RETRACTED ARTICLE: Development of a novel class of glucose transporter inhibitors

Article abstract of DOI:10.1021/jm300015m

On the basis of our finding that the antitumor effect of 5-{4-[(1-methylcyclohexyl)methoxy]benzyl}thiazolidine-2,4-dione, a thiazolidinedione peroxisome proliferator-activated receptor (PPAR)γ agonist, was, in part, attributable to its ability to block glucose uptake independently of PPARγ, we used its PPARγ-inactive analogue to develop a novel class of glucose transporter (GLUT) inhibitors. This lead optimization led to compound 30 {5-(4-hydroxy-3-trifluoromethylbenzylidene)-3- [4,4,4-trifluoro-2-methyl-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-2,4-dione} as the optimal agent, which exhibited high antitumor potency through the suppression of glucose uptake (IC₅₀, 2.5 μM), while not cytotoxic to prostate and mammary epithelial cells. This glucose uptake inhibition was associated with the inhibition of GLUT1 (IC₅₀, 2 μM). Moreover, the mechanism of antitumor action of compound 30 was validated by its effect on a series of energy restriction-associated cellular responses. Homology modeling analysis suggests that the inhibitory effect of compound 30 on glucose entry was attributable to its ability to bind to the GLUT1 channel at a site distinct from that of glucose.

Full text of DOI:10.1021/jm300015m

Article
pubs.acs.org/jmc
Development of a Novel Class of Glucose Transporter Inhibitors
Dasheng Wang,†,‡ Po-Chen Chu,†,‡ Chia-Ning Yang,§ Ribai Yan,† Yu-Chung Chuang,§ Samuel K. Kulp,†
and Ching-Shih Chen*,†,∥
†Division of Medicinal Chemistry, College of Pharmacy, The Ohio State University, Columbus, Ohio 43210, United States
§Institute of Biotechnology, National Kaohsiung University, Kaohsiung, Taiwan 811
∥Institute of Basic Medical Sciences, National Cheng-Kung University, 1 Ta-Hsueh Road, Tainan, 70101, Taiwan
ABSTRACT: On the basis of our finding that the antitumor
effect of 5-{4-[(1-methylcyclohexyl)methoxy]benzyl}thiazolidine-
2,4-dione, a thiazolidinedione peroxisome proliferator-activated
receptor (PPAR)γ agonist, was, in part, attributable to its ability
to block glucose uptake independently of PPARγ, we used its
PPARγ-inactive analogue to develop a novel class of glucose trans-
porter (GLUT) inhibitors. This lead optimization led to compound
30 {5-(4-hydroxy-3-trifluoromethylbenzylidene)-3-[4,4,4-tri-
fluoro-2-methyl-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-2,4-
dione} as the optimal agent, which exhibited high antitumor po-
tency through the suppression of glucose uptake (IC50, 2.5 μM),
while not cytotoxic to prostate and mammary epithelial cells.
This glucose uptake inhibition was associated with the inhibition
of GLUT1 (IC50, 2 μM). Moreover, the mechanism of antitumor action of compound 30 was validated by its effect on a
series of energy restriction-associated cellular responses. Homology modeling analysis suggests that the inhibitory effect of
compound 30 on glucose entry was attributable to its ability to bind to the GLUT1 channel at a site distinct from that of
glucose.
including p53 mutations,21 upregulated Akt signaling,22 and
INTRODUCTION
Cancer cells gain growth advantages in the microenvironment
■
hypoxia.23 To date, a number of small-molecule agents capable
of suppressing the activity/expression of GLUT1 and/or other
by shifting cellular metabolism from oxidative phosphorylation
GLUT members have been reported, including resveratrol,24
naringenin,25 phloretin,26 fasentin,27 8-aminoadenosine,28 and
STF-31.29 Exposure of cancer cells to these agents gave rise to
reduced cell proliferation and/or chemosensitization, providing
a proof-of-concept that targeting GLUT1 is a viable therapeutic
strategy for cancer treatment.
to glycolysis, the so-called Warburg effect.1−4 This glycolytic shift
enables cancer cells to adapt to low-oxygen microenvironments,
to generate biosynthetic building blocks for cell proliferation, to
acidify the local environment to facilitate tumor invasion, and
to generate NADPH and glutathione through the pentose phos-
phate shunt to increase resistance to oxidative stress.2,5,6 Thus,
the Warburg effect is considered as a fundamental property of neo-
plasia, thereby constituting the basis for tumor imaging by [18F]2-
fluoro-2-deoxyglucose positron emission tomography.7 From
a therapeutic perspective, targeting glycolysis by blocking
glucose uptake represents a clinically relevant approach for
cancer treatment, which has constituted the focus of many
investigations.
Substantial evidence indicates that increased glucose uptake
in malignant cells is associated with dysregulated expression of
glucose transporter proteins, especially glucose transporter 1
(GLUT1).8,9 GLUT1 is a class I facilitative sugar transporter re-
sponsible for basal glucose import required to maintain cellular
respiration. GLUT1 overexpression has been reported in many
types of human cancers, including those of brain,10 breast,11,12
cervix,13 colon,14 kidney,15 lung,16 ovary,17 prostate,18 thyroid,19
and skin,20 and is correlated with advanced cancer stages and
poor clinical outcomes. This GLUT1 upregulation might be at-
tributable to genetic alterations or environmental factors,
Previously, we demonstrated that the suppressive effects of
the peroxisome proliferator-activated receptor (PPAR)γ agonist
5-{4-[(1-methylcyclohexyl)methoxy]benzyl}thiazolidine-2,4-
dione (1) on various signaling pathways, including those mediated
by cyclin D1, Sp1, and androgen receptor (AR), in prostate
cancer cells was attributable to its ability to block glucose entry
independently of PPARγ.30,31 This finding provides a mech-
anistic rationale for the present study of using the PPARγ-
inactive analogues of 1 as a scaffold to develop a novel class of
glucose transporter inhibitors. The proof-of-concept of this lead
optimization was provided by compound 30, which exhibited
high potency in inducing apoptotic death in LNCaP cells through
the suppression of glucose uptake (IC50, 2.5 μM). Evidence sug-
gests that this suppression of glucose entry was associated with
the inhibition of GLUT1 (IC50, 2 μM), the predominant
Received: January 4, 2012
Published: April 2, 2012
© 2012 American Chemical Society
3827
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
GLUT isoform expressed in LNCaP cells. Moreover, the mech-
anism of antitumor action of compound 30 was validated by its
ability to elicit a series of energy restriction-associated cellular
responses, reminiscent of that of its parent compound.30,31
target proteins. On the basis of this consideration, compound 5
was selected as the lead agent for further modifications.
Evidence indicates that the ligand binding entailed hydro-
philic interactions with the polar substituents on the terminal
phenyl ring. For example, masking of the -OH substituent of
the terminal phenyl ring of compound 5 with a methyl group
(compound 10) abrogated the inhibitory activity. Moreover,
the adjacent -CF3 function could only be replaced with -NO2
(13), but not -OH (11), -CH3 (12), or -NH2 (14), without
compromising the drug activity, suggesting the involvement of
electronegative function in protein−ligand interactions. This
premise was also supported by lack of inhibitory activity in
compounds 15 and 16, both of which lacked an electronegative
substituent on the phenyl ring.
Introduction of an additional electron-withdrawing group,
such as -F (17), -Br (18), or -NO2 (21), or a -OH function
(19) on the 5-position led to a modest decrease in the glucose
uptake activity compared to the parent compound 5. However,
substitution with -OCH3 (20) or -NH2 (22) resulted in a
complete loss of activity. Compound 19’s regioisomers, 23 and
24, showed similar potency as their parent molecule, indicating
flexibility in ligand recognition. This premise was supported by
replacement of the -CF3 of compound 23 with a -Br atom (25)
which substantially reduced the inhibitory activity, but was
contradicted by the similarity in the potencies of 24 and 26.
This discrepancy suggested the role of the catechol moiety of
26 in interacting with target protein(s), which was corroborated
by the ability of its regioisomer 27 to block glucose uptake with
similar potency. These catechols, however, were not amenable
to drug development due to intrinsic chemical/metabolic
instability.
