Synthesis and evaluation of [18F]-FEAnGA as a PET tracer for β-glucuronidase activity

Article abstract of DOI:10.1021/bc9004602

To increase the therapeutic index of chemotherapeutic drugs, prodrugs have been investigated as anticancer agents, as they may present fewer cytotoxic side effects than conventional cytotoxic drugs, while therapeutic efficacy is maintained or even increas

Full text of DOI:10.1021/bc9004602

Bioconjugate Chem. 2010, 21, 911–920
911
Synthesis and Evaluation of [18F]-FEAnGA as a PET Tracer for
ꢀ-Glucuronidase Activity
Inês F. Antunes,*,† Hidde J. Haisma,‡ Philip H. Elsinga,† Rudi A. Dierckx,† and Erik F. J. de Vries†
Department of Nuclear Medicine and Molecular Imaging, University Medical Center Groningen, University of Groningen, P.O.
Box 30.001, 9700 RB Groningen, The Netherlands, and Department of Pharmaceutical Gene Modulation, University Center for
Pharmacy, University of Groningen, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands. Received October 21, 2009;
Revised Manuscript Received January 29, 2010
To increase the therapeutic index of chemotherapeutic drugs, prodrugs have been investigated as anticancer
agents, as they may present fewer cytotoxic side effects than conventional cytotoxic drugs, while therapeutic
efficacy is maintained or even increased. Extracellular ꢀ-glucuronidase (ꢀ-GUS) in the tumors has been
investigated as a target enzyme for prodrug therapy, as it can convert nontoxic prodrugs into cytostatic
drugs. To optimize ꢀ-GUS-based prodrug therapies, PET imaging could be a useful tool by providing
information regarding the localization and quantification of ꢀ-GUS. Here, we describe our first PET tracer
for extracellular ꢀ-GUS, [18F]-FEAnGA, which consists of a 2-[18F]fluoroethylamine ([18F]-FEA) group bound
to a glucuronic acid via a self-immolative nitrophenyl spacer. [18F]-FEAnGA was synthesized by alkylation
of its imidazole carbamate precursor with [18F]-FEA, followed by deprotection of the sugar moiety with
NaOH in 10-20% overall radiochemical yield. [18F]-FEAnGA is about 10-fold more hydrophilic than the
cleavage product [18F]-FEA, and it is stable in PBS and rat plasma for at least 3 h. In the presence of either
Escherichia coli ꢀ-GUS or bovine liver ꢀ-GUS, in vitro cleavage of [18F]-FEAnGA with complete release
of [18F]-FEA was observed within 30 min. C6 glioma cells incubated with the tracer and Escherichia coli
ꢀ-GUS or bovine liver ꢀ-GUS showed a 4- and 1.5-fold higher uptake of radioactivity, respectively, as
compared to control C6 cells without ꢀ-GUS. Incubation of CT26 murine colon adenocarcinoma cells or the
genetically engineered CT26mꢀGUS cells, which expressed membrane-anchored GUS on the outer cell
membrane, with the tracer, resulted in a 3-fold higher uptake into GUS-expressing cells as compared to
control cells. In a preliminary microPET study in mice bearing both CT26 and CT26mꢀGUS tumors, [18F]-
FEAnGA exhibited a 2-fold higher retention of radioactivity in the tumor expressing ꢀ-GUS than in the
control tumor. [18F]-FEA did not show any difference in tracer uptake between tumors. These results suggest
that [18F]-FEAnGA may be a suitable PET tracer for evaluation of ꢀ-GUS activity, since it is specifically
cleaved by ꢀ-GUS and the released [18F]-FEA remains attached to targeted cells.
various inflammatory processes (2). It is also known that tumors
have high extracellular levels of ꢀ-GUS, due to invading
monocytes and granulocytes (3). Extracellular ꢀ-GUS is also
associated with tumor invasiveness and metastatic potential (3).
Therefore, ꢀ-GUS seems to be a promising candidate for a PD-
converting enzyme.
INTRODUCTION
To overcome the limitation of toxicity associated with
chemotherapeutic drugs, local conversion of a nontoxic prodrug
(PD) to the active cytotoxic drug in or in the vicinity of tumors
provides a means for delivering active drugs selectively to tumor
cells while sparing the normal cells of a host.
Glucuronide-drug conjugates exhibit improved water
solubility, stability, and lower toxicity toward normal cells
than the native drug. However, glucuronide-drug conjugates
are susceptible to enzymatic cleavage by ꢀ-GUS, resulting
in the in situ generation of highly cytotoxic drugs that can
selective incorporate into targeted tumor cells (1, 4, 5). A
number of prodrugs for ꢀ-GUS have been designed. These
prodrugs generally consist of three distinct parts: a trigger,
a linker, and an effector unit. For ꢀ-GUS targeted prodrugs,
the trigger and effector units consist of a glucuronide moiety
and the cytostatic drug, respectively. They are joined together
by the linker unit, which should release the drug as fast as
possibleafterprodrugactivationviaachemicalbreakdown(6-8).
