Conformationally constrained spiro C-arylglucosides as potent and selective renal sodium-dependent glucose co-transporter 2 (SGLT2) inhibitors

Article abstract of DOI:10.1002/cmdc.201000051

(Chemical Equation Presented) Sodium glucose co-transporter 2 (SGLT2) is an emerging target for the treatment of type 2 diabetes mellitus (DM2). Here, the synthesis and preliminary biological evaluation of a series of potent, selective SGLT2 inhibitors, derived from a rigid spiro C-arylglucoside scaffold, is described.

Full text of DOI:10.1002/cmdc.201000051

DOI: 10.1002/cmdc.201000051
Conformationally Constrained Spiro C-Arylglucosides as Potent and
Selective Renal Sodium-Dependent Glucose Co-transporter 2 (SGLT2)
Inhibitors
Binhua Lv,*[a, b] Yan Feng,[b] Jiajia Dong,[b] Min Xu,[b] Baihua Xu,[a, b] Wenbin Zhang,[b] Zelin Sheng,[b]
Ajith Welihinda,[c] Brian Seed,[c] and Yuanwei Chen*[a, b]
Diabetes is a highly prevalent modern disease with over 246
million people afflicted worldwide in 2007.[1] A failure of glyce-
mic homeostasis secondary to nutritional imbalance is consid-
ered to be the principle explanation for the alarming and in-
creasing incidence of type 2 diabetes mellitus (DM2) in both
the high selectivity could potentially reduce the gastrointesti-
nal side effect.[8] Hence inhibitors selective for SGLT2 over
SGLT1 are attractive candidates for development.
Following the initial disclosure of T-1095A, a selective and
potent SGLT2 inhibitor designed based on the naturally occur-
ring inhibitor phlorizin, by Tanabe Seiyaku Co., Ltd. (Osaka,
Japan),[9–11] multiple classes of SGLT2 inhibitors have been re-
ported, including O- and C-glucosides.[12–14] The most advanced
inhibitors currently undergoing clinical development in pha-
se III trials, dapagliflozin (1)[15,16] and canagliflozin,[17] are C-aryl-
glucosides.
developed and developing countries.[2] Although
a large
number of antihyperglycemic agents have been developed to
treat the disease, 63% of DM2 patients fail to achieve the
target levels of glycosylated hemoglobin (HbA1C<7%) recom-
mended by the American Diabetes Association,[3,4] and conse-
quently these individuals are at risk of developing complica-
tions, such as accelerated cardiovascular disease, diabetic
nephropathy, retinopathy and ulceration.[1] Recently, renewed
emphasis on the development of safe oral antidiabetic agents
with a favorable cardiovascular profile has highlighted the at-
tractions of inhibition of renal glucose resorption as a thera-
peutic mechanism.
The studies described here were directed at identifying met-
abolically robust agents with high selectivity towards SGLT2.
Information gained from modeling studies and analysis of the
crystal structure of dapagliflozin (1)[18] suggested the possibility
of creating novel and conformationally constrained chemo-
types with improved potency for SGLT2 by cyclizing the 1- and
6’-positions of the glucose moiety and glucose-proximal
phenyl ring (Figure 1). Preliminary studies showed that reten-
tion of a chlorine substituent at the 4’-position on the proximal
phenyl ring is critical for activity.[19] The synthesis and evalua-
tion of three series of novel analogues, which have a different
scaffold than previously reported inhibitors, are described
here.
