Discovery of the imidazole-derived GPR40 agonist AM-3189

Article abstract of DOI:10.1016/j.bmcl.2015.11.050

As a follow-up to the GPR40 agonist AMG 837, which was evaluated in clinical trials for the treatment of type II diabetes, further optimization led to the discovery of AM-3189 (13k). AM-3189 is representative of a new class of compounds with minimal CNS penetration, superior pharmacokinetic properties and in vivo efficacy comparable to AMG 837.

Full text of DOI:10.1016/j.bmcl.2015.11.050

Bioorganic & Medicinal Chemistry Letters 26 (2016) 15–20
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of the imidazole-derived GPR40 agonist AM-3189
b
a
a
a
a
a
Zhihua Ma a, , Daniel C.-H. Lin , Rajiv Sharma , Jinqian Liu , Liusheng Zhu , An-Rong Li , Todd Kohn ,
Yingcai Wang a, Jiwen (Jim) Liu a, Michael D. Bartberger a, Julio C. Medina a, Run Zhuang b, Frank Li b,
Jane Zhang b, Jian Luo b, Simon Wong c, George R. Tonn c, Jonathan B. Houze a
⇑
a Department of Therapeutic Discovery, Amgen Inc., 1120 Veterans Boulevard, South San Francisco, CA 94080, USA
b Department of Metabolic Disorders, Amgen Inc., 1120 Veterans Boulevard, South San Francisco, CA 94080, USA
c Department of Pharmacokinetics & Drug Metabolism, Amgen Inc., 1120 Veterans Boulevard, South San Francisco, CA 94080, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
As a follow-up to the GPR40 agonist AMG 837, which was evaluated in clinical trials for the treatment of
type II diabetes, further optimization led to the discovery of AM-3189 (13k). AM-3189 is representative
of a new class of compounds with minimal CNS penetration, superior pharmacokinetic properties and
in vivo efficacy comparable to AMG 837.
Received 10 August 2015
Revised 9 November 2015
Accepted 16 November 2015
Available online 17 November 2015
Ó 2015 Elsevier Ltd. All rights reserved.
Keywords:
GPR40
FFAR1
GPCR
Agonist
Type II diabetes
Insulin secretagogue
AMG 837
AM-3189
Type 2 diabetes mellitus (T2DM) is a disease characterized by
defects in insulin secretion from pancreatic b-cells and/or insulin
resistance in target tissues of insulin.1–3 Insulin secretagogues,
such as sulfonylureas and glinides are commonly used to stimulate
insulin secretion in diabetic patients.4 However, these drugs pro-
mote insulin secretion independent of blood glucose levels,
thereby leading to the risk of hypoglycemia.4,5 GPR40 (also known
as FFAR1) is a novel G protein-coupled receptor (GPCR) that is
expressed in pancreatic b-cells and responds to free-fatty acid
(FFA) concentrations.6 Activation of GPR40 potentiates glucose-
stimulated insulin secretion and lowers plasma glucose concentra-
tions in multiple animal models of insulin resistance and obesity.6–8
Because GPR40 mediated insulin secretion is glucose-dependent, it
is believed that pharmacologic activation of the receptor should
not induce hypoglycemia in either fed or fasted states. Conse-
quently, a GPR40 agonist has potential to be a safe and effective
alternative to currently available therapies for T2DM.9–11 Thera-
peutic efficacy has been demonstrated in several clinical trials tar-
geting the GPR40 pathway, for example, LY-2881835 (entered
Phase I clinical trial),11 JTT-851 (entered Phase II clinical trial,
structure not disclosed)11 and TAK-875 (entered Phase III clinical
trial).12
O
OH
O
N
LY-2881835
O
O2
OH
O
S
O
O
TAK-875
We previously described the development of the GPR40 agonist
AMG 837 from a series of beta-substituted propionic acids, which
were identified in a high-throughput screen.13 In multiple animal
models, AMG 837 enhances glucose-stimulated insulin secretion
and lowers plasma glucose levels.14 Because GPR40 activity is
glucose-dependent, AMG 837 did not induce hypoglycemia in any
⇑
Corresponding author.
http://dx.doi.org/10.1016/j.bmcl.2015.11.050
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

16
Z. Ma et al. / Bioorg. Med. Chem. Lett. 26 (2016) 15–20
of the models tested.14 Encouraged by these results, we initiated a
Phase I clinical trial of AMG 837.
groups to the tail group and/or the head group of the AMG 837 class
of GPR40 agonists.
