Discovery of non-glucoside SGLT2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.02.056

A series of benzothiazinone and benzooxazinone derivatives were discovered as SGLT2 inhibitors. The optimization led to the discovery of compounds 31 and 32, which exhibited similar potency and better SGLT1 selectivity compared to dapagliflozin. These compounds may provide novel promising scaffolds, which are different from phlorizin-based SGLT2 inhibitors.

Full text of DOI:10.1016/j.bmcl.2011.02.056

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2472–2475
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of non-glucoside SGLT2 inhibitors
⇑
An-Rong Li , Jian Zhang, Joanne Greenberg, TaeWeon Lee, Jiwen Liu
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:
A series of benzothiazinone and benzooxazinone derivatives were discovered as SGLT2 inhibitors. The
optimization led to the discovery of compounds 31 and 32, which exhibited similar potency and better
SGLT1 selectivity compared to dapagliflozin. These compounds may provide novel promising scaffolds,
which are different from phlorizin-based SGLT2 inhibitors.
Received 6 January 2011
Revised 5 February 2011
Accepted 14 February 2011
Available online 24 February 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Non-glucoside
SGLT2
Inhibitor
Glucose
Diabetes
Sodium glucose co-transporter 2 (SGLT2) is primarily expressed
in the S1 segment of renal proximal tubules and accounts for 90%
of renal glucose reabsorption.1 Inhibition of SGLT2 prevents renal
glucose reuptake from the glomerular filtrate and subsequently
lowers the blood glucose level. The efficacy and safety of this insu-
lin independent treatment has been validated by human genetic
evidence and late stage clinical trials with SGLT2 inhibitors.
Human with genetic defects of SGLT2 excrete as much as
140 g/day of glucose in urine (180 g/day filtered in the kidney),
but live a normal lives without adverse effect on carbohydrate
metabolism.2 Phase II and III clinical trials of dapagliflozin, the
leading SGLT2 inhibitor in development, showed dose-dependent
increases in urinary glucose excretion, reduction in fasting and
postprandial blood glucose levels and moderate weigh loss.3 All
the evidence has demonstrated SGLT2 to be the promising target
for treatment of type 2 diabetes mellitus.
Another well-characterized subtype in the renal glucose
reuptake is SGLT1, which is responsible for the other 10% of renal
glucose reabsorption. In addition to its expression in the kidney,
SGLT1 is also highly expressed in small intestine and responsible
for absorption of dietary glucose and galactose. Inhibition of SGLT1
is likely to cause severe side effect, as demonstrated by mutations
of SGLT1, which have been associated with glucose-galactose mal-
absorption syndrome in humans. The syndrome is characterized by
severe diarrhea and dehydration on a glucose or galactose-contain-
ing diet.4 Therefore achieving SGLT1 selectivity for SGLT2 is
important.
With few exceptions,5–7 most of the reported SGLT2 inhibitors
are glucoside derivatives and can be traced back to a naturally-
occurring O-aryl glucoside phlorizin (1). It is a fairly potent inhib-
itor of SGLT2 (IC50 = 36 nM), with some selectivity against SGLT1
(IC50 = 330 nM).8 It is well documented that administration of 1
caused glucosuria in humans and animals.9 However, the lack of
selectivity and metabolic instability prevent its further application.
Currently, dapagliflozin (2),3 BI10773,10 and canagliflozin11 as
C-aryl glucosides are undergoing phase III clinical trials. Several
other SGLT2 inhibitors, including ASP194,12 TA-7284,13 CSG45214
and LX421115,16 are in phase II clinical trials.
It is somewhat puzzling to see the overwhelming majority of
the reported SGLT2 inhibitors are phlorizin related glucosides.
They may be unique for SGLT2 inhibition, but structurally diverse
SGLT2 inhibitors may offer unforeseen advantages. In this Letter,
we describe our efforts to discover a series of novel non-glucoside
SGLT2 inhibitors with high potency and SGLT1 selectivity. To our
knowledge, this is the first description of an industry effort from
HTS to optimization for non-glucoside SGLT2 inhibitors in a peer
reviewed Letter (see Fig. 1).
High-throughput screening identified compound 3 as a SGLT2
inhibitor (Table 1). Compound
14C]-AMG17 into SGLT2 expressed CHO cell with IC50 of 1.4
for human, 0.4 M for rat and 1.2 M for mouse. The selectivity
3 inhibited transportation of
[
lM
l
l
of SGLT2 versus SGLT1 is less than 25-fold across different species
as shown in Table 1. Herein, we describe our lead optimization
efforts to improve both activity and selectivity of compound 3.
