Design, synthesis and structure?activity relationship studies of GPR40 agonists containing amide linker

Article abstract of DOI:10.1016/j.ejmech.2018.04.023

Free fatty acid receptor 1 (FFAR1/GPR40) attracted significant attention as a potential target for developing novel antidiabetic drugs because of its unique mechanism in glucose homeostasis. Several reports have expressed concerns about central nervous system (CNS) penetration of GPR40 agonists, which is possibly attributed to their high lipophilicity and low total polar surface area. Herein, we report our efforts to improve the physicochemical properties and pharmacokinetic profiles of LY2881835, a GPR40 agonist that had undergone Phase I clinical trial, through a series of structural optimizations. We identified an orally efficacious compound, 15k, which possessed increased plasma exposure, prolonged half-life and reduced CNS exposure and liver to plasma distribution ratio compared with LY2881835. 15k is a potentially valuable lead compound in the development of safe and efficacious GPR40-targeted drugs to treat type 2 diabetes mellitus.

Full text of DOI:10.1016/j.ejmech.2018.04.023

Accepted Manuscript
Design, synthesis and structure−activity relationship studies of GPR40 agonists
containing amide linker
Tingting Chen, Mengmeng Ning, Yangliang Ye, Kai Wang, Ying Leng, Jianhua Shen
PII:
S0223-5234(18)30357-X
10.1016/j.ejmech.2018.04.023
EJMECH 10372
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 31 January 2018
Revised Date: 22 March 2018
Accepted Date: 10 April 2018
Please cite this article as: T. Chen, M. Ning, Y. Ye, K. Wang, Y. Leng, J. Shen, Design, synthesis and
structure−activity relationship studies of GPR40 agonists containing amide linker, European Journal of
Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.04.023.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Design, Synthesis and Structure−activity Relationship Studies of GPR40
Agonists Containing Amide Linker
§
,#,†
#,†
#
#
#,
Tingting Chen , Mengmeng Ning , Yangliang Ye , Kai Wang , Ying Leng *,
#,
Jianhua Shen *
#
State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica (SIMM), Chinese
Academy of Sciences, 555 Zuchongzhi Road, Shanghai 201203, China
§
University of Chinese Academy of Sciences, No.19A Yuquan Road, Beijing 100049, China
ABSTRACT: Free fatty acid receptor 1 (FFAR1/GPR40) attracted significant
attention as a potential target for developing novel antidiabetic drugs because of its
unique mechanism in glucose homeostasis. Several reports have expressed concerns
about central nervous system (CNS) penetration of GPR40 agonists, which is possibly
attributed to their high lipophilicity and low total polar surface area. Herein, we report
our efforts to improve the physicochemical properties and pharmacokinetic profiles of
LY2881835, a GPR40 agonist that had undergone Phase I clinical trial, through a
series of structural optimizations. We identified an orally efficacious compound, 15k,
which possessed increased plasma exposure, prolonged half-life and reduced CNS
exposure and liver to plasma distribution ratio compared with LY2881835. 15k is a
potentially valuable lead compound in the development of safe and efficacious
GPR40-targeted drugs to treat type 2 diabetes mellitus.
Keywords: GPR40, FFAR1, insulin secretion, type 2 diabetes mellitus, CNS
1

ACCEPTED MANUSCRIPT
1
. Introduction
Type 2 diabetes mellitus (T2DM) is a metabolic disorder disease characterized
by both defective insulin secretion and insulin action in maintaining glucose
homeostasis [1]. T2DM often leads to multiple macrovascular and microvascular
complications, including cardiovascular diseases, nephropathy, neuropathy and
retinopathy. [2-4]. The incidence of T2DM is increasing at an alarming rate. By 2017,
the number of individuals affected worldwide was about 425 million and this figure
may reach 629 million in 2045 [5]. This severe situation, along with the undesirable
side effects caused by current hypoglycemic agents, such as high risk of
hypoglycemia, body weight gain and gastric symptoms [6-8], stimulate urgent needs
for novel drugs with improved safety and glycemic control.
Free fatty acid receptor 1 (FFAR1, also known as GPR40) belongs to the family
of G protein-coupled receptors (GPCRs) and was de-orphanized in 2003 as a receptor
for medium- to long-chain free fatty acids (FFAs) [9, 10]. Mainly expressed in beta
cells of pancreatic islets, GPR40 can be activated by the endogenous FFAs, which
subsequently elicit amplified glucose-dependent insulin secretion without high risk of
hypoglycemia [9, 11-15]. GPR40 is also expressed in the incretin-producing
enteroendocrine L cells of the intestinal tract and its activation results in secretion of
incretins such as glucagon like peptide 1 (GLP-1) [16, 17]. Therefore, the ability of
GPR40 to regulate glucose homeostasis by two mechanisms, glucose-stimulated
insulin secretion and incretin secretion, makes it an excellent target for T2DM drug
2