Herein, we report the discovery of AM-3189 (13k). AM-3189
maintains the in vivo efficacy of AMG 837 while displaying a supe-
rior pharmacokinetic profile and minimal CNS exposure.
O
As we reported previously, activity on the GPR40 receptor var-
ies significantly with substitution at the b-carbon relative to the
carboxylate.13 We hoped to modify the beta-substitution of AMG
837 in order to improve potency and desired physiochemical prop-
erties. The racemate of AMG 837 (8a in the Table 1) displayed the
expected twofold decrease in potency on GPR40 compared to the
active enantiomer AMG 837 and its activity crossed over to the
rat and mouse forms of GPR40.13 Due to the lack of an efficient
method for the asymmetric synthesis of b-substituted 2-phenyl-
propanoates at the time, racemic compounds were assayed for
structure–activity relationship (SAR) studies and early screening
purposes. Using the tail group from AMG 837 as a probe, the syn-
thetic route for modification of the beta-substitutions of the car-
boxylic acids is shown in Scheme 1. Treatment of aldehydes 1
with (4-((tetrahydro-2H-pyran-2-yl)oxy)phenyl) magnesium bro-
mide afforded secondary alcohols 2, which were oxidized by pyri-
dinium chlorochromate (PCC) to yield ketones 3. The ketones 3
F3C
OH
O
tail group
head group
AMG 837
While the clinical evaluation of AMG 837 was ongoing, we turned
our attention to the development of a structurally distinct GPR40
agonist. Although AMG 837 is a carboxylic acid, its physicochemical
properties, including low polar surface area (tPSA 47, Table 1),
suggest a reasonable possibility of achieving CNS exposure. Addi-
tional support comes from a structurally close analog of AMG 837
(with 40-chloro-20-ethoxy-(1,10-biphenyl)-4-yl replacing 40-(trifluo-
romethyl)-(1,10-biphenyl)-3-yl of AMG 837) which showed a brain
to plasma ratio of 0.6 3 h after an oral dose of 5 mg/kg in rats. Given
that the efficacy of GPR40 agonists is derived peripherally and that
they are likely to be dosed chronically, we sought molecules with
minimal brain penetration.15 In general, increasing the polar surface
area (PSA) of a molecule tends to decrease its blood–brain barrier
permeability.16 In keeping with this principle, we focused on
increasing the PSA of AMG 837, while maintaining its potency
and metabolic stability. The approach taken was to introduce polar
were converted to a,b-unsaturated esters 4 through Peterson-ole-
fination17 followed by hydrogenation to afford beta-substituted 3-
arylpropanoates 5. After removing the THP protecting group with
TFA, the resulting para-substituted phenols 6 were alkylated with
3-(bromomethyl)-40-(trifluoromethyl)-1,10-biphenyl in the pres-
ence of cesium carbonate to afford the esters 7, which were hydro-
lyzed under basic conditions to yield the desired carboxylic acids 8.
Since GPR40 is a G q-coupled GPCR, compounds were assayed
a
for GPR40 activity in a chemiluminescent system (aequorin) mon-
itoring calcium flux in CHO cells transiently transfected with
GPR40 as reported previously (Table 1).13,14 In the later stage of
the project, a CHO cell line stably transfected with human GPR40
was developed. GPR40 aequorin assay was then run in buffer con-
taining 0.01% human serum albumin or in 100% human serum to
test activity of compounds in the absence or presence of plasma
proteins (Table 2). AMG 837 was included with all assays as a pos-
itive control and as a reference compound for benchmarking SAR.
The SAR of the b-substituted 3-phenylpropanoate lead series
with AMG 837 tail group is shown in Table 1. Activity on the
GPR40 receptor varied significantly with substitution at the b-car-
bon of the carboxylate. Simple phenyl (8b) reduces GPR40 potency
10-fold relative to the methyl acetylene compound (8a, racemate
of AMG 837). Introducing a nitrogen to the phenyl ring in various
positions does not restore the lost potency, thus pyridin-2-yl and
pyridin-3-yl compounds (8c, 8d) are equally potent to the simple
phenyl analog (8b), while the pyridin-4-yl compound (8e) is
slightly more potent. Five-membered ring heterocycles retain
potency slightly more than the phenyl analogs, as evidenced by
2-thiophenyl (8f) compared to the simple phenyl (8b). Replacing
the thiophene ring with a thiazole ring or 2-methylimidazole ring
also showed slightly improved potency (8f vs. 8g and 8h).
