.4.1483.
[3]. Ahmed K, Tunaru S, Offermanns S. GRP109A, GPR109B and GRP81, a family of
hydroxy-carboxylic acid receptors. Trends Pharmacol Sci. 2009;30:557–562.
https://doi.org/10.1016/j.tips.2009.09.001.
Declaration of Competing Interest
[4]. Tunaru S, Kero J, Schaub A, et al. PUMA-G and HM74 are receptors for nicotinic
acid and mediate its anti-lipolytic effect. Nat Med (NY, NY, US). 2003;9:352–355.
https://doi.org/10.1038/nm824.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
[5]. (a) Boden G, Shulman GI. Free fatty acids in obesity and type 2 diabetes: defining
their role in the development of insulin resistance and beta-cell dysfunction. Eur J
Clin Invest. 2002;32:14–23. https://doi.org/10.1046/j.1365-2362.32.s3.3.x
(b) DeFronzo RA. Insulin resistance, lipotoxicity, type 2 diabetes and
atherosclerosis: the missing links. The Claude Bernard Lecture 2009. Diabetologia.
2010;53:1270–1287. https://doi.org/10.1007/s00125-010-1684-1
(c) Unger RH. Lipotoxicity in the pathogenesis of obesity-dependent NIDDM.
Genetic and clinical implications. Diabetes. 1995;44:863–870. https://doi.org/10.
2337/diab.44.8.863.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.126953.
[6]. (a) Lauring B, Taggart Andrew KP, Tata JR, et al. Niacin lipid efficacy is in-
dependent of both the niacin receptor GPR109A and free fatty acid suppression. Sci
Transl Med. 2012;4:148ra115. https://doi.org/10.1126/scitranslmed.3003877
(b) Benyo Z, Gille A, Bennett CL, Clausen BE, Offermanns S. Nicotinicacid-induced
flushing is mediated by activation of epidermal Langerhans cells. Mol Pharmacol.
2006;70:1844–1849. https://doi.org/10.1124/mol.106.030833
(c) Benyo Z, Gille A, Kero J, et al. GPR109A (PUMA-G/HM74A) mediates nicotinic
acid-induced flushing. J Clin Investig. 2005;115:3634–3640. https://doi.org/10.
1172/JCI23626
References
[1]. (a) Blad CC, Ahmed K, IJzerman AP, Offermanns S. Biological and pharmacolo-
gical roles of HCA receptors. Adv Pharmacol. 2011;62:219–250. https://doi.org/10.
1016/B978-0-12-385952-5.00005-1
(b) Cai T, Ren N, Jin L, et al. Role of GPR81 in lactate-mediated reduction of
adipose lipolysis. Biochem Biophys Res Commun. 2008;377:987–991. https://doi.
org/10.1016/j.bbrc.2008.10.088
(d) Offermanns S. The nicotinic acid receptor GPR109A (HM74A or PUMA-G) as a
new therapeutic target. Trends Pharmacol Sci. 2006;27:384–390. https://doi.org/
10.1016/j.tips.2006.05.008
(c) Ahmed K, Tunaru S, Tang C, et al. An autocrine lactate loop mediates insulin-
dependent inhibition of lipolysis through GPR81. Cell Metab. 2010;11:311–319.
https://doi.org/10.1016/j.cmet.2010.02.012.
(e) Blad CC, Ahmed K, IJzerman AP, Offermanns S. Biological and pharma-
cological roles of HCA receptors. Adv Pharmacol. 2011;62:219–250. https://doi.
org/10.1016/B978-0-12-385952-5.00005-1
[2]. (a) Liu C, Wu J, Zhu J, et al. Lactate Inhibits lipolysis in fat cells through activation
of an orphan G-protein-coupled receptor, GPR81. J Biol Chem.
2009;284:2811–2822. https://doi.org/10.1074/jbc.M806409200
(b) Wise A, Foord SM, Fraser NJ, et al. Molecular identification of high and low
affinity receptors for nicotinic acid. J Biol Chem. 2003;278:9869–9874. https://doi.
org/10.1074/jbc.M210695200
(f) Budd TJ, Atkinson J, Buttery PJ, Salter AM, Wiseman J. Effect of insulin and
isoproterenol on lipid metabolism in porcine adipose tissue from different depots.