ꢀ-glucuronidase (ꢀ-GUS) (EC 3.2.1.31) belongs to a group
of isoenzymes that catalyze the hydrolysis of ꢀ-D-glucuronides
to the corresponding alcohol and D-glucuronate. The enzyme
is present in virtually all tissues and can be found in microsomes
(endoplasmic reticulum) and lysosomes (Fishman). The human
enzyme has a pH optimum of 4 to 5. Under normal tissue pH,
only about 10% of the maximum activity is retained. ꢀ-GUS
concentrations are low in human serum and the extracellular
compartment of normal tissue (1). However, increased levels
of extracellular ꢀ-GUS have been reported in a number of
pathological conditions, such as rheumatoid arthritis, leprosy,
myocardial infarction, liver fibrosis, bacterial infection, and
Encouraging results were obtained with the prodrugs DOX-
GA3 and HMR1826 both carrying para-substituted aromatic
spacers, which were able to release doxorubicin after enzymatic
cleavage of the glucuronide moiety by extracellular ꢀ-GUS
(Scheme 1).
As follows from the above, bioactivation of the glucuronide
prodrug by ꢀ-GUS is an attractive approach in prodrug therapy.
* Author for correspondence: Ineˆs Antunes, Department of
Nuclear Medicine and Molecular Imaging, University Medical Center
Groningen, University of Groningen, PO Box 30.001, 9700 RB
Groningen, The Netherlands, Tel.: +31-50-3619309, FAX: +31-
50-3611687, E-mail: i.farinha.antunes@ngmb.umcg.nl.
† University Medical Center Groningen.
‡ University Center for Pharmacy, University of Groningen.
10.1021/bc9004602  2010 American Chemical Society
Published on Web 04/23/2010

912 Bioconjugate Chem., Vol. 21, No. 5, 2010
Antunes et al.
Scheme 1. Mechanism of Activation of the Prodrugs HMR1826 and DOX-GA3 by ꢀ-GUS
Scheme 2. General Structure of the PET Tracer and Its Mechanism of Activation
However, interindividual and intertissue variability of ꢀ-GUS
activity, as well as factors modulating the enzyme’s expression
and activity, are key determinants for the success of these
therapeutic strategies (9, 10).
For PET imaging of ꢀ-GUS activity, a suitable tracer is
necessary (20). Therefore, we aimed to develop a PET tracer
that allows repetitive noninvasive imaging of ꢀ-GUS activity
in living subjects. Thus, we designed a potential PET tracer
with a structure based on HMR1826. In the PET tracer, the
doxorubicin moiety is replaced by a [18F]-fluoroalkylamine
group.
Due to the highly hydrophilic character of the glucuronide,
the intact tracer will be rapidly cleared from plasma and from
tissues that do not have extracellular ꢀ-GUS. However, in tissues
with high levels of ꢀ-GUS, the enzyme should cleave the
glucuronic acid moiety from the PET tracer. As a result, the
less hydrophilic [18F]-fluoroethylamine will be released after
self-immolation of the spacer, and we hypothesized that [18F]-
fluoroethylamine will be trapped in the target tissue (Scheme
2).
For a better understanding of the mechanisms of action and
metabolic conversion of these PD, some in vivo optical
measurements of gene/enzyme expression have been done using
substrates with fluorescent or bioluminescent properties. These
optical imaging modalities are useful in superficial tissues and
have extensive application in small animals, but application to
larger animals or humans is limited by the depth of light
penetration (11-13). Monitoring enzyme or gene expression
in larger animals and ultimately in humans is possible with the
use of specific radiotracers combined with nuclear imaging
techniques like PET or SPECT (14-17). Among these tech-
niques, PET is of particular interest for drug development
because of its high sensitivity and ability to provide quantitative
kinetic data (18, 19).
This study aimed to provide a proof of principle of this tracer
design. The tracer may be further optimized by replacing the

[18F]-FEAnGA for ꢀ-Glucuronidase Activity Imaging
Bioconjugate Chem., Vol. 21, No. 5, 2010 913
[18F]-fluoroethylamine moiety with, for example, a more lipo-
philic or targeted moiety in order to increase trapping efficiency
as a result of increased nonspecific binding.
ArH). 19F NMR (200 MHz, DMSO-d6) δ -217.84. ESI-MS
m/z 592.3 [M+NH4]+, 597.3 [M+Na]+.
Methyl 1-O-(4-((2-Fluoroethylcarbamoyloxy)methyl)-2-nitro-
phenyl)-ꢀ-D-glucopyronuronate (3). 12.6 µL 30% w/v NaOMe
in MeOH was added to a solution of 2 (50 mg, 0.087 mmol) in
anhydrous MeOH at 0 °C. After 2 h of stirring under nitrogen
atmosphere, the reaction was quenched with acetic acid and
concentrated in vacuo. The crude residue was purified by flask
column chromatography (silica, 10% MeOH in CH2Cl2). The
resulting solid was dissolved in acetone and filtered through a
0.2 µm PTFE filter to remove any remaining silica. Trituration
of the solid in diethyl ether yielded 30.3 mg of 3 (78%) as a
white solid.
1H NMR (300 MHz, DMSO-d6) δ 3.2 (m, 2H, CH2N, 3.63
(s, 3H, COOCH3), 4.10 (d, 1H, J ) 9.1 Hz, H-5), 4.38 (dt, 2H,
J ) 47.6 Hz, J ) 10 Hz, CH2F), 5.00 (s, 2H, benzylic CH2),
5.3 (d, 2H, J ) 7 Hz, H-2, H-4), 5.51 (s, 2H, H-3, H-1), 7.42
(d, 1H, J ) 8.8 Hz, ArH), 7.56 (m, 1H, NH), 7.65 (m, 2H, NH,
ArH), 7.84 (s, 1H, ArH). 19F NMR (200 MHz, DMSO-d6) δ
-232. ESI-MS m/z 455.3 [M+Li]+.
