Identification of hydroxybenzoic acids as selective lactate receptor (GPR81) agonists with antilipolytic effects

Article abstract of DOI:10.1021/ml3000676

Following the characterization of the lactate receptor (GPR81), a focused screening effort afforded 3-hydroxybenzoic acid 1 as a weak agonist of both GPR81 and GPR109a (niacin receptor). An examination of structurally similar arylhydroxy acids led to the identification of 3-chloro-5-hydroxybenzoic acid 2, a selective GPR81 agonist that exhibited favorable in vivo effects on lipolysis in a mouse model of obesity.

Full text of DOI:10.1021/ml3000676

Letter
pubs.acs.org/acsmedchemlett
Identification of Hydroxybenzoic Acids as Selective Lactate Receptor
(GPR81) Agonists with Antilipolytic Effects
Curt A. Dvorak,* Changlu Liu, Jonathan Shelton, Chester Kuei, Steven W. Sutton,
Timothy W. Lovenberg, and Nicholas I. Carruthers
Janssen Research & Development, LLC, San Diego, California 92121, United States
S
* Supporting Information
ABSTRACT: Following the characterization of the lactate receptor (GPR81), a focused screening
effort afforded 3-hydroxybenzoic acid 1 as a weak agonist of both GPR81 and GPR109a (niacin
receptor). An examination of structurally similar arylhydroxy acids led to the identification of 3-
chloro-5-hydroxybenzoic acid 2, a selective GPR81 agonist that exhibited favorable in vivo effects
on lipolysis in a mouse model of obesity.
KEYWORDS: lactate receptor, GPR81, HCA1, nicotinic acid receptor, GPR109a, agonist, lipolysis, flushing, FFA
PR81 (HCA1) is a member of a subfamily of recently
deorphanized G-protein-coupled receptors (GPCRs)
anchor point for the carboxylic acid moiety of the endogenous
ligand.8 As recently reported, this led to the identification of 3-
hydroxybenzoic acid (1; Chart 1), which exhibited moderate
G
consisting of GPR81, GPR109a (HM74A), and GPR109b
(HM74).1 All three receptors are predominantly expressed on
adipocytes and are activated by the endogenous ligands 2-
hydroxypropanoate (lactate), 3-hydroxybutyrate, and 3-hydrox-
yoctanoate, respectively. These GPCRs may be considered to
be a family of hydroxy-carboxylic acid receptors and in turn
share homology with the orphan receptor GPR31, the 5-oxo-
6,8,11,14-eicosatetraenoic acid (5-oxo-ETE) receptor OXER1,
the succinate receptor (GPR91), and the α-ketoglutarate
receptor (GPR99). This family of receptors all exhibits a
highly conserved arginine residue in transmembrane helix 3
that presumably interacts with the corresponding endogenous
ligands.
Stimulation of GPR81, GPR109a, or GPR109b in adipose
tissue leads to a decrease in the lipolysis of triglycerides by
hormone-sensitive lipase, which results in reduced transport of
fatty acids to the liver and a decrease in hepatic triglycerides.2,3
The beneficial effects of niacin (3) result from its interaction
with GPR109a, leading to a host of beneficial outcomes
including the lowering of triglycerides (35−45%), increasing
HDL-C (30−40%), and lowering LDL-C. However, patient
compliance is generally limited due to niacin's target related
side effects, which include pruritus and flushing, that are
attributed to a GPR109a-dependent production of PDG2 and
PDE2 by immune cells.4,5 The development of a small molecule
agonist of GPR81, expressed predominantly on adipocytes,
with selectivity over other GPCRs, may provide a novel
treatment for dyslipidemia without the undesirable side effects
associated with the stimulation of skin immune cells.6
Consequently, there have been reports of 3H-imidazo[4,5-
b]pyridin-5-ol derivatives as GPR81 agonists but with unknown
selectivity.7 To this end, we embarked on the search for small
molecule GPR81 agonists that were selective over GPR109a.
We chose to evaluate several hundred low molecular weight
organic acids for GPR81 functional activity given the conserved
arginine residue in transmembrane helix 3, which provides an
Chart 1
activation of both GPR81 and GPR109a.9 The presence of a 3-
hydroxyl is essential for GPR81 activity, and thus, removal
eliminates activity (e.g., benzoic acid), and relocation to the 2-
or 4-position on the aromatic ring eliminates activity at both of
these receptors. However, in contrast to a report describing
phenolic acids capable of activating GPR109a, we were pleased
to discover that the addition of a second hydroxyl in the
orientation as found in 3,5-dihydroxybenzoic acid (4) was
indeed selective for GPR81, albeit slightly less potent than 1.10
The other dihydroxybenzoic acid isomers were determined to
be inactive at GPR81 (vida supra).
