Development of a Cy3-labeled glucose bioprobe and its application in bioimaging and screening for anticancer agents

Article abstract of DOI:10.1002/anie.200604364

(Chemical Equation Presented) Tracing glucose uptake: Fluorescent glucose analogues were synthesized, and the importance of stereochemistry for cellular uptake efficiency was demonstrated. The chiral bioprobe 1 showed superior properties as a glucose-upta

Full text of DOI:10.1002/anie.200604364

Communications
DOI: 10.1002/anie.200604364
Fluorescent Bioprobes
Development of a Cy3-Labeled Glucose Bioprobe and Its Application
in Bioimaging and Screening for Anticancer Agents**
Jongmin Park, Hyang Yeon Lee, Myung-Haing Cho, and Seung Bum Park*
Glucose is the most important energy source for cell growth;
therefore, a fast-growing cancer cell requires more glucose
than a normal cell. One of the biochemical markers in tumor
malignancy is enhanced tumor glycolysis, primarily due to the
overexpression of glucose transporters (GLUTs) and the
increased activity of mitochondria-bound hexokinases in
tumors.[1] The in vitro and in vivo assessment of glucose
utilization has been of considerable interest. One of the
applications of this assessment is tumor diagnosis by using
[18F]-2-fluoro-2-deoxyglucose (18FDG) based positron-emis-
sion tomography (PET).[2a,b] PET with 18FDG is a molecular-
imaging modality that monitors metabolic perturbation in
tumor cells and allows imaging of the exact positions of
tumors in the human body; therefore, it is widely applied in
the diagnosis of various tumors.[2c,d] A fluorescent 2-deoxy-
glucose analogue, 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-
amino]-2-deoxy-d-glucose (2-NBDG), was developed and
extensively studied, primarily by Yoshioka et al.[3] 2-NBDG
has been widely applied in various studies, especially for
tumor imaging and the examination of GLUT-related cell
Scheme 1. Structure of fluorescence-labeled glucose bioprobes: FITC-
labeled cellobiose (1); 2-NBDG (2); Cy3-labeled a-glucose (3); Cy3-
labeled b-glucose (4).
metabolism.[3–8] In addition, some other 2-deoxyglucose
analogues have been reported.[9,10] However, these analogues
have several limitations; for instance, 2-NBDG is applicable
only in a non-physiological sugar-depleted environment. T;
therefore, we designed and synthesized a novel fluorescence-
labeled glucose analogue, a Cy3-linked O-1-glycosylated
glucose, which was not N-2 glycosylated like the previous
analogues (Scheme 1).
In our initial approach, the natural disaccharide cellobiose
was directly labeled with fluorescein isothiocyanate (FITC)
by reductive amination[11] with sodium borohydride (see the
Supporting Information). However, this approach has several
disadvantages: first, the yield is low and the purification
process is tedious; second, only the b-anomer form of the
bioprobe was prepared by using this procedure. In addition,
FITC was not the appropriate fluorescent dye for this study
due to its photochemical properties, especially fast photo-
bleaching during experimentation with a confocal laser
scanning microscope (CLSM) or fluorescence microscope.
Therefore, a second approach was pursued to improve the
reaction yields by using simple purification steps and to
synthesize both anomers (a and b) of d-glucose labeled with
photoresistant dye.
The cyclic form of d-glucose is recognized to be a
hemiacetal ring. Owing to the special reactivity of the
anomeric hydroxy group, 2-bromoethanol can be regioselec-
tively introduced into the anomeric position by acid-catalyzed
Fischer glycosylation with a 2:1 a:b ratio, as confirmed by
NMR spectroscopy.[12] The resulting reaction mixture was
benzoylated for isolation of the a and b anomers (5a and 5b,
respectively; only the synthesis with the a anomer is shown in
Scheme 2).[13] After the coupling of the poly(ethylene glycol)-
amine linker N-Boc-3,6-dioxaoctane-1,8-diamine[14] with both
anomers (5a and 5b) separately, the obtained sugars (6a and
[*] J. Park,[+] H. Y. Lee,[+] Prof. S. B. Park
Department of Chemistry
College of Natural Science
Seoul National University
Seoul 151-747 (Korea)
Fax: (+82)2-889-1568
E-mail: sbpark@snu.ac.kr
Prof. M.-H. Cho
Laboratory of Toxicology
College of Veterinary Toxicology
Seoul National University
Seoul, 151-747 (Korea)
[+] These authors contributed equally to this work.