Sodium glucose co-transporter 2 (SGLT2) is a 672-amino
acid, high-capacity, low-affinity transporter expressed nearly ex-
clusively in the S1 and S2 segments of the renal proximal
tubule and believed to mediate the majority of renal glucose
resorption from the glomerular filtrate.[5] Because the etiology
of type 2 diabetes mellitus (DM2) depends on a hypertrophic
adipose reservoir, mechanisms that promote glucose disposal
by urinary output are therapeutically attractive compared to
mechanisms that promote increased glucose assimilation by
adipocytes. Selective inhibitors of SGLT2 are expected to be
safe because individuals homozygous or compound heterozy-
gous for mutations in SLC5A2, the gene encoding SGLT2, ex-
hibit no significant morbidities.[6] In contrast, penetrant alleles
leading to SGLT1 deficiency are the genetic cause of glucose–
galactose malabsorption syndrome, which is associated with
severe neonatal diarrhea and failure to thrive.[7] In particular,
The synthesis of spiro[isobenzofuran-1,2’-pyran] analogues
12a–e was addressed first (Scheme 1). Persilylated gluconolac-
tone 3 was prepared in 89% yield by the slow addition of tri-
methylsilyl chloride (TMSCl) to commercially available glucono-
lactone 2 in the presence of N-methylmorpholine.[20,21] Benzoic
acid 4 was subjected to bromination with N-bromosuccinimide
(NBS) followed by esterification to yield aniline 6. Sandmeyer
reaction and subsequent oxidation of 7 provided the key elec-
tron-deficient tetra-substituted benzene 8. Friedel–Crafts acyla-
tion of R1 substituted benzenes generated the benzophenone
9. Selective reduction of the resulting ketone with triethylsilane
and further reduction of the methyl ester gave the correspond-
ing benzyl alcohol 10. Protection of the primary hydroxy
group with chloromethyl methyl ether produced bromide 11.
Lithium–halogen exchange and subsequent coupling with lac-
tone 3 gave a mixture of lactols,[22] which were converted
in situ to the desired spiro[isobenzofuran-1,2’-pyran] deriva-
tives 12a–e in 40 to 63% yield after purification by preparative
thin layer chromatography (TLC).[23]
[a] B. Lv, B. Xu, Prof. Dr. Y. Chen
Chengdu Institute of Organic Chemistry (CIOC)
Chinese Academy of Sciences (CAS)
No. 16, Southion 2, the first circle road, 610041 Chengdu (P.R. China)
Fax: (+86)28-8525-9387
E-mail: 51popo51@163.com
chenyw@cioc.ac.cn
[b] B. Lv, Dr. Y. Feng, Dr. J. Dong, Dr. M. Xu, B. Xu, W. Zhang, Dr. Z. Sheng,
Prof. Dr. Y. Chen
Egret Pharma (Shanghai) Company, Ltd.
Halei Road 1118, 201203 Shanghai (P.R. China)
The synthesis of spiro[indane-1,2’-pyran] glucosides 19a–e
was more challenging than that of O-spiroketal C-arylgluco-
sides analogues (Scheme 2). Benzyl alcohol 13 was oxidized
with Dess–Martin reagent and subsequently subjected to
[c] Dr. A. Welihinda, Prof. Dr. B. Seed
Theracos Inc., 550 Del Rey Avenue, Sunnyvale, CA 94805-3528 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000051.
ChemMedChem 2010, 5, 827 – 831
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
827

MED
Figure 1. Origin and design of spiro C-arylglucosides as SGLT2 inhibitors.
Compound 19 f was obtained
through a similar synthetic se-
quence to that of 19a starting
from
(2-bromo-5-chloro-4-(4-
ethoxybenzyl)phenyl)metha-
nol.[23] Treatment with an excess
amount of a proton source to
yield cation 16 followed by a
regio- and stereoselective cycli-
zation should generate carbocy-
cle 17, a compound which is fa-
vored, in principle, by neighbor-
ing group steric hindrance and
energy minimization. Lactols 15
were exposed to trifluoroacetic
acid (TFA) in dichloromethane at
ꢀ208C for 6 h, followed by hy-
drolysis with lithium hydroxide
to give alcohol 20 in 37% yield.