Furthermore, compounds 18 and 21 are more acidic than the
other derivatives examined, including 5, 17, 19, 20, and 22, due
to the inductive effect of the Br and NO2 substituents ortho to
the -OH group. By using a computational protocol in Discovery
Studio 3.1, the pKa values of these derivatives were calculated as
follows: 5, 8.0; 17, 6.0; 18, 6.6; 19, 11.5; 20, 9.6; 21, 8.8; and
22, 8.9. These pKa values, however, did not show a correlation
with the respective potencies of these compounds in sup-
pressing glucose uptake, that is, 5 > 19 > 17, 18 > 21 > 22. This
finding suggests that acidity of the phenolic moiety was not a
primary determinant of the ligand binding.
Considering the enhancing effect of the CF3 moiety on the
activity and metabolic stability of drug candidates in the course
of lead optimization,34−36 we replaced the two terminal methyl
functions at the hydrophobic side chain of compound 5 with
CF3 groups with or without substitution at the tertiary carbon,
leading to 28−30. All of these derivatives showed substantially
improved potency relative to compound 5, in the relative order
30 > 28 > 29.
CHEMISTRY
■
Previously, we reported the pharmacological exploitation of the
PPARγ-inactive analogue of compound 1, (Z)-5-{4-[(1-
methylcyclohexyl)methoxy]benzylidene}thiazolidine-2,4-dione
(Δ2CG, 2), as a scaffold to develop AR-ablative agents via
its permuted isomer 3, which led to (Z)-5-[4-hydroxy-3-(trifluoro-
methyl)benzylidene]-3-[(1-methylcyclohexyl)methyl]-
thiazolidine-2,4-dione (CG-12, 4) as the optimal compound
(Figure 1A).32 Our recent studies demonstrated that the sup-
pressive effect of compound 4 on AR expression was associated
with its ability to mimic glucose starvation through the inhibi-
tion of glucose uptake and the subsequent increase in the
expression level of the E3 ligase β-transducin repeat-containing
protein (β-TrCP).30,31,33 This upregulation facilitated the
proteasomal degradation of the transcription factor Sp1, leading
to the transcriptional repression of AR. Thus, compound 4
was used as a starting point to generate potent glucose uptake
inhibitors.
On the structural modification of compound 4, we hypo-
thesized that there exists interplay between the polar
substituents on the phenyl ring and the terminal hydrophobic
moiety in regulating its glucose uptake inhibitory activity.
Accordingly, the methylcyclohexyl moiety of compound 4 was
replaced by a series of hydrophobic moieties with varying
degrees of bulkiness, generating compounds 5−9, among which
compound 5 exhibited the most potent suppressive effect on
the uptake of [3H]-2-deoxyglucose (2-DG) into LNCaP cells
(Figure 2A).
Compound 5 was subjected to further modifications via three
different strategies: (i) replacement of the electronegative -CF3
function with various substituents (compounds 11−14) or
rearrangement of the disubstituents on the phenyl ring
(compounds 15 and 16), (ii) substitution at the 5-position
with various functional groups (compounds 17−22) or
rearrangement of the trisubstituents on the phenyl ring
(compounds 23−27), and (iii) replacement of the terminal
-CH3 functions of the hydrophobic side arm with -CF3 to
enhance electronegativity (compound 28) in conjunction with
substitution of the tertiary proton with a F atom or -CH3 group
(compounds 29 and 30, respectively). General procedures for
the synthesis of these compounds are depicted in Figure 1B.
RESULTS
■
Screening of the Focused Compound Library to
Identify Lead Glucose Uptake Inhibitors. The aforemen-
tioned derivatives (5−30) along with the parent compounds
(1−4), each at 5 μM, were assessed for their abilities to block
the uptake of [3H]-2-DG into LNCaP cells after 30 min of
exposure, which revealed a subtle structure−activity correlation
(Figure 2A).
Furthermore, MTT assays indicated that the abilities of these
thiazolidinedione derivatives to suppress the viability of LNCaP
cells paralleled their respective inhibitory activities on glucose
uptake (Figure 2B), suggesting a potential causal relationship
between these two cellular events.
Suppression of Glucose Uptake through the Inhib-
ition of Glucose Transporters. Dose−response analysis
confirmed the high potency of compound 30 in blocking [3H]-
2-DG uptake into LNCaP cells with IC50 of 2.5 μM, while the
IC50 values of other compounds examined were as follows: 28,
3.5 μM; 5, 6 μM; 4, 9 μM; 2, 52 μM; and 1, 78 μM (Figure 3A).
Glucose transport across the cytoplasmic membrane is mediated
by members of the facilitative glucose transporter/solute carrier
The role of the hydrophobic side chain in regulating the
glucose uptake-inhibitory potency was manifested by the
differential activities among compounds 4−9, which showed
an inverse correlation with the bulkiness of the hydrophobic
moiety. Especially, the large discrepancy in inhibitory potency
between compounds 5−8 and 4/9 underscored the preferential
recognition of ligands with smaller hydrophobic side chains by
3828
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
Figure 1. (A) Chemical structures of compounds 1−30 in the 5-{4-[(1-methylcyclohexyl)methoxy]benzyl}thiazolidine-2,4-dione-based focused compound
library. (B) General synthetic procedure for compounds 5−30. Reaction conditions: (a) DIPAD, TPP/dry THF; (b) piperidine, ethanol/reflux.
(GLUT/SLC2A) family.37 To date, a total of 14 members have
been identified, which are divided into three classes: class I,
GLUT1−4 and GLUT14; class II, GLUT5, GLUT7, GLUT9,
and GLUT11; class III, GLUT6, GLUT8, GLUT10, GLUT12,
and H+-coupled myo-inositol transporter.37 As information
regarding the expression patterns of individual GLUT members
in LNCaP cells was lacking, we used quantitative real-time
polymerase chain reaction to assess the mRNA levels of the
3829
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
Figure 2. (A) Inhibitory effects of compounds 1−30, each at 5 μM, on the uptake of [3H]-2-DG into LNCaP cells in Krebs−Ringer phosphate
buffer at 37 °C after 30 min of drug treatment. Column, mean; bars, SD (N = 3). (B) Corresponding effects of compounds 1−30 on the viability of
LNCaP cells by MTT assays in 10% FBS-containing RPMI 1640 medium after 72 h of drug treatment. Column, mean; bars, SD (N = 6).
hypoxia-responsive GLUT1 and GLUT3 and three other re-
presentative GLUT members, including the class I GLUT2 and
GLUT4 and the class II GLUT9. Among these five members,
LNCaP cells expressed GLUT1 and, to a much lesser extent,
GLUT9, while the mRNA levels of GLUT2−4 were negligible
(Figure 3B).
To determine if the ability of compound 30 to inhibit glucose
uptake in cancer cells could be attributed to the modulation of
GLUT function, we examined the effects of 30 and its parent
compound 5 on glucose uptake in LNCaP cells ectopically ex-
pressing GLUT isoforms. To assess the isoform specificities of
the compounds, LNCaP cells were transfected with plasmids
of compounds 28 and 30 versus compounds 1, 2, 4, and 5 on the
viability of LNCaP cells revealed differential antiproliferative
potencies that paralleled the respective inhibitory activities in
glucose uptake (Figure 4A). After 72 h of exposure in 10% fetal
bovine serum (FBS) -containing medium, the IC50 values for
individual compounds were 30, 1.5 μM; 28, 2.2 μM; 5, 4.2 μM;
4, 6 μM; 2, 28 μM; 1, 60 μM. It is noteworthy that despite the
high potency of the optimal agent compound 30 in suppressing
the viability of LNCaP cells, normal human prostate epithelial
cells (PrECs) and human mammary epithelial cells (HMECs)
were resistant to the cytotoxic effect of compound 30 even at
10 μM (Figure 4B).