A series of compound 3 analogues was prepared as illustrated in
Scheme 1. Compound 5 was synthesized by heating hydra-
zinecarbothioamide 4 and 2-hydroxyacetate in the presence of
base. Alkylation of 5 led to key intermediate 6, which was then
⇑
Corresponding author.
E-mail address: anrongl@amgen.com (A.-R. Li).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.02.056

A.-R. Li et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2472–2475
2473
ing paraformaldehyde with 13. Final compounds 23–32 were ob-
tained using previously described Mitsunobu conditions (step c
of Scheme 1).
Initially, N-substitutions of the triazole were explored. Replace-
ment of methyl by ethyl resulted in 3-fold of increased activity of
SGLT2 and slightly improved selectivity versus SGLT1. Larger
groups, such as phenyl, cyclopropyl and isopropyl were also toler-
ated (Table 2, 8–10).
HO
O
OH
O
OH
O
HO
HO
OH
OH
Phlorizin
Compound 3 showed high metabolic instability with less than
2% remaining after incubation in human or rat microsomes for
30 min. In order to increase the microsomal stability, the replace-
ment of the sulfur atoms was studied. We were pleased to discover
that benzothiazinone can be replaced by benzooxazinone without
a loss in activity (compare 7 and 15 in Table 3). Therefore, ben-
zooxazinones were used for subsequent SAR studies.
1
Cl
OEt
O
HO
HO
OH
OH
Evaluation of phenyl substitutions of benzooxazinone demon-
strated that a variety of lipophilic groups can be tolerated on 6, 7
or 8 position (Table 3). 6-Chloro, bromo, methyl and 8-chloro ana-
logues (17, 18, 19 and 22) displayed an approximately 2- to 3-fold
increased activity compared to 6-trifluoromethyl counterpart (15).
However, replacement of 6-trifluoromethyl by a 6-nitro group re-
sulted in 10-fold loss of potency (16). Interestingly, variation of
these substitutions (at the 6 or 8 position) has no impact in SGLT1
activity. Introduction of halogen at the 7-position (Cl or Br) slightly
decreased the SGLT2 activity (20 and 21).
Dapagliflozin
2
Figure 1.
Table 1
Activity of the HTS hit (3) across species
S
Replacement of sulfur next to triazole by a carbon as in com-
pound 15 provided compound 23, which is only 2-fold less potent
than 15 for SGLT2 (Table 4).
F
F
N
O
N
F
N
With the two sulfur atoms replaced, we decided to revisit N-
substitutions of the triazole as shown in Table 4. n-Propyl (25)
N
S
was as optimal as other bigger substituents (26, 27, 29).
Branched substituent resulted in significant loss of activity (30).
a-
Cl
hSGLT1 inhibition
3
Species
hSGLT2 inhibition
Since both 6-chloro and 8- chloro were beneficial for SGLT2
activity (Table 3, compounds 18 and 22), 6, 8-dichlorosubstituent-
ed compounds 31 and 32 (Table 5) were synthesized and indeed
exhibited much better potency in hSGLT2 glucose uptake assay,
comparable to that of dapagliflozin (2). They also displayed more
than 1000-fold selectivity toward SGLT2 versus SGLT1 and SGLT6,
which were better than Dapagliflozin (2). In addition, compounds
31 and 32 displaced tritium labeled dapagliflozin with IC50 of
16 nM and 20 nM respectively.18 However, these two compounds
had poor microsomal stability with 4% remaining after incubation
in human or rat microsomes for 30 min.
a
a
IC50
(
l
M)
IC50 (lM)
Human
Rat
Mouse
1.40
0.40
1.20
7.00
9.00
3.00
Inhibition of [14C]AMG transportation into the CHO cells that express hSGLT2 or
a
hSGLT1. Assay conducted in a buffer containing 0.1% BSA. See Ref.17 for assay
protocol. Values are means of three experiments, standard deviation is 30%.
reacted with benzothiazinone or benzooxazinnone to provide com-
pounds 3, 7–10, 15–22.
In summary, we discovered and optimized a series of non-glu-
coside SGLT2 inhibitors. Very potent and highly selective SGLT2
inhibitors were obtained. N-substituents of triazole and phenyl
substituents of benzooxazinone play a key role in activity and
selectivity. Further optimization to improve solubility and pharma-
cokinetic properties will be reported in due course.
In addition, the synthesis of another series of analogues with
different substituents on the right-hand side is shown in Scheme
2. Acid 11 was coupled with amines to provide various amides,
which were then converted to intermediates 12 by reacting with
Lawesson’s reagent. Cyclization with formic acid hydrazide yielded
13. The installation of hydroxymethyl group was achieved by heat-
HO
R
N
N
HO
N
H
N
H
N
b
a
N
N
S
N
R
NH2
S
R
N
S
4
c
H
C
l
6
O
5
N
N
X
RN
S
Cl
R1
3, 7-10,15-22
Scheme 1. Reagents and conditions: (a) methyl 2-hydroxyacetate, NaOH, EtOH, reflux, 95%; (b) 1-(bromomethyl)-4-chlorobenzene, potassium carbonate, DMF, rt, 1 h, 90%;
(c) DEAD, Ph3P, THF, rt, 50%–65%.