This intermediate was smoothly
converted into hexahydroxy spi-
ro[indane-1,2’-pyran] 19b by re-
Scheme 1. Synthesis of spiro[isobenzofuran-1,2’-pyran] analogues 12a–e. Reagents and conditions: a) TMSCl, N-
methylmorpholine, THF, 208C, overnight, 89%; b) NBS, DMF, 58C, 1 h, 87%; c) SOCl2, MeOH, reflux, 6 h, 99%;
d) NaNO2, concd HCl, CuCl, 1,4-dioxane, H2O, 08C, 40 min, 93%; e) KMnO4, tBuOH, 18-crown-6, H2O, reflux, over-
night, 73%; f) (COCl)2, DMF, CH2Cl2; g) AlCl3, R1 substituted phenyl, CH2Cl2; h) Et3SiH, CF3SO3H, TFA; i) NaBH4,
MeOH, THF; j) MOMCl, DIPEA, CH2Cl2, 08C!RT; k) nBuLi, THF, toluene, ꢀ788C, lactone 3; l) CH3SO3H, MeOH,
ꢀ788C!RT, 40–63% (two steps).
ductive
debenzylation.
The
indane hydroxy group in com-
pound 20 was modified using
diethylaminosulfur
trifluoride
(DAST) fluorination, iodome-
thane alkylation or Dess–Martin
periodinane-mediated oxidation,
Wittig olefination to give styrene 14 in 72% overall yield. Lith-
iation of 14 via lithium–halogen exchange, followed by addi-
tion of the aryl-lithium salt to the known 2,3,4,6-tetra-O-
benzyl-b-d-glucolactone, gave a mixture of lactols 15 (b/a=
10:1).
followed by deprotection to give 19c–e, respectively, in 75 to
81% overall yield.
Our initial attempts to synthesize spiro[tetrahydronaphtha-
lene-1,2’-pyran] glucosides 29 by allylation of lactols 15 fol-
lowed by cyclization did not construct the carbocyclic skeleton
since the allyl anion acted as a carbon nucleophile and directly
attacked the cation 17 rather than the anomeric oxonium
ion.[23] Thus, spiro[tetrahydronaphthalene-1,2’-pyran] was pre-
pared from trimethylsilyl ethyne 23, which was generated from
aldehyde 21 through Corey–Fuchs alkyne synthesis and subse-
quent protection with trimethylsilyl chloride (Scheme 3). Lithi-
um–halogen exchange with 23 followed by its addition to the
known 2,3,4,6-tetra-O-benzyl-b-d-glucolactone gave a mixture
of 24. Subsequent deprotection using tetra-n-butylammonium
fluoride (TBAF) provided lactol 25 (b/a=1:1) in 71% overall
yield. Allylation in the presence of boron trifluoride etherate fa-
vored axial nucleophilic attack at the oxocarbenium in the
Two routes to the the spiro[indane-1,2’-pyran] scaffold from
precursor 15 were explored. In the initial test, lactols 15 were
exposed to a mixture of triethylsilane and boron trifluoride
etherate at ꢀ308C and subsequent deprotection by hydroge-
nolysis. These reaction conditions predominantly generated
spiro[indane-1,2’-pyran] 19a in 40% yield after purification by
preparative LC–MS. The configuration of 19a was assigned
using nuclear Overhauser effects (NOEs) in which correlations
were observed between the chemical shifts assigned to 8’-H
and 3-H, 8’-H and 5-H, 2’-H and 2-H of the chair conformation
of 19a (see Supporting Information for characterization).
828
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 827 – 831

Scheme 2. Synthesis of spiro[indane-1,2’-pyran] glucosides 19a–e. Reagents and conditions: a) Dess–Martin reagent, CH2Cl2, ꢀ58C!RT, 2 h, 96%; b) Ph3PCH3I,
KN(TMS)2, toluene, RT, 4 h, 75%; c) nBuLi, 2,3,4,6-tetra-O-benzyl-b-d-gluconolactone, dry THF, toluene, ꢀ788C, 3.5 h, 59%; d) Et3SiH, BF3·OEt2, CH2Cl2, ꢀ308C,
40%; e) 10% Pd/C, H2, THF, MeOH, RT; f) 1. TFA, CH2Cl2, ꢀ208C, 6 h; 2. LiOH·H2O, THF/MeOH/H2O (4:1), RT, 5 h, 37% (two steps); g) DAST (Et2NSF3), CH2Cl2,
ꢀ788C, 3 h, 87%; h) CH3I, NaH (60% in mineral oil), dry THF, RT, 1 h, 85%; i) Dess–Martin reagent, CH2Cl2, 08C!RT, 2 h, 90%.
