Design, syntheses, and SAR studies of carbocyclic analogues of sergliflozin as potent sodium-dependent glucose cotransporter 2 inhibitors

Article abstract of DOI:10.1002/anie.201302543

Combating diabetes: A small-molecule carbohydrate mimic, pseudo-sergliflozin, was synthesized effectively by a regio- and stereoselective allylic substitution reaction. It was found to be a potent and selective inhibitor of a transporter protein - sodium-dependent glucose cotransporter 2 (SGLT2) - which is responsible for glucose reabsorption in the human body. It could be a lead compound for further development into an antidiabetic agent. Copyright

Full text of DOI:10.1002/anie.201302543

Angewandte
Chemie
Communications
Drug Design
Combating diabetes: A small-molecule
carbohydrate mimic, pseudo-sergliflozin,
was synthesized effectively by a regio- and
stereoselective allylic substitution reac-
tion. It was found to be a potent and
selective inhibitor of a transporter pro-
tein—sodium-dependent glucose
cotransporter 2 (SGLT2)—which is
responsible for glucose reabsorption in
the human body. It could be a lead
compound for further development into
an antidiabetic agent.
T. K. M. Shing,* W.-L. Ng, J. Y.-W. Chan,
C. B.-S. Lau
&&&& — &&&&
Design, Syntheses, and SAR Studies of
Carbocyclic Analogues of Sergliflozin as
Potent Sodium-Dependent Glucose
Cotransporter 2 Inhibitors
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
DOI: 10.1002/anie.201302543
Drug Design
Design, Syntheses, and SAR Studies of Carbocyclic Analogues of
Sergliflozin as Potent Sodium-Dependent Glucose Cotransporter 2
Inhibitors**
Tony K. M. Shing,* Wai-Lung Ng, Judy Y.-W. Chan, and Clara B.-S. Lau
Type 2 diabetes mellitus (T2DM) or noninsulin-dependent
DM, is the most common type of diabetes.[1] The number of
patients suffering from T2DM worldwide is expected to rise
to 380 million by 2025.[2] Unfortunately, current therapeutic
agents are not effective enough such that only less than 36%
of the patients have been treated satisfactorily;[3] in addition
many side effects are associated with the medicine.[4] Thus, we
set out to develop novel small-molecule carbohydrate mimics
as potential antidiabetic agents to supplement the existing
medication. Our endeavors in the synthesis of carbohydrate
mimics have already produced potent a-glucosidase inhib-
itors as potential antidiabetic molecules.[5,6] We herein report
the design, syntheses, and structure–activity relationship
(SAR) studies of another class of potential antidiabetic
agents—transporter protein inhibitors.
Sodium-dependent glucose cotransporter 2 (SGLT2),
found mainly in the proximal tubule of kidneys, is responsible
for about 90% of renal glucose reabsorption.[7] On the other
hand, sodium glucose cotransporter 1 (SGLT1) is located not
only in kidneys, but also in small intestines and other tissues.[8]
Inhibition of SGLT2 can help reduce blood glucose level in
patients by promoting urinary glucose excretion,[9] whereas
inhibition of SGLT1 may lead to delayed absorption of
carbohydrates[9] or diarrhea.[10] Selective inhibition of SGLT2
over SGLT1 is thus highly desirable and SGLT2 has emerged
as a promising tool to combat type 2 diabetes.[9–11] Research
towards SGLT2 inhibition was initially prompted by the SAR
studies[12] of phlorizin (1), an aryl b-d-glucopyranoside and
a nonselective SGLT inhibitor.[13] Synthetic aryl O-glucosides
such as Sergliflozin (2a) and Remogliflozin (3a) (Figure 1)
Figure 1. Examples of some SGLT2 inhibitors.
had been developed but were later abandoned during clinical
trials due to their poor selectivity/efficacy and metabolic
instability.[14] In recent years, the corresponding metabolically
stable C-glucosides have received great attention.[11] Out of
this class of compounds, Dapagliflozin (4),[14] Empagliflozin
(5),[15] and Canaglifozin (6)[16] are the three most promising
synthetic therapeutic candidates.[10]
However, results from recent clinical trials indicated that
none of these SGLT2 inhibitors could inhibit more than 50%
of renal glucose reabsorption in humans.[17] Moreover, an
increase in cancer risks was reported during the late develop-
ment phase of the C-glucoside 4.[9b] Thus, the quest for safer
SGLT2 inhibitors with increased effectiveness and metabolic
stability continues.