2474
A.-R. Li et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2472–2475
R
N
S
c
Cl
O
a, b
Cl
Cl
HN
N
N
N
HO
R
12
11
13
O
R
d
N
N
e
N
Cl
HO
X
RN
R1
N
N
Cl
23-32
14
Scheme 2. (a) EDC, HOBt, RNH2, DMF, rt, 100%; (b) Lawesson’s reagent, toluene, 85 °C, 3 h, 99%; (c) formic acid hydrazide, mercuric acetate, dioxane, reflux, 60%–70%; (d)
paraformaldehyde, xylene, 125 °C, 2 h, 80%–90%; (e) DEAD, Ph3P, THF, rt, 65%.
Table 2
Table 4
N-substituents of triazole
N-substituents of triazole
S
N
O
N
F
F
R1
O
N
O
N
F
N
N
N
N
R
R2
S
Cl
Cl
Compound
R
hSGLT2 inhibition
hSGLT1 inhibition
Compound
R1
R2
hSGLT2 IC50
M)
hSGLT1 IC50
M)
a
a
a
a
IC50
(lM)
IC50
(lM)
(
l
(l
3
7
8
9
10
Me
Et
CH(CH3)2
Cyclopropyl
Ph
1.40
0.37
1.20
1.10
2.11
7.00
4.75
2.00
3.75
3.47
23
24
25
26
27
28
29
30
CF3
CF3
Cl
Cl
Cl
Cl
Cl
Cl
CH2CH3
0.64
0.18
0.04
0.03
0.03
0.16
0.07
1.58
46.00
9.88
13.60
11.80
7.28
6.90
13.80
25
(CH2)2CH3
(CH2)2CH3
(CH2)3CH3
CH2CH(CH2)2
CH2CH(CH3)2
CH2CH2CF3
CH(CH2)2
Inhibition of [14C]AMG transportation into the CHO cells that express hSGLT2 or
a
hSGLT1. Assay run in buffer containing 0.1% BSA. See Ref.17 for assay protocol.
Values are means of three experiments, standard deviation is 30%.
Inhibition of [14C]AMG transportation into the CHO cells that express hSGLT2 or
a
hSGLT1. Assay run in buffer containing 0.1% BSA. See Ref.17 for assay protocol.
Values are means of three experiments, standard deviation is 30%.
Table 3
Phenyl substitutions of benzooxazinone and benzothiazinone
8
X
N
Table 5
7
Activity and selectivity of top two compounds
R
6
Cl
O
N
5
O
7
N
N
Cl
N
R
O
N
S
5
Cl
hSGLT1 inhibition
N
N
Compound
R
X
hSGLT2 inhibition
a
a
IC50
(
l
M)
IC50 (lM)
Cl
7
15
16
17
18
19
20
21
22
6-CF3
6-CF3
6-NO2
6-Br
6-Cl
6-Me
7-Cl
S
0.37
0.25
2.10
0.09
0.15
0.08
0.43
0.52
0.15
4.75
6.64
5.61
5.32
7.34
7.13
3.29
2.53
7.34
O
O
O
O
O
O
O
O
Compound
R
hSGLT2
hSGLT1
M) IC50 (lM) IC50 (l
hSGLT6
hSGLT2
M) plasma IC50
M)
a
a
a
b
IC50
(l
(
l
31
32
2
(CH2)2CH3 0.009
(CH2)2CF3 0.010
0.004
9.14
6.29
0.37
14.40
9.20
0.38
0.04
0.019
0.022
7-Br
8-Cl
Inhibition of [14C]AMG transportation into the CHO cells that express hSGLT2 or
a
hSGLT1. Assay run in buffer containing 0.1% BSA. See Ref.17 for assay protocol.
Values are means of three experiments, standard deviation is 30%.
Inhibition of [14C]AMG transportation into the CHO cells that express hSGLT2 or
hSGLT1. Assay run in buffer containing 0.1% BSA. See Ref. 17 for assay protocol.
Values are means of three experiments, standard deviation is 30%.
a
b
Assay conducted in the presence of 100% plasma.

A.-R. Li et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2472–2475
2475
14. Kobayashi, T.; Sato, T.; Nishimoto, M. WO2006080421, 2006; Chem. Abstr.
2006, 145, 189115.