ACCEPTED MANUSCRIPT
development [18, 19].
Reports described a range of synthetic GPR40 agonists with desirable safety
and efficacy profiles, compared with existing antidiabetic therapies such as insulin
and sulfonylureas [20-28]. Among them, several small molecules were tested in the
clinical trial such as TAK-875 [20, 29], AMG 837 [21, 30], LY2922083 and
LY2881835 [23, 31] (Figure 1). The most advanced compound TAK-875 (Fasiglifam,
1
) from Takeda, while providing proof of concept, was withdrawn from phase III trial
due to liver toxicity [32]. The exact reason for its hepatotoxicity is currently unclear,
but some reports indicated that it may be linked with high distribution of TAK-875
and TAK-875 conjugations (TAK-875 acyl glucuronide, TAK-875 glucuronide and
TAK-875 taurine conjugate) in liver, as well as their ability to alter the bile acid
homeostasis [33-35]. Moreover, GPR40 is highly expressed in the brain and can be
activated by polyunsaturated FFAs [9]. The exact function of GPR40 in the brain
remains unclear, but it might have correlation with antinociception, adult
neurogenesis and neurovascular degeneration [36-41]. Therefore, to balance the blood
glucose levels, it would be more preferable to directly target the GPR40 in the
pancreatic beta cells and enteroendocrine L cells rather than to elicit the undesired
effects in the central nervous system (CNS) or liver.
In general, properties such as total polar surface area (tPSA), LogP, molecular
weight, hydrogen bond, pKa and rigidity of the molecule are closely linked to
blood-brain barrier (BBB) penetration [42]. Most reported GPR40 agonists possess
3

ACCEPTED MANUSCRIPT
relatively high lipophilicity [43-45]. Amgen described an analogue of AMG 837
(
LogP = 6.8, tPSA = 47) with significant CNS exposure (total brain-to-plasma drug
distribution ratio, B/P = 0.6). Incorporation of polar groups, as in AMG-3189 (B/P =
.04) and AMG-4668 (B/P = 0.02) (Figure 1), substantially decreased BBB
0
penetration [46, 47]. LY2881835 (2), which finished Phase I clinical trial in 2011,
possessed high in vitro potency and selectivity, as well as significant insulin and
GLP-1 secretion in both in vitro and in vivo assays [23, 31]. However, 2 (LogP = 6.5,
tPSA = 50), like AMG-837, had a certain amount of CNS exposure (B/P = 0.14). In
addition, 2 displayed high clearance in vitro (t1/2 = 18 min in human liver microsomes,
t1/2 = 2 min in mouse liver microsomes) and short half-life in vivo (t1/2 = 1.3 h). We
sought to modify the pharmacokinetic (PK) and physicochemical properties of 2 by
introducing polar atoms, blocking the metabolically susceptible position and cleaving
the indene ring. After a systematic structure-activity relationship (SAR) study and
further optimization, we identified an orally active compound 15k, which exhibited
improved metabolic stability and high plasma exposure, along with lowered CNS
exposure and liver to plasma distribution ratio (L/P).
4

ACCEPTED MANUSCRIPT
a
a
Figure 1. Selected examples of reported GPR40 agonists. tPSA and LogP values
were calculated from ChemBioDraw Ultra 12.0.
2
. Chemistry
The general synthetic approach for amide analogues is summarized in scheme 1.
The carboxylic acids of 7a-d were protected by tert-butyl group, yielding
intermediates 8a-d, and subsequent bromination of 8a-d gave intermediates 9a-d. A
nucleophilic
substitution
-(4-hydroxyphenyl)hex-4-ynoic
-(4-hydroxyphenyl)hex-4-ynoate
-cyclopropyl-3-(4-hydroxyphenyl)propanoate
-(6-hydroxy-2,3-dihydrobenzofuran-3-yl)acetate
reaction
between
9a-b
(16),
and
methyl
(S)-methyl
methyl
3
3
3
2
3
acid
(17),
(18),
(19)
methyl
or
methyl
-(4-hydroxyphenyl)propanoate (20) in the presence of K CO produced
2
3
intermediates 10a-g. Deprotection of the tert-butyl group with CF COOH yielded
3
5

ACCEPTED MANUSCRIPT
acids 11a-g. 11a-g were then condensed with corresponding amines using HATU, and
the resultant amide 12a-v, 14f-i and 15l-v were hydrolyzed to afforded 13a-v, 14a-d
and 15a-k.
Scheme 1. Synthesis of 13a-v, 14a-d and 15a-k.
R1
O
O
O
O
COOMe
X
X
X
X
a
b
c
Br
O
HO
O
O
O
Z
Z
Z
Z
7a-d
8a-d
9a-d
10a-h
1
1a: X = S, Z = CH, R1 = propanyl
R1
11b: X = S, Z = N, R1 = propanyl
11c: X = O; Z = CH, R1 = propanyl
d
O
COOMe
X
1d: X= CH=CH; Z = CH, R1 = propanyl
1e: X = S, Z = CH, R1 = cyclpropyl;
1
1
1
1
1
O
HO
Z
1f: X = S, Z = CH, R1 = S-propanyl
1
1a-h
1g: X= CH=CH; Z = CH, R1 = S-propanyl
1
1h: X=S, Z=CH, R =H
O
O
S
S
N
e
N
f
11a
N
O
N
O
R2
COOMe
R2
COOH
12c-v
13c-v
3c R2=
3h R2=
13d R2=
13i R2=
13e R2=
13j R2=
13f R2=
13k R2=
13g R2=
13l R2=
N
1
N
N
O
1
F
F
13o R2=
F
13p R2=
13q R2=
1
3m R2=
NO
NO2
2 13n R2
=
Cl
O2N
13s R2
=
1
3r R2=
R1
R1
e
O
f
O
X
11a-h
X
COOH
COOMe
R3
R3
O
O
Z
Z
12a-b,12t-v,14f-i,15l-v
13a-b,13t-v,14a-d,15a-k
1
1
13a,15a,15d,15g:X=CH=CH, Z=CH, R =propanyl
1
3b,13t-u: X=S, Z=CH, R =propanyl
3b R3= P1 13t R3= P4
3u R3= P3
13v R3= P5
P1=
P2=
N
N
3
3
3
3
1
1
1
3a R =P1 15a R =P2 15d R =P3 15g R =P5
1
1
P4=
N
5b,15e,15h:X=S, Z=N, R =propanyl
1
3
3
3
N
F
5b R =P2
15e R =P3
15h R =P5
1
4a, 14b, 14d: X=S, Z=CH, R3 =P2
1
1
5c,15f,15i:X=O, Z=CH, R =propanyl
1
14b R1=
3
3
3
1
4a R = H
15c R =P2
15f R =P3
15i R =P5
1
3
N
NH
4d R1=
1
5j,15k:X=CH=CH, Z=CH, R =S-propanyl,R = P5
P3=
P5=
1
6