[*] Prof. Dr. T. K. M. Shing, W.-L. Ng
Department of Chemistry and Center of Novel Functional Molecules
The Chinese University of Hong Kong
Shatin, New Territories, Hong Kong SAR (China)
E-mail: tonyshing@cuhk.edu.hk
Homepage: http://www.chem.cuhk.edu.hk/faculty_shing_tony.htm
Recent research works on SGLT2 inhibitors focused
mainly on the modifications of the aglycone unit of C-
glucoside analogues.[9–11,15–19] Departing from this trend, we
envisioned that the metabolic instability of the O-glucoside
SGLT2 inhibitors can also be greatly improved by modifying
the sugar core. In this project, SARs on the sugar core of
Sergliflozin-A (2b) was performed in view of its high SGLT2
inhibitory potency and selectivity.[20]
J. Y.-W. Chan, Prof. Dr. C. B.-S. Lau
Institute of Chinese Medicine and State Key Laboratory of
Phytochemistry and Plant Resources in West China
The Chinese University of Hong Kong
Shatin, New Territories, Hong Kong SAR (China)
[**] This work was financially supported by a CUHK direct grant and
a special equipment grant (project no. SEG/CUHK09) from the
University Grants Committee of Hong Kong SAR. We thank Dr.
S. H. L. Kok for the gift of some ruthenium catalysts and for helpful
discussions. SAR=structure–activity relationship.
The metabolic instability of 2b was attributed to its acid-
and glucosidase-labile glycosidic bond.[21] To address this
shortcoming, its endocyclic oxygen atom would be replaced
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201302543.
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
by a methylene unit to render the molecule free from
glycosidase degradation, and hence it should be able to
exert a more long-lasting blood-glucose-lowering effect
(Scheme 1).
The carba-sugar core of 7 was effectively constructed from
inexpensive d-d-gluconolactone (14). On the basis of our
recent synthetic endeavors,[24] allylic alcohol 15 was obtained
in just five steps from 14 (Scheme 3). It was then mesylated
and immediately displaced with chloride anion to give allylic
chloride 16 in 84% overall yield.
Scheme 1. Design of a carbocyclic SGLT2 inhibitor, pseudo-sergliflozin
(7).
We present herein the syntheses and SARs of seven small-
molecule carbohydrate mimics, including a metabolically
stable, potent, and selective SGLT2 inhibitor referred to
herein as pseudo-sergliflozin (7).[22] The present synthetic
approach features a stereodivergent transition-metal-cata-
lyzed allylic substitution protocol, in which the unexpected
stereochemical outcome of a ruthenium-catalyzed allylic
substitution reaction is particularly noteworthy.
To perform the key allylic substitution reaction, the
aglycone of 7 and a carba-sugar core were prepared as the
coupling precursors. The preparation of aglycone 13 was
achieved by a novel one-pot reaction sequence starting with
a Baylis–Hillman reaction between p-methoxybenzaldehyde
(8) and cyclohexenone (9) (Scheme 2). The Baylis–Hillman
adduct 10 was obtained by employing DBU as the base.
