Optimization of triazoles as novel and potent nonphlorizin SGLT2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.04.053

Previous efforts have led to the identification of a potent, selective, and nonphlorizin based SGLT2 inhibitor 1. This Letter describes efforts to further optimize the potency, microsomal stability, solubility and pharmacokinetic properties of this series of SGLT2 inhibitors. From these efforts, compounds 28 and 32 have improved solubility and pharmacokinetic properties compared to compound 1.

Full text of DOI:10.1016/j.bmcl.2011.04.053

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3774–3779
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimization of triazoles as novel and potent nonphlorizin SGLT2 inhibitors
⇑
Xiaohui Du , Mike Lizarzaburu, Simon Turcotte, Taeweon Lee, Joanne Greenberg, Bei Shan, Peter Fan,
Yun Ling, Julio C. Medina, Jonathan Houze
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:
Received 9 February 2011
Revised 6 April 2011
Accepted 12 April 2011
Available online 20 April 2011
Previous efforts have led to the identification of a potent, selective, and nonphlorizin based SGLT2 inhib-
itor 1. This Letter describes efforts to further optimize the potency, microsomal stability, solubility and
pharmacokinetic properties of this series of SGLT2 inhibitors. From these efforts, compounds 28 and
32 have improved solubility and pharmacokinetic properties compared to compound 1
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Noncompetitive SGLT2 inhibitors
Triazoles
SAR
Solubility
Benzooxazinone head groups
Indane tail group
Type 2 diabetes mellitus (T2DM) is a prevalent disease. It is esti-
mated that the patient population has reached 285 million and is ex-
pected to increase to 438 million by 2030.1 It is characterized by
hyperglycemia, which is a major risk factor for microvascular and
macrovascular complications including retinopathy, nephropathy,
neuropathy, and accelerated cardiovascular disease.2 Despite vari-
ous therapeutic options, many patients demonstrate inadequate
glycemic control and remain at risk for chronic complications.3
A novel approach to reduce plasma glucose in T2DM patients is
by inhibition of glucose reabsorption in the kidney.4 Plasma glucose
is normally filtered in the kidney in the glomerulus and actively
reabsorbed in the proximal tubules. Sodium glucose co-transporter
2 (SGLT2)5 is predominantly expressed in the early S1 segment of
the renal proximal tubules and is responsible for an estimated
90% of the glucose reuptake in the kidney. SGLT1,6 a related
sodium-dependent glucose transporter is expressed in small intes-
tine and in the more distal S3 segment of the renal proximal tubule.
The remaining 10% of the glucose reuptake is likely mediated by
SGLT1 and other subtypes. However, inhibition of SGLT1 could have
adverse consequences in the GI tract illustrated by the hereditary
syndrome glucose/galactose malabsorption (GGM). Therefore,
selectivity over SGLT1 is preferred.
known to cause glucosuria in animals and humans.8 However,
phlorizin was not selective against SGLT1 and had poor metabolic
stability. Subsequent research efforts by numerous institutions
identified many phlorizin-based SGLT2-selective inhibitors9 such
as dapagliflozin10 (Fig. 1).
Recent phase III clinical trial results showed that dapagliflozin
significantly improved fasting plasma glucose levels, lowered
HbA1c level, and the body weight of the patients, thus demonstrat-
ing the utility of SGLT2 inhibitors as anti-diabetic agents.11
Our previous efforts12 resulted in a series of potent nonphlorizin
based SGLT2 inhibitors exemplified by compound 1. It had a 10 nM
IC50 against the human SGLT2 transporter in our [14C] glucose
transporter assay13 and was selective against hSGLT1 (IC50 = 9
lM).
