C-Aryl glycoside inhibitors of SGLT2: Exploration of sugar modifications including C-5 spirocyclization

Article abstract of DOI:10.1016/j.bmcl.2010.01.075

Modifications to the sugar portion of C-aryl glycoside sodium glucose transporter 2 (SGLT2) inhibitors were explored, including systematic deletion and modification of each of the glycoside hydroxyl groups. Based on results showing activity to be quite to

Full text of DOI:10.1016/j.bmcl.2010.01.075

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1569–1572
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
C-Aryl glycoside inhibitors of SGLT2: Exploration of sugar modifications
including C-5 spirocyclization
*
Ralph P. Robinson , Vincent Mascitti, Carine M. Boustany-Kari, Christopher L. Carr, Patrick M. Foley,
Emi Kimoto, Michael T. Leininger, Andre Lowe, Michelle K. Klenotic, James I. MacDonald, Robert J. Maguire,
Victoria M. Masterson, Tristan S. Maurer, Zhuang Miao, Jigna D. Patel, Cathy Préville, Matthew R. Reese,
Li She, Claire M. Steppan, Benjamin A. Thuma, Tong Zhu
Pfizer Global Research & Development, Groton Laboratories Eastern Point Rd, Groton, CT 06340, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Modifications to the sugar portion of C-aryl glycoside sodium glucose transporter 2 (SGLT2) inhibitors
were explored, including systematic deletion and modification of each of the glycoside hydroxyl groups.
Based on results showing activity to be quite tolerant of structural change at the C-5 position, a series of
novel C-5 spiro analogues was prepared. Some of these analogues exhibit low nanomolar potency versus
SGLT2 and promote urinary glucose excretion (UGE) in rats. However, due to sub-optimal pharmacoki-
netic parameters (in particular half-life), predicted human doses did not meet criteria for further
advancement.
Received 14 December 2009
Revised 9 January 2010
Accepted 13 January 2010
Available online 21 January 2010
Keywords:
Glucose
Transporter
SGLT
Ó 2010 Elsevier Ltd. All rights reserved.
SGLT2
C-Aryl glycoside
Diabetes
Inhibitor
Sodium glucose co-transporter 2 (SGLT2) is a sugar transport
protein located on the luminal side of proximal tubules in the kid-
ney. It is the major contributor to re-uptake of filtered glucose from
pro-urine. By promoting urinary glucose excretion (UGE), inhibi-
tors of SGLT2 show promise for the treatment of diabetes by atten-
uation of postprandial plasma glucose levels and by promoting
weight loss via loss of calories.1,2 With few exceptions,3 inhibitors
of SGLT2 owe their discovery to phlorizin (1), a naturally-occurring
O-aryl glycoside that has long been known to cause glucosuria in
animals and humans (Fig. 1).4 Phlorizin is a fairly potent inhibitor
of SGLT2 (IC50 = 36 nM),5 with some selectivity against SGLT1, a re-
lated sugar transporter (IC50 = 330 nM). Unlike SGLT2, SGLT1 is not
only expressed in the kidney, but is also found in the intestine and
other tissues. Inhibition of SGLT1 may lead to undesirable gastroin-
testinal effects.6 Early drug discovery efforts on SGLT2 focused on
O-aryl glycosides related to phlorizin with the aim of increasing
potency for SGLT2 and minimizing activity versus SGLT1.1 Because
of their susceptibility to the action of glycosidases, few O-aryl gly-
cosides have survived beyond early stages of development due to
poor oral bioavailability. The advent of C-aryl glycoside SGLT2
inhibitors, such as dapagliflozin (2)5,7 solved the problem of glyco-
sidase susceptibility by excising the glycoside oxygen linker
(Fig. 1). Dapagliflozin is currently in Phase 3 clinical trials for the
treatment of diabetes, leading a field of other C-aryl glycoside
SGLT2 inhibitors in development.1
Continuing research to discover structurally distinct SGLT2
inhibitors has been devoted largely to modification of the diarylm-
ethylene side chain of dapagliflozin.1,8 In contrast, changes to the
synthetically more challenging glycoside portion of the molecule
have been less well explored9–13 and developing structure–activity
relationships (SAR) with respect to glycoside modification has been
difficult owing to the limited amount of data typically provided in
HO
O
OH
O
OH
O
HO
HO
1
Cl
OEt
OH
O
HO
HO
OH
OH
2
OH
* Corresponding author. Tel.: +1 860 441 4923; fax: +1 860 686 6010.
