Discovery of 3-aryl-4-isoxazolecarboxamides as TGR5 receptor agonists

Article abstract of DOI:10.1021/jm901434t

A series of 3-aryl-4-isoxazolecarboxamides identified from a high-throughput screening campaign as novel, potent small molecule agonists of the human TGR5 G-protein coupled receptor is described. Subsequent optimization resulted in the rapid identificatio

Full text of DOI:10.1021/jm901434t

7962 J. Med. Chem. 2009, 52, 7962–7965
DOI: 10.1021/jm901434t
Discovery of 3-Aryl-4-isoxazolecarboxamides
as TGR5 Receptor Agonists
Karen A. Evans,*,† Brian W. Budzik,† Sean A. Ross,
David D. Wisnoski,† Jian Jin,† Ralph A. Rivero,†
Mythily Vimal,‡ George R. Szewczyk,
Channa Jayawickreme,§ David L. Moncol,§
Thomas J. Rimele,§ Susan L. Armour,§ Susan P. Weaver,§
Robert J. Griffin, Sarva M. Tadepalli, Michael R. Jeune,
Todd W. Shearer, Zibin B. Chen, Lihong Chen,
Donald L. Anderson, J. David Becherer,
Figure 1. In vitro profile of HTS hit 1.
Maite De Los Frailes,^ and Francisco Javier Colilla^
monocytes and macrophages, lung, spleen, and the intestinal
tract. In response to bile acids, namely, lithocolic acid (LCAa),
TGR5 has been shown to cause a dose dependent elevation in
intracellular concentrations of cAMP in cells that express the
receptor.3 Bile acids and compounds that affect TGR5 have
also been shown to increase glucagon like peptide-1 (GLP-1)
secretion from primary intestinal cells.4 It has therefore been
suggested that bile acids induce GLP-1 secretion by increasing
intracellular cAMP levels via the TGR5 receptor.5
†Discovery Medicinal Chemistry, Discovery Research,
GlaxoSmithKline Pharmaceuticals, 1250 South Collegeville Road,
P.O. Box 5089, Collegeville, Pennsylvania 19426-0989,
‡Discovery Medicinal Chemistry, Discovery Research,
GlaxoSmithKline Pharmaceuticals, New Frontiers Science
Park, Third Avenue, Harlow, Essex, CMI9 5AD, U.K.,
§Discovery Research, Metabolic Diseases Center of Excellence for
Drug Discovery, Five Moore Drive, P.O. Box 13398, Research
Triangle Park, North Carolina 27709-3398, and ^Department of
Screening and Compound Profiling, Discovery Research,
GlaxoSmithKline Pharmaceuticals, C/Santiago Grisolia 4,
28760 Tres Cantos, Madrid, Spain
GLP-1 is a peptide secreted from enteroendocrine L cells
and has a wide variety of physiological effects.6 Recent FDA
approved therapies for type II diabetes based on GLP-1 are
GLP-1 receptor agonist peptides (GLP-1 mimetics) and DPP-
IV inhibitors, which enhance insulin action by prolonging the
half-life of endogenous GLP-1.7 These approaches are effec-
tive because while type II diabetics show impaired GLP-1
secretion, they maintain normal responsiveness to GLP-1.
However, the use of a peptide in clinical treatment is severely
limited because of the requirement for parenteral administra-
tion and low in vivo stability. Therefore, a small molecule
agonist that increases GLP-1 secretion from the gut during a
meal is a potentially useful therapeutic for metabolic disorders
such as type II diabetes and its associated complications. A
number of bile acid derivatives and triterpenoid TGR5 ago-
nists have been disclosed recently in the peer-reviewed litera-
ture.8 In addition, a variety of small molecule agonists are also
known but are currently described only in the patent litera-
ture.9,10 Herein is presented the discovery of a series of 3-aryl-
4-isoxazolecarboxamides as novel, potent small molecule
agonists of the human TGR5 G-protein-coupled receptor.