However, this compound showed poor microsomal stability and
high in vivo clearance in rat. Administration of compound 1 in
100 mg/Kg sc in mouse showed inconsistant results. Attempted
Cl
Cl
OEt
O
N
O
HO
HO
Cl
O
N
Cl
OH
N
OH
N
It has been demonstrated that selective inhibition of SGLT2 re-
sulted in urinary glucose excretion and exhibited a glucose-lower-
ing effect.7 The majority of current SGLT2 inhibitors were based on
phlorizin, a naturally-occurring O-aryl glycoside that has long been
F3C
Dapagliflozin
3 nM
1
hSGLT2
IC50
10 nM
⇑
Corresponding author.
E-mail address: xdu@amgen.com (X. Du).
Figure 1. Structure of dapagliflozin and our nonphlorizin based inhibitor triazole 1.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.04.053

X. Du et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3774–3779
R1
3775
R1
O
R1
OH
O
O
a
b
OEt
R2
NO2
R2
NO2
R2
N
H
O
I
II
O
O
NH
S
c
d
R3
R3
R3
OH
NH
R4
R4
III
IV
R4
N
R4
N
R3
e
f
N
R3
g
N
HO
N
N
VI
V
R1
O
R4
R2
h
N
O
N
N
R3
Cl
N
N
N
N
R3
VIII
VII
R4
Scheme 1. General synthesis of the triazole SGLT2 inhibitors. Representative yields in parentheses are for the preparation of compound 32. Reagents and conditions: (a) ethyl
2-bromoacetate, K2CO3, CH3CN (R1 = R2 = Me, 97%); (b) Fe, acetic acid, 75 °C (R1 = R2 = Me, 82%); (c) SOCl2, DCM, DMF (catalytic), amine (for R3COOH = dihydro-indene
carboxylic acid, R4 = propylamine, 80%); (d) Lawesson’s reagent, toluene, 80 °C, 2 h. (74%); (e) formohydrazide, Hg(OAc)2, reflux, overnight (22%); (f) paraformaldehyde,
xylene, 125 °C (57%); (g) SOCl2, DCM; (h) II, K2CO3, NaI, DMF, 65 °C (53%).
Table 1
Modifications of the C8-substituents
X
8
O
N
2
7
Cl
O
N
Cl
N
N
F3C
Compound #
X group
hSGLT2 inhibition
Microsomal stabilityb
(r/h) % remaining
a
IC50 (nM)
% Inhibition
1
2
3
4
5
6
7
8
9
–Cl
10
57
10000
36
9800
21
22
100
100
95
100
95
100
100
100
99
<5/4
–COOMe
–COOH
–NH2
–NHCOMe
–N(SO2Me)2
–NO2
<5/<5
68/95
14/10
0.1/0.4
1/0.1
10/1
1.2/0.8
0.4/0.2
–CN
–SO2Me
59
660
a
Inhibition of [14C] AMG transportation into the CHO cells that express hSGLT2. See Ref. 13 for assay protocol.
See Ref. 14 for assay protocol. r: rat; h: human.
b
formulation for high dose delivery into mouse also failed. Both re-
sults are likely due to the poor solubility of this compound (1.6 lg/
mL in PBS buffer). In this Letter, we would like to report our further
efforts in optimizing the SAR of this series to improve the phar-
mocokinetic properties of the inhibitors.
Incubation of compound 1 with rat microsomes showed that
oxidation occurred both on the benzooxazinone head group and
the chloro-phenylethyl tail group. Both these areas were targeted
for modification with our initial goal being improving the micro-
somal stability and consequently the in vivo clearance.

3776
X. Du et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3774–3779
Table 2
SAR of the tail aromatic groups
Cl
O
Cl
N
O
N
N
R
N
F3C
Compound #
R group
hSGLT2 inhibition
Microsomal stabilityb
(r/h) % remaining
a
IC50 (nM)
% Inhibition
100
1
10
<5/4
Cl
10c
11
12
89
480
73
100
100
98
<5/<5
<5/<5
<5/<5
O
O
Cl
13
14
8
100
100
0/<5
5/<5
23
Cl
Cl
N
N
15
16
13
3
100
100
<5/7
2/20
N
17
10
12
100
100
3/3
S
18c
19c
<5/8
14
13
100
100
<5/<5
<5/<5
20
a
Inhibition of [14C] AMG transportation into the CHO cells that express hSGLT2. See Ref. 13 for assay protocol.