References and Notes
15. Goodwin, N. C.; Harrison, B. A.; Kimball, S. D.; Mabon, R.; Rawlins, D. B. US
Patent 20080221164, 2008; Chem. Abstr. 2008, 149, 356147.
16. Goodwin, N. C.; Mabon, R.; Harrison, B. A.; Shadoan, M. K.; Almstead, Z. Y.; Xie,
Y.; Healy, J.; Buhring, L. M.; DaCosta, C. M.; Bardenhagen, J.; Mseeh, F.; Liu, Q.;
Nouraldeen, A.; Wilson, A. G. E.; Kimball, S. D.; Powell, D. R.; Rawlins, D. B. J.
Med. Chem. 2009, 52, 6201–6204.
1. Kanai, Y.; Lee, W.-S.; You, G. J. Clin. Invest. 1994, 93, 397.
2. Van den Heuvel, L. P.; Assink, K.; Willemsen, M.; Monnens, L. Hum. Genet. 2002,
111, 544.
3. Komoroski, B.; Vachharajani, N.; Boulton, D.; Kornhauser, D.; Geraldes, M.; Li,
L.; Pfister, M. Clin. Pharmacol. Ther. 2009, 85, 520.
4. Wright, E. M.; hirayama, B. A.; Loo, D. F. J. Int. Med. 2007, 261, 32.
5. Klein, M.; Gericke, R.; Beier, N.; Cezanne, B.; Tsaklakidis, C.; Mederski, W. PCT.
Int. Appl. WO2008/046497 A1.
6. Mederski, W.; Beier, N.; Burgdorf, L. T.; Gericke, R.; Klein, M.; Tsaklakidis, C.
PCT. Int. Appl. WO2008/071288 A1.
17.
[
14C] Glucose transporter assay: Chinese hamster ovary (CHO) cells stably
expressing human SGLT1 and
2 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
7. Burgdorf, L. T.; Cezanne, B.; Klein, M.; Gericke, R.; Tsaklakidis, C.; Mederski, W.;
Beier, N. PCT. Int. Appl. WO2008/101586 A.
addition of ligands to the assay plate,
[ a-methyl glucopyranoside
14C]
8. Meng, W.; Ellsworth, B. A.; Nirschl, A. A.; McCann, P. J.; Patel, M.; Girotra, R. N.;
Wu, G.; Sher, P. M.; Morrison, E. P.; Biller, S. A.; Zahler, R.; Deshpande, P. P.;
Pullockaran, A.; Hagan, D. L.; Morgan, N.; Taylor, J. R.; Obermeier, M. T.;
Humphreys, W. G.; Khanna, A.; Discenza, L.; Robertson, J. G.; Wang, A.; Han, S.;
Wetterau, J. R.; Janovitz, E. B.; Flint, O. P.; Whaley, J. M.; Washburn, W. N. J. Med.
Chem. 2008, 51, 1145.
9. Ehrenkranz, J. R. L.; Lewis, N. G.; Kahn, C. R.; Roth, J. Diabetes Metab. 2005, 21,
31.
10. Eckhardt, M.; Eickelmann, P.; Himmelsbach, F.; Barsoumian, E. L.; Thomas, L.
WO20050092877, 2005; Chem. Abstr. 2005, 143, 286629.
11. Nomura, S.; Kawanishi, E.; Ueta, K. WO2005012326, 2005; Chem. Abstr. 2005,
143, 387313.
([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).
18. Radioligand binding affinity assay: For competition binding studies, compounds
were diluted serially in the assay buffer (10 mM HEPES, 5 mM Tris, 140 mM
NaCl, 2.5 mM KCl, 1.25 mM CaCl2, 1.25 mM MgCl2, pH7.4) via BioMek 3000. To
each well of assay plate, diluted compounds, 8 nM of radioligand
([3H]dapagliflozin, 39.7 Ci/mmol; Moravek, Irvine, CA) and CHO cell
membrane protein stably expressing human SGLT2 (10–20
added and incubated at 37 °C for 60 min. Plates were harvested in GF/B
filtermat with three washes in cold wash buffer (10 mM HEPES). 50
scintillant was added to each well of plate and read on TopCount (Perkin
lg/well) were
ll
12. Imamura, M.; Murakami, T.; Shiraki, R.; Ikegai, K.; Sugane, T.;Iwasaki, F.;
Kurosaki, E.; Tomiyama, H.; Noda, A.; Kitta, K.; Kobayashi, Y. WO2004080990,
2004; Chem. Abstr. 2004, 141, 296242.
Elmer). IC50 and Ki values were calculated using GRAPHPAD PRISM
5 software
(GraphPad, San Diego, CA).
13. Clinical trial status obtained from http://www.clinicaltrials.gov.