encoding GLUT1, GLUT3, GLUT4, or GLUT9 vis-a-
̀
vis the
This drug-induced cell death was, at least in part, attributable
to apoptosis, as evidenced by a dose-dependent increase in
poly(ADP-ribose) polymerase (PARP) cleavage in response to
compound 30 (Figure 4C). Equally important, compound 30
shared the reported activities of compound 4, 2-DG, and glu-
cose starvation in eliciting energy restriction-associated cellular
responses in LNCaP cells, including β-TrCP-facilitated protein
degradation, adenosine monophosphate-activated protein kinase
(AMPK) activation, and endoplasmic reticulum (ER) stress.30,38
Western blot analysis indicates that compound 30 dose-
dependently increased β-TrCP expression, leading to the
downregulated expression of its substrates cyclin D1 and Sp1, as
well as the Sp1 target AR (Figure 4C). Furthermore, as AMPK
negatively regulates the activation status of mammalian homo-
logue of target of rapamycin (mTOR) and p70S6K signal-
ing,39 the drug-facilitated increases in AMPK phosphorylation
pCMV control vector so that the increased glucose uptake in
GLUT-transfected cells relative to pCMV control cells was
indicative of the activity of the ectopically expressed GLUT
protein. Among the four GLUT members examined, GLUT1
was preferentially inhibited by compounds 5 and 30 at 5 μM
(53% and 73%, respectively), followed by GLUT3 (41% and
48%, respectively), GLUT4 (32% and 42%, respectively), and
GLUT9 (26% and 34%, respectively) (Figure 3C). The IC50
values for compounds 5 and 30 in inhibiting GLUT1-mediated
[3H]-2-DG uptake were 5 μM and 2 μM, respectively (Figure 3D),
similar to those determined for suppression of glucose uptake
in LNCaP cells (Figure 3A).
The high antiproliferative potency of compound 30 is asso-
ciated with its ability to elicit energy restriction-associated cellular
responses. Examinations of the dose-dependent suppressive effects
3830
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
and GLUT1’s binding pocket. Compared to that shown in
Figure 5A, the deprotonated form of compound 30 adopted a
different mode of ligand binding such that the phenoxide
moiety lay in close proximity to Arg126 to mediate ionic/hydrophilic
interactions with its guanidine side chain (Figure 5B). Moreover, the
amino hydrogen on the Trp412 side chain interacted with the sulfur
atom and the carbonyl oxygen atom of the thiazolidinedione ring of
compound 30. In spite of the slightly upward shift in this binding
mode, the π−π interaction with Tyr28 and the electrostatic
interaction of the terminal -CF3 groups with His160 and Thr137
were maintained.
DISCUSSION
■
In the course of malignant transformation, tumor cells gain
growth advantage by increasing glucose consumption through
aerobic glycolysis.1−4 This reprogramming of energy metabo-
lism is manifested by increased glucose uptake through the up-
regulation of glucose transporters, especially GLUT1. In this
study, we report the use of thiazolidinediones as a scaffold to
develop a novel class of glucose transporter inhibitors. The op-
timal agent, compound 30, exhibited high potency in sup-
pressing the [3H]-2-DG uptake and viability of LNCaP cells,
with IC50 values of 2.5 μM and 1.5 μM, respectively, which
represents a 40-fold improvement over that of compound 1.
Equally important, compound 30 exhibited no appreciable
cytotoxicity in PrECs and HMECs, indicating the ability to
discriminate between malignant and normal epithelial cells.
Among the four GLUT isoforms examined, compound 30
preferentially inhibited GLUT1-mediated [3H]-2-DG uptake
with IC50 of 2 μM versus that of ≥5 μM for GLUT3, GLUT4,
and GLUT9. The effectiveness of compound 30 in GLUT1
inhibition underlies its high potency in triggering energy
restriction-associated cellular responses in LNCaP cells, leading
to changes in the functional status of an array of signaling pro-
teins governing cell cycle progression and apoptosis (Figure 6).
Docking modeling analysis suggests that the inhibitory effect
of compound 30 on glucose entry is attributable to its ability to
bind to the GLUT1 channel at a site distinct from that of
glucose. This docking analysis provides a structural basis to
account for the subtle structure−activity relationship among
various derivatives of compound 5. For example, compounds
28−30 exhibited higher potencies than compound 5 in GLUT1
inhibition, in part due to the additional electrostatic interac-
Figure 3. (A) Dose-dependent inhibitory effects of compounds 1, 2, 4,
5, 28, and 30 on the uptake of [3H]-2-DG into LNCaP cells in Krebs−
Ringer phosphate buffer at 37 °C after 30 min of drug treatment.
Points, mean; bars, SD (N = 3). (B) Quantitative real-time PCR anal-
ysis of the differential expression of GLUT1−4 and GLUT9 in LNCaP
cells. Column, mean; bars, SD (N = 3). (C) Suppressive effects of
compounds 5 and 30, each at 5 μM, on [3H]-2-DG uptake into
LNCaP cells overexpressing GLUT1, GLUT3, GLUT4, or GLUT9.
The analysis was carried out in Krebs−Ringer phosphate buffer at
37 °C after 30 min of drug treatment. Column, mean; bars, SD (N = 3).
(D) Dose-dependent suppressive effects of compound 5 and 30 on
[3H]-2-DG uptake into LNCaP cells overexpressing GLUT1. Column,
mean; bars, SD (N = 3).
was accompanied by concomitant decreases in the levels of
p-mTOR and p-p70S6K. Compound 30-induced ER stress was
manifested by increased expression of two ER stress markers,
glucose-regulated protein (GRP) 78 and growth arrest- and
DNA damage-inducible gene (GADD) 153. Moreover, re-
miniscent of the demonstrated effect of compound 4 on the
epigenetic activation of KLF6,38 compound 30 increased the
expression of this tumor suppressor protein in a dose-dependent
manner.
Modeling Analysis of Ligand Binding. We also
performed modeling analysis to envisage the mode of ligand
binding using the homology-modeled structure of the human
GLUT1 protein [Protein Data Bank (PDB) code 1SUK],
which was developed by use of glycerol phosphate transporter
as a template.40 Blind docking simulations revealed that com-
pound 30 and glucose bound to distinct sites in GLUT1’s
intermembrane channel for glucose passage (Figure 5A, left
panel). While the glucose recognition site was located near the
channel opening, compound 30 bound to the central segment
of the channel. Docking analysis indicates that compound 30
interacted with the putative binding site through electrostatic
tions of the two terminal -CF3 functions with His160 and Thr137
.
Similarly, relative to compounds 28 and 29, the -CH3 sub-
stituent on the tertiary carbon of compound 30 might have a
steric effect on the configuration the two -CF3 functions to
allow closer interactions with His160 and Thr137 for tighter
binding. Also noteworthy is the role of the -CF3 function on the
phenyl ring in mediating electrostatic interactions with Arg126
,
which might account for the loss of glucose uptake inhibitory
activity when this electronegative moiety in compound 5 was
replaced by -CH3 (compound 12) or -NH2 (compound 14).
CONCLUSION
and π−π stacking interactions with Tyr28, Arg126, Thr137, His160
,
■
and Gln282, as depicted in the close-up view (Figure 5A, right
panel).
Data from this and other laboratories have demonstrated that
targeting aerobic glycolysis via the inhibition of glucose trans-
porters represents a therapeutically relevant strategy for cancer
treatment. In light of the high potency of compound 30 in
suppressing glucose uptake, it serves as a useful agent to shed
light onto the signaling pathways, at both cellular and epi-
genetic levels, by which caloric restriction induces cancer cell
As compound 30 exhibits a calculated pKa value of 8.0, the
deprotonated form accounts for 13% of the total population of
molecules in a physiological environment of pH 7.2. Therefore,
we conducted another series of docking simulations to gain a
better understanding of the interactions between the phenoxide
3831
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
Figure 4. (A) Dose-dependent suppressive effects of compounds 1, 2, 4, 5, 28, and 30 on the viability of LNCaP cells by MTT assays in 10% FBS-
containing RPMI 1640 medium after 72 h of drug treatment. Points, mean; bars, SD (N = 6). (B) Dose-dependent effect of compound 30 on the
viability of normal prostate epithelial cells (PrECs) and human mammary epithelial cells (HMECs) after 72 h of treatment. Points, mean; bars, SD
(N = 6). (C) Western blot analysis of the dose-dependent effects of compound 30 on markers associated with apoptosis (PARP cleavage), β-TrCP-
mediated protein degradation (β-TrCP, cyclin D1, Sp1, and AR), AMPK activation (p-AMPK, p-mTOR, and p-p706SK), ER stress (GRP78 and
GADD153), and epigenetic activation of KLF6.