Subsequent addition of saturated ammonium chloride solu-
tion promoted the elimination of a water molecule from the
allylic alcohol 10 and gave the diene intermediate 11. A
plausible reaction mechanism is that a 1,5-hydride shift
occurred and resulted in the formation of intermediate 12,
which rapidly tautomerized under acidic conditions to give
the ortho-substituted phenol 13.[23]
Scheme 3. Synthesis of the carba-sugar core and allylic substitution
reaction between b-allylic chloride 16 and phenol 13. a) MsCl, Et3N,
3 ꢀ MS, nBu4NCl, CH2Cl2, 81%; b) [Pd(dba)2], dppe, K2CO3, CH3CN,
RT, 94%; c) [RuCl2(cod)]n, dppf, K2CO3, CH3CN, 608C, 72%; d) TFA,
H2O, RT, 18: 85%; 20: 84%. cod=1,5-cyclooctadiene, dba=dibenzyli-
deneacetone, dppe=1,2-bis(diphenylphosphino)ethane, dppf=1,1’-
bis(diphenylphosphino)ferrocene, EOM=ethoxymethyl, Ms=methane-
sulfonyl, TFA=trifluoroacetic acid.
With the coupling precursors b-allylic chloride 16 and
ortho-substituted phenol 13 in hand, allylic substitution
reactions were thoroughly studied. This seemingly simple
transformation proved very challenging. The use of either
monodentate phosphine ligands or bidentate phosphine
ligands with a large bite angle led to the formation of
a significant amount of diene side products, presumably due
to the steric bulk of both the nucleophile and b-elimination-
prone electrophile. The difficulty of coupling an ortho-
substituted phenol and a sterically hindered cyclic allylic
electrophile was also reported previously by Trost et al.[25]
After extensive screening, we found that dppe, a bidentate
phosphine ligand with a small bite angle,[26] effectively
suppressed the formation of the diene side products. Thus,
the combination of [Pd(dba)2] and dppe gave b-allylic ether
17 in 94% yield with excellent regio- and stereoselectivities
(Scheme 3).
Extensive screening of the experimental conditions also
resulted in the discovery of a novel ruthenium-catalyzed
allylic substitution reaction. The versatile polymeric ruthe-
nium(II) complex [RuCl2(cod)]n catalyzed the reaction regio-
and stereospecifically with an overall inversion of configu-
ration at C-1, yielding a-allylic ether 19. The stereochemical
outcome was completely unexpected since an overall reten-
tion of configuration at C-1 had been anticipated.[27,28] It is the
first ruthenium-catalyzed allylic substitution reaction with
Scheme 2. One-pot synthesis of aglycone of pseudo-sergliflozin (7).
DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, brsm=based on recovered
starting material.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
simultaneous inversion of the configuration of a chiral allylic
electrophile.[28] This reaction would supplement its corre-
sponding palladium-catalyzed substitution reaction, provid-
ing facile access to different diastereomeric compounds
during natural product syntheses and drug discovery process-
es. Research towards the scope and generality of this novel
transformation is underway.
25, whereas diimide reduction stereospecifically gave a-d-
glucose-configured cyclohexane 26. Upon deprotection, tet-
raol 27 was obtained in 74% overall yield.
Under the action of Pearlmanꢀs catalyst, hydrogenolysis
followed by hydrogenation of allylic tetraol 18 occurred and
yielded the C-6 deoxygenated analogue 28. The absence of
a primary hydroxy group in 28 made it an important candidate
for subsequent SAR studies, owing to the replacement of
a hydrogen-bond donor with a hydrophobic methyl group.
Thus, seven novel carbocyclic analogues of sergliflozin
with diverse conformations were synthesized in a convergent
and stereodivergent approach. The stereochemistry of the
analogues was verified by 1D 1H NMR and 2D 1H–1H COSY
experiments. The structure of 7 was further confirmed by X-
ray crystallographic analysis (Figure 2).[31]
Subsequently, the EOM ether protecting groups in allylic
ether 17 and 19 were hydrolyzed under acidic conditions. The
corresponding tetraol 18 and 20 were obtained as the first two
carbocyclic analogues of Sergliflozin. We speculated that the
distorted-chair conformation of 18 and 20 could provide us
a better understanding of the SAR of this class of SGLT2
inhibitors. The configuration at C-1 was unambiguously
confirmed by 1D 1H NMR and 2D 1H–1H COSYexperiments.