ACCEPTED MANUSCRIPT
Reagents and conditions: (a) (BOC) O, DMAP, DIPEA, DCM, reflux, 24h; (b) NBS,
2
CCl , 75°C, 4h; (c) K CO , DMF, rt, overnight; (d) CF COOH, DCM, rt, overnight;
4
2
3
3
(
e) HATU, DIPEA, DMF, amine, rt, 4-8h; (f) MeOH/H O, LiOH, rt, 0.5-12h.
2
3
. Results and Discussion
A diverse set of compounds were synthesized to improve the PK and
physicochemical properties of 2. Compounds’ agonist effects on hGPR40 were
assessed by calcium mobilization assay using hGPR40-HEK293 cell line which stably
expressed human GPR40.
Figure 2. Design of new GPR40 agonists based on the LY2881835 scaffold
Preliminary explorations focused on the linker of 2 (Table 1). Lilly reported that the
methylene between the phenyl and the piperidyl group was metabolically susceptible
to N-dealkylation, so we replaced the methylene with a carbonyl moiety to block the
metabolic site [23, 48] (Figure 2). In addition, the carbonyl replacement increased
7

ACCEPTED MANUSCRIPT
tPSA and reduced LogP, which would be expected to reduce the CNS penetration.
Direct replacement with the carbonyl in the linker gave 13a. 13a was less potent but
possessed lower LogP (5.9) and higher tPSA (67) than 2. Based on the structure of
another compound from Lilly, LY2922083, we produced 13b with a thiophene ring in
the linker 13b displayed comparable activity to that of 13a. Further liver microsomal
.
stability and Caco-2 permeability assays (Table 2) showed that, 13a and 13b retained
the permeability of 2. Introduction of the carbonyl group gave a slight increase in
metabolic stability in mouse liver microsomes. As 13b was more preferred for better
metabolic stability on both human and mouse liver microsomes than 2 and 13a,
5
-carbonylthiophene moiety was chosen as linker for the subsequent compounds.
In the optimization of both the stability and activity, we modified the indene ring in
the tail to improve stability because the nonaromatic double bond of indene could be
susceptible to oxidative metabolism in liver microsomes [49]. First, we opened the
five-member ring of indene in two ways (Figure 2) and introduced a nitrogen atom to
the piperidyl group, affording 13c and 13d with 4-phenylpiperazine and
4
-benzylpiperazine as the tail, respectively. 4-phenylpiperazine (13c) was more
favorable and displayed a 2.2-fold increase in potency than 13d. As expected, 13c
revealed promoted stability in human and mouse liver microsomes. Furthermore,
conversion of the benzene ring into a methyl of cyclohexyl group gave 13e and 13f
respectively. These two compounds had either markedly decreased potency or
complete loss of activity, suggesting that the aromatic ring was critical for
maintaining the agonistic activity. Also, we replaced the benzene ring with pyridyl
8

ACCEPTED MANUSCRIPT
rings (13g, 13h, 13i), giving lower LogP and higher tPSA. Alteration of the benzene
ring into 2-pyridyl (13i) was well tolerated, while 3-pyridyl (13h) and 4-pyridyl (13g)
replacement resulted in more than 16-fold and 70-fold loss in potency compared with
1
3c respectively, demonstrating that moving the nitrogen atom to different positions
greatly impacted the activity. In view of these outcomes, 13c was selected for further
exploration.
1
Table 1. Variations of R (the tail) and ring A (the linker)
hGPR40
1
b
b
Compd
R
Ring A
tPSA
LogP
5.9
a
EC (nM)
50
13a
179
67
1
3b
172
72
67
70
70
70
70
5.9
5.4
5.1
3.3
4.9
1
3c
1
3d
155
1
3e
>5000
820
1
3f
9

ACCEPTED MANUSCRIPT
1
3g
>5000
82
82
4.8
4.0
13h
1136
138
1
3i
82
50
4.1
6.5
2
134
a
b
EC50 was the average of at least 3 determination tPSA and LogP values were
,
Calculated from ChemBioDraw Ultra 12.0.
Table 2. Metabolic stability and Caco-2 permeability of 2, 13a and 13b
a
Liver microsomal stability (t , min)
Caco-2 permeability
1
/2
Compd
A to B
10 cm/s)
human
mouse
Efflux Ratio
-6
(
2
18
9
2
6
15.0
0.4
0.3
0.3
1
3a
3b
3c
10.9
19.3
1
25
105
7
b
b
1
55
ND
ND
a
b
‘
A to B’ means from apical to basolateral, ND means not determined.
The SAR of 13c was then investigated according to three parts of the molecule, the
-phenylpiperazine tail, the central linker, and the β-propynyl substituted
-benzenepropanoic acid head. We first investigated the influence of various
4
3
substituents para to the phenyl of the tail considering that the para-position was
potentially the major site of oxidative metabolism for the phenylpiperazine moiety [50]
(Table 3).
1
0