See Ref. 14 for assay protocol. r: rat; h: human.
Compounds are racemic.
b
c
Scheme 1 shows the general synthetic sequences to obtain our
basic conditions with benzooxazinones II to make the final inhib-
itor VIII.
SGLT2 inhibitors. Benzooxazinone head groups II were generally
made by reacting di-substituted nitro phenol with ethyl 2-bromo-
acetate followed by reductive cyclization of the nitro ester I. The
synthesis of the remainder of the inhibitor started from the tail
carboxylic acid. Coupling with a suitable amine generated the
amide III which was converted to thioamide IV with Lawesson’s
reagent. Triazole formation of IV with formohydrazide was cata-
lyzed by mercury acetate. Subsequent hydroxymethylation of
the triazole V and conversion to the chloride with thionyl chlo-
ride led to compound VII. Compound VII was then coupled under
Polar groups were introduced onto the benzooxazinone to see if
potency could be maintained and to improve microsomal stability
(Table 1). Since C7 substitution was not well-tolerated12, attention
was focused on C8 position. As shown in Table 1, the C-8 position
tolerated a variety of polar groups including methylester (2), ani-
line (4), bis-sulfonamide (6), nitro (7) and cyano (8) groups. Com-
pounds with highly polar carboxylic acid (3), acetamide (5), and
sulfone (9) substitutions showed much weaker inhibition. Most
compounds were not stable in rat and human liver micorsomes

X. Du et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3774–3779
3777
Table 3
Combinations of head benzooxazinones groups and tail aromatic groups improved microsomal stability
R1
O
Cl
N
O
N
N
N
F3C
N
R2
a
Compound #
R1 group
R2 group
IC50 (nM) (hSGLT2 inhibition)
Microsomal stabilityb % remaining (m/r/h)
21
22
23
24
–NO2
–NO2
–NO2
–CN
–H
41
7
7
78/19/73
77/6/51
41/29/68
70/28/80
–Cl
–CF3
–H
110
a
Inhibition of [14C] AMG transportation into the CHO cells that express hSGLT2. See Ref. 13 for assay protocol.
See Ref. 14 for assay protocol. m: mouse; r: rat; h: human.
b
except carboxylic acid 3. Selective compounds were tested for its
stability in mouse liver micorsome for further comparison. Com-
pounds 7 and 8 with strongly electron-withdrawing nitro and cya-
no groups showed slightly better mouse liver microsome stability
with 25% and 16% remaining compared to less than 5% remaining
of compound 1.
Simultaneously, methyl substitutions and oxygen were intro-
duced into the tail aromatic group area in an attempt to slow down
benzylic oxidation by microsomes (Table 2). Mono-substitution at
the benzylic position (10) was tolerated but di-substitution signif-
icantly decreased the potency as shown with compound 11. Oxy-
gen replacement of the benzylic carbon in compound 12 also
resulted in a significant loss of potency.
Linker length variation between triazole and tail aromatic group
was investigated. Compound 13 with a three atom linker main-
tained potency, but was not stable in microsomes. Therefore, the
linker was truncated to see whether potency could be maintained
and stability improved. Compound 14 with the aryl group directly
attached to the triazole maintained potency. The chloro-phenyl
group in compound 14 could be replaced by a more polar pyridine
group (15, 16) and other heterocycles such as thiazole (17) with
improvement in potency.
Furthermore we sought to conformationally constrain the tail
linker in an attempt to reduce the flexibility of the structures. A
number of constrained analogs such as cyclopropyl linker 18, tetra-
hydronaphthalene 19 and indane 20 all showed potent SGLT2
inhibition.
than compound 21 with a simple pyridine tail. All four combina-
tions brought significant improvement in mouse and human liver
microsomal stability.