With allylic ether 17 and 19 in hand, reduction of their
alkene moiety was carried out using different hydrogenation
methods to yield four diastereoisomeric carbocyclic ana-
logues of Sergliflozin (Scheme 4). Hydrogenation of 17
catalyzed by Pearlmanꢀs catalyst[29] followed by acid hydrol-
ysis provided pseudo-sergliflozin (7) in 66% overall yield for
two steps. Diimide reduction[30] and subsequent acid hydrol-
ysis furnished a mixture of C-5 epimeric tetraols, 7 and 23, in
nearly 1:1 ratio. Similarly, Raney-Nickel catalyzed hydro-
genation of 19 followed by global deprotection gave tetraol
Figure 2. X-ray crystallographic structure of pseudo-sergliflozin (7).
In vitro 14C-a-methyl-d-glucopyranoside (14C-AMG)
uptake assay was used to determine the SGLT2/SGLT1
inhibitory activities of these newly synthesized compounds.
The results are summarized in Figure 3. Pseudo-sergliflozin
(7) was found to be a very potent and selective SGLT2
inhibitor. It demonstrated an IC50 value of 2.45 nm and an
over 200000-fold SGLT2/SGLT1 selectivity. It is noteworthy
that 7 was seven times more potent than the reference
compound Dapagliflozin (4), which is one of the most potent
SGLT2 inhibitors to date.[15] Thus, the combination of high
metabolic stability,[22] potency, and selectivity makes 7 an
excellent candidate for further development into an antidia-
betic agent.
The SARs on the sugarlike core of other carbocyclic
Sergliflozin analogues were further explored. The C-6 hy-
droxy group was found to be very important for SGLT2/
SGLT1 selectivity as this ratio dropped dramatically from
over 200000-fold to about 129-fold in the C-6 deoxygenated
analogue 28. The b-stereogenic centers at both C-5 and C-
1 were essential for low-nanomolar SGLT2 inhibition as
indicated by the significant loss of SGLT2 inhibitory activity
in 23, 25, and 27. Allylic analogues 18 and 20 did not show any
nanomolar-range inhibition towards either SGLT2 or SGLT1,
hinting that a subtle change of the chair conformation of the
sugarlike core can lead to a complete loss in inhibitory activity
in this class of compounds.
Scheme 4. Syntheses of diastereoisomeric carbocyclic analogues of
Sergliflozin. a) Pd(OH)2/C, EtOH, H2, room temperature; b) p-
TsNHNH2, NaOAc, THF/H2O=1:1, reflux; c) Raney Ni, H2O/1,4-
dioxane=1:1, H2, RT; d) HCl, H2O, EtOH, 508C, 7: 66% from 17; 23:
47% from 17; 25: 87% from 19; 27: 74% from 19; e) Pd(OH)2/C,
EtOH, RT, 58%. Ts=toluenesulfonyl.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[7] M. Isaji, Kidney Int. 2011, 79, S14.
[8] W. S. Lee, Y. Kanai, R. G. Wells, M. A. Hediger, J. Biol. Chem.
1994, 269, 12032.
[9] a) E. C. Chao, R. R. Henry, Nat. Rev. Drug Discovery 2010, 9,
551; b) C. Sheridan, Nat. Biotechnol. 2012, 30, 899.
[10] D. Jones, Nat. Rev. Drug Discovery 2011, 10, 645.
[11] S. Nair, J. P. H. Wilding, J. Clin. Endocrinol. Metab. 2010, 95, 34.
[12] 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.
[13] J. R. Ehrenkranz, N. G. Lewis, C. R. Kahn, J. Roth, Diabetes
Metab. Res. Revs. 2005, 21, 31.
[14] a) C.-H. Yao, J.-S. Song, C.-T. Chen, T.-K. Yeh, M.-S. Hung, C.-C.