ACCEPTED MANUSCRIPT
Among these substituents, 13k with a tert-butyl group showed the greatest loss of
potency (5.8-fold), while the smaller methyl group (13j) had a 3-fold decrease in
potency only. This indicated that a bulky substituent at the para position of the phenyl
group was deleterious to activity. Neither an electron-donating methoxy group (13l)
nor a strong electron-withdrawing nitro group (13m) was beneficial for activity. A
2
comparison of 13l and 13m with 13k suggested that the size of R contributed to the
major loss of potency. Similar outcomes were observed with 13n and 13o, that the
compound with the bulkier chlorine group (13n) was less potent than that with the
smaller fluorine group (13o), and that 13o had higher potency than 13c. These SAR
findings led to our speculation that small substituents at the para position of the
phenyl would be favorable.
Further explorations at the ortho- and meta- postions of the phenyl in the tail
proceeded on the basis of the preliminary results. First, we investigated the fluorine
group at the meta- (13p) and othro- positions (13q) of the phenyl group, with both
compounds being slightly less potent than 13o. Surprisingly, conversion of fluorine
into the strong electron withdrawing nitro group at the ortho- position (13s) displayed
a potency similar to that of 13q, while 13r (nitro group at the meta- position) was less
active than 13s but more potent than 13m. We further examined chlorine and methoxy
substitutions ortho and meta to the phenyl, and a similar trend of potency to that of
1
3r and 13s (data not shown). These results imply that the ortho- position is well
tolerated with electron-withdrawing/donating and sterically hindered substituents,
while the length and size of the substituents at the meta- position may impact activity.
11

ACCEPTED MANUSCRIPT
In addition, we assessed the viability of 1,2,3,4-tetrahydroquinoline (13t) and
,2,3,4-tetrahydroisoquinoline (13u) as the tail on account of our previous SAR
1
results. To our delight, 13t and 13u each had comparable activity to that of 13c, with
1
3u being slightly more potent than 13c. Further structural expansion of 13t led to
3v, a compound with slightly decreased potency. These result indicated that a linear
1
fused-ring tail was more favorable.
Table 3. Variations of the terminal carboxamide group
hGPR40
2
b
b
Compd
R
tPSA
LogP
5.9
a
EC (nM)
50
1
3j
211
417
289
231
133
54
70
1
3k
70
79
6.8
5.3
1
3l
c
13m
121
70
--
1
3n
6.0
5.6
1
3o
70
1
3p
97
70
5.6
1
2

ACCEPTED MANUSCRIPT
1
3q
113
70
5.6
c
1
3r
3s
187
94
122
122
--
c
1
--
1
3t
80
57
67
67
76
5.6
5.3
5.9
1
3u
1
3v
109
a
EC was the average of at least 3 determination, the EC of 13j was the average of
5
0
50
b
2
determinations. tPSA and LogP was calculated from ChemBioDraw Ultra 12.0.
c
tPSA can not be calculated by ChemBioDraw Ultra 12.0.
Altogether, these SAR results revealed that the potency was strongly impacted by
the size of the substituents and steric hindrance around the phenyl ring in the tail. The
incorporation of a small fluorine group at para-position of the tail gave the most
potent compound 13o. Further exploration of the tail led to identification of another
potent compound, 13u, with a shorter fused-ring as the tail. These results were
followed by continued structural modifications to the head and linker of the chemical
structure.
1
3

ACCEPTED MANUSCRIPT
Modifications to the β-substituted benzenepropanoic acid in the head of 13o (Table
) did not yield any improvement in potency. For example, replacement of the
-propynyl group with a cyclopropyl group (14b), direct removal of the propynyl
4
1
group
(14a)
and
further
alteration
of
the
head
moiety
into
3
-((2,3-dihydrobenzofuran-3-yl))acetate acid (14c) all resulted in markedly eroded
potency. Next, the stereochemistry at the β-position was examined on account of
previous reports that the S-isomer was more favorable [23]. The S-isomer (14d)
indeed had a slightly improved potency over 13o (racemate), while the R-isomer (14e)
had markedly lower potency. These results supported that (S)-β-propynyl substituted
3
-benzenepropanoic acid was an optimal head for the amide structure.
Table 4. Variations of the β-substituents of benzenepropanoic acid in the head
hGPR40
3
b
b
Compd
R
tPSA
LogP
5.0
a
EC (nM)
50
1
4a
610
70
1
4b
>1667
>1667
70
79
5.5
4.3
1
4c
1
4

ACCEPTED MANUSCRIPT
1
4d
47
70
70
5.6
5.6
1
4e
>5000
a
b
EC50 was the average of at least 3 determination tPSA and LogP was calculated
,
from ChemBioDraw Ultra 12.0.
The SAR exploration of the linker aimed to replace the thiophene ring with
bioisosteres. Some moieties found to be favorable for potency in our previous SAR
findings were combined with the linker in an effort to produce the most potent
compound(s). (S)-β-Propynyl substituted 3-benzenepropanoic acid was chosen as the
head and, for the tail, 4-fluorine phenylpiperazine, 1,2,3,4-tetrahydroisoquinoline and
1
,2,3,4-tetrahydro-1-naphthylamine were selected, with focus on both potency and
structural diversity. Thiazole, benzene and furan rings were introduced to replace the
thiophene ring of the linker (Table 5). The thiazole ring was deleterious for potency
with three types of tails (15b, 15e, 15h). The combinations of
a
1
,2,3,4-tetrahydroisoquinoline
and
a
furan
linker
(15f)
or
1
,2,3,4-tetrahydro-1-naphthylamine and a benzene linker (15g) afforded high potency,
with 15g exhibiting the best activity.
Table 5. SAR analysis of ring B in the linker
1
5