These observations prompted an examination of the SAR
around 4, which is the subject of this current report. The
objectives were to improve affinity at GPR81 while maintaining
selectivity over GPR109a and ultimately determine the effects
on adipocyte lipolysis in vivo with a compound that provided
suitable exposure after oral administration. Two approaches
were pursued; the first was to scan a range of substituted 3-
hydroxybenzoic acids to determine the effects of replacing one
of the hydroxyl groups, and the second was to evaluate
common carboxylic acid isosteres. These structures are
summarized in Table 1 and Figure 1, respectively. The
Received: March 23, 2012
Accepted: June 11, 2012
Published: June 11, 2012
© 2012 American Chemical Society
637
dx.doi.org/10.1021/ml3000676 | ACS Med. Chem. Lett. 2012, 3, 637−639

ACS Medicinal Chemistry Letters
Letter
condensation of 3,5-dihydroxybenzonitrile and Bu3SnN3 in 1,4-
dioxane followed by aqueous hydrochloric acid. The N-oxide of
5-hydroxynicotinic acid (6) was obtained by treatment of 5-
hydroxynicotinic acid (5) with 30% H2O2 in glacial acetic acid.
An examination of the SAR around 4 with holding the 3-
hydroxybenzoic acid core constant is shown in Table 1. The
introduction of a hydroxyl onto the nicotinic acid structure to
afford 5 significantly reduces GPR109a activity (EC50 = 368
μM) without imparting any GPR81 activity. The corresponding
N-oxide (6), which places the oxygen atoms in relatively the
same orientation as 4, was found to have no activity at either
receptor. Capping one of the hydroxyl groups with a methyl
group thereby giving the methyl ether 7 was not tolerated and
led to a loss of activity. However, the replacement of the
hydroxyl with a simple methyl group, as in 8, exhibited
moderate GPR81 activity at ∼42 μM. Simple hydrophobic
groups of varying sizes were then explored. The larger tert-butyl
and aryl set in compounds 9 and 10 resulted in a loss of
potency at GPR81. The more lipophilic trifluoromethyl, 11,
and the more polar nitrile 12 were also not tolerated.
Gratifyingly, the halogenated series of compounds, 13, 2, and
14 all exhibited a significant increase of GPR81 activity over
that of 3,5-dihydroxybenzoic acid (4). The chloro derivative, 3-
chloro-5-hydroxybenzoic acid (2), was determined to have the
best potency with an EC50 ∼ 16 μM. All of the compounds in
Table 1 were also counter screened against GPR109a and
found to be inactive at the concentrations tested (EC50 > 1
mM). An exploration of carboxylic acid replacements was less
fruitful with compounds 15−18, found to be inactive at both
GPR81 and GPR109a. The lack of potency seen for 18 was
somewhat surprising given the occurrence of tetrazoles as a
carboxylic acid replacement for GPR109a agonists.11 Com-
pound 2 was then selected for further profiling and examined in
more detail.
Table 1. Potency of 3-Hydroxy-5-Substituted Benzoic Acids
in the hGPR81 Assay
a
compd
R
EC50 (μM)
5
N−
N(O)−
OH
>1000
>1000
6
4
377 88
>1000
7
OCH3
CH3
tBu
8
42 16
>1000
9
10
11
12
13
2
Ph
>1000
CF3
CN
>1000
>1000
Br
129 26
Cl
16
38
9
1
14
15−18
F
>1000
GTPγS binding assays were performed as described previously.12
a
Results are reported as an average of at least two determinations with
standard deviations.
Figure 1. Carboxylic acid isosteres.
compounds evaluated were commercially available with the
exception of those whose syntheses are depicted in Scheme 1.