3.3 equiv) dropwise at 0 °C. The reaction mixture was stirred at room
temperature for 16 h, concentrated, dissolved in ethyl acetate, washed in
tandem with water and brine, dried, and concentrated. The residue was
purified by column chromatography (hexane/ethyl acetate) to afford N-
substituted thiazolidine-2,4-diones (i−viii).
3-(2-Ethylbutyl)thiazolidine-2,4-dione (i). Light yellow oil; 81%
yield. 1H NMR (CDCl3) δ 0.88 (t, J = 7.2 Hz, 6H), 1.28 (m, 4H), 1.70
(m, 1H), 3.51 (d, J = 7.2 Hz, 2H), 3.96 (s, 2H).
death through apoptosis and autophagy. From a translational
perspective, research aimed at understanding these underlying
mechanisms will help foster novel strategies for using this
potent glucose transporter inhibitor, alone or in combination,
in cancer therapy, which is currently underway.
EXPERIMENTAL SECTION
■
3-(2-Ethyl-2-methylbutyl)thiazolidine-2,4-dione (ii). Light yellow
Unless otherwise indicated, all anhydrous solvents and chemical
reagents were purchased at the highest grade available from Sigma−
Aldrich (St. Louis, MO) and used without further purification. Flash
column chromatography was performed with silica gel (230−400
mesh; Sorbent Technologies, Norcross, GA). Antibodies against various
proteins were obtained from the following sources: Sp1, AR, cyclin D1,
phospho-p70S6K (T389), p70S6K, GRP78, GADD153, and KLF6 were
from Santa Cruz (Santa Cruz, CA); β-TrCP was from Invitrogen;
phospho-Thr-172-AMPK, AMPK, phospho-Ser-2448-mTOR, mTOR,
and PARP were from Cell Signaling Technology (Danvers, MA); and
β-actin was from MP Biomedicals (Irvine, CA).
1
oil; 85% yield. H NMR (CDCl3) δ 0.81 (s, 3H), 0.86 (t, J = 7.2 Hz,
6H), 1.29 (m, 4H), 3.50 (s, 2H), 3.98 (s, 2H).
3-(1-Methylcyclobutylmethyl)thiazolidine-2,4-dione (iii). Light
1
yellow oil; 80% yield. H NMR (CDCl3) δ 1.19 (s, 3H), 1.70 (m,
2H), 1.92 (m, 2H), 2.06 (m, 2H), 3.60 (s, 2H), 3.97 (s, 2H).
3-(1-Methylcyclopentylmethyl)thiazolidine-2,4-dione (iv). Light
1
yellow oil; 81% yield. H NMR (CDCl3) δ 0.96 (s, 3H), 1.36 (m,
2H), 1.55 (m, 2H), 1.70 (m, 4H), 3.61 (s, 2H), 3.97 (s, 2H).
3-(1-Methylcycloheptylmethyl)thiazolidine-2,4-dione (v). Light
1
yellow oil; 78% yield. H NMR (CDCl3) δ 0.93 (s, 3H), 1.36 (m,
Nuclear magnetic resonance spectra (1H NMR) were measured
on a Bruker DPX 300 model spectrometer. Chemical shifts (δ) are
reported in parts per million (ppm) relative to the tetramethylsilane
(TMS) peak. Electrospray ionization mass spectrometric analyses were
performed with a Micromass Q-Tof II high-resolution electrospray
mass spectrometer. Melting points were determined by capillary
melting point apparatus (Thomas-Hoover). The purity of all tested
compounds was determined to be greater than 95% by elemental
analyses, which were performed by Atlantic Microlab, Inc. (Norcross,
GA) and were reported within 0.4% of calculated values. Compounds
2−4 were prepared as previously described.32 General procedures for
the synthesis of compounds 5−30 are depicted in Figure 1B.
2H), 1.54 (m, 10H), 3.60 (s, 2H), 3.97 (s, 2H).
3-[4,4,4-Trifluoro-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-2,4-
dione (vi). Light yellow oil; 82% yield. 1H NMR (CDCl3) δ 2.32 (m, 4H),
2.65 (m, 1H), 3.87 (d, J = 7.2 Hz, 2H), 4.05 (s, 2H).
3-[2,4,4,4-Tetrafluoro-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-
2,4-dione (vii). Light yellow oil; 81% yield. 1H NMR (CDCl3) δ 2.83−
2.67 (m, 4H), 4.04 (d, J = 20.4 Hz, 2H), 4.07 (s, 2H).
3-[4,4,4-Trifluoro-2-methyl-2-(2,2,2-trifluoroethyl)butyl]-
thiazolidine-2,4-dione (viii). Light yellow oil; 81% yield. H NMR
(CDCl3) δ 1.19 (s, 3H), 2.47−2.25 (m, 4H), 3.77 (s, 2 H), 4.03 (s,
2H).
Step b. A mixture of individual di- and trisubstituted benzaldehydes
(1.0 equiv), the corresponding N-substituted thiazolidine-2,4-dione
(1.15 equiv), and a catalytic amount of piperidine in ethyl alcohol was
refluxed until the reaction was completed, as monitored by thin-layer
1
Step a. To a mixture of individual alcohols (1.1 equiv), thiazolidine-
2,4-dione (1.0 equiv), and triphenyl phosphine (3.5 equiv) in dry tetra-
hydrofuran (THF) was added diisopropyl azodicarboxylate (DIPAD;
3832
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
Figure 5. (A, left panel) Schematic representation of the predicted binding mode of compound 30 vis-a-
̀
vis that of glucose in the docking analysis of
human GLUT1, showing that these two molecules bind to the transmembrane channel region at distinct sites. (A, right panel) Representation of the
GLUT1 residues surrounding the docked compound 30, showing the potential electrostatic and π−π stacking interactions (dashed lines). (B)
Representation of the GLUT1 residues surrounding the docked phenoxide species of compound 30, showing the potential electrostatic and π−π
stacking interactions (dashed lines).
Figure 6. Schematic diagram depicting the mode of action of compound 30 in eliciting energy restriction-associated cellular responses.
chromatography (TLC), and concentrated. The residue was purified
by column chromatography (hexane/ethyl acetate) to afford com-
pounds 5−30.
3-(2-Ethylbutyl)-5-(4-hydroxy-3-trifluoromethylbenzylidene)-
thiazolidine-2,4-dione (5). White solid; 82% yield; mp 158−160 °C.
1H NMR (CDCl3) δ 0.92 (t, J = 7.2 Hz, 6H), 1.34 (m, 4H), 1.81 (m,
7.61 (d, J = 8.64 Hz, 1H), 7.70 (s, 1H), 7.83 (s, 1H). HRMS exact
mass of C17H18F3NO3S (M + Na)+, 396.0857 amu; found, 396.0859 amu.
Anal. Calcd, C 54.68, H 4.86, N 3.75; found, C 54.68, H 4.77, N 3.76.
3-(2-Ethyl-2-methylbutyl)-5-(4-hydroxy-3-trifluoromethylbenzylidene)-
thiazolidine-2,4-dione (6). White solid; 78% yield. 1H NMR (CDCl3) δ
0.90 (m, 9H), 1.34 (m, 5H), 3.69 (s, 2H), 7.10 (d, J = 8.58 Hz, 1H),
7.55 (d, J = 8.64 Hz, 1H), 7.68 (s, 1H), 7.81 (s, 1H).
1H), 3.69 (d, J = 7.35 Hz, 2H), 6.19 (s, 1H), 7.10 (d, J = 8.58 Hz, 1H),
3833
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-(1-methylcyclo-
3-(2-Ethylbutyl)-5-(4-hydroxy-3-nitro-5-trifluoromethylbenzylidene)-
1
1
thiazolidine-2,4-dione (21). Yellow solid; 75% yield. H NMR (CDCl3)
butylmethyl)thiazolidine-2,4-dione (7). White solid; 76% yield. H
δ0.91 (t, J = 7.4 Hz, 6H), 1.30 (m, 4H), 1.84 (m, 1H), 3.70 (d, J = 7.5 Hz,
2H), 5.69 (s, 1H), 7.81 (s, 1H), 8.05 (s, 1H), 8.48 (s, 1H).