Comp Biochem Physiol Pharmacol Toxicol Endocrinol. 1994;108:137–143. https://
doi.org/10.1016/1367-8280(94)90024-8
(c) Ge H, Weiszmann J, Reagan JD, et al. Elucidation of signaling and functional
activities of an orphan GPCR, GPR81. J Lipid Res. 2008;49:797–803. https://doi.
org/10.1194/jlr.M700513-JLR200
(g) Cheng K, Wu TJ, Wu KK, et al. Antagonism of the prostaglandin D2 receptor1
suppresses nicotinicacid-induced vaso dilation in mice and humans. Proc Natl Acad
Sci USA. 2006;103:6682–6687. https://doi.org/10.1073/pnas.0601574103
(h) Hanson J, Gille A, Zwykiel S, et al. Nicotinic acid-and monomethylfumarate-
induced flushing involves GPR109A expressed by keratinocytes and COX-2-
dependent prostanoid formation in mice. J Clin Investig. 2010;120:2910–2919.
https://doi.org/10.1172/JCI42273
(d) Abi-Saab WM, Maggs DG, Jones T, et al. Striking differences in glucose and
lactate levels between brain extracellular fluid and plasma in conscious human
subjects: effects of hyperglycemia and hypoglycemia. J Cereb Blood Flow Metab.
2002;22:271–279. https://doi.org/10.1097/00004647-200203000-00004
(e) Berkowitz BA, Bansal N, Wilson CA. Non-invasive measurement of steady-state
7

Ö. Davidsson, et al.
Bioorganic&MedicinalChemistryLetters30(2020)126953
(i) Jacobson TA. A “hot” topic in dyslipidemia management “how to beat a flush”:
sensor GPR81 in cardiovascular control. JCI Insight. 2017;19:1–18. https://doi.org/
Optimizing niacin tolerability to promote long-term treatment adherence and
coronary disease prevention. Mayo Clin Proc. 2010;85:365–379 doi: 10.
4065%2Fmcp.2009.0535
10.1172/jci.insight.92564.
[9]. Almqvist J, Hovdal D, Ahlström C, Fjellström O, Gennemark P, Sundqvist M.
Overexpressing cell systems are a competitive option to primary adipocytes when
predicting in vivo potency of dual GPR81/GPR109A agonists. Eur J Pharm Sci.
2017;114:155–165. https://doi.org/10.1016/j.ejps.2017.11.022.
(j) Pike NB. Flushing out the role of GPR109A (HM74A) in the clinical efficacy of
nicotinicacid. J Clin Investig. 2005;115:3400–3403 doi: 10.1172%2FJCI27160
(k) Pike NB, Wise A. Identification of a nicotinic acid receptor: is this the
molecular target for the oldest lipid – lowering drug? Curr Opin Investig Drugs.
2004;5:271–275.
[10]. McCoull W, Barton P, Brown AJ, et al. Identification, optimization, and pharma-
cology of acylurea GHS-R1a inverse agonists. J Med Chem. 2014;57:6128–6140.
https://doi.org/10.1021/jm500610n.
[7]. (a) Bobileva O, Bokaldere R, Gailite V, et al. Synthesis and evaluation of (E)-2-
(acrylamido)cyclohex-1-enecarboxylic acid derivatives as HCA1, HCA2, and HCA3
receptor agonists. Bioorg Med Chem. 2014;22:3654–3669. https://doi.org/10.1016/
j.bmc.2014.05.011
[11]. Geyer M, Baus JA, Fjellström O, Wellner E, Gustafsson L, Tacke R. Synthesis and
pharmacological properties of silicon-containing GPR81 and GPR109A agonists.
ChemMedChem. 2015;10:2063–2070. https://doi.org/10.1002/cmdc.201500343.
[12]. Allen FH. The Cambridge structural database: a quarter of a million crystal struc-
tures and rising. Acta Cryst. 2002;B58:380–388. https://doi.org/10.1107/
s0108768102003890 The Identifier BAQJUY and FUHQEF where used in the
search.
(b) Liu C, Kuei C, Zhu J, et al. 3,5-Dihydroxybenzoic acid, a specific agonist for
hydroxycarboxylic acid 1, inhibits lipolysis in adipocytes. J Pharmacol Exp Ther.
2012;341:794–801. https://doi.org/10.1124/jpet.112.192799
(c) Dvorak CA, Liu C, Shelton J, et al. Identification of Hydroxybenzoic Acids as
Selective Lactate Receptor (GPR81) Agonists with Antilipolytic Effects. ACS Med
Chem Lett. 2012;3:637–639. https://doi.org/10.1021/ml3000676.
[8]. (a) Sakurai T, Davenport R, Stafford S, et al. Identification of a novel GRP81-se-
lective agonist that suppresses lipolysis in mice without cutaneous flushing. Eur J
Pharm. 2014;727:1–7. https://doi.org/10.1016/j.ejphar.2014.01.029
(b) Wallenius K, Thalén P, Björkman J-A, et al. Involvement of the metabolic
[13].. Corina, version 3.60. Maestro, version 9.8, Schrödinger, LLC, New York, NY, 2014.
3D Conformations where generated using Corina and a mixed torsional/Low-mode.
Sampling conformational search was performed within Maestro (OPLS2.1, energy
window 54kJ/mol, in water). The initial minimal energy conformation did not display the
expected intramolecular hydrogen bond, which was, however, present within the con-
formational ensemble with a relative energy difference to the minimal conformation up to
6.5 kJ/mol.
8