E-mail address: ralph.p.robinson@pfizer.com (R.P. Robinson).
Figure 1.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.075

1570
R. P. Robinson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1569–1572
Table 214
patent applications. Here, we report the SAR associated with sys-
tematic deletion and modification of each of the glycoside hydro-
xyl groups and the use of this information to design novel
spirocyclic analogues in the C-aryl glycoside series.
Y
Z
Me
Et
X
O
O
O
Structures of final target compounds and associated SGLT2 and
SGLT1 inhibition data are shown in Tables 1 and 2.14 We explored
the SAR using 4-methyl-3-(4-ethylbenzyl) and 4-chloro-3-(4-
methoxybenzyl), side chains B and C, respectively (Table 1). Based
on the very similar potencies of compounds 2, 3, and 4 against
SGLT2 (IC50s ꢀ1–2 nM), we assumed that these aryl side chains
were more or less interchangeable for the purpose of generating
SAR versus SGLT2.
16
HO
OH
HO
OH
OH
OH
Compd
X
Y
Z
SGLT2 IC50 (nM)
SGLT1 IC50 (nM)
16
17
18
19
20
21
22
23
24
25
Me
Et
Et
OMe
Et
Et
53 (1)
1100 (1)
Bond
Bond
NH
SO2
O
Me
Me
Me
Me
Me
Me
Cl
3.0 0.5 (5)
7.9 3.5 (7)
5100 (1)
190 40 (5)
370 180 (6)
>10,000 (1)
>10,000 (1)
1500 (2)
1540 180 (7)
>9600 (5)
5600 930 (4)
Reference compounds 2,5 3,15 and 4,15 as well as analogue 5
were prepared by addition of the appropriately substituted 3-ben-
14 (1)
Et
3.4 0.8 (3)
6.6 2.5 (8)
23 22 (5)
32 79 (5)
O
O
O
OMe
OMe
OEt
zylphenyllithium to 2,3,4,6-tetra-O-trimethylsilyl-b-D-glucolac-
tone.16 Modifications at C-2 and C-4, providing compounds 6, 7,
10,9 11,9 and 12, were carried out as shown in Scheme 1, using a
protecting group sequence described for 4 in the patent literature.9
Analogues 8 and 9, with changes at C-3, were synthesized as
shown in Scheme 2. Compounds 13 and 14 were prepared as pre-
viously described.10 Spiro analogue 16 was synthesized as shown
in Scheme 3, the key step involving an intramolecular olefin
metathesis reaction. Spirocyclopropyl compounds 17–18 were pre-
pared via cyclopropanation of the corresponding 6-methylene
intermediate (Scheme 4). Compound 15 and spirocycles 19–24
were obtained as recently described by Mascitti.17
Cl
Me
Bond
(C@O)Me
a
1100 nM and 1900 nM.
Y
Y
Z
O
SEMO
c-e
f, e
H
O
6
7
OH
MeO
H
O
a, b
OMe
Z
3, 4
The only C-1-modified compound prepared was 5, substituting
the benzylic hydrogen atom for OMe. The activity of this compound
versus SGLT2 was significantly lower (ꢀ700-fold) than 3. Changes at
positions C-2, C-3, and C-4 (Table 1 and entries 6–12), including
OH deletion (6, 8, and 10), conversion of OH to OMe (7 and 12) and
inversion of stereochemistry (9 and 11) gave similar results, though
the magnitude to which activity decreased due to C-4 modification
was less (ꢀ20–180-fold). In distinctcontrast tothe changes explored
at positions C-1–C-4, SGLT2 activity appeared relatively tolerant of
modifications to the C-5 hydroxymethyl group as demonstrated by
C-5 methyl analogue 13 (IC50 = 2.4 nM) and C-5 fluoromethyl ana-
logue 14 (IC50 = 5.1 nM). Attachment of a second hydroxymethyl
a, b
SEMO
HO
O
H
c-e
g, h, e
10
O
H
OMe
O
11
12
MeO
f, e
Scheme 1. Reagents and conditions: (a) 2,3-butanedione/HC(OMe)3/BF3ÁOEt2/
60 °C/3 h; (b) SEMCl/iPr2NEt/CH2Cl2/rt/overnight; (c) 1,1-thiocarbonyl-diimidaz-
ole/toluene/reflux/18 h; (d) AIBN/(TMS)3SiH/120 °C/overnight; (e) TFA/H2O/rt/1 h;
(f) MeI/NaH/THF/rt/overnight; (g) Dess–Martin reagent/CH2Cl2/rt/overnight; (h)
Na–Selectride/THF, À78 °C to rt/3 h.