Chang, Y.-W. Liu, M.-C. Yuan, C.-J. Hsieh, C.-Y. Huang, M.-H.
Wang, C.-H. Chiu, T.-C. Hsieh, S.-H. Wu, W.-C. Hsiao, K.-F. Chu,
C.-H. Tsai, Y.-S. Chao, J.-C. Lee, J. Med. Chem. 2011, 54, 166;
b) J. Rosenstock, M. Vico, L. Wei, A. Salsali, J. F. List, Diabetes
Care 2012, 35, 1473.
[15] R. Grempler, L. Thomas, M. Eckhardt, F. Himmelsbach, A.
Sauer, D. E. Sharp, R. A. Bakker, M. Mark, T. Klein, P.
Eickelmann, Diabetes Obes. Metab. 2012, 14, 83.
[16] a) S. Nomura, S. Sakamaki, M. Hongu, E. Kawanishi, Y. Koga, T.
Sakamoto, Y. Yamamoto, K. Ueta, H. Kimata, K. Nakayama, M.
Tsuda-Tsukimoto, J. Med. Chem. 2010, 53, 6355; b) D. Denvin,
L. Morrow, M. Hompesch, D. Skee, A. Vandebosch, J. Murphy,
K. Ways, S. Schwartz, Diabetes Obes. Metab. 2012, 14, 539.
[17] J. Liu, T. Lee, R. A. DeFronzo, Diabetes 2012, 61, 2199.
[18] For SARs on the modifications of the aglycone unit, please see
the following papers and references cited therein: a) S. H. Lee,
M. J. Kim, S.-H. Lee, J. Kim, H.-J. Park, J. Lee, Eur. J. Med.
Chem. 2011, 46, 2662; b) S. Y. Kang, M. J. Kim, J. S. Lee, J. Lee,
Bioorg. Med. Chem. Lett. 2011, 21, 3759; c) B. Xu, Y. Feng, H.
Cheng, Y. Song, B. Lv, Y. Wu, C. Wang, S. Li, M. Xu, J. Du, K.
Peng, J. Dong, W. Zhang, T. Zhang, L. Zhu, H. Ding, Z. Sheng,
A. Welihinda, J. Y. Roberge, B. Seed, Y. Chen, Bioorg. Med.
Chem. 2011, 21, 4465.
[19] For SARs on the modifications of the sugar core, please see:
Refs. [14a], [22], and: V. Mascitti, T. S. Maurer, R. P. Robinson,
J. Bian, C. M. Boustany-Kari, T. Brandt, B. M. Collman, A. S.
Kalgutkar, M. K. Klenotic, M. T. Leininger, A. Lowe, R. J.
Maguire, V. M. Masterson, Z. Miao, E. Mukaiyama, J. D. Patel,
J. C. Pettersen, C. Prꢁville, B. Samas, L. She, Z. Sobol, C. M.
Steppan, B. D. Stevens, B. A. Thuma, M. Tugnait, D. Zeng, T.
Zhu, J. Med. Chem. 2011, 54, 2952.
Figure 3. Structures and IC50 values of carbocyclic analogues of Sergli-
flozin.
In conclusion, we have synthesized and evaluated seven
small-molecule carbohydrate mimics as SGLT2 inhibitors.
The synthetic route towards the most potent compound,
pseudo-sergliflozin (7), featured a regio- and stereoselective
allylic substitution reaction in which the formation of
elimination side products was minimized. This short and
facile synthetic route, involving only nine steps from d-d-
gluconolactone in 24% overall yield, should facilitate further
preclinical trials of the highly selective and low-nanomolar
SGLT2 inhibitor 7.
[20] K. Katsuno, Y. Fujimori, Y. Takemura, M. Hiratochi, F. Itoh, Y.
Komatsu, H. Fujikura, M. Isaji, J. Pharmacol. Exp. Ther. 2007,
320, 323.
[21] J. F. List, J. M. Whaley, Kidney Int. 2011, 79, S20.