Scheme 3. Synthesis of spiro[naphthalene-1,2’-pyran] glucosides 29. Reagents and conditions: a) CBr4, PPh3, CH2Cl2, RT, overnight, 60%; b) LDA, THF, ꢀ788C,
1.5 h; c) TMSCl, RT, 4 h, quantitative; d) nBuLi, 2,3,4,6-tetra-O-benzyl-b-d-gluconolactone, dry THF, toluene, ꢀ788C, 2.5 h; e) TBAF, THF, 08C, 1 h, 71% (two
steps); f) BF3·OEt2, allyl TMS, ꢀ58C, overnight, 50%; g) Lindlar catalyst, H2, EtOAc, RT, 3 h, 69%; h) Second generation Grubbs’ catalyst, CH2Cl2, reflux, overnight,
78%; i) 10% Pd/C, H2, THF, MeOH, RT, 1 h,73%.
gluco configuration to form only one isomer of bis-C,C-gluco-
side 26.[24] Reduction of the resulting alkyne using Lindlar cata-
lyst followed by ring-closing metathesis using second genera-
tion Grubbs’ catalyst formed carbocyclic compound 28, which
was finally converted to spiro[tetrahydronaphthalene-1,2’-
pyran] 29 by hydrogenolysis in 73% yield.
ChemMedChem 2010, 5, 827 – 831
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
829

MED
All derivatives were screened for their inhibitory activity in
cell-based SGLT functional assays (Table 1). As indicated by the
IC50 value for compound 12a, the first generation spiro[isoben-
hydroxy group at the indane 7’-position led to compound 19b
and subsequent oxidation generated ketone 19e, both of
which showed diminished SGLT2 inhibition compared with the
parent compound. However,
good activity was observed
Table 1. Structures, binding affinities (IC50) of spiro C-arylglucosides on human SGLT2 and SGLT1.
when the 7’-position was substi-
tuted with a fluorine atom or a
methoxy group (compounds
19c and 19d, respectively). To
the best of our knowledge, com-
pounds 19c and 19d are among
the most selective SGLT2 inhibi-
tors, with potency ratios in the
order of 5ꢁ104. Interestingly,
compound 19 f, derived from
the replacement of the ethyl
substituent with an ethoxy
group on the distal ring of 19a,
exhibited reduced inhibitory ac-
tivity against SGLT2 but retained
the high selectivity for SGLT2
over SGLT1. To further under-
stand the influence of the size of
the spiro ring between the car-
bohydrate moiety and the proxi-
mal aryl ring of the aglycon
moiety on selectivity, spiro[tetra-
hydronaphthalene-1,2’-pyran]
[b]
Compound
L[a]
R
IC50
Selectivity
hSGLT2 [nm]
hSGLT1 [mm]
hSGLT1/hSGLT2
12a
12b
12c
12d
12e
19a
19b
19c
19d
19e
19 f
29
ꢀOCH2ꢀ
OEt
Et
iPr
tBu
OCF3
Et
Et
Et
Et
Et
6.5
6.6
7.1
13
0.3
0.3
40
0.9
1.8
1.3
0.6
2.5
9.7
3.1
5.6
138
51
99
ND
200
91
352
ꢀOCH2ꢀ
ꢀOCH2ꢀ
ꢀOCH2ꢀ
746
ꢀOCH2ꢀ
10333
18667
3450
56667
55000
ND
ꢀCH2CH2ꢀ
ꢀCH2CH(OH)ꢀ
ꢀCH2CHFꢀ
ꢀCH2CH(OCH3)ꢀ
ꢀCH2COꢀ
ꢀCH2CH2ꢀ
ꢀCH2CH2CH2ꢀ
>100
11
OEt
Et
10-100
>200
0.89
909–9090
>6060
132
33
6.7
Dapagliflozin (1)[c]
[a] Linker (L) connecting the anomeric carbon to the proximal phenyl ring. [b] Values given are the mean of
two independent experiments conducted in duplicate; a reference standard was always included in the assays.