In an attempt to quantify the likely improvement in the balance
between physiochemical properties and potency, the lipophilic
efficiency (LipE)18 of this set of compounds was calculated and
shown in Table 1. 8h was revealed to be the best of replacements
in terms of lipophilic efficiency among the set shown in Table 1
and it was comparable to that of 8a (racemate of AMG 837). By
incorporating both acidic and basic functional groups, analogs
derived from 8h should be less likely to penetrate the BBB. There-
fore, the imidazole arylpropionate was used as the scaffold for the
next set of compounds. In designing our analogs, emphasis was
placed on increasing polarity and conformational rigidity. We
had established in previous work13 that meta-biaryls yielded the
Table 1
SAR of the b-substituted 2-phenylpropanoate lead series with AMG 837 tail group
R1
b
O
F3C
OH
a
O
C ring
A ring
B ring
Head Group
AMG 837 Tail Group
Compounda
R1
hGPR40b
EC50 (lM)
LogD (pH 7.4)
LipEc
tPSA (Å2)
AMG 837
0.013
0.025
3.66
3.66
4.23
3.94
47
47
8a
8b
8c
0.33
0.33
4.84
3.12
1.64
3.36
47
59
N
S
8d
0.40
3.17
3.23
59
N
N
8e
8f
0.19
0.19
3.13
4.42
3.59
3.30
59
47
S
N
8g
8h
0.12
0.13
3.79
3.01
3.13
3.89
59
62
N
N
a
All compounds reported in this table are racemates except AMG 837.
GPR40 activity was assayed in a chemiluminescent system (aequorin) moni-
b
toring calcium flux in CHO cells transiently transfected with GPR40 as reported
previously13,14
.
c
Lipophilic efficiency, LipE = pEC50 – LogD.

Z. Ma et al. / Bioorg. Med. Chem. Lett. 26 (2016) 15–20
17
O
R
O
O
OH
OEt
OEt
c
O
a
b
d
R
R
R
R
H
THP
THP
R
THP
THP
O
O
O
O
O
O
O
1
2
3
4
5
R
R
O
f
g
e
F3C
F3C
OEt
OH
OEt
O
HO
6
7
8
Scheme 1. General chemistry approach for modification of the beta-substitution of AMG 837. Reagents and conditions: (a) (4-((tetrahydro-2H-pyran-2-yl)oxy)
phenyl)magnesium bromide, THF, ꢀ78 to ꢀ20 °C, 70–90%; (b) PCC, CH2Cl2, 23 °C, 75–90%; (c) TMSCH2CO2Et, LiHMDS, THF, ꢀ78–23 °C, 60–85%; (d) H2, Pd/C, EtOAc, 23 °C,
90–100%; (e). TFA, CH2Cl2, 23 °C, 80–100%; (f) (1) chiral separation when needed, hexane/i-PrOH, OD or AD column, 40–50%; (2) 3-(bromomethyl)-40-(trifluoromethyl)-1,10-
biphenyl, Cs2CO3, DMF, 23 °C; (g) LiOH, THF/water (3/1), 23 °C, 80–99%.