[22] When our carbocyclic mimics were undergoing biological
evaluation, Ohtake et al. disclosed their research work on
some carba-glucopyranose derivatives, including our target
molecule 7. To our delight, the pharmacokinetics data obtained
by Ohtake et al. shows that pseudo-sergliflozin (7) has a three-
fold longer half-life and longer hypoglycemic action than
sergliflozin-A (2b) in db/db mice. For details, see: Y. Ohtake,
T. Sato, H. Matsuoka, M. Nishimoto, N. Taka, K. Takano, K.
Yamamoto, M. Ohmori, T. Higuchi, M. Murakata, T. Kobayashi,
K. Morikawa, N. Shinma, M. Suzuki, H. Hagita, K. Ozawa, K.
Yamaguchi, M. Kato, S. Ikeda, Bioorg. Med. Chem. 2011, 19,
5334.
Received: March 27, 2013
Published online: && &&, &&&&
Keywords: allylation · antidiabetics · carbohydrate mimics ·
.
diastereoselectivity · SGLT2
[1] S. H. Golden, K. A. Robinson, I. Saldanha, B. Anton, P. W.
Ladenson, J. Clin. Endocrinol. Metab. 2009, 94, 1853.
[2] International Diabetes Federation. Diabetes Atlas, 3rd ed.,
Hoorens Printing NV, Heule, Belgium, 2006.
[23] For a stepwise reaction pathway towards ortho-substituted
phenols, see: K. Y. Lee, J. E. Na, J. N. Kim, Bull. Korean
Chem. Soc. 2003, 24, 409.
[3] W. N. Washburn, J. Med. Chem. 2009, 52, 1785.
[4] O. J. Phung, J. M. Scholle, M. Talwar, C. I. Coleman, JAMA J.
Am. Med. Assoc. 2010, 303, 1410.
[24] For our synthesis of the precursor of the pseudo-sugar core, see:
T. K. M. Shing, Y. Chen, W. L. Ng, Synlett 2011, 1318.
[25] B. M. Trost, W. Tang, F. D. Toste, J. Am. Chem. Soc. 2005, 127,
14785.
[5] T. K. M. Shing, H. M. Cheng, W. F. Wong, C. S. K. Kwong, J. Li,
C. B. S. Lau, P. S. Leung, C. H. K. Cheng, Org. Lett. 2008, 10,
3145.
[6] T. K. M. Shing, H. M. Cheng, Org. Lett. 2008, 10, 4137.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
[26] M. N. Birkholz nꢁe Gensow, Z. Freixa, P. W. N. M. van Leeu-
wen, Chem. Soc. Rev. 2009, 38, 1099.
1101; Angew. Chem. Int. Ed. 2002, 41, 1059; b) Y. Morisaki, T.
Kondo, T. A. Mitsudo, Organometallics 1999, 18, 4742; c) M.
Kawatsura, M. Sato, H. Tsuji, F. Ata, T. Itoh, J. Org. Chem. 2011,
76, 5485.
[27] As a control experiment, the allylic substitution reaction was
also carried out in the absence of the ruthenium catalyst. No
coupled product 19 could be obtained even though the experi-
ment was conducted in refluxing MeCN for 2 days. Thus, we
concluded that the allylic substitution was a ruthenium-catalyzed
reaction and not a base-promoted SN2 reaction.
[28] For literature reports on ruthenium-catalyzed allylic substitution
reactions with overall retention of configuration, please see:
a) B. M. Trost, P. L. Fraisse, Z. T. Ball, Angew. Chem. 2002, 114,
[29] W. M. Pearlman, Tetrahedron Lett. 1967, 8, 1663; R. Adams, V.
Voorhees, R. L. Shriner, Org. Synth. Coll. Vol. 1941, 1, 463.
[30] D. T. Hog, P. Mayer, D. Trauner, J. Org. Chem. 2012, 77, 5838.
[31] CCDC 930197 (7) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
6
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!