NMR (CDCl3) δ 1.19 (s, 3H), 1.86 (m, 4H), 2.08 (m, 2H), 3.78 (s,
2H), 6.21 (s, 1H), 7.11 (d, J = 6.6 Hz, 1H), 7.62 (d, J = 6.6 Hz, 1H),
7.71 (s, 1H), 7.85 (s, 1H).
5-(3-Amino-4-hydroxy-5-trifluoromethylbenzylidene)-3-(2-
1
ethylbutyl)thiazolidine-2,4-dione (22). Brown solid; 70% yield. H
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-(1-
NMR (CDCl3) δ 0.93 (t, J = 7.5 Hz, 6H), 1.35 (m, 4H), 1.81 (m, 1H),
2.36 (s, 2H), 3.67 (d, J = 7.2 Hz, 2H), 7.36 (s, 1H), 7.57 (s, 1H), 7.74
(s, 1H), 8.53 (s, 1H).
5-(2,4-Dihydroxy-5-trifluoromethylbenzylidene)-3-(2-ethylbutyl)-
thiazolidine-2,4-dione (23). Off-white solid; 78% yield. 1H NMR
(CDCl3) δ 0.93 (t, J = 7.5 Hz, 6H), 1.35 (m, 4H), 1.81 (m, 1H), 3.68
(d, J = 5.4 Hz, 2H), 6.32 (s, 1H), 6.56 (s, 1H), 7.20 (s, 1H), 7.68 (s,
1H), 8.26 (s, 1H).
methylcyclopentylmethyl)thiazolidine-2,4-dione (8). White solid;
1
80% yield. H NMR (CDCl3) δ 1.01 (s, 3H), 1.39 (m, 2H), 1.71
(m, 6H), 3.75 (s, 2H), 6.06 (s, 1H), 7.11 (d, J = 6.6 Hz, 1H), 7.62 (d,
J = 6.0 Hz, 1H), 7.70 (s, 1H), 7.84 (s, 1H).
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-(1-
methylcycloheptylmethyl)thiazolidine-2,4-dione (9). White solid;
1
75% yield. H NMR (CDCl3) δ 0.93 (s, 3H), 1.38 (m, 2H), 1.54
(m, 10H), 3.75 (s, 2H), 6.08 (s, 1H), 7.11 (d, J = 6.6 Hz, 1H), 7.63 (d,
J = 6.0 Hz, 1H), 7.70 (s, 1H), 7.86 (s, 1H).
3-(2-Ethylbutyl)-5-(4-methoxy-3-trifluoromethylbenzylidene)-
thiazolidine-2,4-dione (10). White solid; 88% yield. 1H NMR (CDCl3)
δ 0.92 (t, J = 7.2 Hz, 6H), 1.34 (m, 4H), 1.81 (m, 1H), 3.69 (d, J = 7.35
Hz, 2H), 4.02 (s, 3H), 7.10 (d, J = 8.58 Hz, 1H), 7.61 (d, J = 8.64 Hz,
1H), 7.70 (s, 1H), 7.83 (s, 1H).
5-(3,4-Dihydroxybenzylidene)-3-(2-ethylbutyl)thiazolidine-2,4-
dione (11). Light brown solid; 80% yield. 1H NMR (CDCl3) δ 0.94 (t,
J = 7.5 Hz, 6H), 1.34 (m, 4H), 1.81 (m, 1H), 3.68 (d, J = 7.2 Hz, 2H),
5.15 (s, 1H), 5.26 (s, 1H), 7.07 (d, J = 6.6 Hz, 1H), 7.10 (s, 1H), 7.18
(d, J = 8.58 Hz, 1H), 8.14 (s, 1H).
5-(2,3-Dihydroxy-5-trifluoromethylbenzylidene)-3-(2-ethylbutyl)-
1
thiazolidine-2,4-dione (24). Light brown solid; 74% yield. H NMR
(CDCl3) δ 0.92 (t, J = 7.5 Hz, 6H), 1.33 (m, 4H), 1.79 (m, 1H), 3.67
(d, J = 7.2 Hz, 2H), 6.26 (br s, 2H), 7.21 (s, 1H), 7.30 (s, 1H), 8.12 (s,
1H).
5-(5-Bromo-2,4-dihydroxybenzylidene)-3-(2-ethylbutyl)-
thiazolidine-2,4-dione (25). Off-white solid; 76% yield. 1H NMR
(CDCl3) δ 0.93 (t, J = 7.5 Hz, 6H), 1.35 (m, 4H), 1.81 (m, 1H), 3.67
(d, J = 7.6 Hz, 2H), 6.52 (s, 1H), 6.66 (s, 1H), 6.76 (s, 1H), 7.54 (s,
1H), 8.14 (s, 1H).
5-(5-Bromo-2,3-dihydroxybenzylidene)-3-(2-ethylbutyl)-
1
thiazolidine-2,4-dione (26). Light brown solid; 72% yield. H NMR
3-(2-Ethylbutyl)-5-(4-hydroxy-3-methylbenzylidene)thiazolidine-
(CDCl3) δ 0.94 (t, J = 7.5 Hz, 6H), 1.34 (m, 4H), 1.81 (m, 1H), 3.68
(d, J = 7.2 Hz, 2H), 5.55 (s, 1H), 5.90 (s, 1H), 7.07 (s, 1H), 7.17 (s,
1H), 8.14 (s, 1H).
1
2,4-dione (12). White solid; 86% yield. H NMR (CDCl3) δ 0.92 (t,
J = 7.2 Hz, 6H), 1.34 (m, 4H), 1.81 (m, 1H), 2.21 (s, 3H), 3.69 (d, J =
7.35 Hz, 2H), 6.19 (s, 1H), 7.10 (d, J = 8.58 Hz, 1H), 7.61 (d, J = 8.64
Hz, 1H), 7.70 (s, 1H), 7.83 (s, 1H).
5-(2-Bromo-3,4-dihydroxybenzylidene)-3-(2-ethylbutyl)-
1
thiazolidine-2,4-dione (27). Light yellow solid; 83% yield. H NMR
3-(2-Ethylbutyl)-5-(4-hydroxy-3-nitrobenzylidene)thiazolidine-
(CDCl3) δ 0.94 (t, J = 7.5 Hz, 6H), 1.34 (m, 4H), 1.81 (m, 1H), 3.68
(d, J = 7.5 Hz, 2H), 5.85 (s, 1H), 6.07 (s, 1H), 7.02 (d, J = 8.7 Hz,
1H), 7.11 (d, J = 8.4 Hz, 1H), 8.12 (s, 1H).
1
2,4-dione (13). Yellow solid; 73% yield. H NMR (CDCl3) δ 0.93 (t,
J = 5.4 Hz, 6H), 1.35 (m, 4H), 1.80 (m, 1H), 3.67 (d, J = 5.4 Hz, 2H),
5.21 (s, 1H), 7.27 (d, J = 6.6 Hz, 1H), 7.32 (d, J = 6.8 Hz, 1H), 7.68
(s, 1H), 8.52 (s, 1H).
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-[4,4,4-trifluoro-2-
(2,2,2-trifluoro-ethyl)butyl]thiazolidine-2,4-dione (28). Off-white
1
5-(3-Amino-4-hydroxybenzylidene)-3-(2-ethylbutyl)thiazolidine-
solid; 80% yield; mp 165−167 °C, H NMR (CDCl3) δ 2.31 (m,
1
2,4-dione (14). Brown solid; 68% yield. H NMR (CDCl3) δ 0.93 (t,
4H), 2.68 (m, 1H), 3.86 (d, J = 7.4 Hz, 2H), 5.99 (br s, 1H), 7.10 (d,
J = 8.1 Hz, 1H), 7.63 (d, J = 8.2 Hz, 1H), 7.72 (s, 1H), 7.89 (s, 1H).
