Design of potent and selective GPR119 agonists for type II diabetes

Article abstract of DOI:10.1016/j.bmcl.2010.12.086

Screening of the Merck sample collection identified compound 1 as a weakly potent GPR119 agonist (hEC₅₀ = 3600 nM). Dual termini optimization of 1 led to compound 36 having improved potency, selectivity, and formulation profile, however, modest

Full text of DOI:10.1016/j.bmcl.2010.12.086

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2665–2669
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design of potent and selective GPR119 agonists for type II diabetes
Jason W. Szewczyk a, , John Acton a, Alan D. Adams a, Gary Chicchi c, Stanley Freeman a, Andrew D. Howard b,
⇑
Yong Huang a, Cai Li b, Peter T. Meinke a, Ralph Mosely a, Elizabeth Murphy c, Rachel Samuel a,
Conrad Santini a, Meng Yang a, Yong Zhang a, Kake Zhao a, Harold B. Wood a
a Department of Medicinal Chemistry, Merck & Co., Inc., PO Box 2000 Rahway, NJ 07065, USA
b Lead Optimization Pharmacology, Merck & Co., Inc., PO Box 2000 Rahway, NJ 07065, USA
c Preclinical Drug Metabolism and Pharmacokinetics, Merck & Co., Inc., PO Box 2000 Rahway, NJ 07065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Screening of the Merck sample collection identified compound 1 as a weakly potent GPR119 agonist
(hEC50 = 3600 nM). Dual termini optimization of 1 led to compound 36 having improved potency, selec-
tivity, and formulation profile, however, modest physical properties (PP) hindered its utility. Design of a
new core containing a cyclopropyl restriction yielded further PP improvements and when combined with
the termini SAR optimizations yielded a potent and highly selective agonist suitable for further preclinical
development (58).
Received 17 November 2010
Accepted 16 December 2010
Available online 22 December 2010
Keywords:
GPR119 agonist
GPCR
Ó 2011 Elsevier Ltd. All rights reserved.
Diabetes
GDIS
Agonism of the GPR119 receptor has recently emerged as a
promising new approach for the treatment of type 2 diabetes mel-
litus.1 GPR119 is a X-linked class A GPCR2 whose expression in hu-
mans, mice and rats is highly restricted to pancreatic islets and
specific intestinal regions.3 GPR119 agonists have been implicated
in the control of both incretin secretions (e.g., GLP-1)1b,3b and islet
hormone release (i.e., insulin)4 in mediating glucose homeostasis.
Additionally, dosing of a synthetic GPR119 agonist has been re-
ported to lower plasma glucose levels in an oral glucose tolerance
test (OGTT) both preclinically4 and clinically.5
The promise of GPR119 agonists to modulate glucose homeo-
stasis through both an incretin-dependent (i.e., L-cell GLP-1 re-
lease) and an incretin-independent mechanism (i.e., beta cell
insulin release) has attracted considerable attention as evidenced
by many new patent filings1 and presentations.6 Recently, two de-
tailed reports have appeared on the synthesis and characterization
of synthetic GPR119 agonists.7 Herein we report the synthesis, SAR
optimization and in vivo effects on glucose plasma levels during an
OGTT for a novel family of GPR119 agonists.
agonists was optimizing the combination of solubility and selectiv-
ity while generating robust in vivo efficacy.
To develop agonists with a suitable profile, two SAR optimiza-
tions were simultaneously initiated. First, for the bis-piperidine
core, the SAR of eastern and western amino moieties were investi-
gated, focusing on identifying polar, solubilizing substituents. This
initial strategy was to identify suitably polar and metabolically sta-
ble groups for either termini and thus improve both physical prop-
erties and hERG selectivity. A second focus, was modification of the
agonist core to identify a new scaffold with an improved profile.
Knowing that design of new cores would be more time intensive
than termini optimization, we utilized molecular modeling and
overlayed the core of 1 with another GPR119 agonist7a to aid de-
sign. The premise was that if the new core mapped similar space
as the other agonists it would be possible to merge the information
from these diverse exercises into a single, optimized molecule.