Me
b-d
Et
H
H
O
a
O
3
Table 114
O
OH
Me
Et
Ph
H
OH
O
5
O 1 Ar
HO
O
O
e-g
h- j
4
2
8
9
OBn
Ph
H
HO
OH
3OH
OH
Cl
OEt
Me
Et
Cl
OMe
Scheme 2. Reagents and conditions: (a) PhCH(OMe)2/HBF4ÁOEt2/DMF/rt/overnight;
(b) Bu2SnO/toluene/reflux/4 h, then PMBBr/Bu4N+IÀ/reflux/overnight; (c) BnBr/
Bu4N+IÀ/DMF/0 °C to rt/overnight; (d) DDQ/CH2Cl2/rt/3 h; (e) NaH/imidazole/THF/
reflux/3 h, then CS2/reflux/0.75 h, then MeI/reflux/0.5 h; (f) Bu3SnH/AIBN/toluene/
reflux/1.25 h; (g) H2 (3 atm)/Pd(OH)2/HCl/EtOH/THF/rt/3 days; (h) Dess–Martin
reagent/CH2Cl2/rt/overnight; (i) NaBH4/MeOH/THF/0–10 °C/0.33 h; (j) as for Step g
except 1 day reaction time.
A
B
C
Compd
Sugar modification
Ar
SGLT2 IC50 (nM)
SGLT1 IC50 (nM)
1200 120 (4)
2
3
4
5
6
7
8
9
10
11
12
13
14
15
None
None
None
A
B
C
B
B
C
B
B
C
B
C
C
C
C
1.4 0.3 (5)
1.5a (2)
1.0 0.4 (11)
1000 (1)
>10,000 (1)
140 (1)
240 40 (4)
C-1 H to OMe
C-2 OH to H
C-2 OH to OMe
C-3 OH to H
C-3 epi-OH
C-4 OH to H
C-4 epi-OH
C-4 OH to OMe
C-5 CH2OH to Me
C-5 CH2OH to CH2F
C-5 H to CH2OH
2000b (2)
3700 (1)
21 (1)
group at C-5 (15) produced a modest loss of activity (SGLT2
IC50 = 59 nM). Based on these results, we focused subsequent efforts
on further structural exploration at C-5, including novel spiro ana-
logues (Table 2). Whereas the result for spirotetrahydrofuran 16
was somewhat disappointing (SGLT2 IC50 = 53 nM), spirocyclopro-
pyl compounds 17 and 18 displayed reasonably potent activity
against the target (IC50 = 3.0 nM and 7.9 nM, respectively), warrant-
ing further profiling including counter screening against SGLT1,
in vivo rat pharmacokinetics (PK)18 and measurement of 24 h UGE
following oral administration in rat. Compared to 2, neither ana-
logue was particularly selective for SGLT2 over SGLT1 (17: ꢀ63-fold;
139 23 (3)
180 (1)
2.4 0.5 (5)
5.1 (1)
73c (2)
59 (1)
6800d (2)
a
b
c
2.0 nM and 1.2 nM.
1650 nM and 2500 nM.
55 nM and 95 nM.
d
5600 nM and 8400 nM.

R. P. Robinson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1569–1572
1571
Me
Et
a, b, c
O
3, 4
HO
BnO
OBn
d-g
Me
OBn
Et
Me
Et
O
O
O
O
BnO
OBn
OBn
h- j
BnO
OBn
OBn
k
16
Scheme 3. Reagents and conditions: (a) PhCH(OMe)2/HBF4ÁOEt2/DMF/rt/overnight;
(b) BnBr/Bu4N+IÀ/DMF/0 °C to rt/overnight; (c) LiAlH4/Et2O/CH2Cl2/0 °C then AlCl3/
reflux/3 h; (d) TsCl/Et3N/DMAP/CH2Cl2/rt/24 h; (e) NaI/2-butanone/80 °C/overnight;
(f) CsCO3/DMF/80 °C/25 h; (g) O3/CH2Cl2/À78 °C then Me2S; (h) CH2CHMgBr/THF/
À78 °C/2 h; (i) CH2CHCH2OH/4 Å sieves/montmorillonite K-10 clay/0.5 h; (j)
(PCy3)2(PhCH)RuCl2/CHCl3/rt/0.75 h; (k) Pd/10% HCO2H in MeOH/rt/overnight.