HRMS exact mass of C17H12F9NO3S (M + Na)+, 504.0292 amu;
found, 504.0298 amu. Anal. Calcd, C 42.42, H 2.51, N 2.91; found, C
42.35, H 2.49, N 2.99.
J = 5.4 Hz, 6H), 1.35 (m, 4H), 1.80 (m, 1H), 2.26 (s, 2H), 3.67 (d, J =
5.4 Hz, 2H), 6.97 (d, J = 6.6 Hz, 1H), 7.22 (d, J = 6.8 Hz, 1H), 7.69
(br s, 1H), 7.87 (s, 1H), 8.76 (s, 1H).
5-(2,4-Dihydroxybenzylidene)-3-(2-ethylbutyl)thiazolidine-2,4-
1
dione (15). Light yellow solid; 78% yield. H NMR (CDCl3) δ 0.93
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-[2,4,4,4-tetra-
(t, J = 7.5 Hz, 6H), 1.35 (m, 4H), 1.81 (m, 1H), 3.67 (d, J = 7.6 Hz,
2H), 6.52 (s, 1H), 6.60 (s, 1H), 6.76 (d, J = 6.6 Hz, 1H), 7.18 (d, J =
6.8 Hz, 1H), 7.54 (s, 1H), 8.14 (s, 1H).
fluoro-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-2,4-dione (29).
1
White solid; 76% yield; mp 173−176 °C. H NMR (DMSO-d6) δ
3.00−2.82 (m, 2H), 3.30−3.10 (m, 2H), 4.09 (d, J = 21 Hz, 2H), 7.21
(d, J = 9 Hz, 1H), 7.73 (d, J = 9 Hz, 1H), 7.87 (s, 1H), 7.96 (s, 1H),
11.50 (br s, 1H). HRMS exact mass of C17H11F10NO3S (M + Na)+,
522.0198 amu; found, 522.0199 amu. Anal. Calcd, C 40.89, H 2.22, N
2.81; found, C 41.11, H 2.57, N 2.64.
5-(5-Bromo-2-hydroxybenzylidene)-3-(2-ethylbutyl)thiazolidine-
1
2,4-dione (16). Light yellow solid; 84% yield. H NMR (CDCl3) δ
0.93 (t, J = 7.5 Hz, 6H), 1.35 (m, 4H), 1.81 (m, 1H), 3.67 (d, J = 7.6
Hz, 2H), 6.52 (s, 1H), 7.09 (d, J = 6.8 Hz, 1H), 7.21 (d, J = 6.6 Hz,
1H), 7.54 (s, 1H), 8.14 (s, 1H).
3-(2-Ethylbutyl)-5-(3-fluoro-4-hydroxy-5-trifluoromethylbenzylidene)-
thiazolidine-2,4-dione (17). White solid; 81% yield. 1H NMR (CDCl3)
δ 0.92 (t, J = 7.5 Hz, 6H), 1.34 (m, 4H), 1.80 (m, 1H), 3.68 (d, J = 7.2
Hz, 2H), 6.35 (br s, 1H), 7.46 (d, J = 10.8 Hz, 1H), 7.52 (s, 1H), 7.76
(s, 1H).
5-(4-Hydroxy-3-trifluoromethylbenzylidene)-3-[4,4,4-trifluoro-2-
methyl-2-(2,2,2-trifluoroethyl)butyl]thiazolidine-2,4-dione (30). Off-
white solid; 86% yield; mp 164−166 °C. 1H NMR (DMSO-d6) δ 1.17
(s, 3H), 2.58−2.50 (m, 4H), 3.76 (s, 2H), 7.21 (d, J = 9 Hz, 1H), 7.74
(d, J = 9 Hz, 1H), 7.87 (s, 1H), 7.95 (s, 1H), 11.54 (br s, 1H). HRMS
exact mass of C18H14F9NO3S (M + Na)+, 518.0448 amu; found,
518.0441 amu. Anal. Calcd, C 43.64, H 2.85, N 2.83; found, C 43.64,
H 2.79, N 2.82.
3-(2-Ethylbutyl)-5-(3-bromo-4-hydroxy-5-trifluoromethylbenzylidene)-
1
thiazolidine-2,4-dione (18). Light yellow solid; 79% yield. H NMR
(CDCl3) δ 0.91 (t, J = 7.5 Hz, 6H), 1.34 (m, 4H), 1.84 (m, 1H), 3.67
(d, J = 6.0 Hz, 2H), 6.05 (br s, 1H), 7.68 (d, J = 10.8 Hz, 1H), 7.76 (s,
1H), 7.82 (s, 1H).
Cells and Cell Culture. LNCaP prostate cancer cells were
obtained from the American Type Culture Collection (Manassas, VA).
Cells were maintained in 10% FBS-supplemented RPMI 1640 medium
(Invitrogen). Normal HMECs and PrECs were obtained from Lonza
(Walkersville, MD) and were maintained in mammary epithelial cell
growth medium (MEGM) and prostate epithelial cell growth medium
(PrEGM) (Lonza, Walkersville, MD), respectively.
Glucose Uptake Assay. LNCaP cells were seeded in six-well
plates (3 × 105 cells/well) for 24 h. Cells were washed twice with
Krebs−Ringer phosphate buffer (126 mM NaCl, 2.5 mM KCl, 25 mM
NaHCO3, 1.2 mM NaH2PO4, 1.2 mM MgCl2, and 2.5 mM CaCl2, pH
7.4) and were then treated with individual agents in Krebs−Ringer
phosphate buffer. After 0.5 h, glucose uptake was initiated by adding
5-(3,4-Dihydroxy-5-trifluoromethylbenzylidene)-3-(2-ethylbutyl)-
1
thiazolidine-2,4-dione (19). Light brown solid; 76% yield. H NMR
(CDCl3) δ 0.91 (t, J = 7.4 Hz, 6H), 1.34 (m, 4H), 1.84 (m, 1H), 3.67
(d, J = 8.4 Hz, 2H), 6.48 (br s, 1H), 6.53 (s, 1H), 7.18 (s, 1H), 7.35 (s,
1H), 7.81 (s, 1H).
3-(2-Ethylbutyl)-5-(4-hydroxy-3-methoxy-5-trifluoromethylbenzylidene)-
thiazolidine-2,4-dione (20). White solid; 86% yield. 1H NMR (CDCl3) δ
0.91 (t, J = 7.4 Hz, 6H), 1.34 (m, 4H), 1.84 (m, 1H), 3.67 (d, J = 8.4
Hz, 2H), 4.05 (s, 3H), 6.48 (br s, 1H), 7.16 (s, 1H), 7.35 (s, 1H), 7.81
(s, 1H).
3834
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
1 mL pf Krebs−Ringer buffer containing 1 mCi/mL [3H]-2-DG
(PerkinElmer Life Science) and 100 mM nonradioactive 2-DG and
was terminated by washing with cold phosphate-buffered saline (PBS).
The cells were lysed in 500 mL of lysis buffer (10 mM Tris-HCl, pH
8.0, and 0.1% sodium dodecyl sulfate, SDS) and aliquots were taken
for measurement of radioactivity on a scintillation counter (Beckman
LS6500).
Cell Viability Assay. Cell viability was determined by the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay.
Cancer cells were seeded at 5000 cells/well and normal cells were
seeded at 8000 cells/well in 96-well plates, and the plates were in-
cubated in 10% FBS-supplemented medium for 24 h. Cells were then
treated with individual agents for 72 h. Drug-containing medium was
replaced with medium containing MTT (0.5 mg/mL), followed by
incubation at 37 °C for 1 h. After removal of medium, the reduced
MTT dye in each well was solubilized in 100 μL of dimethyl sulfoxide
(DMSO) and absorbance at 570 nm was measured.