Table 2
Tail group SAR of the imidazole derived analogs
N
N
O
C ring
B ring
2
OH
b
R3
a
O
A ring
13
6
4
R2
5
tail group
head group
a
a
Compound
R2
R3
hGPR40 Aeq EC50 in cLogD (pH 7.4) LipEb tPSA (Å2) hGPR40 Aeq EC50 in serum (
l
M)
buffer (lM)
AMG 837
13a
13b
13c
13d
13e
13f
13g
13h
N/A
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
N/A
2-MeSO2
2-CN
2-MeO
3-MeO
4-MeO
2-CF3O
2-Me
0.080
23
0.69
0.41
0.75
3.66
0.32
1.3
1.716
1.81
1.86
2.74
2.5
2.52
3.03
2.96
3.07
3.07
2.44
2.52
3.1
3.43
4.32
4.86
4.67
4.31
4.50
4.21
4.87
4.85
4.29
4.28
4.41
4.14
4.79
4.92
4.06
3.11
3.50
47
96
86
71
71
71
71
62
62
62
62
62
62
62
62
62
62
62
8.8
ND
ND
ND
ND
ND
61
89
56
65
29
10
22
21
42
30
23
9.8
0.44
0.112
0.043
0.043
0.048
2-Cl
2-Et
13i
13j
2-Me, 6-Me 0.058
13k (AM-3189)
13l
13m
13n
13o
13p
2-Me, 4-Cl
2-Me, 3-Cl
2-Me, 4-F
2-Me, 5-F
2-Cl, 4-Cl
0.033
0.061
0.059
0.036
0.069
0.24
2-Me 2-Me, 4-Cl
4-Me 2-Me, 4-Cl
3.51
3.51
13q
0.097
a
GPR40 aequorin assay in CHO cells stably transfected with human GPR40. Assay was run in buffer containing 0.01% human serum albumin or
in 100% human serum.
b
Lipophilic efficiency, LipE = pEC50 – LogD.
most potent molecules so we excluded ortho- and para-biaryls
from our designs.
ethers 11, which were converted to compounds 12 using Suzuki
cross coupling. The desired carboxylic acids 13 were obtained by
hydrolysis of esters 12 under basic conditions. The ee value of com-
pound 10 was found to be preserved during these transformations
as all final products 13 had ee% > 98%.
Tail group SAR of the imidazole derived analogs is shown in
Table 2. The previous SAR study from the discovery of AMG 837
revealed that a simple un-substituted meta-biphenyl tail group
was less favorable in terms of potency.13 Efforts to introduce polar-
ity to the tail group were not successful. When a methylsulfonyl,
cyano or methoxy group were added to the different positions of
the C ring, all yielded various level of diminished activity (e.g.,
13a–13e). This was not entirely unexpected. As endogenous
GPR40 ligands are FFAs, the ligand binding site in the receptor
is hydrophobic in nature.21 It was encouraging to find that a
The synthesis of meta-biaryl analogs of 8h is depicted in
Scheme 2. Compound 9 was synthesized as a racemic mixture,
using the commercially available 1-methyl-1H-imidazole-2-car-
baldehyde with the procedure described in Scheme 1. The sepa-
rated single enantiomer (compound 10), which yielded more
potent compounds, was advanced through the synthesis. The abso-
lute configuration of the compound was later determined by com-
parison of experimental and calculated vibrational circular
dichroism (VCD) spectra19 and optical rotations20 to be in the
S-configuration, which is consistent with the finding from the
discovery of AMG 837.13 The enantiomerically pure material 10
(ee > 98%) was alkylated with substituted 3-bromobenzyl bro-
mides or chlorides in the presence of cesium carbonate to afford

18
Z. Ma et al. / Bioorg. Med. Chem. Lett. 26 (2016) 15–20
N
N
N
N
N
N
O
O
O
b
a
OEt
OEt
OEt
Br
O
HO
HO
N
9
10
11
R2
N
N
N
O
O
d
c
OH
OEt
R3
R3
O
O
12
13
R2
R2
Scheme 2. Chemistry approach for modification of the tail groups of the imidazole head group derived analogs. Reagents and conditions: (a) chiral separation, hexane/i-PrOH,
the later-eluting enantiomer on AD or AD-H column, 48%; (b) substituted benzyl halide, Cs2CO3, DMF, 23 °C, 90–100%; (c) Pd(OAc)2, S-Phos, K3PO4, dioxane, 100 °C, 80–95%;
(d) LiOH, THF/water (1/3), 23 °C, 90–99%.
Table 3
more potent than AMG 837. Adding one more methyl at the other
ortho-position of the C ring or increasing the size of the substitu-
Pharmacokinetic properties of compound 13k
a
Property
Mouse (n = 2)a Rat (n = 2)
Dog (n = 3)
Cyno (n = 3)
tion relative to the methyl group does not further improve potency
(13i and 13j vs. 13g). With the encouraging results from com-
pounds 13g and 13h, the combination of 2-methyl with a methyl,
chloro or fluoro were screened (13k–13o). Compound 13k (AM-
3189) stands out by displaying greater potency both in buffer
and in 100% human serum. 13k (AM-3189) also possesses favor-
able drug-like properties with a calculated logD (pH 7.4) of 3.01,
tPSA of 64 Å2 and more favorable lipophilic efficiency (4.4 com-
pared to 3.4 for AMG 837). Adding a methyl group to the ortho-
positions of the B ring proximal to the C ring of 13k, targeting
restricted rotation of the rotable bond between the B ring and C
ring to reduce possible conformers, yielded less potent compounds
13p and 13q.
i.v.