ACCEPTED MANUSCRIPT
hGPR40
4
b
b
Compd
5a
R
Ring B
tPSA
LogP
5.5
a
EC (nM)
50
1
114
203
70
1
5b
82
79
67
79
76
5.0
4.2
5.2
4.7
3.9
1
5c
308
106
531
50
1
5d
1
5e
1
5f
1
5g
42
76
88
85
5.8
5.3
4.5
1
5h
144
133
1
5i
a
b
EC50 was the average of at least 3 determination tPSA and LogP was calculated
,
from ChemBioDraw Ultra 12.0.
1
6

ACCEPTED MANUSCRIPT
1
5g, which was the most active agonist for hGPR40 among the analogues we
synthesized, has two chiral centers and hence exists as four diastereomers.
Considering the more favorable S-configuration at the β-position of the head, the
impact of the absolute configuration at the tail was evaluated (Figure 3). The
S,R-isomers (15k) had 2-fold promotion in potency over the S,S-isomers (15j) and the
racemate (15g). Furthermore, 15k, 15j and 14d exhibited good agonistic activities for
both rat and mouse GPR40, indicating low species-dependent differences (Table 6).
Figure 3. Effect of the stereochemistry of the 1,2,3,4-tetrahydro-1-naphthylamine
moiety of 15g on the hGPR40 activity
Table 6. Agonistic activities for GPR40 from different species
hGPR40
rGPR40
mGPR40
Compd
4d
5j
5k
a
a
a
EC (nM)
EC (nM)
EC (nM)
50
50
50
1
47
25
28
14
31
42
23
1
39
20
1
a
EC was the average of at least 3 determination
.
5
0
1
5k, 15j and 14d, were then submitted to metabolic stability studies in human and
mouse liver microsomes and permeability test with Caco-2 cells (Table 7). 15k, 15j
-6
and 14d all retained excellent permeability (Papp value greater than 15.2 × 10 cm/s)
1
7

ACCEPTED MANUSCRIPT
and displayed modestly improved metabolic stability, compared with 2. 14d with
4
-fluorine phenylpiperazine as tail demonstrated the best stability in mouse liver
microsomes. When the tail and linker were replaced with
S/R)-1,2,3,4-tetrahydro-1-naphthylamine and a benzene ring (15k) there was a
(
markedly improved stability in human liver microsomes. The S,R-isomeres (15k) was
more stable than the S,S-isomeres (15j) in both human and mouse liver microsomes,
indicating that the stereochemistry of the tail had little effect on potency but great
impact on metabolic stability.
Table 7. In vitro metabolic stability and permeability of 15k, 15j and 14d
Liver microsomal stability
a
Caco-2 permeability
(
t , min)
1/2
Compd
A to B
10 cm/s)
human
mouse
Efflux Ratio
-6
(
1
4d
5j
5k
64
120
27
20.6
0.8
1.2
0.9
1
214
387
15.2
18.8
1
74
a
‘A to B’ means from apical to basolateral
The PK and the tissue distribution profiles of 15k, 15j, 14d and 2 (Table 8) in male
ICR mice were assessed. 15k, 15j, 14d all exhibited prolonged half-lives and
improved plasma exposure. 15k achieved the highest concentration (15.45µg/mL) in
plasma among four compounds. The observation that 15j (S,S-isomers) and 15k
(S,R-isomers) differed substantially in peak times and peak plasma concentrations
indicated that the stereochemistry of the tail significantly affected absorption.
1
8

ACCEPTED MANUSCRIPT
Compared with 2, 15k had 5.4-fold, 4.1-fold and 15.5-fold improvement in plasma
exposure, half-life time and AUC (area under the curve), respectively.
a
Table 8. In vivo PK and distribution parameters of selected compounds in ICR Mice
AUC0
-8
Cmax
a
a
b
b
compd
T (h) Tmax (h)
B/P
L/P
a
1/2
a
(
h*µg/mL)
(µg/mL)
2
5.15
1.3
3.8
3.0
5.3
0.7
0.8
2.85
13.31
3.77
0.14
0.018
0.006
0.011
2.5
15.1
0.9
1
4d
5j
5k
28.45
3.76
1
0.25
1.2
1
79.98
15.45
0.3
a
Dose orally to male ICR mice at 30mg/kg (n = 3), vehicle was 0.1% Tween-80 in 1%
b
hydropropylcellulose sodium. Dose orally to male ICR mice at 30mg/kg (n = 3). All
mice was sacrificed at the time point 0.75h (15j, 2), 1h (14d, 15k), respectively. B/P
means total brain-to-plasma drug distribution ratio, L/P means the liver-to-plasma
drug distribution ratio.
Tissue distribution assays showed that 14d, 15j and 15k had much lower CNS
exposure (B/P = 0.018, 0.006 and 0.011 respectively) than 2. We also examined the
distribution of compounds to the liver, based on the report that TAK-875 had 3 times
higher distribution to the liver than to plasma [35]. 15j and 15k displayed lower L/P
compared with 2. However, 14d was not studied any further because of safety
concerns associated with its 15-times higher distribution to the liver than to plasma.
1
5k, with promising PK and safety profiles, was progressed to in vivo efficacy studies,
where it demonstrated a significant decrease (P < 0.001) in blood glucose level
1
9

ACCEPTED MANUSCRIPT
following a 2.5 g/kg oral glucose load in the oral glucose tolerated test (OGTT)
(Figure 4).
Figure 4. OGTT assay of 15k (100mg/kg) in ICR mice. (A) blood glucose
concentration of 15k during OGTT; (B) blood glucose AUC0-120 after
min
administraion of 15k. Compound 15k was administered orally to ICR mice (n=8) at
0min prior to a oral glucose load (2.5 g/kg). Blood glucose levels were measured
3
berefore and after glucose load. ** p<0.01, *** p<0.001 versus control were analyzed
by Student’s t-test. Error bar indicateds SEM.
4
. Conclusion
We developed a series of amide derivatives driven by the moderate CNS exposure,
high in vitro clearance and limited oral exposure of 2. By inserting a carbonyl at the
position prone to N-dealkylation metabolism and opening up the indene, as well as
introducing a nitrogen atom to reduce LogP and elevate tPSA, we obtained 13c with
improved physicochemical properties and liver microsomal stability. Further SAR
2
0