The synthesis of the bis-piperidine agonists has been previously
reported with full experimental and spectral characterization.8 The
Boc containing bis-piperidines were synthesized by reaction of the
commercially available Boc-amine with either: an appropriate aryl
halide using the palladium catalyzed coupling protocol of Buch-
wald,9 or SNAr of a suitable heteroaryl halide at elevated tempera-
tures to afford the desired analogs (1–21, Scheme 1). The Boc group
was removed with TFA and the resulting amine further functional-
ized using the two methods described above to provide agonists
22–36 (Scheme 1).
A new GPR119 agonist (1) was identified by screening the
Merck sample collection (Table 1). This bis-piperidine class of ago-
nists was characterized by poor PP (i.e., low aqueous solubility and
high plasma protein binding), high mouse clearance, insufficient
potency, and poor hERG selectivity. A key challenge for these
To discover novel core scaffolds a series of racemic cyclopropa-
nol intermediates was prepared (Scheme 2). The synthesis of all
the cyclopropyl derivatives have been previously reported with full
⇑
Corresponding author.
E-mail address: jason_szewczyk@merck.com (J.W. Szewczyk).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.086

2666
J. W. Szewczyk et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2665–2669
Boc
Boc
N
O
N
N
Boc
OBn
Boc
N
a or b
a, b
47
OBn
HN
N
1-21
HO
Ar
N
Ar
(Me)2NOC
(Me)2NOC
O
O
c
N
B
nBu
c, a
22-36
N
Boc
N
Boc
d, e, f
Ar
49
48
OBn
Scheme 1. Synthesis of various bis-piperidine agonists. Reagents and conditions:
(a) Het-X, Cs2CO3 or DBU, 50–130 °C; (b) Ar-X, Pd(OAc)2, Cs2CO3, toluene, 90 °C,
PhDavePhos; (c) TFA (neat), rt.
HO
HO
Scheme 3. Chiral cyclopropanol scaffold synthesis. Reagents and conditions: (a)
nBuLi, BF3ÁEt2O, epoxide; (b) 5% Pd on CaCO3 poisoned with Pb, quinoline, EtOAc,
H2; (c) Et2Zn, CH2I2, dioxaborolane, DCM, Et2O (1 equiv), À20 °C to À0 °C; (d) 20%
Pd(OH)2, H2, 1:1 EtOAc/EtOH, 50 psi; (e) NaIO4, DCM, water, 0 °C; (f) NaBH4, EtOH,
rt.
OBn
CIS
Boc
Boc
O
N
a, b, c, d
N
H
n
n = 0, 1
n
37, 38
cyclopropanation using the (S, S) dioxaborolane as a chiral auxil-
iary afforded the chiral diastereomeric cyclopropanation products
48 (93:7 dr).10 Subsequent debenzylation with palladium hydrox-
ide, periodate oxidation of the diol and reduction of the resulting
aldehyde furnished the desired chiral cyclopropane. A final chiral
chromatographic separation gave the desired cis-cyclopropanol
49 in >99:1 er. In contrast, not using the chiral ligand or not incor-
porating the chiral alcohol gave much lower enantioselection
(83:17 dr and 78:22 er, respectively).
The final cyclopropane agonists were synthesized in a route
analogous to the bis-piperidines (Scheme 4). The Boc groups (39–
41, 44–46, 49) were removed with TFA and coupling with an
appropriate heterocyclic-halide provided the desired aryl interme-
diates. Mistunobu reaction of the alcohol gave agonists 51–57 and
59. Alternatively, deprotonation of the alcohol with sodium hy-
dride and reaction with a heteroaryl-halide gave compounds 58,
60, 61. In some cases the racemic compounds were resolved using
chiral chromatography (54, 55) or the single enantiomers were
prepared directly (58–61) from 49 (Scheme 4).
After identification of 1 as a weakly potent GPR119 agonist the
termini optimization of the bis-piperidine core began with the SAR
of the western aryl (Table 1). Compound efficacy was evaluated for
the human and mouse receptors in vitro using a GPR119 HTRF
cAMP activation assay.8,10 Removal of the pyridyl nitrogen gave
the phenyl analog 2, which formed a platform for rapid investiga-
tion. A methyl group was substituted around the phenyl ring (3–5).