Dose (mg/kg)b
Cl (L/h/kg)
T1/2 (h)
1
0.5
0.5
0.5
0.17
1.5
0.78
0.31
1.6
0.73
0.15 0.040 0.27 0.030
5.1 3.8 7.6 0.78
0.58 0.079 1.6 0.13
Vdss (L/kg)
p.o.
Dose (mg/kg)c
5
2
2
5
%F
ca. 100
4.89
43
0.77
66 0.14
4.40 0.67
34 0.13
1.19 0.33
Cmax (lM)
a
Plasma concentration data was determined from two mice at each time point.
IV formulations were prepared in 10% N,N-dimethylacetamide, 10% ethanol,
b
30% propylene glycol and 50% sterile water.
c
Oral formulations were prepared as suspensions using an aqueous vehicle
containing 1% Tween 80 in 1% methylcellulose.
All tested compounds in Table 2 have similar or lower Emax than
that of AMG 837, which indicates that, like AMG 837, these imida-
zole-derived compounds are partial agonists on the GPR40 receptor.
To confirm the activity difference between a pair of enan-
tiomers that was observed previously,13 the opposite enantiomer
of 13k (AM-3189) was also synthesized and was found to be
>100ꢁ less active in the GPR40 aequorin assay.
Table 4
P-glycoprotein transport properties for compound 13k determined in MDCK–MDR1
cells
Efflux
ratio
Permeability, apical to
basal (10–6 cm/s)
Permeability, basal to
apical, (10ꢀ6 cm/s)
Similar to AMG 837, while highly potent on GPR40, 13k (AM-
3189) was highly selective over the closely related GPCRs, GPR41
and GPR43. 13k showed no significant activity in cell-based assays
13k (AM- 22
3189)
1.4
31
against PPAR-a, -d, and -c, despite a possible structural resem-
blance to some PPAR agonists. An external panel of 64 receptors
(MDS Pharma) also revealed no significant activity. Overall, 13k
was both highly potent and selective in vitro.
In addition to its favorable profile in vitro, compound 13k
(AM-3189) distinguished itself by displaying an excellent pharma-
cokinetic profile in multiple species. As shown in Table 3, 13k
demonstrated low clearance, moderate volume of distribution,
trifluoromethoxy group at the ortho-position of the C ring gave an
imidazole head group compound with similar potency as AMG 837
(13f vs. AMG 837). However, the activity of 13f is significantly
reduced in the presence of 100% human serum. A breakthrough
was revealed when a methyl or chlorine was introduced at the
ortho-position of the C ring (13g, 13h), yielding compounds 2-fold
Figure 1. Stimulation of insulin secretion in human islet and human GPR40 knock-in mouse islets by compound 13k (AM-3189) and AMG 837.

Z. Ma et al. / Bioorg. Med. Chem. Lett. 26 (2016) 15–20
19
Figure 2. Effect of compound 13k (AM-3189) during OGTT in human islet transplanted mice and hGPR40 knock-in mice.
and good oral bioavailability. 13k (AM-3189) does not penetrate
Supplementary data
the rat CNS as indicated by a rat brain to plasma ratio of 0.04 at
3 h after an oral dose of 5 mg/kg. The reduced CNS penetration
may be due to P-glycoprotein mediated efflux as shown in Table 4.
The activity of 13k (AM-3189) was examined in in vitro insulin
secretion assays, first using human islets. As was observed in
recombinant cell systems, activity of 13k and AMG 837 was similar
in this assay (Fig. 1A). We also evaluated the activity of 13k in
islets isolated from human GPR40 knock-in mice.22 In these mice,
the mouse GPR40 gene has been replaced with the human GPR40
gene using genetic recombination technology. 13k increased glu-
cose stimulated insulin secretion from human GPR40 knock-in
islets (Fig. 1B).