MATERIALS AND METHODS
Reagents and solvents were obtained from commercial
suppliers (Aldrich, Fluka, Sigma, and Merck) and used without
further purification. A glucose, magnesium, and calcium solution
(GMC-PBS) was prepared by adding 5.6 mM D-glucose, 0.49
mM MgCl2, and 0.68 mM CaCl2 to 100 mL phosphate-buffered
saline (PBS). Flash chromatography was performed on silica
gel 60 (0.040-0.063, Merck). All reactions were monitored by
thin layer chromatography on Merck F-254 silica gel plates.
Detection of the compounds on the TLC plates was performed
with UV light (254 nm). Sugar compounds were additionally
visualized by staining the TLC strip with 8% H2SO4 followed
by heating. For radiolabeled compounds, the detection on the
TLC was performed with Cyclone phosphor storage screens
(multisensitive, Packard). These screens were exposed to the
TLC strips for a few seconds and subsequently read out using
a Cyclone phosphor storage imager (PerkinElmer) and analyzed
with OptiQuant software. 1H and 19F NMR were recorded on a
Varian 300 and 200 MHz, respectively. Chemical shifts were
determined relative to the signal of the solvent, converted to
the TMS scale, and expressed in δ units (ppm) downfield from
TMS. Coupling constants were reported in Hertz (Hz). Splitting
patternsaredefinedass(singlet),d(doublet),dd(double-doublet),
t (triplet), dt (double-triplet), or m (multiplet). HPLC purifica-
tions were performed with an Elite LaChrom VWR Hitachi
L-2130 pump system using a Phenomenex Prodigy C18-column,
connected to a UV-spectrometer (Elite LaChrom VWR Hitachi
L-2400 UV detector) set at 254 nm and a Bicron frisk-tech
radiation detector. Radioactivity measurements for log D
determination and cell studies were done using an automated
gamma counter (Compugamma, LKB Wallac). Absorbance
measurements in the enzyme assays were obtained with a UV
spectrophotometer (Waters 2487 dual wavelength absorbance
detector). CT26mꢀGUS murine colon adenocarcinoma cells
were kindly provided by Steve Roffler (11), Institute of
Biomedical Science, Academia Sinica, Academia Road, Section
2 No128, 11529 Taipei, Taiwan.
Synthesis. Methyl 1-O-(4-((2-Fluoroethylcarbamoyloxy)
methyl)-2-nitrophenyl)-2,3,4-tri-O-acetyl-ꢀ-D-glucopyronur-
onate (2). Methyl 1-O-(4-(1H-imidazole-carbamoyloxymethyl)-
2-nitrophenyl)-2,3,4-tri-O-acetyl-ꢀ-D-glucopyronuronate (1) was
synthesized in 3 steps according to the procedure described by
Duimstra et al. (21). Carbamate 1 (0.2 g, 0.35 mmol) was
dissolved in 3 mL anhydrous CH2Cl2 under a N2 atmosphere
and cooled to 0 °C. Methyl triflate (45 µL, 0.40 mmol) was
added to the reaction mixture. After stirring for 30 min at 0 °C,
the reaction mixture was diluted with 2 mL Et2O and cooled to
-20 °C to allow the methylated product to precipitate. The white
solid was collected by filtration, washed with Et2O, and dried
in vacuo. The methylated compound was resuspended in 2.5
mL CH2Cl2 under a N2 atmosphere, and 2-fluoroethylamine
hydrochloride (0.05 g, 0.54 mmol) and tryethylamine (75 µL,
0.54 mmol) were added. The slurry was stirred for 5 h at room
temperature and then washed with water and brine. The organic
layer was dried on Na2SO4, concentrated in vacuo, and purified
by flash chromatography (silica, EtOAc-Hex 1:1 + 0.1% Et3N)
to give 81 mg (41%) of 2 as a white solid. 1H NMR (300 MHz,
DMSO-d6) δ 1.97 (s, 3H, OAc), 1.99 (s. 6H, OAc), 3.23 (m,
2H, CH2N), 3.61 (s, 3H, COOCH3), 4.38 (dt, 2H, J ) 47.2 Hz,
J ) 10 Hz, CH2F), 4.71 (d, 1H, J ) 9.5 Hz, H-5), 5.01-5.12
(m, 4H, benzylic CH2, H-2, H-4), 5.43 (t, 1H, J ) 19 Hz, H-3),
5.71 (d, 1H, J ) 7.7 Hz, H-1), 7.40 (d, 1H, J ) 8.4 Hz, ArH),
7.56 (m, 1H, NH), 7.65 (d, 1H, J ) 9.5 Hz, ArH), 7.87 (s. 1H,
1-O-(4-(2-Fluoroethyl-carbamoyloxymethyl)-2-nitrophenyl)-
O-ꢀ-D-glucopyronuronate (FEAnGA) (4). A solution of 3 (20
mg, 0.047 mmol) in acetone (1 mL) was stirred in the presence
of a 1 M NaOH (380 µL) solution at 0 °C. After 10 min, the
mixture was neutralized with 1 M HCl, concentrated in vacuo,
and purified by flash chromatography (Silica, MeCN/H2O 7:3)
to give 17.3 mg of a light yellow solid (91%).