Substitution at the meta (4) and para (5) positions improved po-
tency, while ortho (3) substitution reduced activity. In general,
for the agonists described herein, ortho substitution was poorly
For n =0
e, f, g, h, i
For n =1
e, f
j, k,d
For n =0
e, f, g, h, i
TRANS
BnO
Boc
Boc
HO
p
N
N
For n =1
e, f
n
n
42, 43
39-41; 44-46
p = 1, 2
n = 0, 1
Scheme 2. Synthesis of various cyclopropanol core scaffolds. Reagents and
conditions: (a) PPh3, CBr4; (b) nBuLi, (CH2O)n; (c) 5% Pd on CaCO3 poisoned with
Pb, quinoline, EtOAc, H2; (d) BnBr, NaH, DMF; (e) Et2Zn, ClCH2I, DCE, À18 °C to
À10 °C; (f) 20% Pd(OH)2, H2, 1:1 EtOAc/EtOH, 50 psi; (g) TPAP, NMO, DCM, rt; (h)
nBuLi, Ph3PCH2I, THF, À78 °C; (i) (a) BH3Me2S, THF; (b) NaOH; (c) H2O2; (j)
P(O)(OEt)2CH2CO2Me, LiCl, DBU, rt; (k) DIBAL-H, DCM, À78 °C.
experimental and spectral characterization.10 For the preparation
of the cis-cyclopropanols, the aldehyde (n = 0, 1) was reacted with
carbon tetrabromide and triphenyl phosphine and the resulting di-
bromo-olefin treated with n-butyl lithium to afford an alkyne.
Reduction with Lindlar’s catalyst gave the necessary cis-alkenes
which were protected as the benzyl ethers (37, 38; n = 0, 1). Cyclo-
propantion using a modification of the Charette cyclopropanation
followed by hydrogenolysis gave the desired cis-cyclopropanols
(39, 40; n = 0, 1; p = 1) (Scheme 2).10,11 Additionally, 39 (n = 0,
p = 1) could be homologated by a sequence of TPAP oxidation, Wit-
tig olefination, hydroboration and boronate oxidation to give the
homologated cis-cyclopropanols (41; n = 0, p = 2).
Similarly the trans-cyclopropanols were obtained via an analo-
gous sequence. The aldehyde (n = 0, 1) was transformed under
Horner–Emmons olefination to provide the trans olefin. The ester
was reduced with DIBAL-H and the resulting alcohol protected as
the benzyl ether (42, 43; n = 0, 1). Cyclopropanation and benzyl
deprotection proceeded as described for the cis route to give the
desired trans-cyclopropanols (44, 45; n = 0, 1; p = 1) (Scheme 2).
Additionally the trans-cyclopropanol 44 (n = 0, p = 1) could be
homologated as described above to give the extended trans-cyclo-
propanol (46, n = 0, p = 2).
Cyclopropane scaffold 49 was of particular interest and to ob-
tain large quantities of this chiral intermediate we developed a
large scale double, diastereoselective protocol10 based on Cha-
rette’s enantioselective cyclopropanation reaction11 (Scheme 3).
The starting alkyne was reacted with the (R)-glycidyl epoxide to
provide the (R)-glycidyl 47. Reduction with Lindlar’s catalyst and
Ar
Boc
HO
N
HO
N
p
p
n
n
a or b
Ar
c or d
39-41; 44-46; 49
p = 1, 2
n = 0, 1
Ar
O
N
p
51-61
n
Scheme 4. Synthesis of various cyclopropane agonists. Reagents and conditions: (a)
TFA (neat), rt; (b) Het-X, Cs2CO3, NMP; (c) DBAD, (polymer)-Ph3P, DCM, THF; (d) (i)
NaH, 0 °C, (ii) Het-X, rt.