13k (AM-3189) was significantly less potent and efficacious on
the rodent GPR40 receptor compared to the human receptor. In the
presence of 0.01% HSA, the EC50 value of 13k on rat GPR40 was
640 60 nM and the Emax (% of AMG 837) value was 41%. Similar
weak activity was seen on receptors from other non-primate pre-
clinical species, including mouse, guinea pig, hamster, and pig
(data not shown). 13k (AM-3189) was a potent agonist of dog
GPR40 (EC50 = 37 nM) but had low efficacy compared to AMG
837 (Emax = 37% of AMG 837). 13k (AM-3189) had good activity
against the cynomolgus monkey receptor, with EC50 and Emax
(% of AMG 837) values of 29 nM and 96%, respectively.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.11.
050.
References and notes
1. Taylor, S. I.; Accili, D.; Imai, Y. Diabetes 1994, 43, 735.
2. Groop, L. C.; Widen, E.; Ferrannini, E. Diabetologia 1993, 36, 1326.
3. Wajchenberg, B. L. Endocr. Rev. 2007, 28, 187.
4. Krentz, A. J.; Bailey, C. J. Drugs 2005, 65, 385.
5. Groop, L. Int. J. Clin. Pract. Suppl. 2000, 113, 3.
6. Itoh, Y.; Kawamata, Y.; Harada, M.; Kobayashi, M.; Fujii, R.; Fukusumi, S.; Ogi,
K.; Hosoya, M.; Tanaka, Y.; Uejima, H.; Tanaka, H.; Maruyama, M.; Satoh, R.;
Okubo, S.; Kizawa, H.; Komatsu, H.; Matsumura, F.; Noguchi, Y.; Shinohara, T.;
Hinuma, S.; Fujisawa, Y.; Fujino, M. Nature 2003, 422, 173.
7. Briscoe, C. P.; Tadayyon, M.; Andrews, J. L.; Benson, W. G.; Chambers, J. K.;
Eilert, M. M.; Eillis, C.; Elshourbagy, N. A.; Goetz, A. S.; Minnick, D. T.; Murdock,
P. R.; Sauls, H. R., Jr.; Shabon, U.; Spinage, L. D.; Strum, J. C.; Szekeres, P. G.; Tan,
K. B.; Way, J. M.; Ignar, D. M.; Wilson, S.; Muir, A. I. J. Biol. Chem. 2003, 278,
11303.
8. Kotarsky, K.; Nilsson, N. E.; Flodgren, E.; Owman, C.; Olde, B. Biochem. Biophys.
Res. Commun. 2003, 301, 406.
9. Medina, J. C.; Houze, J. B. In Annual Reports in Medicinal Chemistry; Macor, J. E.,
Ed.; Academic Press: San Diego, 2008; Vol. 43, pp 75–85.
10. Houze, J. B. In New Therapeutic Strategies for Type 2 Diabetes: Small Molecule
Approaches; Jones, R. M., Ed.; The Royal Society of Chemistry, 2012; pp 403–
415.
11. Poitout, V.; Lin, D. C.-H. Drug Discovery Today 2013, 18, 1301.
12. Takeda news release: <http://www.takeda.com/press/article_43205.html>.
13. Houze, J. B.; Zhu, L.; Sun, Y.; Akerman, M.; Qiu, W.; Zhang, A.; Sharma, R.;
Schmitt, M.; Wang, Y.; Liu, J.; Liu, J.; Medina, J. C.; Reagan, J. D.; Luo, J.; Tonn, G.;
Zhang, J.; Lu, J.; Chen, M.; Lopez, E.; Nguyen, K.; Yang, L.; Tang, L.; Tian, H.;
Shuttleworth, S.; Lin, D. Bioorg. Med. Chem. Lett. 2011, 22, 1267.
14. Lin, D. C.-H.; Zhang, J.; Zhuang, R.; Li, F.; Nguyen, K.; Chen, M.; Tran, T.; Lopez,
E.; Lu, J. Y. L.; Li, X. N.; Tang, L.; Tonn, G. R.; Swaminath, G.; Reagan, J. D.; Chen,
J.-L.; Tian, H.; Lin, Y.-J.; Houze, J. B.; Luo, J. PLoS One 2011, 6, e27270.