[**] This work is supported by the Korea Science and Engineering
Foundation (KOSEF), the Molecular and Cellular BioDiscovery
Research Program from the Ministry of Science and Technology
(MOST), and MarineBio21, Ministry of Maritime Affairs and
Fisheries, Korea (MOMAF). M.-H.C. is supported by NSI-NCRC
through KOSEF. H.Y.L. is grateful for the award of a Seoul Science
Fellowship award. J.P. and H.Y.L. are grateful for the award of a BK21
fellowship award.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
2018
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2018 –2022

Angewandte
Chemie
Scheme 2. Synthetic scheme for Cy3-labeled glucose analogues: a) 2-
bromoethanol, Dowex 50WX8-400 ion-exchange resin, 708C reflux;
b) benzoyl chloride, pyridine, DMAP; c) N-Boc-3,6-dioxaoctane-1,8-di-
amine, triethylamine, DMF, 508C; d) 1. NaOMe, MeOH; 2. 50% TFA/
DCM; 3. Cy3-COOH, EDC, DIPEA, DMF. Bz: benzoyl, Boc: tert-
butoxycarbonyl, DMAP: 4-dimethylaminopyridine, DMF: N,N-dimethyl-
formamide, TFA: trifluoroacetic acid, DCM: dichloromethane, EDC: 1-
(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride, DIPEA:
diisopropylethylamine.
Figure 1. A) Cy3-Glc-a uptake by A549 cells. Fluorescence intensities Ifl.
were determined by continuous measurement from regions of interest
in five independent cells marked in (B) and based on an unbiased
selection. B) Merged phase-contrast and fluorescence images in A549
cells captured by live-cell imaging with a CLSM: a) after 0 min; b) after
60 min. The scale bar represents 20 mm.
6b) were treated for debenzoylation and Boc deprotection
(Scheme 2).
After basification of the deprotected products with
DIPEA in order to remove any residual TFA, the resulting
chiral glucoside with a primary amine group can be coupled
with any probe, such as fluorescent dyes or biotin. Our choice
for the final modification was Cy3 fluorescent dye instead of
FITC because the former is tolerant to intense light sources
and compatible with various bioassay systems. We used in-
house Cy3 carboxylic acid synthesized with high yields by
well-documented procedures.[15] EDC coupling of Cy3 car-
boxylic acid with the resulting glucosides yielded the desired
Cy3-labeled d-glucoses after preparative HPLC purification.
Both anomers of Cy3-labeled d-glucose were prepared by
using identical procedures with complete control over the
stereochemistry on the glucosides, and the final products were
maximum within 35 min. Previous work with 2-NBDG (2)
reported that an apparent maximum uptake of 2-NBDG was
reached within 30 min, which is consistent with our observa-
tion.[16] Under these optimized conditions, we tested whether
or not our bioprobe acts as a glucose analogue.
Unlike 2-NBDG, our bioprobe is a glucose analogue that
is O-glycosylated at the C-1 position; therefore, we asym-
metrically synthesized both anomers simultaneously under
the assumption that the behavior of these anomers could be
different, because the molecular conformation of Cy3-Glc-a
and Cy3-Glc-b would be quite distinct. To confirm our
hypothesis, we performed real-time uptake measurements
with Cy3-Glc-a by live imaging of A549 cells with an inverted
fluorescence microscope and we compared the findings with
those for Cy3-Glc-b. As shown in Figure 2 A, the uptake of
Cy3-Glc-a was 40% greater than that of Cy3-Glc-b. This led
us to the conclusion that the configuration at the C-1
anomeric position definitely influences the efficiency in
mimicking glucose; this might be due to the binding
orientation of d-glucose in GLUTs.[17] Based on this obser-
vation, further studies in bioimaging and bioapplication were
performed only with the a anomer of the Cy3-labeled glucose
analogue as our choice of the bioprobe. We also measured the
efficiency of Cy3-Glc-a uptake; in particular, we focused our
attention on the differentiation of GLUT-overexpressing
cancer cells (A549: lung carcinoma cell line; HeLa: cervical
carcinoma cell line) from normal cells (WI-38: lung normal
cell line; NIH/3T3: murine fibroblast cell line). Cy3-Glc-a
uptake in NIH/3T3 cells was only 30% of that in A549 cells.