ACCEPTED MANUSCRIPT
research and optimization yielded 15k, which exhibited the best in vitro activity,
excellent permeability and markedly improved in vitro stability. Furthermore, studies
demonstrated that 15k possessed superior PK properties in vivo, lower distribution to
brain and liver than to plasma compared with LY2881835, supporting that 15k is
more likely to avoid the undesired effects in the CNS and liver. In conclusion, our
results disclosed an orally efficacious compound 15k, which represents a promising
lead compound for developing a safe antidiabetic drug that acts via activation of
GPR40.
5
. Experimental section
5
.1 Chemistry
All reagents were purchased from commercial suppliers and used without further
purification. Column chromatography was carried out on silica gel (200-300 mesh) or
with pre-packed silica cartridges (4-40 g) from Bonna-Agela Technologies Inc.
@
(
Tianjin, China) and eluted with a CombiFlash Rf 200 from Teledyne Isco.
Prep-HPLC separation was carried out in Unimicro Easysep-1010 series LC (UV 254
−
1
nM, 25 °C, flow rate = 10 mL min ) with the column of Agilent Prep-C18 (10 µm,
2
1.2×250mm) or CHIRALPAK IG (5µm, 10×250mm), while the mobile phase was
1
13
4
0-90% MeOH /H O or 80% MeOH /DCM. H NMR and C NMR spectra were
2
recorded on a Varian-Mercury Plus-300 or a Bruker Avance III 400 or a Bruker
Avance III 500 NMR spectrometer using tetramethylsilane or solvent signals as an
internal reference. High-resolution mass spectra (ESI) were obtained on a Q-TOF or
2
1

ACCEPTED MANUSCRIPT
Thermo Orbitrap Elite. The purity of tested compounds was determined by HPLC
[
Agilent LC1260, Agilent ChemStation, ZORBAX SB-C18 (5 µm, 4.6 × 150 mm,
−
1
UV 254 nM, 20 °C, flow rate = 1.0 mL min )]. All of the assayed compounds
possess >95% purity. The enantiomeric excess and diastereoisomeric excess were
defined by HPLC [Agilent LC1260, Agilent ChemStation, CHIRALPAK IG (5 µm,
4
.6 mm × 150 mm) or CHIRALPAK IA (5 µm, 4.6 × 250 mm), UV 254 nM, flow
−
1
rate = 1.0 mL min )]. The optical rotation of tested compounds was measured by
Rudoph Autopol VI automatic polarimeter.
5
.1.1. General procedure A for preparation of intermediates 8a-d
To a solution of 7a-d (1 equiv) in DCM was added DMAP (0.1 equiv) and DIPEA
(1.5 equiv) at 0°C, then added di-tert-butyl dicarbonate (2 equiv) dropwise over
1
5min, and the mixture was stirred at 40°C for 24h. The mixture was washed with
brine, dried over MgSO4, and evaporated in vacuo. The residue was purified with
column chromatography (petroleum ether/ethyl acetate = 20:1) to yield intermediate
8
a-d.
5
.1.1.1. tert-butyl 5-methylthiophene-2-carboxylate (8a)
Intermediate 8a was prepared with general procedure
A
using
1
5
-methylthiophene-2-carboxylic acid to a colorless oil in 74.6%. H NMR (400 MHz,
CDCl ) δ 7.54 (d, J = 3.7 Hz, 1H), 6.75 (d, J = 3.6 Hz, 1H), 2.52 (s, 3H), 1.58 (s, 9H).
3
5
.1.1.2. tert-butyl 5-methylthiazole-2-carboxylate (8b)
2
2

ACCEPTED MANUSCRIPT
Intermediate 8b was prepared with general procedure
A
using
1
5
-methylthiazole-2-carboxylic acid to afford a light-yellow oil in 64.9%. H NMR
400 MHz, CDCl ) δ 8.12 (s, 1H), 2.71 (s, 3H), 1.55 (s, 9H).
(
3
5
.1.1.3. tert-butyl 5-methylfuran-2-carboxylate (8c)
Intermediate 8c was prepared with general procedure
A
using
1
5
-methylfuran-2-carboxylic acid to afford a light-yellow oil in 67.2%. H NMR (400
MHz, CDCl ) δ 6.98 (d, J = 3.3 Hz, 1H), 6.08 (d, J = 3.3 Hz, 1H), 2.37 (s, 3H), 1.58
3
(s, 9H).
5
.1.1.4. tert-butyl 4-methylbenzoate (8d)
Intermediate 8d was prepared with general procedure A using 4-methylbenzoic
1
acid to afford a colorless oil in 69.3%. H NMR (400 MHz, CDCl ) δ 7.88 (d, J = 8.2
3
Hz, 2H), 7.22 (d, J = 8.0 Hz, 2H), 2.40 (s, 3H), 1.60 (s, 9H).
5
.1.2. General procedure B for preparation of intermediates 9a-d
To a solution of 8a-d (1 equiv) in CCl was added N-bromobutanimide (0.9 equiv)
4
and benzoyl peroxide (0.1 equiv), then the mixture was refluxed at 75°C for 4h. The
mixture was washed with brine, dried over MgSO4, filtered, and concentrated. The
residue obtained was purified with column chromatography (petroleum ether/ethyl
acetate = 20:1) to yield intermediate 9a-d.
5
.1.2.1. tert-butyl 5-(bromomethyl)thiophene-2-carboxylate (9a)
2
3