In addition to its potency and improved in vitro stability, com-
pound 22 also showed a low clearance of 0.79 L/h/Kg in rat dosed
0.5 mg/Kg. However, when compound 22 was delivered into
mouse or rat through high dose ip administration, it showed low
exposure in both animal species. The poor solubility of this com-
pound (2.2 lg/mL in PBS buffer) could be preventing its absorption
in vivo. This result led us to focus on improving the solubility of our
inhibitors to improve the potential exposure of our compounds.
During our SAR optimization, N-propyl substitution on the tria-
zole was found to be equally potent to the N-trifluoropropyl substi-
tution. The substitution was therefore changed to N-propyl to
reduce the hydrophobicity of the overall structures.
In addition, another strategy to improve the solubility was to
reduce the number of chlorine atoms in the inhibitors (Table 4).
Compound 26 with a methyl group replacing the 6-chloro in
the 6-chloro, 8-nitro bezooxazinone head group in compound
25 and compound 27 with a 6,8-dimethyl bezooxazinone head
group both resulted in equally-potent inhibitors. Removing the
chlorine in the tail pyridine moiety significantly improved solu-
bility. Compound 28 had a marked improvement in solubility
(64 lg/mL) while maintaining a potent inhibition of 12 nM. The
loss of potency due to the removal of the chlorine in the tail moi-
ety was compensated partially by the potent di-methyl bezooxaz-
inone head group.
Compound 16 showed a slight improvement in human liver
microsome stability. Compound 15 showed a slight improvement
in mouse liver microsome stability with 15% remaining compared
to less than 5% remaining of compound 1. The lack of general sta-
bility of these structures regardless of the tail aromatic groups re-
flected the ease of oxidation of thc 6,8-dichlorobenzoxazinone
head group.
Having examined each area of the molecule separately, we com-
bined the most stable benzooxazinone head groups with our most
stable tail aromatic groups (Table 3). Compound 21 with chloro-ni-
tro substituted head group was more potent than compound 24
with chloro-cyano substituted head group. Compounds 22 and
23 with a substituted pyridine tail were consistently more potent
One helpful discovery during the process was shown in com-
pound 29. Though 3-methyl substitution decreased the inhibition
of hSGLT2, the compound’s solubility was markedly increased pre-
sumably due to the greater dihedral angle between the triazole and
the pyridine promoted by the methyl group. This led us to incorpo-
rate the linkers as shown in compounds 30–32 to disrupt the pla-
nar alignment of the triazole and pyridine to improve solubility.
Indeed, compounds 30–32 all showed satisfactory solubility. Apart
from high potency, compounds 28 and 32 were synthetically more
accessible by having no chiral center in their structures compared
to compounds 30 and 31. Both compounds 28 and 32 also had
great selectivity against hSGLT1 with both IC50 being 30
lM on
hSGLT1. These two compounds were evaluated for their rodent PK.

3778
X. Du et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3774–3779
Table 4
More SAR optimizations to improve the solubility of the inhibitors
head
N
N
N
tail
a
Compound #
Head group
Tail group
IC50 (nM)
Solubility (lg/mL)
N
NO2
O
N
25
5
4
1
Cl
Cl
Cl
Cl
O
NO2
N
O
26
2
N
O
O
O
N
N
O
N
27
28
3
1.1
O
N
12
64
NO2
N
O
N
29
5100
80
Cl
O
O
N
30
31
32
8
10
5
19
23
35
O
O
O
O
N
O
N
a
Inhibition of [14C] AMG transportation into the CHO cells that express hSGLT2. See Ref. 13 for assay protocol.