Plasmid Construction, Transient Transfection, and Immuno-
blotting. The full-length GLUT1, GLUT4, and GLUT9 ORF cDNA
clones were purchased from Addgene (Cambridge, MA) and GLUT3
ORF cDNA was purchased from Origene Technologies (Rockville,
MD). GLUT1, GLUT3, and GLUT9 were subcloned into the HindIII/
SalI sites and GLUT4 was subcloned into the EcoRI/SalI sites of the
pEGFP-N2 expression vector (Clontech, Palo Alto, CA). Trans-
fections were performed by electroporation with Nucleofector kit R of
the Amaxa Nucleofector system (Lonza, Walkersville, MD) according
to the manufacturer’s protocol. Immunoblotting was performed with
cell lysates harvested with SDS lysis buffer (1% SDS, 50 mM Tris-HCl,
pH 8.0, and 10 mM ethylenediaminetetraacetic acid, EDTA) contain-
ing protease inhibitor cocktail (Sigma) and phosphatase inhibitor,
electrophoresed in SDS−8−12% polyacrylamide gels, and then trans-
ferred onto nitrocellulose membranes. After being blotted in 5% non-
fat dry milk, the membranes were incubated with primary antibodies at
1:1000 dilution in Tris-buffered saline (TBS)−Tween 20 overnight at
4 °C and then with secondary antibodies conjugated with horseradish
peroxidase at 1:5000 dilution in TBS−Tween 20 for 1 h at room tem-
perature. Protein bands were visualized on X-ray film via an enhanced
chemiluminescence system.
Quantitative Real-Time Polymerase Chain Reaction (PCR).
Total RNA was isolated and reversed-transcribed to cDNA by use of
TRIzol reagent (Invitrogen) and the iScript cDNA synthesis kit (Bio-Rad
Laboratories, Hercules, CA), respectively, according to the vendors’
instructions. Real-time PCR was carried out in the Bio-Rad CFX96
real-time PCR detection system with iQ SYBR green supermix (Bio-
Rad). The sequences of the primers used were as follows: GLUT1 for-
ward, 5′-GCGGAATTCAATGCTGATGA-3′, GLUT1 reverse, 5′-CGAA-
GATGCTCGTGGAGTAA-3′; GLUT2 forward, 5′- ATGTCAGTGG-
GACTTGTGCTGC-3′, GLUT2 reverse, 5′- CACAGTCTCTGTAGC-
TCCTAG-3′; GLUT3 forward, 5′- TTAAAGGATAACTATAATGG-3′,
GLUT3 reverse, 5′- GACATTGGTGGTGGTCTCCT-3′; GLUT4
forward, 5′- CAGAAGGTGATTGAACAGAG-3′, GLUT4 reverse,
5′- AGATGCTGGTCGAATAATAG-3′; GLUT9 forward, 5′-GC-TC-
TTGGAGAAGCACAACGAG-3′, GLUT9 reverse, 5′- AAAGTTGGA-
GAGCCAGTTGA-3′. Relative gene expression was normalized to
that of 18s rRNA and calculated by using the published 2−ΔΔCt
method.41
respective pKa values were calculated by the Molecular Properties
protocol in Discovery Studio 3.1 (Accelrys, San Diego, CA).
AUTHOR INFORMATION
Corresponding Author
*Phone 614-688-4008; fax 614-688-8556; e-mail chen.844@
osu.edu.
■
Author Contributions
‡These authors made equal contributions to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by National Institutes of Health
Grant CA112250 and Department of Defense Prostate Cancer
Research Program Grant W81XWH-09-0198 (to C.-S.C.) and
by Ministry of Economic Affairs (Taiwan) Grant 99-EC-17-A-
17-S1-152 (to C.-N.Y.).
ABBREVIATIONS
■
PPARγ, peroxisome proliferator-activated receptor γ; GLUT,
glucose transporter; AR, androgen receptor; β-TrCP, β-
transducin repeat-containing protein; 2-DG, 2-deoxyglucose;
PrECs, prostate epithelial cells; HMECs, human mammary
epithelial cells; PARP, poly(ADP-ribose) polymerase; AMPK,
adenosine monophosphate-activated protein kinase; ER,
endoplasmic reticulum; mTOR, mammalian homologue of
target of rapamycin; GRP78, glucose-regulated protein 78;
GADD153, growth arrest- and DNA damage-inducible gene
153; PDB, Protein Data Bank
REFERENCES
■
(1) Warburg, O. On the origin of cancer cells. Science 1956, 123,
309−314.
(2) Kroemer, G.; Pouyssegur, J. Tumor cell metabolism: cancer’s
Achilles’ heel. Cancer Cell 2008, 13, 472−482.
(3) Vander Heiden, M. G.; Cantley, L. C.; Thompson, C. B.
Understanding the Warburg effect: the metabolic requirements of cell
proliferation. Science 2009, 324, 1029−1033.
(4) Kim, J. W.; Dang, C. V. Cancer’s molecular sweet tooth and the
Warburg effect. Cancer Res. 2006, 66, 8927−8930.
(5) Vander Heiden, M. G. Targeting cancer metabolism: a
therapeutic window opens. Nat. Rev. Drug Discovery 2011, 10, 671−
684.
(6) Gatenby, R. A.; Gillies, R. J. Why do cancers have high aerobic
glycolysis? Nat. Rev. Cancer 2004, 4, 891−899.
(7) Kelloff, G. J.; Hoffman, J. M.; Johnson, B.; Scher, H. I.; Siegel, B.
A.; Cheng, E. Y.; Cheson, B. D.; O’Shaughnessy, J.; Guyton, K. Z.;
Mankoff, D. A.; Shankar, L.; Larson, S. M.; Sigman, C. C.; Schilsky, R.
L.; Sullivan, D. C. Progress and promise of FDG-PET imaging for
cancer patient management and oncologic drug development. Clin.
Cancer Res. 2005, 11, 2785−2808.
Molecular Docking Experiment. Docking was carried out with
AutoDock 4.2. The molecular structure of compound 30 was prepared
by the SYBYL 8.1 program (Tripos International, St. Louis, MO) via
MMFF94 molecular mechanics force-field calculation. The coordinates
for GLUT1 (PDB code 1SUK) were obtained by homology modeling
based on glycerol phosphate transporter as a template.40 The initial
blind docking used a grid box of 100 × 100 × 126 points in three
dimensions with a spacing of 0.6 Å centered on the whole GLUT1 and
indicated that the major interacting region was located in the channel.
Accordingly, further docking simulations, centered at the channel with
a grid box of 70 × 70 × 92 points in three dimensions with a spacing
of 0.375 Å, were applied to explore the binding behavior.
(8) Macheda, M. L.; Rogers, S.; Best, J. D. Molecular and cellular
regulation of glucose transporter (GLUT) proteins in cancer. J. Cell.
Physiol. 2005, 202, 654−662.
(9) Calvo, M. B.; Figueroa, A.; Pulido, E. G.; Campelo, R. G.;
Aparicio, L. A. Potential role of sugar transporters in cancer and their
relationship with anticancer therapy. Int. J. Endocrinol. 2010, 2010.
(10) Nishioka, T.; Oda, Y.; Seino, Y.; Yamamoto, T.; Inagaki, N.;
Yano, H.; Imura, H.; Shigemoto, R.; Kikuchi, H. Distribution of the
glucose transporters in human brain tumors. Cancer Res. 1992, 52,
3972−3979.
(11) Brown, R. S.; Wahl, R. L. Overexpression of Glut-1 glucose
transporter in human breast cancer. An immunohistochemical study.
Cancer 1993, 72, 2979−2985.
Calculation of pKa. Compounds 5, 17−22, and 28−30 were
retrieved from those optimized for docking modeling and the
3835
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836

Journal of Medicinal Chemistry
Article
(12) Younes, M.; Brown, R. W.; Mody, D. R.; Fernandez, L.;
Laucirica, R. GLUT1 expression in human breast carcinoma:
correlation with known prognostic markers. Anticancer Res. 1995, 15,
2895−2898.
(13) Rudlowski, C.; Becker, A. J.; Schroder, W.; Rath, W.; Buttner,
R.; Moser, M. GLUT1 messenger RNA and protein induction relates
to the malignant transformation of cervical cancer. Am. J. Clin. Pathol.
2003, 120, 691−698.
(14) Haber, R. S.; Rathan, A.; Weiser, K. R.; Pritsker, A.; Itzkowitz, S.