[c] The IC50 value of dapagliflozin (CAS: 461432-26-8) was determined by in vitro assay (see Supporting Informa-
tion for more details on the binding assay used in this report).
analogue
29
(shown
in
Scheme 3) was synthesized. Al-
zofuran-1,2’-pyran] exhibited comparable potency against
human SGLT2 (hSGLT2) and 1.5-fold improvement of selectivity
over human SGLT1 (hSGLT1) compared with dapagliflozin
(1).[25,26] With respect to the alkyl substituent at the para-posi-
tion of the distal aryl ring (compounds 12b–d), the ethyl-sub-
stituted derivative (12b) showed an IC50 value of 6.6 nm for
hSGLT2 and only 91-fold selectivity over hSGLT1, whereas the
iso-propyl and tert-butyl derivatives (12c and 12d, respective-
ly) exhibited increased selectivity and slightly reduced inhibito-
ry activity against hSGLT2 indicating a trend towards greater
selectivity for hSGLT2 over hSGLT1 with increasing substituent
size. Replacement of the ethoxy group with a trifluoromethoxy
moiety in this position (compound 12e) led to a pronounced
increase in inhibition and selectivity towards hSGLT2 compared
with dapagliflozin (1). Overall, all analogues 12a–e with small,
lipophilic para-substituents on the distal aryl ring in this series
had equivalent or increased potency and selectivity for SGLT2
inhibition relative to dapagliflozin (1), possibly because of the
greater conformational constraints imposed by the spiro[iso-
benzofuran-1,2’-pyran] scaffold.
though compound 29 exhibited a fivefold decrease in potency
towards SGLT2 compared to dapagliflozin (1), the selectivity
over SGLT1 was still greater than 6060-fold. A detailed explan-
ation for the observed selectivity is not possible without defini-
tive determination of the active-site geometry, however, we
speculate that the binding pocket of SGLT2 may accomodate
the rigid conformation of the spiro[indane-1,2’-pyran] and spi-
ro[naphthalene-1,2’-pyran] better.
As compound 19a showed the highest potency against
hSGLT2, it was evaluated for its ability to induce urinary glu-
cose excretion in healthy Sprague–Dawley rats. Oral adminis-
tration of 19a to rats in a single dose of 0.33, 1, 3, 9,
27 mgkgꢀ1 induced urinary glucose excretions of 110, 370, 650,
1470 and 1450 mg, respective, per 200 g of body weight over
24 h, resulting in a 10- to 100-fold elevation relative to the ve-
hicle control. Figure 2 compares the dose-dependent glucosu-
ric response with previously reported data for oral dosing
(1.0 mgkgꢀ1) of dapagliflozin (1).[27] As glucosuria excretion in-
duced by 19a was less than that observed with dapagliflozin
(1) at the same dose in rats (1 mgkgꢀ1), we speculated that
19a may be liable to metabolic degradation. In a separated ex-
periment, the predicted elimination half-life for 19a was 1.2,
10.5, 1.5 and 6.4 h following incubation with liver microsomes
and hepatocytes from rat, dog, monkey and human, respec-
tively. The calculated hepatic clearance rate in rats was
33 mL/minkgꢀ1 compared to low in vitro metabolic rate
(<2.2 mL/minkgꢀ1) of dapagliflozin.[27,28] The SGLT2 inhibition,
Since the ketal structure in the spiro[isobenzofuran-1,2’-
pyran] analogues might be metabolically labile, the spiro[in-
dane-1,2’-pyran] analogues 19a–f were investigated. Substitu-
tion of a methylene group for an oxygen atom in the isoben-
zofuran analogue 12b gave a more potent inhibitor 19a,
which exhibited an IC50 value of 0.3 nm and greater than
18600-fold selectivity for SGLT2 over SGLT1. Introduction of a
830
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 827 – 831

[3] AACE Diabetes Mellitus Guidelines, Endocr. Pract. 2007, 13 (Suppl. 1),
S4–S68.