1H NMR (300 MHz, DMSO-d6) δ 3.2 (m, 2H, CH2N), 3.55
(d, 1H, J ) 9.4 Hz, H-5), 4.38 (dt, 2H, J ) 47.6 and 10 Hz,
CH2F), 5.01 (s, 2H, benzylic CH2), 5.11 (d, 2H, J ) 6.8 Hz,
H-2, H-4), 5.30 (d, 2H, J ) 4.4 Hz, H-3, H-1), 7.41 (d, 1H, J
) 9.1 Hz, ArH), 7.60 (m, 2H, NH, ArH), 7.83 (s. 1H, ArH).
19F NMR (200 MHz, DMSO-d6) δ -225; m/z 457.2 [M+Na]+.
Radiolabeling. Radiosynthesis of [18F]-FEAnGA ([18F]-4).
Aqueous [18F]-fluoride was produced by irradiation of [18O]
water with a Scanditronix MC-17 cyclotron via the 18O
(p,n)18F nuclear reaction. The [18F]-fluoride solution was
passed through a SepPak Light Accell plus QMA anion
exchange cartridge (Waters) to recover the 18O-enriched
water. The [18F]-fluoride was eluted from the cartridge with
1 mL of K2CO3 (5 mg/mL) and collected in a vial with 20
mg Kryptofix 2.2.2. To this solution, 1 mL acetonitrile was
added and the solvents were evaporated at 130 °C. The 18F-
KF/Kryptofix complex was dried 3 times by the addition of
0.5 mL acetonitrile, followed by evaporation of the solvent.
[18F]-FEA ([18F]-7) was prepared by fluorination of N-[2-
(toluene-4-sulfonyloxy)ethyl]phthalimide (30 mg, 0.09 mmol)
in acetonitrile at 110 °C, followed by deprotection with
hydrazine hydrate (50 µL, 1.03 mmol) as described in the
literature (22). The resulting amine was distilled at 75 °C
into an ice-cold solution of 1 (5 mg, 0.009 mmol) in
acetonitrile (0.2 mL). The mixture was allowed to react at
room temperature for 30 min. The product was diluted with
20 mL of water and passed through a tC18 cartridge
(activated with 5 mL EtOH and 10 mL H2O). Deprotection
of the hydroxyl groups was done on the tC18 cartridge by
addition of 1 mL 2 M NaOH and incubation for 7 min at
room temperature. The product was eluted from the cartridge
with 4 mL of water and purified by HPLC (10% EtOH in
NaH2PO4 0.025 M (pH 6.3); 4 mL/min; retention time: [18F]-
FEAnGA ) 10 min).
The radiolabeled compound [18F]-FEAnGA ([18F]-4) was
obtained in 5-20% radiochemical yield from [18F]-fluoride
(decay-corrected) in 150 min. At the end of synthesis (EOS),
the specific activity was 5-10 GBq/µmol and the radiochemical
purity was >95%.

914 Bioconjugate Chem., Vol. 21, No. 5, 2010
Antunes et al.
Stability of [18/19F]-FEAnGA. Samples of [18/19F]-FEAnGA
were taken for stability testing. The sample of FEAnGA was
dissolved in PBS (1 mL) and incubated at 37 °C. After 60 min,
the solution was analyzed by HPLC (retention time: FEAnGA
) 3.2 min, HNBA ) 10.4 min) (20% CH3CN/NaH2PO4 0.025
M (pH 6.3); 4 mL/min).
A sample of radiolabeled [18F]-FEAnGA was dissolved in
PBS (1 mL) and rat plasma (1 mL) and incubated at 37 °C.
After 1 h and 3 h of incubation, respectively, the stability of
the radiolabeled [18F]-FEAnGA was followed by Radio-TLC
(Rf FEA ) 0.0, Rf FEAnGA ) 0.4) (eluent: MeOH/CH2Cl2 2:8
+ 0.1% Et3N). After elution, the TLCs were analyzed by
phosphor storage imaging. The screens were scanned with
Cyclone phosphor storage system (PerkinElmer), and the
percentage of conversion of [18F]-FEAnGA as a function of the
incubation period was calculated by ROI analysis using Opti-
Quant software.
In order to not underestimate the amounts of [18F]-FEA due
to its volatile nature, we created a calibration curve of the
amount of [18F]-FEA evaporated over time in the radio TLC (y
) -0.2141x + 103.11, R2 ) 0.72, P ) 0.03) and corrected all
values. The calibration curve was prepared by pipetting 1.5 µL
of a stock [18F]-FEA solution at different time points in a TLC
plate. The TLC was read out by storage phosphorus imaging.
Images were analyzed by Optiquant software and the percentage
of evaporated activity was calculated. At 60 min, approximately
10% of [18F]-FEA has evaporated.
determined using the same conditions as for FEAnGA. The
experiment was preformed in triplicate.
In Vitro Hydrolysis of [18F]-FEAnGA. To a 0.6 mL solution
of 0.1% BSA in PBS (pH ) 7), 3.7 µg (74U) Escherichia coli
ꢀ-GUS or 95.2 µg (952U) bovine liver ꢀ-GUS and 0.4 mL of
HPLC purified [18F]-FEAnGA (0.68-1.28 MBq) were added.