Table 5
intestinal absorption to saturate first-pass metabolism resulting in
good plasma concentrations. Both compounds were formulated in
high dose and dosed in mouse with sc administration. The PK
parameters were shown in Table 5. At 100 mg/Kg, compound 28
reached nine-fold of IC50 at Cmax calculated by free unbound frac-
tion. At 70 mg/Kg, compound 32 had about six-fold of IC50 at Cmax
calculated by free unbound fraction.
In conclusion, we have further optimized our triazole series of
hSGLT2 inhibitors with a variety of head benzooxazinone and no-
vel tail aromatic groups. Both areas tolerated polar groups such
as a cyano group and a pyidine group. Many inhibitors had better
potency and solubility compared to our previous lead inhibitor
compound 1. More importantly, apart from the structural differ-
ences, compounds 28 and 32 both delivered micromolar plasma
concentrations in rats. Future efforts in this series will focus on
High dose pharmacokinetic parameters of 28 and 32 in rat
Compound #
28
32
SCa (rat)
Cmax
AUC0–24 h
(
l
M)
(
1.05
16.2
8
3%
100
3.5
3.5
61.6
24
0.5%
70
3
l
M/h)b
Tmax (h)
Fraction unbound
Dose (mg/Kg)
Rat SGLT2 IC50 (nM)
Microsomal stability % remaining (rat/human)
<5/20
<5/<5
a
For 28 and 32: n = 3.
Due to the depot effect of both 28 and 32, AUC was calculated up to 24 h.
b
Despite relatively high microsomal turnover, it was felt that the
improved solubility of compounds 28 and 32 would allow sufficient

X. Du et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3774–3779
3779
M.; Kawakita, Y.; Fukasawa, Y.; Iida, I.; Hagima, N.; Takeuchi, H.; Chino, Y.;
Asami, J.; Okumura-Kitajima, L.; Io, F.; Yamamoto, D.; Miyata, N.; Takahashi, T.;
Uchida, S.; Yamamoto, K. J. Med. Chem. 2010, 53, 3247; (c) Lv, B.; Xu, B.; Peng,
K.; Xu, G.; Du, J.; Zhang, L.; Zhang, W.; Zhang, T.; Zhu, L.; Ding, H.; Sheng, Z.;
Welihinda, A.; Seed, B.; Chen, Y. Bioorg. Med. Chem. Lett. 2009, 19, 6877; (d) Kim,
M. J.; Lee, J.; Kang, S. Y.; Lee, S.-H.; Son, E.-J.; Jung, M. E.; Lee, S. H.; Song, K. S.;
Lee, M.; Han, H.-K.; Kim, J.; Lee, J. Bioorg. Med. Chem. Lett. 2010, 19, 6877.
10. Meng, W.; Ellsworth, B. A.; Nirschl, A. A.; McCann, P. J.; Patel, M.; Girotra, R. N.;
Wu, G.; Sher, P. M., et al J. Med. Chem. 2008, 51, 1145.
introducing polar groups to the tail groups for better solubility, free
fraction and potentially lower in vivo clearance.
References and notes
1. Norman, P. Diabetes Pipeline: Intense Activity to Meet Unmet Need Insight Pharma
Reports, Cambridge Health Institute, September 2010.
2. (a) Nazimek-Siewniak, B.; Moczulski, D.; Grzeszczak, W. J. Diabetes
Complications 2002, 16, 271; (b) Kles, K. A.; Vinik, A. I. Curr. Diab. Rev. 2006,
2, 131; (c) Rahman, S.; Rahman, T.; Ismail, A. A.; Rashid, A. R. Diabetes Obes.
Metab. 2007, 9, 767.
11. (a) Komoroski, B.; Vachharajani, N.; Feng, Y.; Li, L.; Kornhauser, D.; Pfister, M.
Clin. Pharmacol. Ther. 2009, 85, 513; (b) List, J. F.; Woo, V.; Morales, E.; Tang, W.;
Fiedorek, F. T. Diabetes Care 2009, 32, 650; (c) Wilding, J. P. H.; Norwood, P.;
J’joen, C.; Bastien, A.; List, J. F.; Fiedorek, F. T. Diabetes Care 2009, 32, 1656.