J. W. Szewczyk et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2665–2669
2667
Table 1
Table 3
Bis-piperidine SAR: western aryl optimizationa,b
Mouse pharmacokinetic parameters for GPR1119 agonistsa
3
2
Compd
Cl (ml/min/kg)
Vdss (L/kg)
t1/2 (h)
AUCn (lM h kg/mg)
O
O
R
4
Ar
N
N
6
27
120
2
6.9
2.3
0.6
13.5
0.09
19.3
5
6
a
2/10 mpk iv/po.
Compd
Ar
R
hGPR119
EC50 (nM)
mGPR119
EC50 (nM)
1
2
3
4
5
6
7
8
3-Pyridyl
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
H
H
2-Me
3-Me
3.6 K
1.4 K
NA
514
338
22
1.9 K
11
25
NA
NA
NA
NA
1.9 K
36
NA
132
341
513
165
NA
2.5 K
379
523
142
11
Next, a series of aromatic heterocycles was investigated, which
in principle, would improve both physical properties and potency
by appropriate display of a nitrogen lone pair. Introduction of
nitrogen at the 2-position (13) was tolerated and equipotent with
phenyl 2 thus a series of related heterocycles were prepared which
displayed a second lone pair at the 2, 3, and 4-positions (14–16). As
the previous analysis indicated, positioning a lone pair in either the
meta or para position gave a >10-fold increase in potency (14, 16)
without increasing molecular weight. A variety of substituted het-
erocycles were surveyed, focusing on substituents with the poten-
tial to be metabolically, non-displaceable (e.g., cyano and methyl).
Addition of a nitrile at either the 3- or 5-position (17, 18) greatly
improved potency (>20-fold). Gratifyingly, addition of a second
methyl substituent (19, 20) was additive and further increased po-
tency to provide single digit nanomolar agonists. Other substituted
heterocycles were investigated (e.g., 21), however, these analogs
were all less potent than the corresponding pyrimidines, especially
on the GPR119 mouse receptor.
4-Me
4-SO2Me
3-SO2Me
4-SOMe
4-C(O)Me
4-SMe
4-C(O)NH2
3-C(O)NH2
H
H
H
H
3-CN
9
10
11
12
13
14
15
16
17
18
19
20
21
600
76
NA
831
116
526
103
6
6
1
2
15
2-Pyridyl
2,4-Pyrimidyl
2,6-Pyrimidyl
2,5-Pyrazinyl
2,4-Pyrimidyl
2,4-Pyrimidyl
2,4-Pyrimidyl
2,4-Pyrimidyl
2,4-Pyridyl
5-CN
31
2
2
119
3-CN, 5-Me
3-Me, 5-CN
3-Me, 5-CN
a
Values are means of at least two independent titrations.
SAR optimization of the eastern group focused on PK improve-
ment. It was postulated that the Boc group was a key driver of the
poor PK observed (Tables 3 and 6) and replacement with a meta-
bolically stable heterocyclic isostere would improve the profile. A
series of heterocycles were surveyed, and pyridines, pyrazines,
and pyrimidines were all well tolerated 22–24. In general, for the
eastern aryl, substitution was allowed at the 2, 4, and 6-positions,
while substitution at the 3- or 5-positions greatly decreased po-
tency. For example, 4-methyl substitution (25) yields a potent ago-
nist (hEC50 = 57 nM), while 3-substitution is completely inactive
(26).
b
NA = not active, <10% activation at 10 lM.
Table 2
Bis-piperidine SAR: eastern aryl optimizationa,b
2
6
3
5
O
R
4
Ar
S
N
N
O
Compd
Ar
R
hGPR119
EC50 (nM)
mGPR119
EC50 (nM)
A substituent survey was conducted for a series of 4-substituted
pyridines, with a bias towards metabolically robust groups (27–
31). It was observed that methyl, chloro, fluoro, cyano and CF3
22
23
24
25
26
27
28
29
30
31
32
2-Pyridyl
H
H
H
39
201
98
57
NA
48
19
18
62
13
5
682
239
887
45
NA
131
206
73
2,5-Pyrazinyl
2,6-Pyrimidyl
2,6-Pyrimidyl
2,6-Pyrimidyl
2-Pyridyl
2-Pyridyl
2-Pyridyl
2-Pyridyl
2-Pyridyl
4-Me
3-Me
4-Cl
4-F
4-Me
4-CN
4-CF3
4-Cl
were acceptable (Table 2).