High-throughput screening (HTS) of the GSK proprietary
compound collection using a MRE/CRE (multiple response
element/cAMP response element) directedreporter gene assay
measuring luciferase production in response to changes in
cAMP (via a Gs-protein-coupled signaling pathway)11 in a
BacMam transduced human osteosarcoma cell line (U2-OS)
led to the identification of isoxazole 1 (Figure 1) as an agonist
with a pEC50 of 5.3 and 100% maximum response.12 The
Received September 26, 2009
Abstract: A series of 3-aryl-4-isoxazolecarboxamides identified
from a high-throughput screening campaign as novel, potent small
molecule agonists of the human TGR5 G-protein coupled receptor
is described. Subsequent optimization resulted in the rapid identi-
fication of potent exemplars 6 and 7 which demonstrated improved
GLP-1 secretion in vivo via an intracolonic dose coadministered
with glucose challenge in a canine model. These novel TGR5
receptor agonists are potentially useful therapeutics for metabolic
disorders such as type II diabetes and its associated complications.
Type II diabetes, also known as diabetes mellitus, is now
internationally recognized as one of the major threats to
human health in the 21st century. According to the Interna-
tional Diabetes Federation (IDF), diabetes is expected to
cause 3.8 million deaths worldwide in 2007, roughly 6% of
total world mortality, about the same as HIV/AIDS and
malaria combined.1 Those that suffer from type II diabetes
have too little insulin or cannot use insulin effectively. As a
result, glucose levels build up in the blood and urine and, if left
untreated, can cause life-threatening complications, including
blindness, kidney failure, and heart disease.
Current oral therapies include insulin secretagogues, such
as sulfonylureas; glucose-lowering effectors, such as metfor-
min; activators of the peroxisome proliferator-activated re-
ceptor-γ (PPAR-γ), such as the thiazolidinediones; and
R-glucosidase inhibitors.2 There are, however, deficiencies
associatedwithcurrentlyavailable treatmentsincluding hypo-
glycemic episodes, weight gain, gastrointestinal problems,
edema, and loss of responsiveness over time.
compound elicited no response in the host cell line (pEC50
<
4.6 in untransfected U2-OS cells) which confirmed its speci-
ficity. In addition, the compound has an excellent physico-
chemical profile that made it of particular interest as a starting
TGR5, also known as BG37, M-BAR, or hGPCR19, is a
bile acid G-protein-coupled receptor primarily expressed in
aAbbreviations: 7TM, seven-transmembrane; BA, bile acid; LCA,
lithocholic acid; GPCR, G-protein-coupled receptor; FXR, farnesoid X
receptor; T2D, type II diabetes; HTS, high-throughput screening;
GLP-1, glucagon-like peptide 1; CYP450, cytochrome P450; PK, phar-
macokinetic; MDCK, Madin-Darby canine kidney.
*To whom correspondence should be addressed. Phone: 1-610-917-
7764. Fax: 1-610-917-6151. E-mail: karen.a.evans@gsk.com.
r
pubs.acs.org/jmc
Published on Web 11/10/2009
2009 American Chemical Society

Letter
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 7963
Table 1. Profile of Isoxazolecarboxamides 6 and 7
Scheme 1. Synthesis of 3-Aryl-4-isoxazolecarboxamidesa
a (a) Various substituted N-methylanilines, Et3N, DCM, room temp;
(b) various substituted anilines, Et3N, DCM, room temp; (c) CH3I,
powdered KOH, DMSO, 50 ꢀC.
compd 6
melanophore
compd 7
melanophore
(human,
rat,
canine)
(human,
rat,
canine)
U2-OS
(human)
U2-OS
(human)
point for optimization: low molecular weight (MW < 400),
good aqueous solubility (134 uM), and high permeability in
both artificial membrane13 (310 nm/s) and MDCK cells14
(passive Papp = 448 nm/s).