The solutions were incubated in a water bath at 37 °C, and 75
µL samples were collected at different time points ranging from
30 s to 30 min after the start of incubation. The enzymatic
activity was stopped by dilution of the samples in 0.5 mL of
acetonitrile and cooling of the samples in ice. The degradation
of the radiolabeled [18F]-FEAnGA was followed by Radio-TLC
(eluent: MeOH/CH2Cl2 2:8 + 0.1% Et3N). After elution, the
TLCs were analyzed by phosphor storage imaging as described
earlier.
Cellular Uptake of [18F]-FEAnGA in C6 Glioma Cells. C6
glioma cells were maintained in 5 mL Dulbecco’s Modified
Eagle Medium (DMEM) supplemented with 7.5% fetal bovine
serum (FBS) in 25 cm3 culture flasks. Cells were grown in a
humidified atmosphere containing 5% CO2 and were passaged
every 3-4 days. For uptake experiments, cells were plated in
triplicate in a 12-well plate at a density of 7 × 105 cells per
well. After 24 h, the medium was discarded and 1 mL of PBS-
GMC buffer per well was added, followed by the addition of
200 µL [18F]-FEAnGA (ca. 250 kBq) or 10 µL [18F]-FEA (ca.
125 kBq). In 2 series of 3 wells, an excess of E. coli ꢀ-GUS
(740U) or bovine liver ꢀ-GUS (340U) was added, respectively.
After 60 min of incubation at 37 °C, the medium was collected
and the cells washed with cold PBS (3 × 1 mL/well), harvested
with trypsin (250 µL), and resuspended in 1750 µL of DMEM.
The cell suspensions as well as the supernatant fractions were
collected separately for each well, and the radioactivity was
measured using a gamma counter. For each well, the medium
was analyzed for conversion of [18F]-FEAnGA into [18F]-FEA
as described earlier.
Cellular Uptake of [18F]-FEAnGA in CT26 Adenocarci-
noma Cells. CT26 control cells as well as the genetically
engineered CT26mꢀGUS murine colon adenocarcinoma cells
that express membrane-anchored GUS on the outer cell mem-
brane were maintained in 10 mL DMEM medium supplemented
with 10% FBS and 1% penicillin-streptomycin in 75 cm3
culture flasks. Cells were grown in a humidified atmosphere
containing 5% CO2 and were passaged every 3-4 days. For
uptake experiments, cells were plated in triplicate in a 12-well
plate at a density of 7 × 105 cells per well. After 24 h, the
medium was discarded and 1 mL of PBS-GMC buffer per well
was added followed by the addition of 200 µL [18F]-FEAnGA
(ca. 250 kBq) or 10 µL [18F]-FEA (ca. 125 kBq). The rest of
the tracer uptake experiment was performed as described above.
Distribution Coefficient (Log D7.4). To determine the Log
D7.4 of [18F]-FEA and [18F]-FEAnGA, an aliquot of 1 mL of
HPLC purified [18F]-FEA or [18F]-FEAnGA solution was added
to a mixture of n-octanol/PBS (5 mL/5 mL) at pH 3 and 7 for
[18F]-FEAnGA and at pH 7 and 10 for [18F]-FEA. The tubes
were vortexed at room temperature for 1 min, followed by 30
min shaking in a water bath at 37 °C. Aliquots of 1000 and 25
µL were drawn from the n-octanol and aqueous phase,
respectively, and the radioactivity was counted using an
automated gamma counter. The experiments were performed
at least in triplicate.
In Vitro Hydrolysis of FEAnGA. To assess the kinetics
parameters for the ꢀ-GUS-catalyzed hydrolysis of FEAnGA,
aliquots of FEAnGA were prepared with concentrations
ranging from 0.001 to 0.01 M in PBS buffer (pH 7.4). To a
0.950 mL solution containing 0.1% BSA in PBS (pH ) 7.4),
3.7 µg (74U) Escherichia coli ꢀ-GUS (E.C.3.2.1.31, 20 000
U/mg Aldrich) and 50 µL of FEAnGA were added. The
solution was incubated at 37 °C in a water bath, and 75 µL
samples were collected at different time points ranging from
30 s to 30 min after the start of incubation. Each sample
was immediately added to 500 µL acetonitrile in ice to stop
the reaction. The release of 4-hydroxy-3-nitrobenzyl alcohol
(HNBA) (ε ) 3460 M-1 cm-1) was monitored by measuring
the UV absorbance at 420 nm. The initial velocities of the
reactions were determined from the slope of the absorbance
curve at time ) 0.The kinetic parameters KM and Vmax were
obtained from a Lineweaver-Burk plot (12, 23, 24).
The kinetic parameters of the hydrolysis of FEAnGA by
bovine liver ꢀ-GUS were determined as described above, but
with minor modifications. Briefly, to a 0.8 mL solution
containing 0.1% BSA in PBS (pH ) 5), 95.2 µg (952U)
bovine liver ꢀ-GUS (B-10, E.C.3.2.1.31, 10 000 U/mg Sigma)
and 200 µL of FEAnGA were added. The same assay was
also performed for a control sample solution (without
enzyme) under the same conditions, and it showed negligible
absorption for the duration of the experiment. For compari-
son, the kinetic parameters of a reference substrate, p-nitro-
phenyl-ꢀ-D-glucuronide (PNPG) (ε ) 7578 M-1 cm-1) were
Efflux Analysis of [18F]-FEA in C6 Glioma Cells. C6
glioma cells were plated in triplicate in a 12-well plate at a
density of 7 × 105 cells per well. After 24 h, the medium was
discarded and 1 mL of PBS-GMC buffer per well was added,
followed by the addition of 10 µL [18F]-FEA (ca. 1.3 MBq).