A potent 4-chloro analog 27
(hEC50 = 48 nM) was identified and profiled in mouse PK to test
this replacement strategy. Comparison of the pyridyl replacement
27 to the Boc parent 6 demonstrates that replacement of the Boc
with a heterocyclic bioisostere greatly improves the PK profile (Ta-
ble 3). Notably, plasma clearance decreased from 120 (ml/min/kg)
for 6 to 2 (ml/min/kg) for agonist 27. From this basis set, the best
substituents were added onto other allowed heteroaryls to identify
other preferred replacements, that is, 4-chloro pyrimidine (e.g.,
32), and 4-methyl pyrimidine/pyrazine (Table 4).
41
8
8
2,6-Pyrimidyl
a
Values are means of at least two independent titrations.
NA = not active, <10% activation at 10 lM.
b
With optimized eastern and western subunits identified, a ma-
trix exercise was initiated to determine the optimal combination of
groups for both termini (Table 4). Compounds were profiled for
their GPR119 potency, hERG selectivity, and efficacy in OGTT.13
The four optimized compounds are shown in Table 4 (33–36). As
predicted, all the agonists possess good to excellent potency, how-
ever, the key issue was insufficient hERG selectivity. Compounds
33 and 34 were disqualified from further profiling due to unaccept-
able hERG activity (33, 34). Surprisingly, while 35 and 36 are
pyrimidine regioisomers, 35 had no in vivo efficacy at 30 mpk. Ulti-
mately, 36 was identified as having the best overall profile from
the bis-piperidines possessing the desired combination of in vitro
potency, OGTT efficacy, and physical properties suitable for high
level dosing (i.e., 750 mpk) and advanced in vivo profiling.
tolerated generally producing inactive compounds; with the
exception being small groups (i.e., fluorine or nitrogen lone pair).
Indications from competitors patents1b led to preparation of a ser-
ies of methyl sulfones. para-Substitution (6) was preferred, how-
ever meta substitution (7) greatly reduced activity. While 6 had
improved in vitro potency by >163-fold, the utility of this com-
pound was limited due to poor solubility in dosing vehicles and
PK (Table 3). To better understand the improvement in potency,
a series of related analogs were prepared 8–10. Sulfoxide 8, and ke-
tone 9 were roughly equipotent to 6, while sulfide 10 lost >27-fold
in potency suggesting that a key contact is formed by an oxygen
lone pair in the para position promoting agonism. Interestingly,
while the primary para amide 11 was a potent agonist, the meta
amide 12 was inactive further supporting the model that a lone
pair at this position is a key feature to potentiate agonism.
While 36 possessed an acceptable profile for early development,
the PP were insufficient to provide a formulation usable for later

2668
J. W. Szewczyk et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2665–2669
Table 4
Optimized agonists combining east and west aryls
R3
N
R4
Ar
N
N
N
R5
Compd
R3, R5
Het
R4
hGPR119
mGPR119
EC50 (nM)a
hERG
OGTTb
Dose (mpk)
OGTTb
EC50 (nM)a
IC50 (nM)a
% glu. Corr.c
33
34
35
36
3-CN, 5-Me
3-Me, 5-CN
3-Me, 5-CN
3-CN, 5-Me
2,6-Pyrimidyl
2,6-Pyrimidyl
2,5-Pyrazinyl
2,5-Pyrazinyl
Cl
2
8
30
19
11
10
82
24
277
30
30
30
37⁄
Me
Me
Me
1.8 K
5.4 K
10.6 K
46⁄
7
1, 3, 10, 30
7, 26⁄, 30⁄, 35⁄
a
b
c
Values are means of at least two independent titrations.
HFF mice (3 weeks on diet) were po dosed qd and the glucose AUC monitored.13
⁄
P values of 60.05 indicated by
.