To explore this novel HTS hit, an efficient one- or two-step
synthesis was utilized to explore substitutions at R1 and R2
(Scheme 1). Commercially available 5-methyl-3-aryl-4-isoxa-
zolecarbonyl chlorides 2 were converted to the desired amides
3 in one step from commercially available, or prepared,
substituted N-alkylanilines.15 Alternatively, the acid chloride
was coupled with an aryl-substituted primary aniline and then
alkylated in a second step.
In Vitro Potencya
TGR5 pEC50
6.8
7.5
5.7
7.2
4.7
6.3
7.1
6.2
7.5
4.5
host
specificity
pEC50
<4.3
<4.3
Rat in Vivo PK Parameters
T
1/2 = 6.6 h b
T1/2 = 8.7 h c
Cl = 148 mL/min/kg
Cl = 85 mL/min/kg
ss = 7.1 L/kg
F = 1%
Our lead optimization work rapidly led to the discovery
of more potent compounds. Compound 6 (R1 = 4-Cl, R2 =
2-Cl) showed improved potency in the U2-OS cell assay
(pEC50 = 6.8) and in melanophore cells (pEC50 = 7.5).16
Likewise, 7 (R1 = 3-Cl-4-Me, R2 = 2-Cl) showed analogous
potency (U2-OS cells pEC50 = 6.3; melanophore cells
pEC50 = 7.1). Compounds 6 and 7 were therefore further
profiled in vitro and in vivo (Table 1) to quickly ascertain
whether a small molecule agonist from this series would have
utility in improving GLP-1 secretion and thereby improving
glucose disposal. These exemplars were fully specific for the
human TGR5 receptor in the U2-OS and melanophore cell
lines and also highly selective against other 7TM receptors.
Compound 6 was profiled againstmorethan100 in-houseand
external 7TM, ion channel, enzyme, transporter, and nuclear
hormone receptor selectivity assays, including FXR, another
bile acid receptor, and showed significant response only
in secretion of the pro-inflammatory cytokine TNFalpha
(pIC50 = 6.8) in human primary monocytes following stimu-
lation with LPS (lipopolysaccharide).17 This result is not
unexpected, as the TGR5 receptor is expressed in monocytes.
In addition, 6 and 7 have good physicochemical properties
and no measurable activity against three of the common
cytochrome P450 (CYP450) isoforms (1A2, 2C9, and 2D6)
or hERG dofetilide binding (pIC50 <4.3). In rat pharmaco-
kinetic (PK) studies, however, both compounds showed high
in vivo clearance (6, Cl = 85 mL/min/kg; 7, Cl = 148 mL/
min/kg) and intrinsic clearance (6, Clint = 48 mL/min/g; 7,
Clint = 54 mL/min/g) which provided a reasonable explana-
tion for the observed poor exposure. Because the TGR5
receptor is expressed in the GI tract at levels that increase
corresponding with L-cell population density, we believe
that agonists such as 6 and 7 possessing poor systemic
exposure are good tool compounds for directly targeting
the TGR5 receptor in the GI tract via local administration
(vide infra) rather than systemic exposure. Our hypothesis
was thatforthis receptor, systemic exposure was notnecessary
to achieve the desired effect of stimulating GLP-1 secretion
in vivo.