H.; Bodian, C.; Slater, G.; Weiss, A.; Burstein, D. E. GLUT1 glucose
transporter expression in colorectal carcinoma: a marker for poor
prognosis. Cancer 1998, 83, 34−40.
(30) Wei, S.; Kulp, S. K.; Chen, C. S. Energy restriction as an
antitumor target of thiazolidinediones. J. Biol. Chem. 2010, 285, 9780−
9791.
(31) Wei, S.; Chuang, H. C.; Tsai, W. C.; Yang, H. C.; Ho, S. R.;
Paterson, A. J.; Kulp, S. K.; Chen, C. S. Thiazolidinediones mimic
glucose starvation in facilitating Sp1 degradation through the up-
regulation of beta-transducin repeat-containing protein. Mol. Pharma-
col. 2009, 76, 47−57.
(32) Yang, J.; Wei, S.; Wang, D. S.; Wang, Y. C.; Kulp, S. K.; Chen,
C. S. Pharmacological exploitation of the peroxisome proliferator-
activated receptor gamma agonist ciglitazone to develop a novel class
of androgen receptor-ablative agents. J. Med. Chem. 2008, 51, 2100−
2107.
(15) Nagase, Y.; Takata, K.; Moriyama, N.; Aso, Y.; Murakami, T.;
Hirano, H. Immunohistochemical localization of glucose transporters
in human renal cell carcinoma. J. Urol. 1995, 153, 798−801.
(16) Younes, M.; Brown, R. W.; Stephenson, M.; Gondo, M.; Cagle,
P. T. Overexpression of Glut1 and Glut3 in stage I nonsmall cell lung
carcinoma is associated with poor survival. Cancer 1997, 80, 1046−51.
(17) Cantuaria, G.; Fagotti, A.; Ferrandina, G.; Magalhaes, A.; Nadji,
M.; Angioli, R.; Penalver, M.; Mancuso, S.; Scambia, G. GLUT-1
expression in ovarian carcinoma: association with survival and
response to chemotherapy. Cancer 2001, 92, 1144−1150.
(18) Stewart, G. D.; Gray, K.; Pennington, C. J.; Edwards, D. R.;
Riddick, A. C.; Ross, J. A.; Habib, F. K. Analysis of hypoxia-associated
gene expression in prostate cancer: lysyl oxidase and glucose
transporter-1 expression correlate with Gleason score. Oncol. Rep.
2008, 20, 1561−1567.
(19) Haber, R. S.; Weiser, K. R.; Pritsker, A.; Reder, I.; Burstein, D. E.
GLUT1 glucose transporter expression in benign and malignant
thyroid nodules. Thyroid 1997, 7, 363−367.
(20) Baer, S. C.; Casaubon, L.; Younes, M. Expression of the human
erythrocyte glucose transporter Glut1 in cutaneous neoplasia. J. Am.
Acad. Dermatol. 1997, 37, 575−577.
(33) Yang, C. C.; Wang, Y. C.; Wei, S.; Lin, L. F.; Chen, C. S.; Lee, C.
C.; Lin, C. C. Peroxisome proliferator-activated receptor gamma-
independent suppression of androgen receptor expression by
troglitazone mechanism and pharmacologic exploitation. Cancer Res.
2007, 67, 3229−3238.
(34) Muller, K.; Faeh, C.; Diederich, F. Fluorine in pharmaceuticals:
looking beyond intuition. Science 2007, 317, 1881−1886.
(35) Hagmann, W. K. The many roles for fluorine in medicinal
chemistry. J. Med. Chem. 2008, 51, 4359−4369.
(36) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Fluorine in
medicinal chemistry. Chem. Soc. Rev. 2008, 37, 320−330.
(37) Joost, H. G.; Thorens, B. The extended GLUT-family of sugar/
polyol transport facilitators: nomenclature, sequence characteristics,
and potential function of its novel members (review). Mol. Membr.
Biol. 2001, 18, 247−256.
(38) Chen, C. H.; Huang, P. H.; Chu, P. C.; Chen, M. C.; Chou, C.
C.; Wang, D.; Kulp, S. K.; Teng, C. M.; Wang, Q.; Chen, C. S. Energy
restriction-mimetic agents induce apoptosis in prostate cancer cells in
part through epigenetic activation of KLF6 tumor suppressor gene
expression. J. Biol. Chem. 2011, 286, 9968−9976.
(39) Inoki, K.; Zhu, T.; Guan, K. L. TSC2 mediates cellular energy
response to control cell growth and survival. Cell 2003, 115, 577−590.
(40) Salas-Burgos, A.; Iserovich, P.; Zuniga, F.; Vera, J. C.; Fischbarg,
J. Predicting the three-dimensional structure of the human facilitative
glucose transporter GLUT1 by a novel evolutionary homology
strategy: insights on the molecular mechanism of substrate migration,
and binding sites for glucose and inhibitory molecules. Biophys. J.
2004, 87, 2990−2999.
(21) Schwartzenberg-Bar-Yoseph, F.; Armoni, M.; Karnieli, E. The
tumor suppressor p53 down-regulates glucose transporters GLUT1
and GLUT4 gene expression. Cancer Res. 2004, 64, 2627−2633.
(22) Young, C. D.; Anderson, S. M. Sugar and fat - that’s where it’s
at: metabolic changes in tumors. Breast Cancer Res. 2008, 10, 202.
(23) Airley, R. E.; Mobasheri, A. Hypoxic regulation of glucose
transport, anaerobic metabolism and angiogenesis in cancer: novel
pathways and targets for anticancer therapeutics. Chemotherapy 2007,
53, 233−256.
(41) Livak, K. J.; Schmittgen, T. D. Analysis of relative gene
expression data using real-time quantitative PCR and the 2(-ΔΔC(T))
method. Methods 2001, 25, 402−408.
(24) Park, J. B. Inhibition of glucose and dehydroascorbic acid
uptakes by resveratrol in human transformed myelocytic cells. J. Nat.
Prod. 2001, 64, 381−384.
(25) Harmon, A. W.; Patel, Y. M. Naringenin inhibits glucose uptake
in MCF-7 breast cancer cells: a mechanism for impaired cellular
proliferation. Breast Cancer Res. Treat. 2004, 85, 103−110.
(26) Cao, X.; Fang, L.; Gibbs, S.; Huang, Y.; Dai, Z.; Wen, P.; Zheng,
X.; Sadee, W.; Sun, D. Glucose uptake inhibitor sensitizes cancer cells
to daunorubicin and overcomes drug resistance in hypoxia. Cancer
Chemother. Pharmacol. 2007, 59, 495−505.
(27) Wood, T. E.; Dalili, S.; Simpson, C. D.; Hurren, R.; Mao, X.;
Saiz, F. S.; Gronda, M.; Eberhard, Y.; Minden, M. D.; Bilan, P. J.; Klip,
A.; Batey, R. A.; Schimmer, A. D. A novel inhibitor of glucose uptake
sensitizes cells to FAS-induced cell death. Mol. Cancer Ther. 2008, 7,
3546−3555.
(28) Shanmugam, M.; McBrayer, S. K.; Qian, J.; Raikoff, K.; Avram,
M. J.; Singhal, S.; Gandhi, V.; Schumacker, P. T.; Krett, N. L.; Rosen,
S. T. Targeting glucose consumption and autophagy in myeloma with
the novel nucleoside analogue 8-aminoadenosine. J. Biol. Chem. 2009,
284, 26816−26830.
(29) Chan, D. A.; Sutphin, P. D.; Nguyen, P.; Turcotte, S.; Lai, E. W.;
Banh, A.; Reynolds, G. E.; Chi, J. T.; Wu, J.; Solow-Cordero, D. E.;
Bonnet, M.; Flanagan, J. U.; Bouley, D. M.; Graves, E. E.; Denny, W.
A.; Hay, M. P.; Giaccia, A. J. Targeting GLUT1 and the Warburg effect
in renal cell carcinoma by chemical synthetic lethality. Sci. Transl. Med.
2011, 3, 94ra70.
3836
dx.doi.org/10.1021/jm300015m | J. Med. Chem. 2012, 55, 3827−3836