ACCEPTED MANUSCRIPT
Intermediate 9a was prepared with general procedure B using 8a to afford a
1
colorless oil in 85.8%. H NMR (400 MHz, CDCl ) δ 7.78 (d, J = 3.4Hz, 1H), 7.22 (d,
3
J = 3.4 Hz, 1H), 4.96 (s, 2H), 1.42 (s, 9H).
5
.1.2.2. tert-butyl 5-(bromomethyl)thiazole-2-carboxylate (9b)
Intermediate 9b was prepared with general procedure B using 8b to afford a
1
light-yellow oil in 71.6%. H NMR (400 MHz, CDCl ) δ 8.12 (s, 1H), 5.30 (s, 2H),
3
1
.42 (s, 9H).
5
.1.2.3. tert-butyl 5-(bromomethyl)furan-2-carboxylate (9c)
Intermediate 9c was prepared with general procedure B using 8c to afford a
1
light-yellow oil in 86.9%. H NMR (400 MHz, CDCl ) δ 7.00 (d, J = 3.2 Hz, 1H),
3
6
.47 (d, J = 3.1 Hz, 1H), 4.49 (s, 2H), 1.58 (s, 9H).
5
.1.2.4. tert-butyl 4-(bromomethyl)benzoate (9d)
Intermediate 9d was prepared with general procedure B using 8d to afford a
1
colorless oil in 84.5%. H NMR (400 MHz, CDCl ) δ 7.97 (d, J = 8.3 Hz, 2H), 7.44
3
(d, J = 8.3 Hz, 2H), 4.50 (s, 2H), 1.60 (s, 9H).
5
.1.3. General procedure C for preparation of intermediates 10a-i
To a solution of 16-20 (1 equiv) in DMF was added K CO (2 equiv), stirred for
2
3
5
min then add the intermediates 9a-d (1.1 equiv), then the mixture was stirred at room
temperature overnight. Water was added to the mixture and extracted with ethyl
acetate, then washed with brine, dried over MgSO4, filtered, and concentrated. The
2
4

ACCEPTED MANUSCRIPT
residue obtained was purified with column chromatography (petroleum ether/ethyl
acetate = 8:1) to yield intermediate 10a-i.
5
.1.3.1.
tert-butyl
5
-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)thiophene-2-carboxylate (10a)
Intermediate 10a was prepared with general procedure C using 9a and methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (16) to afford a yellow oil in 78.3%. H NMR (400
MHz, CDCl ) δ 7.62 (d, J = 3.7 Hz, 1H), 7.31 (d, J = 8.6 Hz, 2H), 7.05 (d, J = 3.8 Hz,
3
1
H), 6.93 (d, J = 8.7 Hz, 2H), 5.20 (s, 2H), 4.11 – 4.05 (m, 1H), 3.68 (s, 3H), 2.81 –
.63 (m, 2H), 1.85 (d, J = 2.4 Hz, 3H), 1.59 (s, 9H).
2
5
.1.3.2.
tert-butyl
5
-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)thiazole-2-carboxylate (10b)
Intermediate 10b was prepared with general procedure C using 9b and methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (16) to afford a yellow oil in 82.1%. H NMR (400
MHz, CDCl ) δ 8.29 (s, 1H), 7.33 (d, J = 8.6 Hz, 2H), 6.96 (d, J = 8.6 Hz, 2H), 5.35
3
(s, 2H), 4.12 – 4.06 (m, 1H), 3.68 (s, 3H), 2.78 (dd, J = 15.3, 8.3 Hz, 1H), 2.67 (dd, J
=
15.3, 7.0 Hz, 1H), 1.85 (d, J = 2.4 Hz, 3H), 1.59 (s, 9H).
5
.1.3.3.
tert-butyl
5
-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)furan-2-carboxylate (10c)
Intermediate 10c was prepared with general procedure C using 9c and methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (16) to afford a yellow oil in 75.2%. H NMR (400
2
5

ACCEPTED MANUSCRIPT
MHz, CDCl ) δ 7.31 (d, J = 8.6 Hz, 2H), 7.06 (d, J = 3.4 Hz, 1H), 6.93 (d, J = 8.6 Hz,
3
2
H), 6.49 (d, J = 3.4 Hz, 1H), 5.05 (s, 2H), 4.13-4.05 (m, 1H), 3.68 (s, 3H), 2.78 (dd,
J = 15.3, 8.3 Hz, 1H), 2.67 (dd, J = 15.2, 7.0 Hz, 1H), 1.85 (d, J = 2.3 Hz, 3H), 1.59
s, 9H).
(
5
.1.3.4. tert-butyl 4-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)benzoate
(10d)
Intermediate 10d was prepared with general procedure C using 9d and methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (16) to afford a colorless oil in 63.5%. H NMR
(
400 MHz, CDCl ) δ 8.01 (d, J = 8.3 Hz, 2H), 7.48 (d, J = 8.3 Hz, 2H), 7.30 (d, J =
3
8.6 Hz, 2H), 6.92 (d, J = 8.7 Hz, 2H), 5.12 (s, 2H), 4.10 – 4.05 (m, 1H), 3.68 (s, 3H),
2.78 (dd, J = 15.3, 8.3 Hz, 1H), 2.67 (dd, J = 15.3, 7.0 Hz, 1H), 1.84 (d, J = 2.4 Hz,
3H), 1.62 (s, 9H).
5
5
.1.3.5.
tert-butyl
-((4-(1-cyclopropyl-3-methoxy-3-oxopropyl)phenoxy)methyl)thiophene-2-carboxylat
e (10e)
Intermediate 10e was prepared by general procedure C using 9a and methyl
-cyclopropyl-3-(4-hydroxyphenyl)propanoate (17) to afford a brown oil in 67.7%.
3
1
H NMR (400 MHz, CDCl ) δ 7.61 (d, J = 3.7 Hz, 1H), 7.39 (d, J = 15.0 Hz, 2H),
3
7
.04 (d, J = 3.7 Hz, 1H), 6.86 (d, J = 14.9 Hz, 2H), 5.19 (s, 2H), 3.60 (s, 3H), 2.79 –
2
.67 (m, 2H), 2.34 (dd, J = 17.3, 7.6 Hz, 1H), 1.57 (s, 9H), 1.05 – 0.95 (m, 1H), 0.62
2
6