[4] American Diabetes Association, Diabetes Care 2007, 30 (Suppl. 1), S4–
S41; DOI: 10.2337/dc07-S004.
[5] Y. Kanai, W. S. Lee, G. You, D. Brown, M. A. Hediger, J. Clin. Invest. 1994,
93, 397–404.
[6] L. P. van den Heuvel, K. Assink, M. Willemsen, L. Monnens, Hum. Genet.
2002, 111, 544–547.
[7] E. M. Wright, Am. J. Physiol. 1998, 275, G879–882.
[8] J. M. Pascual, D. Wang, B. Lecumberri, H. Yang, X. Mao, R. Yang, D. C.
De Vivo, Eur. J. Endocrinol. 2004, 150, 627–633.
[9] J. R. Ehrenkranz, N. G. Lewis, C. R. Kahn, J. Roth, Diabetes/Metab. Res.
Rev. 2005, 21, 31–38.
[10] A. Oku, K. Ueta, K. Arakawa, T. Ishihara, M. Nawano, Y. Kuronuma, M.
Matsumoto, A. Saito, K. Tsujihara, M. Anai, T. Asano, Y. Kanai, H. Endou,
Diabetes 1999, 48, 1794–1800.
[11] K. Tsujihara, M. Hongu, K. Saito, H. Kawanishi, K. Kuriyama, M. Matsumo-
to, A. Oku, K. Ueta, M. Tsuda, A. Saito, J. Med. Chem. 1999, 42, 5311–
5324.
[12] M. Isaji, Curr. Opin. Invest. Drugs 2007, 8, 285–292.
[13] W. N. Washburn, J. Med. Chem. 2009, 52, 1785–1794.
[14] W. N. Washburn, Expert Opin. Ther. Pat. 2009, 19, 1485–1499.
[15] B. Komoroski, N. Vachharajani, D. Boulton, D. Kornhauser, M. Geraldes, L.
Li, M. Pfister, Clin. Pharmacol. Ther. 2009, 85, 520–526.
[16] B. Komoroski, N. Vachharajani, Y. Feng, L. Li, D. Kornhauser, M. Pfister,
Clin. Pharmacol. Ther. 2009, 85, 513–519.
Figure 2. Dose-dependent glucosuric response to treatment with 19a over
24 h (per 200 g body weight) following a single oral dose (Sprague–Dawley
rats). Data shown are the mean ꢁSD (n=5).
[17] Clinical trial: CANagliflozin Treatment and Trial Analysis-Sulfonylurea
(CANTATA-SU) SGLT2 Add-on to Metformin Versus Glimepiride, 2009;
details can be found here: http://clinicaltrials.gov/ct2/show/
NCT00968812 (Last accessed: April 6, 2010).
[18] J. Z. Gougoutas, (Bristol-Myers Squibb Co., Princeton, NJ, USA),
WO 2002/083066 A2, 2002; [Chem. Abstr. 2002, 137, 311199].
[19] B. Lv, B. Xu, Y. Feng, K. Peng, G. Xu, J. Du, L. Zhang, W. Zhang, T. Zhang,
L. Zhu, H. Ding, Z. Sheng, A. Welihinda, B. Seed, Y. Chen, Bioorg. Med.
Chem. Lett. 2009, 19, 6877–6881.
[20] D. Horton, W. Priebe, Carbohydr. Res. 1981, 94, 27–41.