Interestingly, Cy3-Glc-a uptake in HeLa cells was almost
identical to that in A549 cells, which confirms the selective
1
completely characterized by H and 13C NMR spectroscopy,
MALDI-TOF MS, and HRMS (see the Supporting Informa-
tion for detailed procedures).
After completion of chiral Cy3-labeled d-glucose syn-
thesis, we proceeded to evaluate the applicability of our
bioprobes in a bioassay system. First, we measured the
efficiency of Cy3-Glc-a (3) uptake into cells by using a CLSM.
To decide the optimum concentration of Cy3-Glc-a, we
allowed A549 cells (human lung carcinoma cell line) to
incorporate our bioprobe at Cy3-Glc-a concentrations of 100,
50, 25, 12.5, and 6 mm. Based on the repeated tests, we selected
12.5 mm as the optimum concentration for cellular uptake
experiments (see the Supporting Information). With an
optimized concentration for live-cell imaging, we used a
CLSM to measure the optimum incubation time required to
achieve the maximum uptake of our bioprobe. As shown in
Figure 1, the uptake of our bioprobe by A549 cells reaches the
Angew. Chem. Int. Ed. 2007, 46, 2018 –2022
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2019

Communications
Figure 3. Dose-dependent inhibition of Cy3-Glc-a uptake in A549 cells
in the absence or presence of 10 or 50 mm d-glucose. To demonstrate
specific inhibition by d-glucose, an identical experiment was performed
in the presence of 50 mm l-glucose, and this resulted in no uptake
inhibition. Fluorescence intensities were determined by fluorometry,
and the data are the mean measurements from 35–50 cells in an
experimental representative of at least two independent experiments.
Figure 2. A) Cy3-Glc-a and Cy3-Glc-b uptake by A549 cells. Fluores-
cence intensities were determined by fluorometry, and data are the
mean measurements from 35–50 cells in an experimental representa-
tive of at least two independent experiments. B) Efficiency of Cy3-Glc-a
uptake in cancer cells (A549, HeLa) versus normal cells (WI-38, NIH/
3T3). C) Comparison of Cy3-Glc-a uptake by A549 cells (a, e), HeLa
cells (b, f), WI-38 cells (c, g), and NIH/3T3 cells (d, h); a–d) phase-
contrast images; e–h) fluorescence images. The scale bar represents
40 mm.
diffusion. Therefore, Cy3-Glc-a can function as a d-glucose
mimic and can be applied as a research tool in the study of
glucose metabolism.
Having confirmed that Cy3-Glc-a acts as a d-glucose
analogue, we compared Cy3-Glc-a with a fluorescent ana-
logue of 2-deoxyglucose, namely, 2-NBDG.[1,3–8] The cellular
uptake of 2-NBDG cannot be detected under identical
experimental conditions to those used for Cy3-Glc-a. To
achieve a fluorescence intensity with 2-NBDG that is up to
80% that of Cy3-Glc-a, a 10-fold increase in 2-NBDG
concentration and a 20-fold increase in lens exposure time in
d-glucose-depleted medium were needed with identical
experimental setups. In addition, the cellular uptake of 2-
NBDG in normal media (containing 10 mm d-glucose) is
extremely low (> 60% uptake reduction in normal media)
and can hardly be detected by using fluorescence-based
imaging methods.[16] It is detectable only in glucose-depleted
media, and this is a critical limitation on its application in
biologically significant environments. In contrast to the
results with 2-NBDG, the reduction in Cy3-Glc-a uptake in
glucose-containing media is 5% compared to that in glucose-
depleted media (Figure 4). Therefore, Cy3-Glc-a can be
applied in a bioassay system without glucose starvation.