ACCEPTED MANUSCRIPT
0.53 (m, 1H), 0.46 – 0.38 (m, 1H), 0.25 (td, J = 9.6, 4.9 Hz, 1H), 0.14 (td, J = 9.8,
.1 Hz, 1H).
–
5
5
.1.3.6.
(S)-tert-butyl
5
-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)thiophene-2-carboxylate (10f)
Intermediate 10f was prepared by general procedure C using 9a and (S)-methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (18) to afford a light-yellow oil in 78.8%. H NMR
(
400 MHz, CDCl ) δ 7.61 (d, J = 3.7 Hz, 1H), 7.30 (d, J = 8.6 Hz, 2H), 7.04 (d, J =
3
3.7 Hz, 1H), 6.91 (d, J = 8.7 Hz, 2H), 5.19 (s, 2H), 4.09 – 4.04 (m, 1H), 3.67 (s, 3H),
2.76 (dd, J = 15.3, 8.3 Hz, 1H), 2.66 (dd, J = 15.3, 7.0 Hz, 1H), 1.84 (d, J = 2.4 Hz,
3H), 1.57 (s, 9H).
5
.1.3.7. (S)-tert-butyl 4-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)benzoate
(10g)
Intermediate 10g was prepared by general procedure C using 9d and (S)-methyl
1
3
-(4-hydroxyphenyl)hex-4-ynoate (18) to afford a light-yellow oil in 75.9.2%. H
NMR (400 MHz, CDCl ) δ 8.00 (d, J = 8.4 Hz, 2H), 7.47 (d, J = 8.4 Hz, 2H), 7.29 (d,
3
J = 8.6 Hz, 2H), 6.91 (d, J = 8.7 Hz, 2H), 5.11 (s, 2H), 4.09 – 4.03 (m, 1H), 3.67 (s,
3
H), 2.76 (dd, J = 15.3, 8.4 Hz, 1H), 2.66 (dd, J = 15.3, 6.9 Hz, 1H), 1.83 (d, J = 2.4
Hz, 3H), 1.60 (s, 9H).
5
5
.1.3.8.
tert-butyl
-((4-(3-methoxy-3-oxopropyl)phenoxy)methyl)thiophene-2-carboxylate (10h)
2
7

ACCEPTED MANUSCRIPT
Intermediate 10h was prepared by general procedure C using 9a and methyl
1
3
-(4-hydroxyphenyl)propanoate (19) to afford a light-yellow oil in 79.4%. H NMR
(
400 MHz, CDCl ) δ 7.61 (d, J = 3.7 Hz, 1H), 7.13 (d, J = 8.6 Hz, 2H), 7.04 (d, J =
3
3
.7 Hz, 1H), 6.89 (d, J = 8.6 Hz, 2H), 5.18 (s, 2H), 3.67 (s, 3H), 2.90 (t, J = 7.8 Hz,
H), 2.61 (t, J = 7.8 Hz, 2H), 1.57 (s, 9H).
2
5
.1.3.9.
tert-butyl
5
-(((3-(2-methoxy-2-oxoethyl)-2,3-dihydrobenzofuran-6-yl)oxy)methyl)thiophene-2-
carboxylate (10i)
Intermediate 10i was prepared by general procedure C using 9a and methyl
2
7
1
2
5
1
-(6-hydroxy-2,3-dihydrobenzofuran-3-yl)acetate (20) to afford a light-yellow oil in
1
4.3%. H NMR (400 MHz, CDCl ) δ 7.61 (d, J = 3.7 Hz, 1H), 7.04 (d, J = 4.0 Hz,
3
H), 7.03 (s, 1H), 6.47 (dd, J = 8.1, 2.3 Hz, 1H), 6.45 (d, J = 2.2 Hz, 1H), 5.16 (s,
H), 4.77 (t, J = 9.0 Hz, 1H), 4.28 (dd, J = 9.2, 6.1 Hz, 1H), 3.81 (ddd, J = 14.9, 9.1,
.9 Hz, 1H), 3.72 (s, 3H), 2.76 (dd, J = 16.5, 5.5 Hz, 1H), 2.57 (dd, J = 16.5, 9.3 Hz,
H), 1.57 (s, 9H).
5
.1.4. General procedure D preparation of intermediates 11a-i
To a solution of 10a-i (1 equiv) in DCM was added CF COOH (4 equiv), and the
3
mixture was stirred at room temperature overnight. The solvent was evaporated under
reduced pressure without further purification to yield intermediate 11a-i.
5
.1.4.1. 5-((4-(1-methoxy-1-oxohex-4-yn-3-yl)phenoxy)methyl)thiophene-2-carboxylic
acid (11a)
2
8