V
Vss = 40 L/kg
F < 1%
Dog in Vivo PK Parametersd
not determined
T1/2 = 5.2 h
Cl = 39 mL/min/kg
Vss = 8.6 L/kg
F = 0.3%
In Vitro PK Parameterse
rat
dog
Clint = 48 mL/min/g
Clint = 90 mL/min/g
Clint = 54 mL/min/g
Clint = 111 mL/min/g
CYP450 pIC50
6.5 (2C19) 5.9 (3A4),
<5.0 (1A2,
2C9 and 2D6)
6.5 (2C19) 5.9 (3A4),
<5.0 (1A2,
2C9 and 2D6)
Permeability
422 nm/s
257 nm/s
Solubility
246 μM
112 μM
hERG Dofetilide Binding
pIC50 <4.3 pIC50 <4.3
a Standard deviation is less than (0.7. b Rat PK parameters: n = 4,
male CD rats; iv, 2.5 mg/kg; po, 5 mg/kg; vehicle formulation
5% DMSO/2% EtOH/20% SBECD. c Rat PK parameters: n = 3; iv,
3 mg/kg, formulated in 5% DMSO/10% Solutol/85% 25 mM citrate
buffer (pH 4); po, 30 mg/kg, formulated in 0.5% HPMC/0.1% Tween-
80. d Dog PK parameters: n = 3; iv, 3 mg/kg, formulated in 5% DMSO/
10% Solutol/85% saline; po, 30 mg/kg, formulated in 0.5% HPMC/
0.1% Tween-80. e Liver hepatocytes preparation.
conscious canine in vivo model was found to be more robust
than an anesthetized rat in vivo model for measurement of
GLP-1 levels. Moreover, the in vitro melanophore data
(Table 1) suggested that the compounds were a full log unit
less potent at the rat receptor but were equipotent at the
human and canine receptors. The dog PK also showed a very
similar profile to the observed rat PK, in vivo and in vitro
for 7, with reasonable half-life, high clearance, and poor oral
exposure. Therefore, in the conscious dog model, an intraje-
jeunal injection of glucose (0.125 g/kg) was coadministered
To test this hypothesis, 6 and 7 were studied in an acute
conscious dog animal model to evaluate GLP-1 secretion. The

7964 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Evans et al.
complications. A full account of the synthesis and structure-
activity relationships of this novel series will be the subject of a
future publication.
Acknowledgment. The authors thank Walter Johnson,
Chad Quinn, and Bill Leavens for their assistance with the
LCMS spectra and high resolution mass spectra respectively,
Victoria Magaard and Carl Bennett for their assistance with
compound purification, and Katrina L. Creech and Ray V.
Merrihew for the FXR screening data.
Figure 2. Effect of 6 and 7 on GLP-1 secretion in vivo via an
intrajejeunal dose (1 mg/kg) coadministered with a glucose chal-
lenge at t = 0 (0.125 g/kg) in conscious dogs. Vehicle formulation is
0.5% HPMC/0.1% Tween-80. Data from time points at 5 and
60 min are statistically significant in their difference from vehicle
control.
Supporting Information Available: Synthetic procedures and
characterization data for all compounds; procedures for pri-
mary assays and in vivo canine model. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) IDF Diabetes Atlas, 3rd ed. http://www.diabetesatlas.org/.
(2) Ashiya, M.; Smith, R. E. T. Non-insulin therapies for type 2
diabetes. Nat. Rev. Drug Discovery 2007, 6, 777–778.
(3) Maruyama, T.; Miyamoto, Y.; Nakamura, T.; Tamai, Y.; Okada,
H.; Sugiyama, E.; Nakamura, T.; Itadani, H.; Tanaka, K. Identi-
fication of membrane-type receptor for bile acids (M-BAR).
Biochem. Biophys. Res. Commun. 2002, 298, 714–719.
(4) Thomas, C.; Pellicciari, R.; Pruzanski, M.; Auwerx, J.; Schoonjans,
K. Targeting bile-acid signalling for metabolic diseases. Nat. Rev.
Drug Discovery 2008, 7, 678–693.
(5) Katsuma, S.; Hirasawa, A.; Tsujimoto, G. Bile acids promote
glucogon-like peptide-1 secretion through TGR5 in a murine
enteroendocrine cell line STC-1. Biochem. Biophys. Res. Commun.
2005, 329, 386–390.