Table 5
restricted analogs remained active providing an avenue for further
design. Addition of a second methylene gave the homologated cis-
(53) and trans-cyclopropanes (54, 55). The two trans-enantiomers
were separated, however both these analogs suffered from either
loss of potency (54) or poor selectivity (55) (Table 5). Interestingly,
the cis-cyclopropane scaffold is highly optimized for GPR119 agon-
ism (53), possessing picomolar activity and excellent hERG selec-
tivity. It was determined that most of the potency resided in only
one enantiomer, and the absolute configuration for the more po-
tent antipode is shown 60.12 For example, 61 is >250-fold more po-
Core SAR: scaffold optimizationa
N
N
NH
N
Cl
O
O
F
p
n
Compd
n, p
Stereob
hGPR119
EC50 (nM)
mGPR119
EC50 (nM)
hERG
IC50 (nM)
50c
51
52
53
54
55
56
57
–(CH)3–
0, 1
NA
cis
0.2
38
20
0.3
14
3
0.8
9
23
0.4
5
2
36
18
>30 K
1.5 K
1.9 K
>30 K
>30 K
2.6 K
1.2 K
>30 K
0, 1
0, 2
trans
cis
tent (GPR119 hEC50) than the less active isomer (Table 6).
Additionally, the corresponding cis and trans-cyclopropyl regioi-
somers were prepared (56, 57; n = 1, p = 1), however these cores
had insufficient potency and selectivity (Table 5).
0, 2
trans (+)
0, 2
1, 1
trans (À)
cis
18
9
1, 1
trans
A second matrix optimization was conducted combining this
newly identified scaffold and the best eastern/western termini
from the prior SAR. Compounds were selected based on hGPR119
in vitro potency, hERG selectivity, PP (quantified by measured sol-
ubility in PEG400) and efficacy in OGTT. The four best analogs iden-
tified are shown in Table 6. All compounds 58–61 possessed
excellent in vitro potency (Table 6) and mouse PK.15 hERG selectiv-
ity was identified as a key issue, thus compounds were profiled in a
patch clamp hERG assay and determined to have acceptable (61) to
excellent selectivities (58–60). For the western SAR, three new
groups were developed as extensions of the prior SAR, and in-
cluded in the matrix (58–60). It was determined that 2-pyridyl
analogs were equipotent with phenyl, and these pyridyl analogs
had improved log D’s (e.g., 58 and 60) and improved solubilities
(Table 6). Additionally, the potent cyclopropyl amide SAR (Table 5)
was optimized by bioisosteric replacement of the amide with sta-
ble, five-membered heterocycles. While a variety of heterocycles
was explored (e.g., pyrazole, triazoles, tetrazoles, and oxadiazoles),
the 1,2,3-triazoles were ultimately identified as superior (59, 60).
a
b
c
Values are means of at least two independent titrations.
All compounds are racemic except where noted.
Acyclic lead with no cyclopropyl restriction in the methylene chain. Prepared
from commercially available SM.
stage testing, thus a new core with improved PP was required.
There was a correlation between poor formulatability of certain
bis-piperidine analogs and high crystallinity of those solids. It
was postulated that a new core with increased flexibility and/or
less symmetry would improve PP. The initial design was simple
opening of a piperidine ring to form the corresponding acyclic 4
atom chain (with an oxygen for nitrogen replacement) 50. While
this agonist had the desired combination of potency and PP, shortly
after being identified a competitors patent14 appeared necessitat-
ing further core development.
Next, a cyclopropane restriction was incorporated into the
chain and both the (rac)-cis (51) and (rac)-trans (52) isomers were
prepared. While both potency and selectivity were decreased, the
Table 6
Optimized cyclopropyl agonists combining the east and west aryl SAR
N
N
R4
R
Ar
N
O
Compd
Ar
R
R4
hGPR119
mGPR119
EC50 (nM)a
hERG(patchclamp)
IC50 (nM)
Solubilityc
PEG400 (mg/ml)
OGTTb
MEDmax (mpk)
EC50 (nM)a
58
59
60
61
2-Pyridyl
Ph
2-Pyridyl
2,6-Pyrimidyl
4-SO2Me
Cl
Cl
Cl
Me
4
4
2
2
4
6
4
2
19 K
19 K
44 K
4 K
7.1
5.0
11.4
24.8
1–3
10
1–3
0.01–0.03
4-(1,2,3-triazolyl)
4-(1,2,3-triazolyl)
3-Me, 5-CN
a
b
c
Values are means of at least two independent titrations.