After 60 min of incubation at 37 °C, the medium was discarded,
the cells washed with cold PBS (3 × 1 mL/well), and 1 mL of
PBS-GMC buffer per well was added. At different time points,
the medium was collected and the cells harvested with trypsin
(250 µL) and resuspended in 1750 µL of DMEM. The cell
suspensions as well as the supernatant fractions were collected
separately for each well, and the radioactivity was measured
using a gamma counter.
MicroPET Imaging in a ꢀ-GUS Expressing Tumor
(CT26mꢀGUS) Mouse Model. The animals were provided with
standard laboratory chow and tap water ad libitum. All studies
were carried out in compliance with the local ethical guidelines
for animal experiments. The protocol was approved by the local

[18F]-FEAnGA for ꢀ-Glucuronidase Activity Imaging
Bioconjugate Chem., Vol. 21, No. 5, 2010 915
Scheme 3. Synthesis of Compound FEAnGA
Animal Ethics Committee. CT26 and CT26mꢀGUS cells
[(1-1.6) × 106, in a 1:1 mixture of Matrigel and DMEM with
10% FBS] were subcutaneously injected into the right and left
shoulders of two male Balb/c mice (6-8 weeks old), respec-
tively. On day 7 after inoculation of the mice with CT26 and
CT26mꢀGUS cells, the mice were anesthetized with 2%
Isoflurane and fixed to the bed in the small animal PET camera
in transaxial position. A 20 min transmission scan with a Co-
57 point source was obtained for the correction of attenuation
of 511 keV photons by tissue. After the transmission scan was
completed, the PET tracer [18F]-FEAnGA (4.4 MBq) or [18F]-
FEA (4.2 MBq) was injected via the penile vein of the animal.
A 60 min static image was acquired with a Focus 220 rodent
scanner (CTI Siemens, Munich, Germany). Emission sinograms
were iteratively reconstructed (OSEM2D) after being normal-
ized, corrected for attenuation, and corrected for decay of
radioactivity. Subsequent to the µPET imaging, the animal fixed
to the bed was positioned in the microCT scanner (MicroCT
II, CTI Siemens) and a microCT image (exposure time ) 1050
ms; X-ray voltage ) 55 kvp; anode current ) 500 µA; number
of rotation steps ) 500; total rotation ) 360°) was acquired
for 15 min for anatomic localization. From the difference in
bed positions from the PET to the CT, a transformation matrix
was constructed, which describes the translation between both
images. This transformation matrix was used for precise fusion
of both images, using Inveon Research Workplace (Siemens)
software.
cleavage of the PET tracer by ꢀ-GUS, the labeled lipophilic
part of the molecule is supposed to be trapped at the cleavage
site by nonspecific binding.
Compound 1 was prepared in 3 steps from methyl 1-bromo-
2,3,4-tri-O-acetyl-R-D-glucopyronuronate according to the pro-
cedure described by Duimstra et al. (21). Activation of the
imidazole intermediate with methyltriflate followed by the
reaction with fluoroethylamine gave compound 2 in 30% yield.
It was observed that the amount of methyl triflate used was
critical in this step, since higher amounts than 1.1 mol equiv
led to lower yields of compound 2. This could be due to the
reaction of fluorethylamine with the excess of methyl triflate
instead of compound 1. The deprotection of the sugar moiety
in 2 in a single step did not give product 4 in the desired yield,
as the reaction was incomplete and led to the formation of
multiple side products. Moreover, the purification of product
FEAnGA by flash chromatography proved difficult, and con-
sequently, it was decided to perform the synthesis of FEAnGA
by a two-step procedure as depicted in Scheme 3. First,
deacetylation of the hydroxyl groups was achieved using sodium
methoxide. Then, the basic hydrolysis of the carboxylic ester
was performed with NaOH, which afforded the unlabeled
reference compound FEAnGA in 50% yield. The overall yield
of this procedure was 23%. The identity of the compound was
confirmed by low-resolution mass spectroscopy (LRMS) and
the disappearance of the methoxy and acetyl groups of the
1
glucuronic moiety in H NMR.
Data Analysis. Apparent kinetic constants (Vmax and KM) were
determined from Lineweaver-Buck plots of the initial rates (1/
V0) against the reciprocal substrate concentration 1/[S] generated
in Sigmaplot software. The comparison between cells with and
without ꢀ-GUS was performed using a two-sided t test that was
paired for the different conditions in a single experiment. In all
cases, P e 0.05 was considered significant.
The anomeric ꢀ-configuration, essential for substrate affinity
for ꢀ-GUS, was confirmed by the NMR, which showed the H-1
proton as a doublet at a chemical shift δH-1 of 5.71 ppm with a
characteristic coupling constant JH1-H2 of 7.7 Hz. Typically, the
H1-H2 coupling constants of the R anomer are 2-5 Hz,
whereas the coupling constant of the ꢀ anomer is between 6
and 10 Hz (19, 25, 26).