The phosphonic acid 16 and phosphinic acid 17 were both
prepared from a common bromide precursor 19. Palladium-
catalyzed cross-coupling of either diethyl phosphite or ethyl
methylphosphinate with 1-bromo-3,5-dimethoxybenzene (19)
provided the corresponding intermediates 20 and 21 in good
yield. The methyl ethers were then cleaved with BBr3 followed
by hydrolysis of the ethoxy group(s), thus providing 16 and 17,
respectively. The tetrazole analogue 18 was prepared by the
The activity of 3-chloro-5-hydroxybenzoic acid (2) at GPR81
was well conserved across all relevant preclinical species tested
except the dog−human (EC50 = 16 μM), monkey (EC50 = 17
μM), dog (EC50 = 67 μM), rat (EC50 = 7 μM), mouse (EC50
=
22 μM), and hamster (EC50 = 27 μM). The pharmacokinetic
data for compound 2 were obtained to establish the exposure
upon oral dosing in mice (Table 2). The 30 mg/kg po dose
provided a Cmax of 67 μM at approximately 3-fold over the in
a
Scheme 1. Syntheses of Compounds 6, 16−18, 20 and 21
a
(i) Pd(PPh3)4, Et3N, toluene Δ; HP(O)OEt2 (R = OEt) or HP(O)OEt(CH3) (R = CH3). (ii) BBr3, −78 °C, CH2Cl2. (iii) 6 M HCl, 100 °C. (iv)
Bu3SnN3, 1,4-dioxane, 100 °C 24 h; 1 M HCl, 16 h. (v) 30% H2O2, AcOH.
638
dx.doi.org/10.1021/ml3000676 | ACS Med. Chem. Lett. 2012, 3, 637−639

ACS Medicinal Chemistry Letters
Letter
Table 2. Pharmacokinetic Parameters of 3-Chloro-5-
hydroxybenzoic Acid (2) in Mice
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures for assay protocols, in vivo studies,
and the synthesis and characterization of compounds 6, 16−18,
20 and 21. This material is available free of charge via the
Internet at http://pubs.acs.org.
dose
tmax
AUCinf
Cmax
Cmax
t1/2
(mg/kg)
(h)
(h ng/mL)
(ng/mL)
(μM)
(h)
30
0.5
0.5
9356
11689.50
55252.80
67.2
1.47
1.10
100
51312
341.9
AUTHOR INFORMATION
Corresponding Author
*Tel: 858-784-3284. E-mail: cdvorak@its.jnj.com.
Notes
vitro EC50 with an AUCinf of 54 h μM and a short t1/2 of 1.5 h,
and the 100 mg/kg po dose attained a Cmax of 342 μM at
greater than 15-fold over the observed in vitro EC50.
Compound 2 was then tested acutely by measuring
nonesterified free fatty acid cholesterol (NEFAc) as a marker
for lipolysis in vivo at 30, 100, and 300 mg/kg po in overnight
fasted C57Bl6/J mice fed high fat chow for 10 weeks (DIO
mice). Shown in Figure 2, it was determined that compound 2
■
The authors declare no competing financial interest.
REFERENCES
■
(1) Ahmed, K.; Tunaru, S.; Offermanns, S. GPR109A, GPR109B and
GPR81, a family of hydroxy-carboxylic acid receptors. Trends
Pharmacol. Sci. 2009, 30, 557−562.
(2) Tunaru, S.; Kero, J.; Schaub, A.; Wufka, C.; Blaukat, A.; Pfeffer,
K.; Offermanns, S. PUMA-G and HM74 are receptors for nicotinic
acid and mediate its anti-lipolytic effect. Nat. Med. (N. Y., NY, U. S.)
2003, 9, 352−355.
(3) Wise, A.; Foord, S. M.; Fraser, N. J.; Barnes, A. A.; Elshourbagy,
N.; Eilert, M.; Ignar, D. M.; Murdock, P. R.; Steplewski, K.; Green, A.;
Brown, A. J.; Dowell, S. J.; Szekeres, P. G.; Hassall, D. G.; Marshall, F.
H.; Wilson, S.; Pike, N. B. Molecular identification of high and low
affinity receptors for nicotinic acid. J. Biol. Chem. 2003, 278, 9869−
9874.
(4) Benyo, Z.; Gille, A.; Kero, J.; Csiky, M.; Suchankova, M. C.;
Nuesing, R. M.; Moers, A.; Pfeffer, K.; Offermanns, S. GPR109A
(PUMA-G/HM74A) mediates nicotinic acid-induced flushing. J. Clin.
Invest. 2005, 115, 3634−3640.
(5) Benyo, Z.; Gille, A.; Bennett, C. L.; Clausen, B. E.; Offermanns, S.
Nicotinic acid-induced flushing is mediated by activation of epidermal
langerhans cells. Mol. Pharmacol. 2006, 70, 1844−1849.