Figure 2. Rat 24 h UGE versus exposure, correcting for SGLT2 potency (BW = body
weight). Male Sprague–Dawley rats (n = 6/group) were randomized to receive one
of five doses of compound 22 (1, 3, 10, 30, 60 mg/kg) or dapagliflozin (2; 10 mg/kg).
Following compound administration, urine was collected over 24 h for measure-
ment of glucose excretion. Simultaneously, drug exposure was assessed in satellite
animals. Compounds 22 and 2 achieved a maximal UGE of 2000–2500 mg/200 g
BW, respectively.
Z
Y
O
a-c
d
3, 4
17, 18
HO
OH
OH
this shorter half-life, maintaining high percent inhibition of the
target over the dosing interval is much more difficult. Indeed, our
modeling of the pharmacokinetic–pharmacodynamic (PK/PD) rela-
tionship for SGLT2 indicates that the dose required to achieve the
maximal rate of UGE increases exponentially in relation to decreas-
ing half-life, in line with the in vivo results for 22. From in vitro
studies using human liver microsomes and hepatocytes (turnover
rates and metabolite identification), it appears that 22 undergoes
glucuronidation on the sugar moiety at an increased rate relative
to 2, leading to a comparatively short in vivo half-life.
In conclusion, based on these studies, we conclude that struc-
tural changes at the C-5 position in the C-aryl glycoside SGLT2
inhibitor series can be well tolerated.20 The C-5 position is the only
position whereby OH deletion does not lead to appreciable loss of
SGLT2 inhibition. This information allowed the design of a series of
novel C-5 spiro analogues, some of which exhibit low nanomolar
potency versus SGLT2 and promote urinary glucose excretion
(UGE) in rats. However, due to sub-optimal pharmacokinetics
(half-life), predicted human doses did not meet criteria for further
advancement.
Scheme 4. Reagents and conditions: (a) TsCl (1.5 equiv)/py/0 °C/overnight; (b) NaI
(1.5 equiv)/2-butanone/80 °C/overnight; (c) NaOMe (9 equiv)/MeOH/0–45 °C; (d)
CH2I2 (9 equiv)/Et2Zn (1 M in hexane; 6 equiv)/CH2Cl2/À10 °C to rt/overnight.
18 ꢀ47-fold; 2: ꢀ860-fold). In PK studies (Table 3), 17 and 18
showed moderate to very high plasma clearance in rats, a species
thatpredicts thehumanclearanceof 2 inhumans quitewell.18 Unex-
pectedly, despite poor oral bioavailability, 17 (administered orally at
10 mg/kg) promoted 24 h UGE similar to the maximal UGE achieved
with 2 in rats (10 mg/kg). We attribute the pharmacodynamic activ-
ity of 17 in rats to formation of a long-lived active metabolite, possi-
bly ketone 25 (Table 2), which was unambiguously identified as a
major metabolite in rats, as well as in rat, dog and human microsome
preparations.
Concomitant with characterization of 17 and 18, synthesis of
additional C-5 spiro analogues continued. This effort provided azeti-
dine 19 (SGLT2 IC50 = 5100 nM), dioxothietane 20 (SGLT2
IC50 = 14 nM), and oxetanes 21–24. Based on the similar potencies
of compounds 2, 3, and 4, the range of activity within the oxetane
series (SGLT2 IC50 ꢀ3–30 nM) was surprising as was the compara-
tively weak activity of 24, bearing the same aryl side chain as that
in 2. Compound 22, having no potential for metabolism to a methyl
References and notes
ketone and being quite potent and selective for SGLT2 (IC50
6.6 nM), was selected for additional profiling.
=
1. (a) Washburn, W. N. J. Med. Chem. 2009, 52, 1785; (b) Washburn, W. N. Expert
Opin. Ther. Patents 2009, 19, 1485.
2. (a) Marsenic, O. Am. J. Kidney Dis. 2009, 53, 875; (b) Idris, I.; Donnelly, R.
Diabetes Obes. Metab. 2009, 11, 79.
3. Mederski, W.; Beier, N.; Cezanne, B.; Gerike, R.; Klein, M.; Tsaklakidis, C. PCT
Int. Appl. WO2007/147478 A1.