Based on the fact that Cy3-Glc-a could be taken up by
cells as a d-glucose analogue through a d-glucose-specific
transport mechanism in normal glucose-containing medium,
we applied Cy3-Glc-a to the screening of small-molecule
modulators involved in cellular metabolism. We postulated
that the depression of cellular metabolism in cancer cells,
which is closely related to the reduction in glucose uptake, can
be caused by anticancer agents. We intended to monitor this
phenomenon with our fluorescent bioprobe Cy3-Glc-a. As a
proof-of-principle experiment, we treated A549 cancer cells
with taxol, an anticancer agent, and we measured the uptake
of Cy3-Glc-a at t = 3, 6, 12, and 24 h after treatment. When we
treated the cells with 9.8 mm taxol, the cellular uptake of our
bioprobe was reduced as the incubation time increased
(Figure 5); this clearly demonstrates the potential of Cy3-
uptake of Cy3-Glc-a in cancer cells with enhanced glucose
metabolism (Figure 2B). These data demonstrated that the
cellular uptake of Cy3-Glc-a is affected by the higher glucose
metabolism in cancer cells, which in turn relies on the
adenosine triphosphate (ATP) generated from glycolysis in
order to meet the energy requirements of rapidly replicating
tissue. Therefore, glucose metabolism is strongly correlated
with the GLUT/hexokinase expression levels. These data led
us to the application of Cy3-Glc-a in bioimaging and bioassay
in cancer studies.
To confirm whether the intracellular uptake pathway of
glucose analogues is relevant to that of d-glucose, the direct
competition experiment has been utilized in many stud-
ies.[1,8,9,16] If the cellular uptake of a certain glucose analogue
depends on the concentration of d-glucose but not on that of
l-glucose, then that particular glucose analogue would enter
the cell through a GLUT-mediated glucose-uptake system.
Based on prior experiments, we tested the process of the
cellular uptake of Cy3-Glc-a by measuring the efficiency of
Cy3-Glc-a uptake by A549 cells at 378C in RPMI1640
medium lacking glucose or containing 10 mm d-glucose,
50 mm d-glucose, or 50 mm l-glucose. As shown in Figure 3,
the uptake of Cy3-Glc-a decreased as the concentration of d-
glucose in the medium increased. However, the uptake of
Cy3-Glc-a was not influenced by the l-glucose in the
medium. In addition, we measured the uptake of Cy3-Glc-a
by A549 cells in a medium containing 55 mm alanine, in order
to ensure that the osmotic pressure in the medium does not
affect Cy3-Glc-a uptake; we observed no difference in the
uptake efficiency in the medium with or without 55 mm
alanine (see the Supporting Information). The uptake of
Cy3-Glc-a by A549 cells is inhibited by d-glucose but not by
l-glucose; this suggests that Cy3-Glc-a is taken up by the cell
through a glucose-specific transport system, not by passive
2020 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2018 –2022

Angewandte
Chemie
Glc-a for evaluation of the metabolic perturbation caused by
bioactive small molecules in live cells under physiological
conditions. Therefore, Cy3-Glc-a should not affect cell
viability. By measuring the viability of cells treated with
Cy3-Glc-a (12.5 mm) by using a CCK-8 kit (Dojindo, Japan),
we concluded that Cy3-Glc-a treatment does not affect cell
viability under our incubation conditions (see the Supporting
Information).
We also observed the dose dependency of Cy3-Glc-a
uptake by changing the taxol concentration from 490 to 49 nm
(Table 1). Another anticancer agent, combretastatin, also
Figure 4. Uptake images of 2-NBDG and Cy3-Glc-a by A549 cells
under different conditions: a,d) 12.5 mm Cy3-Glc-a with lens exposure
of the CCD camera for 500 ms; b,e) 12.5 mm 2-NBDG with lens
exposure of the CCD camera for 500 ms; c,f) 125 mm 2-NBDG with
lens exposure of the CCD camera for 11000 ms; a–c) phase-contrast
images; d–f) fluorescence images. Fluorescence of 2-NBDG cannot be
detected under the same conditions as those used for Cy3-Glc-a. Even
125 mm 2-NBDG cannot be detected with a lens exposure of the CCD
camera for 500 ms. A 10-fold increase in concentration and 20-fold
increase in lens exposure time was needed to get measurable
fluorescence intensity with 2-NBDG. The scale bar represents 40 mm.