(6) (a) Bojanowska, E. Physiology and pathophysiology of glucagon-
like peptide-1 (GLP-1): the role of GLP-1 in the pathogenesis of
diabetes mellitus, obesity, and stress. Med. Sci. Monit. 2005, 11 (8),
RA271–RA278. (b) Perry, T.; Greig, N. H. Enhancing central nervous
system endogenous GLP-1 receptor pathways for intervention in
Alzheimer's disease. Curr. Alzheimer Res. 2005, 2, 377–85. (c) Meier,
J. J.; Nauck, M. A. Glucagon-like peptide 1(GLP-1) in biology and
pathology. Diabetes Metab. Res. Rev. 2005, 21, 91–1 17.
Figure 3. Effect of 6 and 7 on portal vein glucose levels.
Table 2. Dog Exposure Levels for Compounds 6 and 7a
vein
sampling
location
plasma
concn Cmax
(ng/mL)
AUC0-2h
(ng h/
hEC50
(ng/mL)
3
mL/mg/kg)
compd
6
portal
(local)
cephalic
514 ( 256
22 ( 37
852 ( 814
9 ( 8
142 ( 80
11
11
30
30
(7) Arulmozhi, D. K.; Portha, B. GLP-1 based therapy for type 2
diabetes. Eur. J. Pharm. Sci. 2006, 28, 96–108.
6
7
7
ND
(8) Sato, H.; Macchiarulo, A.; Thomas, C.; Gioiello, A.; Une, M.;
Hofmann, A. F.; Saladin, R.; Schoonjans, K.; Pellicciari, R.;
Auwerx, J. Novel potent and selective bile acid derivatives as
TGR5 agonists: biological screening, structure-activity relation-
ships, and molecular modeling studies. J. Med. Chem. 2008, 51,
1831–1841.
(systemic)
portal
(local)
cephalic
(systemic)
a ND = not detected. Standard deviations are noted.
226 ( 210
ND
(9) (a) Pinkerton, A. B.; Kabakibi, A.; Herbert, M. R.; Siegel, D. L.
Preparation of Quinoline Compounds as Modulators of TGR5 for
Treatment of Disease. WO 2008/097976, 2008. (b) Pinkerton, A. B.;
Kabakibi, A.; Hoffman, T. Z.; Siegel, D. L.; Noble, S. A. Preparation of
Quinazolinone Modulators of TGR5 for Treatment of Metabolic,
Cardiovascular and Inflammatory Diseases. WO 2008/067219, 2008.
(c) Pinkerton, A. B.; Kabakibi, A.; Gahman, T. C. Preparation of
Diazepines and Other Heterocyclic Modulators of TGR5 for Treating
Metabolic, Cardiovascular, And Inflammatory Diseases. WO 2008/
067222, 2008. (d) Szewczyk, J. R.; Laudeman, C. P.; Evans, K. A.; Li,
Y. H.; Dock, S. T.; Chen, Z. Preparation of Bis(arylsulfonyl)piperazine
and Related Dinitrogen Moieties as Antidiabetic Agents. WO 2007/
127505 A2, 2007. (e) Ito, F.; Hard, N.; Kobayashi, H.; Kanzaki, N.
Preparation of N-Heterocyclic Compounds as G Protein-Coupled
Receptor (TGR5) Agonists. JP 2006/063064, 2006. (f) Ito, F.; Tawar-
aishi, T.; Hirohashi, M.; Matsumoto, H. Preparation of Fused Pyridine
Compounds as TGR5 Receptor Agonists. JP 2006/056881, 2006.
(10) Tiwari, A.; Maiti, P. TGR5: an emerging bile acid G-protein
coupled receptor target for the potential treatment of metabolic
disorders. Drug Discovery Today 2009, 14, 523–530.
(11) Fitzgerald, L. R.; Mannan, I. J.; Dytko, G. M.; Wu, H. L.; Nambi,
P. Measurement of responses from Gi, Gs or Gq coupled receptors
by a multiple response element/cAMP response element-directed
reporter assay. Anal. Biochem. 1999, 275, 54–61.
(12) Bological assay results are the mean of at least two determinations
with standard deviation of less than (0.7 unless otherwise noted.