(6) Liu, C.; Wu, J.; Zhu, J.; Kuei, C.; Yu, J.; Shelton, J.; Sutton, S. W.;
Li, X.; Yun, S. J.; Mirzadegan, T.; Mazur, C.; Kamme, F. Lactate
inhibits lipolysis in fat cells through activation of an orphan G-protein-
coupled receptor, GPR81. J. Biol. Chem. 2009, 284 (5), 2811−2822.
(7) Boatman, P. D.; Johnson, B. R.; Jung, J.-K.; Kasem, M.; Schrader,
T. O.; Skinner, P. J.; Colletti, S. L. Preparation of 3H-imidazo[4,5-
b]pyridin-5-ol derivatives useful in the treatment of GPR81 receptor
disorders. WO2010030360A1.
Figure 2. Nonesterified free fatty acid cholesterol (NEFAc) following
treatment with niacin at 30 or compound 2 at 30, 100, and 300 mg/kg
po in overnight fasted C57Bl6/J mice fed high fat chow for 10 weeks
(DIO mice). *P < 0.05 vs vehicle and **P < 0.01 vs vehicle.
(8) Kuei, C.; Yu, J.; Zhu, J.; Wu, J.; Zhang, L.; Shih, A.; Mirzadegan,
T.; Lovenberg, T.; Liu, C. Study of GPR81, the lactate receptor, from
distant species identifies residues and motifs critical for GPR81
functions. Mol. Pharmacol. 2011, 80, 848−858.
(9) Liu, C.; Kuei, C.; Zhu, J.; Yu, J.; Zhang, L.; Shih, A.; Mirzadegan,
T.; Shelton, J.; Sutton, S.; Connelly, M. A.; Lee, G.; Carruthers, N.;
Wu, J.; Lovenberg, T. W. 3,5-Dihydroxybenzoic Acid, a Specific
Agonist for HCA1, Inhibits Lipolysis in Adipocytes. J. Pharmacol. Exp.
Ther. 2012, 341, 794−801.
(10) Ren, N.; Kaplan, R.; Hernandez, M.; Cheng, K.; Jin, L.; Taggart,
A. K. P.; Zhu, A. Y.; Gan, X.; Wright, S. D.; Cai, T.-Q. Phenolic acids
suppress adipocyte lipolysis via activation of the nicotinic acid receptor
GPR109A (HM74a/PUMA-G). J. Lipid Res. 2009, 50, 908−914.
(11) Boatman, P. D.; Schrader, T. O.; Kasem, M.; Johnson, B. R.;
Skinner, P. J.; Jung, J.-K.; Xu, J.; Cherrier, M. C.; Webb, P. J.; Semple,
G.; Sage, C. R.; Knudsen, J.; Chen, R.; Taggart, A. K.; Carballo-Jane,
E.; Richman, J. G. Potent tricyclic pyrazole tetrazole agonists of the
nicotinic acid receptor (GPR109a). Bioorg. Med. Chem. Lett. 2010, 20,
2797−2800.
(12) Liu, C.; Eriste, E.; Sutton, S.; Chen, J.; Roland, B.; Kuei, C.;
Farmer, N.; Joernvall, H.; Sillard, R.; Lovenberg, T. W. Identification
of Relaxin-3/INSL7 as an Endogenous Ligand for the Orphan G-
protein-coupled Receptor GPCR135. J. Biol. Chem. 2003, 278, 50754−
50764.
gave significant reductions in NEFAc at all doses tested. This
resulted in compound 2 having a minimum efficacious dose
equivalent to that of niacin for the reduction of observed
NEFAc. The potential differentiating factor would be that
compound 2 should not cause the flushing that is associated
with niacin due to a lack of GPR109a cross-reactivity.
Comparing the pharmacokinetic data for compound 2 and
the observed pharmacodynamics at the two lower doses of the
lipolysis study demonstrate a good correlation between
potency, exposure, and in vivo efficacy. The reduction in free
fatty acids plateaus at the higher doses examined.
In conclusion, a series of phenolic benzoic acids were
determined to have activity at GPR81 with selectivity over
GPR109a. An examination of the SAR around the 3-
hydroxybenzoic acid afforded the discovery of 3-chloro-5-
hydroxybenzoic acid as a potent and selective GPR81 agonist.
The compound was examined in vivo for effects on lipolysis,
and it was found that 3-chloro-5-hydroxybenzoic acid gave
significant reductions in free fatty acids at all doses tested,
resulting in a minimum efficacious dose of 30 mg/kg. Further
investigations of GPR81 agonists are ongoing and will be the
subject of future publications.
639
dx.doi.org/10.1021/ml3000676 | ACS Med. Chem. Lett. 2012, 3, 637−639