12. An-Rong, Li.; Zhang, K.; Greenberg, J.; Lee, T.; Liu, J. Bioorg. Med. Chem. Lett.
2011, 21, 2472.
3. Blonde, L. Am. J. Manag. Care 2007, 13, S36.
4. Handlon, A. L. Expert Opin. Ther. Pat. 2005, 15, 1431.
5. (a) Wright, E. M.; Hirayama, B. A.; Loo, D. F. J. Intern. Med. 2007, 261, 32; (b)
Santer, R.; Kinner, M.; Lassen, C. L.; Schneppenheim, R.; Eggert, P.; Bald, M.;
Brodehl, J.; Daschner, M.; Ehrich, J. H. H.; Kemper, M.; Volti, S. L.; Neuhaus, T.;
Skovby, F.; Swift, P. G. F.; Schaub, J.; Klaerke, D. J. Am. Soc. Nephrol. 2003, 14,
2873.
6. (a) Wright, E. M.; Hirayama, B.; Hazama, A.; Loo, D. D.; Supplisson, S.; Turk, E.;
Hager, K. M. Soc. Gen. Physiol. Ser. 1993, 48, 229; (b) Wright, E. M.; Turk, E.;
Hager, K.; Lescale-Matys, L.; Hirayama, B.; Supplisson, S.; Loo, D. D. F. Acta
Physiol. Scand. Suppl. 1992, 146, 201.
7. (a) Oku, A.; Ueta, K.; Arakawa, K.; Ishihara, T.; Nawano, M.; Kuronuma, Y.;
Matsumoto, M.; Saito, A.; Tsujihara, K.; Anai, M.; Asano, T.; Kanai, Y.; Endou, H.
Diabetes 1999, 48, 1794; (b) Katsuno, K.; Fujimori, Y.; Takemura, Y.; Hiratochi,
M.; Itoh, F.; Komatsu, Y.; Fujikura, H.; Isaji, M. J. Pharmacol. Exp. Ther. 2007, 320,
323; (c) Ueta, K.; Yoneda, H.; Oku, A.; Nishiyama, S.; Saito, A.; Arakawa, K. Biol.
Pharm. Bull. 2006, 29, 114.
13.
[
14C] Glucose transporter assay Chinese hamster ovary (CHO) cells stably
expressing human SGLT2 cDNA were seeded in 96-well Cytostar scintillant
coated plates (PerkinElmer Biosciences). Cells were rinsed briefly with assay
buffer (10 mM HEPES, 5 mM tris, 140 mM NaCl, 2.5 mM KCl, 1.25 mM CaCl2,
pH 7.4). Ligands were serially diluted via BioMek 3000 (Beckman Coulter, Inc.)
in the assay buffer supplemented with 0.1% BSA. After addition of ligands to
the assay plate, [14C]-methyl glucopyranoside ([14C] AMG, 10 uCi/ml, Moravek)
was added and incubated at 37 °C incubator for 60 min. Assay plate was briefly
washed with the buffer before air drying, and counted in Microbeta plate
counter (PerkinElmer Biosciences).
14. Microsome stability were measured in a high throughput format by incubating
inhibitors at 1 M concentration with rat (r) or human (h) microsomes and the
l
percentage of the parent compounds remaining were measured after 30 min.
of incubation by liquid chromatography/mass spectrometry analysis. Only
selective compounds were measured of its mouse liver microsomal stability
manually.
8. Ehrenkranz, J. R. L.; Lewis, N. G.; Kahn, C. R.; Roth, J. Diabetes Metab. Res. Rev.
2005, 21, 31.
9. Some of the most recent review and papers are: (a) Washburn, W. N. J. Med.
Chem. 2009, 52, 1785; (b) Kakinuma, H.; Oi, T.; Hashimoto-Tsuchiya, Y.; Arai,