The synthesis of [18F]-FEAnGA is a 3-step procedure
consisting of the preparation of [18F]-FEA followed by the
coupling of the amine with compound 1 and the removal of
the protecting groups. [18F]-FEA was prepared according to
the procedure described by Tewson (22) by reaction of the
tosylate 5 with [18F]fluoride and subsequent removal of the
phthalimide group 6 with hydrazine. This last step is of
critical importance with regard to the final labeling yield. In
order to obtain distilled [18F]-FEA in 48 ( 9% (n ) 8) yield,
it is necessary to use hydrated hydrazine in a reaction mixture
that is free of acetonitrile to avoid the evaporation of the
hydrazine associated water, which is required for the hy-
drolysis of the phthalimide moiety. On the other hand,
distillation should not be performed at higher temperatures
than 75 °C or longer than 15 min to avoid the distillation of
hydrazine into the second reaction vial. Hydrazine competes
with [18F]-FEA in the carbamate formation reaction, thus
RESULTS AND DISCUSSION
Prior studies with DOX-GA3 and HMR 1826, which have a
ꢀ-glucuronyl group attached to doxorubicin via a para-
substituted aromatic group showed promising preclinical results
for PD therapy. Their low general toxicity has been clearly
demonstrated in mice and rats studies with human tumor
xenografts such as lungs, colon, breast, ovary, stomach, and
pancreatic cancer. Experiments in ex vivo human lung tumor
specimens have demonstrated that HMR1826 leads to increased
levels of free DOX in tumor tissues and decreased levels in
normal tissues compared to native DOX, while DOX-GA3
showed enhanced efficacy in ovarian cancer xenografts in
comparison to native DOX. Inspired by the structures of the
doxorubicin prodrugs, we designed an 18F-labeled PET tracer
for imaging of ꢀ-GUS (Scheme 2). In this PET tracer, the DOX
moiety is replaced by a lipophilic labeled molecule; after

916 Bioconjugate Chem., Vol. 21, No. 5, 2010
Antunes et al.
Scheme 4. Radiosynthesis of Compound [18F]FEAnGA
decreasing the final yield. The carbamate formation of com-
pound 1 with the distilled [18F]-FEA gives compound [18F]-
FEAnGA in 33 ( 19% (n ) 8) yield after 30 min of reaction
at room temperature. Deprotection of the hydroxyl groups was
done on the tC18 cartridge by addition of 1 mL 2 M NaOH
and incubation for 7 min at room temperature. The product was
eluted from the cartridge with 4 mL of water and was purified
by HPLC in 79 ( 9% (n ) 8) yield. In contrast to the
deprotection of the reference compound 2 which was obtained
in 2 steps, surprisingly it was possible to deprotect [18F]-2 in a
single step giving our final product in good yield. Thus, the
final product [18F]-FEAnGA was obtained in 13 ( 7% (n ) 8)
overall radiochemical yield (corrected for decay, based on [18F]-
fluoride) with a total synthesis time of 150 min (Scheme 4).
The identity of the tracer was confirmed by coelution with an
authentic sample of the nonradioactive compound by RP-HPLC.
The stability of [18/19F]-FEAnGA in PBS was determined at
37 °C. After 1 h of incubation, HPLC analysis showed that 98%
of the tracer was still intact. The stability of [18F]-FEAnGA in
PBS and rat plasma were also determined at 37 °C. After 1 and
3 h of incubation, 98% and 99% (Radio-TLC) of the radioactiv-
ity still corresponded to the intact tracer, respectively. This
indicates that the tracer is highly stable in vitro.
The lipophilicities of [18F]-FEAnGA and [18F]-FEA were
determined by measuring their n-octanol/water distribution
coefficient at pH 7.4; the Log D was found to be -1.61 ( 0.01
for [18F]-FEAnGA and -0.69 ( 0.02 for [18F]-FEA. As
expected, [18F]-FEAnGA is about 10 times more hydrophilic
than [18F]-FEA due to the presence of the glucuronic acid
moiety. Moreover, [18F]-FEAnGA is less hydrophilic at pH 3
(Log D3.0 -1.01 ( 0.03) than at pH 7.4, because the glucuronic
acid is partially protonated at acidic pH (pKa ≈ 3). In contrast,
Figure 1. Mechanism of activation of [18/19F]-FEAnGA and PNPG by ꢀ-GUS.

[18F]-FEAnGA for ꢀ-Glucuronidase Activity Imaging
Bioconjugate Chem., Vol. 21, No. 5, 2010 917
Table 1. Enzyme Kinetic Parameters for FEAnGA Represented by the Mean Values of Triplicate Samples ( SD
enzyme
substrate
KM (µM)
Vmax (µmol min-1 mg-1
)
kcat (s-1
)
kcat/KM (106 M-1 s-1
)
E. coli ꢀ-GUS
PNPG
FEAnGA
PNPG
8.2 ( 1.6
7.1 ( 0.9
34.0 ( 16.5
5.7 ( 2.7
1366 ( 68
348 ( 24
56.9 ( 23
25.8 ( 9
660
134
19
80.3
24.6
0.6
bovine liver ꢀ-GUS
FEAnGA
10
1.8
the log D of [18F]-FEA was increased at pH 10 (-0.42 ( 0.03)
as compared to pH 7.4, due to the fact that [18F]-FEA is no
longer protonated at basic pH (pKa ≈ 8.8) (15, 19, 27-30).