HFF mice (3 weeks on diet) were po dosed qd and glucose AUC monitored. MEDmax is the minimal dose for maximal efficacy.13
Equilibrium values reported after comparison at 3 and 7 days.

J. W. Szewczyk et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2665–2669
2669
Ther. Patents 2009, 19, 1339; (c) Shah, U. Curr. Opin. Drug Disc. Dev. 2009, 12,
519; (d) Fyfe, M. C. T.; McCormack, J. G.; Overton, H. A.; Procter, M. J.; Reynet, C.
Expert Opin. Drug Discov. 2008, 3, 403.
Notably, these agonists (58–61) possessed good in vitro poten-
cies, however, it was their physical properties, selectivity and effi-
cacy in OGTT that differentiated these agonists (Table 6).
Solubilities ranged from modest (59) to excellent (61). Each com-
pound 58–61 was titrated in OGTT and the MEDmax (minimal effi-
cacious dose for maximal efficacy) determined.
2. Fredriksson, R.; Höglund, P. J.; Gloriam, D. I.; Lagerström, M. C.; Schiöth, H. B.
FEBS Lett. 2003, 554, 381.
3. (a) Soga, T.; Ohishi, T.; Matsui, T.; Saito, T.; Matsumoto, M.; Takasaki, J.;
Matsumoto, S.; Kamohara, M.; Hiyama, H.; Yoshida, S.; Momose, K.; Ueda, Y.;
Matsushime, H.; Kobori, M.; Furuichi, K. Biochem. Biophys. Res. Commun. 2005,
326, 744; (b) Chu, Z.; Carroll, C.; Alfonso, J.; Gutierrez, V.; He, H.; Lucman, A.;
Pedraza, M.; Mondala, H.; Gao, H.; Bagnol, D.; Chen, R.; Jones, R. M.; Behan, D.
P.; Leonard, J. Endocrinology 2008, 149, 2038; (c) Sakamoto, Y.; Inoue, H.;
Kawakami, S.; Miyawaki, K.; Miyamoto, T.; Mizuta, K.; Itakura, M. Biochem.
Biophys. Res. Commun. 2006, 351, 474.
4. Chu, Z.; Jones, R. M.; He, H.; Carroll, C.; Gutierrez, V.; Lucman, A.; Moloney, M.;
Gao, H.; Mondala, H.; Bagnol, D.; Unett, D.; Liang, Y.; Demarest, K.; Semple, G.;
Behan, D. P.; Leonard, J. Endocrinology 2007, 148, 2601–2609.
5. (a) Arena Pharmaceutical press release, Dec. 15, 2008, San Diego, CA.; (b)
Prosidion Pharmaceutical press release, May. 11, 2009, Melville, NY.; (c)
Metabolex press release, Nov. 12, 2008, Hayward, CA.
6. (a) Peckham, G. E. Abstracts of Papers, 240th ACS National Meeting, Boston, MA,
United States, 2010; Abstract MEDI-199.; (b) Jones, R. M. Abstracts of Papers,
239th ACS National Meeting, San Francisco, CA, 2010; Abstract MEDI-316.; (c)
Wu, Y.; Kuntz, J.D.; Carpenter, A. Abstracts of Papers, 240th ACS National
Meeting, Boston, MA, 2010; MEDI-203.
7. (a) Semple, G.; Fioravanti, B.; Pereira, G.; Calderon, I.; Uy, J.; Choi, K.; Xiong, Y.;
Ren, A.; Morgan, M.; Dave, V.; Thomsen, W.; Unett, D. J.; Xing, C.; Bossie, S.;
Carroll, C.; Chu, Z.; Grottick, A. J.; Hauser, E. K.; Leonard, J.; Jones, R. M. J. Med.
Chem. 2008, 51, 5172; (b) Wu, Y.; Kuntz, J. D.; Carpenter, A. J.; Fang, J.; Sauls, H.