Table 1: Dose dependence of Cy3-Glc-a uptake by A549 cells in the
presence of various concentrations of anticancer agents.
After 6 h
After 12 h
taxol (9.8 mm)
taxol (490 nm)
taxol (49 nm)
combretastatin (2 mm)
54.8%
88.4%
96.5%
57.4%
50.2%
61.8%
86.9%
45.0%
prohibited cellular proliferation by disturbing cellular metab-
olism, as indicated by a reduction in the cellular uptake of
Cy3-Glc-a. Based on these proof-of-principle experiments,
we are confident that Cy3-Glc-a can be used to evaluate the
behavior of bioactive small molecules in cells in a similar
manner to the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphe-
nyltetrazolium bromide) assay, which measures the mito-
chondrial function.[18] The advantages of a screening system
with Cy3-Glc-a over an MTT assay are as follows: first, the
measurement time is short. MTT assays usually take 24 h and
may take up to 72 h in many cases, that is, when taxol and
combretastatin are used; however, the screening system with
Cy3-Glc-a showed significant differences in 6 h and the
maximum difference in 12 h. Second, the observation channel
of a screening system with Cy3-Glc-a is quite different from
that of a cell-viability assay; the former involves measuring
glucose-uptake efficiency, while the latter involves measuring
mitochondria function. Therefore, we envision that the two
assay systems will compensate for each others limitations.
In conclusion, we designed and completed the asymmetric
synthesis of novel fluorescent glucose analogues and we
demonstrated the importance of stereochemistry at the
anomeric C-1 position of glucose for efficiency of cellular
uptake in A549 lung carcinoma cells. The chiral bioprobe of
our choice, namely Cy3-Glc-a, showed superior properties as
a glucose-uptake tracer compared to the previously reported
2-NBDG. We also developed a novel system for the screening
of anticancer agents through the detection of metabolic
perturbation by measuring glucose uptake in cancer cells with
our bioprobe Cy3-Glc-a. The development of a high-
throughput screening (HTS) system with Cy3-Glc-a is
currently underway, not only for the discovery of novel
anticancer agents but also for that of glucose-uptake regulat-
ing agents for the treatment of obesity or diabetes.[16]
Figure 5. A) Cy3-Glc-a uptake by A549 cells is measured after 0, 3, 6,
12, and 24 h of treatment with taxol (9.8 mm) at 378C. Fluorescence
intensities were determined by fluorometry, and the data are the mean
measurements from 35–50 cells from an experimental representative
of at least two independent experiments. B) Fluorescence images of
Cy3-Glc-a uptake by A549 cells after treatment with taxol (9.8 mm) at
378C for a) 0, b) 3, c) 6, d) 12, and e) 24 h. After incubation with taxol,
each image was captured with a fluorescence microscope after 30 min
of Cy3-Glc-a treatment as described in the Supporting Information.
The fluorescence intensity decreased as the incubation time with taxol
was increased. The scale bar represents 40 mm. f) Phase-contrast
image after 6 h incubation.
Received: October 25, 2006
Published online: February 7, 2007
Angew. Chem. Int. Ed. 2007, 46, 2018 –2022
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2021

Communications
[9] Z. Cheng, J. Levi, Z. Xiong, O. Gheysens, S. Keren, X. Chen, S. S.
Gambhir, Bioconjugate Chem. 2006, 17, 662 – 669.
[10] a) Z. Zhang, H. Li, Q. Liu, L. Zhou, M. Zhang, Q. Luo, J.
Glickson, B. Chance, G. Zheng, Biosens. Bioelectron. 2004, 20,
643 – 650; b) Y. Ye, S. Bloch, S. Achilefu, J. Am. Chem. Soc. 2004,
126, 7740 – 7741.