[21] P. P. Deshpande, B. A. Ellsworth, J. Singh, C. Lai, G. Crispino, M. E. Ran-
dazzo, J. Z. Gougoutas, T. W. Denzel, (Bristol-Myers Squibb Co., Prince-
ton, NJ, USA), WO 2004/063209 A2, 2004; [Chem. Abstr. 2004, 141,
89317].
[22] S. Czernecki, G. Ville, J. Org. Chem. 1989, 54, 610–612.
[23] Y. Chen, Y. Feng, B. Xu, B. Lv, J. Dong, B. Seed, M. J. Hadd, (Theracos,
Inc., Sunnyvale, CA, USA), US 2007/0275907 A1, 2007; [Chem. Abstr.
2007, 147, 542063].
glucosuria and in vitro metabolic rate of compound 19a sug-
gest the spiro[indane-1,2’-pyran] series may be merit further
exploration.
In conclusion, we have identified a rigid spiro C-arylgluco-
side scaffold that gave rise to a viable series of potent, selec-
tive SGLT2 inhibitors that are conformationally constrained
compared to previously reported C- and O-glucosides. The
structure–activity relationship exploration of this series estab-
lished the higher binding affinity and greater selectivity for
SGLT2 of the spiro chemotype compared with previously re-
ported agents.
Experimental Section
[24] A. M. Gꢂmez, C. Uriel, S. Jarosz, S. Valverde, J. C. Lꢂpez, Eur. J. Org.
Chem. 2003, 4830–4837.
For experimental details on the synthesis and biological assays see
the Supporting Information.
[25] The published selectivity of dapagliflozin for hSGLT1/hSGLT2 is 1200-
fold (see Reference [27]). However, in our assays, we observed a 132-
fold selectivity for hSGLT2. See Supporting Information for further de-
tails on the experimental methods used in this report.
[26] N. C. Goodwin, R. Mabon, B. A. Harrison, M. K. Shadoan, Z. Y. Almstead,
Y. Xie, J. Healy, L. M. Buhring, C. M. DaCosta, J. Bardenhagen, F. Mseeh,
Q. Liu, A. Nouraldeen, A. G. E. Wilson, S. D. Kimball, D. R. Powell, D. B.
Rawlins, J. Med. Chem. 2009, 52, 6201–6204.
[27] W. Meng, B. A. Ellsworth, A. A. Nirschl, P. J. McCann, M. Patel, R. N. Giro-
tra, G. Wu, P. M. Sher, E. P. Morrison, S. A. Biller, R. Zahler, P. P. Desh-
pande, A. Pullockaran, D. L. Hagan, N. Morgan, J. R. Taylor, M. T. Ober-
meier, W. G. Humphreys, A. Khanna, L. Discenza, J. G. Robertson, A.
Wang, S. Han, J. R. Wetterau, E. B. Janovitz, O. P. Flint, J. M. Whaley, W. N.
Washburn, J. Med. Chem. 2008, 51, 1145–1149.
All biological experiments involving animals were performed in ac-
cordance with the PR China law and the local ethical committee
guidelines for animal research.
Acknowledgements
We thank Dr. Jacques Y. Roberge for thoughtful comments, Ying
Chen for performing in vitro assays and the analytical group of
Egret Pharma (Shanghai) Ltd. for analytical services.
[28] R. S. Obach, Drug Metab. Dispos. 1999, 27, 1350–1359.
Keywords: C-glycosides
SGLT2 · spiro compounds
· diabetes mellitus · inhibitors ·
[1] IDF Diabetes Atlas, 3rd ed., Hoorens Printing NV, Heule, 2006.
Received: February 8, 2010
Revised: March 23, 2010
[2] National Diabetes Fact Sheet (US, 2005), Centers for Disease Control
and Prevention (CDC); http://apps.nccd.cdc.gov/ddtstrs/template/ndfs_
2005.pdf (Last accessed: April 6, 2010).
Published online on April 22, 2010
ChemMedChem 2010, 5, 827 – 831
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
831