15. Bagal, S. K.; Bungay, P. J. ACS Med. Chem. Lett. 2012, 3, 948.
16. Hitchcock, S. A.; Pennington, L. D. J. Med. Chem. 2006, 49, 7559.
17. Peterson, D. J. J. Org. Chem. 1968, 33, 780.
Because 13k (AM-3189) had limited efficacy on rodent GPR40,
we used two rodent models expressing human GPR40 to demon-
strate efficacy of 13k (AM-3189). First, we used human islet trans-
planted nude mice. In this model, endogenous pancreatic b-cells
were ablated using streptozotocin, and mice were rescued to eug-
lycemia using human islets transplanted under the kidney cap-
sule.23 In these mice, 13k (AM-8139) dosed at 1 and 10 mg/kg po
prior to a glucose challenge lowered glucose AUC levels (Fig. 2A).
The glucose lowering effect was comparable to that of a 100 lg/kg
18. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
19. Stephens, P. J.; Devlin, F. J.; Pan, J.-J. Chirality 2008, 20, 643.
20. Stephens, P. J.; Devlin, F. J.; Cheeseman, J. R.; Frisch, M. J.; Rosini, C. Org. Lett.
2002, 4, 4595.
dose of GLP-1 (Fig. 2A). The efficacy of 13k was also examined in
human GPR40 knock-in mice.24 Again, 13k (30 mg/kg po) lowered
glucose levels during an OGTT (Fig. 2B andC).
In summary, further optimization of beta-substituted
carboxylic acids as GPR40 agonists led to the discovery of 13k
(AM-3189), a potent GPR40 agonist that has good pre-clinical
pharmacokinetic properties and low CNS penetration. 13k
(AM-3189) represents a valuable tool in the exploration of the role
that the GPR40 receptor may play in type II diabetes.
21. Srivastava, A.; Yano, J.; Hirozane, Y.; Kefala, G.; Gruswitz, F.; Snell, G.; Lane, W.;
Ivetac, A.; Aertgeerts, K.; Nguyen, J.; Jennings, A.; Okada, K. Nature 2014, 513,
124.
22. Insulin secretion was assessed using islets isolated from human GPR40 knock-
in mice and was performed as described in Ref. 14. Assays were performed in
buffer containing 16.7 mM glucose and 0.1% human serum albumin.
23. Luo, J.; Nguyen, K.; Chen, M.; Tran, T.; Hao, J.; Tian, B.; Rulifson, I. C.; Zhang, Y.;
Tian, L.; Zhang, Y.; Lopez, E.; Lin, D. C.-H.; Wang, Y.; Ma, Z.; Houze, J.; Guo, Z.
Metabolism 2013, 62, 90.
24. Human GPR40 knock-in animals were generated at Ozgene (Perth, Australia)
by replacing the mouse GPR40 gene (a single exon gene) with the human
GPR40 gene. Briefly, the targeting vector was created by PCR amplifying two
fragments of genomic DNA (from C57Bl/6 mice) corresponding to 5637 bp
upstream and 5126 bp downstream of mouse GPR40 and inserting these into
the plasmid L-Sniper (Ozgene). Human GPR40 was PCR amplified and fused in
between these fragments along with the selectable neomycin resistance
marker driven the PGK promoter. The targeting vector was electroporated into
Acknowledgments
Zhihua Ma thanks Margaret Chu-Moyer, Paul Dransfield,
Richard V. Connors and Larry McGee for their very helpful com-
ments on the manuscript and numerous helpful discussions.

20
Z. Ma et al. / Bioorg. Med. Chem. Lett. 26 (2016) 15–20
ES cells, selected for by culturing in neomycin, and successful insertion of the
performed on female human GPR40 knock-in mice at 28 weeks of age. Mice
were dosed via oral gavage at 10 mg/kg of compound 13k following a 6 h fast.
Compound 13k was formulated in 1% methylcellulose, 1% Tween-80. Glucose
was administered by oral gavage at 2 g/kg 1-hour post drug dose. Blood
glucose measurements were taken from tail vein samples at various time
points using AlphaTRAK (Abbott) blood glucose monitoring system.
human GPR40 was verified by PCR of genomic DNA. Following ES cell injection
into blastocysts, generation of chimeric mice and germline transmission of the
knockin allele, human GPR40 knock-in mice were created and confirmed by
PCR of genomic DNA. mRNA expression of human GPR40, and lack of
expression of mouse GPR40, was confirmed by qPCR using RNA from islets
isolated from human GPR40 knock-in mice. Oral glucose tolerance tests were