[11] K. D. McReynolds, A. Bhat, J. C. Conboy, S. S. Saavedra, J.
Gervay-Hague, Bioorg. Med. Chem. 2002, 10, 625 – 637.
[12] F. Fazio, M. C. Bryan, O. Blixt, J. C. Paulson, C.-H. Wong, J. Am.
Chem. Soc. 2002, 124, 14397 – 14402.
[13] a) M. A. Maier, C. G. Yannopoulos, N. Mohamed, A. Roland, H.
Fritz, V. Mohan, G. Just, M. Manoharan, Bioconjugate Chem.
2003, 14, 18 – 29; b) R. E. Campbell, M. E. Tanner, J. Org. Chem.
1999, 64, 9487 – 9492.
[14] a) M. Trester-Zedlitz, K. Kamada, S. K. Burley, D. Fenyo, B. T.
Chait, T. W. Muir, J. Am. Chem. Soc. 2003, 125, 2416 – 2425; b) Y.
Li, Y.-M. Zhu, H.-J. Jiang, J.-P. Pan, G.-S. Wu, J.-M. Wu, Y.-L.
Shi, J.-D. Yang, B.-A. Wu, J. Med. Chem. 2000, 43, 1635 – 1640.
[15] G. A. Korbel, G. Lalic, M. D. Shair, J. Am. Chem. Soc. 2001, 123,
361 – 362.
Keywords: antitumor agents · bioassays · bioimaging ·
fluorescent probes · glucose
.
[1] M. Zhang, Z. Zhang, D. Blessington, H. Li, T. M. Busch, V.
Madrak, J. Miles, B. Chance, J. D. Glickson, G. Zheng, Biocon-
jugate Chem. 2003, 14, 709 – 714.
[2] a) P. Som, H. L. Atkins, D. Bandoypadhyay, J. S. Fowler, R. R.
MacGregor, K. Matsui, Z. H. Oster, D. F. Sacker, C. Y. Shiue, H.
Turner, C. N. Wan, A. P. Wolf, S. V. Zabinski, J. Nucl. Med. 1980,
21, 670 – 675; b) H. Yorimitsu, Y. Murakami, H. Takamatsu, S.
Nishimura, E. Nakamura, Angew. Chem. 2005, 117, 2768 – 2771;
Angew. Chem. Int. Ed. 2005, 44, 2708 – 2711; c) P. S. Conti, D. L.
Lilien, K. Hawley, J. Keppler, S. T. Grafton, J. R. Bading, Nucl.
Med. Biol. 1996, 23, 717 – 735; d) J. Czernin, M. E. Phelps, Annu.
Rev. Med. 2002, 53, 89 – 112.
[3] K. Yoshioka, H. Takahashi, T. Homma, M. Saito, K. B. Oh, Y.
Nemoto, H. Matsuoka, Biochim. Biophys. Acta 1996, 1289, 5 – 9.
[4] K. Yoshioka, M. Saito, K. B. Oh, Y. Nemoto, H. Matsuoka, M.
Natsume, H. Abe, Biosci. Biotechnol. Biochem. 1996, 60, 1899 –
1901.
[16] R. G. OꢀNeil, L. Wu, N. Mullani, Mol. Imaging Biol. 2005, 7,
388 – 392.
[17] P. W. Hruz, M. M. Mueckler, Mol. Membr. Biol. 2001, 18, 183 –
193.
[5] A. Natarajan, F. Srienc, J. Microbiol. Methods 2000, 42, 87 – 96.
[6] K. B. Oh, H. Matsuoka, Int. J. Food Microbiol. 2002, 76, 47 – 53.
[7] P. G. Lloyd, C. D. Hardin, M. Sturek, Physiol. Res. 1999, 48, 401 –
410.
[18] T. Mosmann, J. Immunol. Methods 1983, 65, 55 – 63.
[8] K. Yamada, M. Nakata, N. Horimoto, M. Saito, H. Matsuoka, N.
Inagaki, J. Biol. Chem. 2000, 275, 22278 – 22283.
2022 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2018 –2022