(13) The artificial phospholipid membrane technique is similar to the
widely used Caco-2 cell monolayer permeation technique. In short,
egg phosphatidyl choline (1.8%) and cholesterol (1%) are dissolved
in n-decane. A small amount of the volatile mixture is applied to the
bottom of the microfiltration filter inserts. Phosphate buffer
(0.05 M, pH 7.05) is quickly added to the donors and receivers,
with compound (1 mg/kg) using a nanomilled suspension in
0.5% HPMC/0.1% Tween-80, and portal and cephalic vein
blood samples were taken at 5, 15, 30, 60, and 120 min. The
results of this experiment clearly demonstrated that 6 and 7
showed significant improvements in hepatic portal vein
GLP-1 secretion (Figure 2) and reduction in portal vein
glucose levels (Figure 3) compared to vehicle for the entire
duration of the study.
Compound exposure levels at Cmax from the portal vein
were 50-fold and 30-fold over hEC50 for 6 and 7, respectively
(Table 2). Significant differences were observed in samples
taken from the cephalic vein, however, with minimal plasma
exposures observed, suggesting that systemic exposure is not
required for acute efficacy, and possibly a local intracolonic
mechanism for TGR5 stimulated GLP-1 release.
In conclusion, novel and highly selective small molecule
TGR5 receptor agonists have been discovered. Isoxazole
carboxamides 6 and 7 have excellent physicochemical proper-
ties, and their in vivo efficacy results in the acute canine model
via intracolonic administration demonstrate that these novel
TGR5receptor agonists are potentially useful therapeutics for
metabolic disorders such as type II diabetes and its associated

Letter
and the lipids are allowed to form self-assembled lipid bilayers
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 7965
dride (Na(OAc)3BH), and acetic acid in dichloroethane (DCE)
for 16 h at room temperature. See also the following: Thorstens-
son, F.; Kvarnstrom, I.; Musil, D.; Nilsson, I.; Samuelsson, B.
Synthesis of novel thrombin inhibitors. Use of ring-closing me-
tathesis reactions for synthesis of P2 cyclopentene- and cyclohex-
enedicarboxylic acid derivatives. J. Med. Chem. 2003, 46, 1165–
1179.
across the small holes in the filter. Permeation experiment is
initiated by spiking the compounds of interest to the donor sides,
and the experiment is stopped at a predetermined elapsed time. The
samples are withdrawn and transferred to appropriate vials for
analysis by HPLC with UV detection (215 nm).
(14) (a) Tran, T. T.; Mittal, A.; Gales, T.; Maleeff, B.; Aldinger, T.;
Polli, J. W.; Ayrton, A.; Ellens, H.; Bentz, J. Exact kinetic analysis
of passive transport across a polarized confluent MDCK cell
monolayer modelled as a single barrier. J. Pharm. Sci. 2004, 93,
2108–2123. (b) Polli, J. W.; Wring, S. A.; Humphreys, J. E.; Huang, L.;
Morgan, J. B.; Webster, L. O.; Serabjit-Singh, C. S. Rational use of
in vitro P-glycoprotein assays in drug discovery. J. Pharmacol. Exp.
Ther. 2001, 299, 620–626.
(16) All lead optimization compounds (post-HTS) were tested in two
TGR5 assays using both U2-OS cells and melanophore cells, with
specificity screening in the corresponding host cell line. Experi-
mental protocols are described in the Supporting Information.
(17) Sample profilingresultsbytargetclass. (a) 7TM:5-HT1A, pEC50
<
5.5; M1, pEC50 < 4.8. (b) Ion channel: Ca2þ channel (verapamil
site), <10% inhiþbition at 1 μM; K þ V channel, <10% inhibi-
tion at 1 μM; Na channel, 16% inhibition at 1 μM. (c) Nuclear
receptors: FXR, pEC50 < 5.
(15) Typical reaction conditions are treatment of the aniline with the
appropriate aldehyde or ketone (R), sodium triacetoxyborohy-