R.; Gomez, D. J.; Ammala, C.; Xu, Y.; Hart, S.; Tadepalli, S. Bioorg. Med. Chem.
Lett. 2010, 20, 2577.
While the in vitro potencies were uniformly excellent (GPR119
hEC50 from 2 to 4 nM) the in vivo efficacy of these agents differed
dramatically (MEDmax from 10 to 0.01 mpk). It is interesting to
compare the correlation of compound solubility in PEG400 and
the MEDmax. Notably, increased solubility correlates with an
impressive increase in OGTT efficacy. Based on the combination
of superior solubility and in vivo potency, 61 was selected for ad-
vanced profiling. Unfortunately, 61 was ultimately disqualified
due to unacceptable QTc prolongation in the CV dog. With this
QTc issue identified 58 was selected for further testing. Compound
58 combines excellent off-target selectivity (i.e., counterscreening
against a panel of 175 receptors identified only two weak, micro-
molar hits) with good in vivo efficacy, and was shown to have no
effects in the CV dog. This systematic combination of eastern, wes-
tern and core SAR investigations lead to a series of agonists that
successfully combine excellent in vivo potency, selectivity, and
PP, and ultimately identified 58 as a candidate suitable for further
preclinical investigation.
8. Wood, H. B.; Adams, A. D.; Freeman, S.; Szewczyk, J. W.; Santini, C.; Huang, Y.;
Mosley, R.; WO Patent 2008076243, 2008.
In summary, a new series of GPR119 agonists is presented. Syn-
thetic methodology was described for the chiral synthesis of the
key chiral cyclopropane scaffold. Information from the systematic
SAR optimization of both terminal ends was successfully merged
onto new agonist scaffolds. These combined optimizations lead
to the identification of highly potent and selective agonists which
successfully combine excellent in vivo potency, selectivity, and
physical properties.
9. Review: Schlummer, B.; Scholtz, U. Adv. Synth. Catal. 2004, 346, 1599. and refs.
therein.
10. Wood, H. B.; Szewczyk, J. W.; Huang, Y.; Adams, A. D. WO Patent 2009129036
A1, 2009.
11. (a) Charette, A. B.; Juteau, H.; Lebel, H.; Molinaro, C. J. Am. Chem. Soc. 1998, 120,
11943; (b) Charette, A. B.; Molinaro, C.; Brochu, C. J. Am. Chem. Soc. 2001, 123,
12160.
12. Absolute configuration determined by X-ray crystallography.
13. Male C57BL6 were fed a high fat diet for 3 weeks prior to the OGTT. OGTT: 4 h
fasted mice were administered compound by oral gavage. One hour later they
received an oral glucose challenge (5 g/kg). Blood glucose was measured via
tail knick using
a glucometer at 0, 20, 40, 60 and 120 min post-glucose
Acknowledgment
challenge. The area under the curve (AUC) for the glucose response was
calculated for each mouse. The glucose AUC of compound treated mice was
compared to that of vehicle-treated mice using an unpaired Student’s t-test or
One-Way ANOVA where appropriate. The data is expressed as the % change of
the Glucose AUC relative to vehicle (% Glu Corr).
We would like to thank Professor Charette for helpful
consultations.
14. Bradley, S. E.; Fyfe, M. C.; Bertram, L. S.; Gattrell, W.; Jeevaratnam, R. P.; Keily,
J.; Procter, M. J.; Rasamison, C. M.; Rushworth, P. J.; Sambrook-Smith, C. P.;
Stonehouse, D. F.; Swain, S. A.; Williams, G. M.; WO Patent 2007003962, 2007
15. Mouse PK for 58 is representative for members of this family: Cl (ml/min/
References and notes
1. For recent reviews: (a) Jones, R. M.; Leonard, J. N. Annu. Rep Med. Chem. 2009,
44, 149; (b) Jones, R. M.; Leonard, J. N.; Buzard, D. J.; Lehmann, J. Expert Opin.
kg) = 2.1; Vdss (L/kg) = 2.1; t1/2 (h) = 12.4; AUCn (lM h kg/mg) = 12.1.