4. Ehrenkranz, J. R. L.; Lewis, N. G.; Kahn, C. R.; Roth, J. Diabetes Metab. Res. Rev.
2005, 21, 31.
5. 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.
In rat PK studies, 22 displayed high bioavailability (100%), and
improved clearance (22.3 mL/min/kg) relative to cyclopropyl ana-
logues 17 and 18. However, in rat acute UGE studies, this com-
pound gave disappointing results (Fig. 2). Despite increasing
exposure with dose, 22 was unable to elicit 24 h UGE equal to 2,
even when dosed as high as 60 mg/kg. We attribute this result to
a half-life that is short (1.2 h) compared to that of 2 (4.4 h).19 With
6. Lam, J. T.; Martín, M. G.; Turk, E.; Hirayama, B. A.; Bosshard, N. U.; Steinmann,
B.; Wright, E. M. Biochim. Biophys. Acta 1999, 1453, 297.
Table 3
Rat pharmacokinetic parameters for key compounds
7. Brooks, A. M.; Thacker, S. M. Ann. Pharmacother. 2009, 43, 1286.
8. Ellsworth, B. A.; Meng, W.; Patel, M.; Girotra, R. N.; Wu, G.; Sher, P. M.; Hagan,
D. L.; Obermeier, M. T.; Humphreys, W. G.; Robertson, J. G.; Wang, A.; Han, S.;
Waldron, T. L.; Morgan, N. N.; Whaley, J. M.; Washburn, W. N. Bioorg. Med.
Chem. Lett. 2008, 18, 4770.
Compd Dose (mg/kg; iv) Cl (mL/min/kg) Vdss (L/kg) t1/2 (h) Fa (%)
219
17
18
22
2
1
1
2
10.0
112
48.9
22.3
3.8
5.7
2.6
2.3
4.4
100
0.59
0.61
1.2
9. Eckhardt, M.; Himmelsbach, F.; Eickelmann, P.; Thomas, L.; Barsoumian, E. L.
U.S. Patent Appl. 2006/0074031 A1.
100
10. Eckhardt, M.; Himmelsbach, F.; Eickelmann, P.; Thomas, L.; Barsoumian, E. L.
U.S. Patent Appl. 2006/0009400 A1.
a
5 mg/kg po.

1572
R. P. Robinson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1569–1572
11. Frick, W.; Glombik, H.; Theis, S.; Elvert, R. PCT Int. Appl. WO2009/100936 A2.
12. Xu, B.; Lv, B.; Feng, Y.; Xu, G.; Du, J.; Welihinda, A.; Sheng, Z.; Seed, B.; Chen, Y.
Bioorg. Med. Chem. Lett. 2009, 19, 5632.
16. Deshpande, P. P.; Ellsworth, B. A.; Buono, F. G.; Pullockaran, A.; Singh, J.;
Kissick, T. P.; Huang, M.-H.; Lobinger, H.; Denzel, T.; Mueller, R. H. J. Org. Chem.
2007, 72, 9746.
13. 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.
17. Mascitti, V.; Robinson, R. P.; Préville, C.; Thuma, B. A.; Carr, C. L.; Reese, M. R.;
Maguire, R. J.; Leininger, M.T.; Lowe, A.; Steppan, C.M. Tetrahedron Lett.,
submitted for publication.
18. Rat PK studies were carried out in Sprague–Dawley rats. Allometric scaling of
rat clearance for 2 provided an estimate of human clearance in good agreement
with the value derived from published PK data for 2.
14. IC50 values were obtained by measuring inhibition of sodium-dependent
methyl-a-D
-[U-14C]glucopyranoside (AMG) uptake in CHO cells stably
expressing human SGLT1 or 2; Han, S.; Hagan, D. L.; Taylor, J. R.; Xin, L.;
Meng, W.; Biller, S. A.; Wetterau, J. R.; Washburn, W. N.; Whaley, J. M. Diabetes
2008, 57, 1723.
19. Although we determined a half-life in rats for 2 similar to that reported in Ref.
3 (4.6 h), our measurements of Cl and Vdss in differ from those reported
(4.8 mL/min/kg and 1.7 L/h, respectively).
15. Ellsworth, B. A.; Washburn, W. N.; Sher, P. M.; Wu, G.; Meng, W. PCT Int. Appl.
WO2001/027128 A1.
20. Additional support for this conclusion is provided by the C-5 modifications
reported in Ref. 13.