The enzymatic processing of FEAnGA was studied using
commercially available bacterial Escherichia coli ꢀ-GUS and
mammalian bovine liver ꢀ-GUS. The water-soluble FEAnGA
is stable in aqueous buffer in the absence of ꢀ-GUS, and as a
result, a solution of FEAnGA remains colorless. However, in
the presence of ꢀ-GUS it undergoes rapid hydrolysis, resulting
in the spontaneous release of [18F]-FEA and the self-immolating
spacer, which immediately colors the solution yellow. The
liberation of 4-hydroxy-3-nitrobenzyl alcohol (HNBA) can be
measured with a UV spectrophotometer at 420 nm. In a similar
manner, the commercially available reference substrate, PNPG,
is cleaved by ꢀ-GUS, producing p-nitrophenol (PNP) which can
also be measured by UV spectrophotometer at 402 nm (Figure
1.)
order of magnitude as those of PNPG, which indicates that
FEAnGA is a good substrate of ꢀ-GUS as well. The maximum
hydrolysis rate (Vmax) of PNPG by Escherichia coli ꢀ-GUS and
bovine liver ꢀ-GUS was 4 and 2 times higher than that of
FEAnGA, respectively (Table 1). FEAnGA and PNPG show
similar affinity for E. coli ꢀ-GUS (KM ≈ 7 µM). However,
FEAnGA (KM ) 5.8 µM) has 6 times higher affinity than PNPG
(KM ) 34 µM) toward bovine liver-ꢀ-GUS. Thus, despite the
slightly lower turnover (Kcat) of FEAnGA by bovine liver-ꢀ-
GUS as compared to PNPG, FEAnGA will still undergo
relatively fast hydrolysis at low concentrations of substrate and
enzyme, as is demonstrated by higher catalytic efficiency, Kcat/
KM. Although FEAnGA is a slightly less efficient substrate for
E. coli ꢀ-GUS compared to PNPG, it proves to be a better
Thus, the kinetic parameters for the hydrolysis of FEAnGA
were determined by measuring the release of HNBA under
physiological conditions. The kinetic parameters of PNPG were
also determined under the same conditions for comparison. In
general, the kinetic parameters of FEAnGA are on the same
Figure 3. Cell-associated radioactivity of [18F]-FEAnGA and [18F]-
FEA in control C6 cells and in C6 cells in the presence of E. coli
ꢀ-GUS and bovine liver ꢀ-GUS. Data are expressed as % radioactivity/
million cells and are mean values of triplicate samples ( SD.* P <
0.05, paired bidirectional Student’s t test.
Figure 4. Cell-associated radioactivity of [18F]-FEAnGA and [18F]-
FEA in control (CT26) and CT26mꢀGUS cells. Data are expressed as
% radioactivity/million cells and are mean values of triplicate samples
( SD. *P < 0.05, paired bidirectional Student’s t test.
Figure 2. Cleavage of [18F]-FEAnGA by (A) E. coli ꢀ-GUS (n ) 1),
(B) bovine liver ꢀ-GUS (n ) 1).

918 Bioconjugate Chem., Vol. 21, No. 5, 2010
Antunes et al.
[18F]-FEA within 10 min of incubation with Escherichia coli
ꢀ-GUS (95% conversion) and 30 min of incubation with bovine
liver ꢀ-GUS (86% conversion) (Figure 2).
C6 glioma cells were incubated with [18F]-FEAnGA or [18F]-
FEA for 60 min at 37 °C, in the presence of either E. coli ꢀ-GUS
(740 U) or bovine liver ꢀ-GUS (340 U) or in the absence of
the enzyme (control). As depicted in Figure 3, the cell associated
radioactivity was 6 times higher when C6 cells are incubated
with [18F]-FEA than when the C6 cells were incubated with
[18F]-FEAnGA in the absence of ꢀ-GUS. This increase in uptake
is most likely due to the small size of [18F]-FEA that enables
the crossing of the cell membrane by passive diffusion. As
expected, there is no significant difference in the cell-associated
radioactivity of C6 cells incubated with [18F]-FEA in the
presence or absence of ꢀ-GUS. However, when C6 cells are
incubated with [18F]-FEAnGA the cell-associated radioactivity
increases 4 and 1.5 times in the presence of E. coli ꢀ-GUS and
bovine liver ꢀ-GUS, respectively. The higher tracer uptake in
the presence of E. coli ꢀ-GUS compared to the uptake in the
presence of bovine liver ꢀ-GUS could be explained by the fact
that the catalytic rate of bovine liver ꢀ-GUS toward [18F]-
FEAnGA is 13 times lower than the catalytic rate of E. coli
ꢀ-GUS.
Figure 5. Cell-associated radioactivity of [18F]-FEA in C6 cells. Data
are expressed as % radioactivity/million cells and are mean values of
triplicate samples ( SD. The efflux follows an exponential decay (y )
51.47 + 50.06 e(-0.18x), R2 ) 0.90; P ) 0.03).
Indeed, the radio-TLC of the medium samples shows that in
the presence of E. coli ꢀ-GUS the conversion of [18F]-FEAnGA
into [18F]-FEA is 99%, while with bovine liver ꢀ-GUS, there is
only 37% of conversion into [18F]-FEA (Supporting Informa-
tion). These results suggest that the increase in cell-associated
radioactivity is associated with the conversion of [18F]-FEAnGA
into [18F]-FEA.
substrate for bovine liver ꢀ-GUS (high Kcat/KM) (Table 1). In a
control experiment under identical conditions, but in the absence
of ꢀ-GUS, no hydrolysis of FEAnGA was observed. These
results indicate that FEAnGA is specifically hydrolyzed by
ꢀ-GUS (21, 27, 31, 32).
Radiolabeled [18F]-FEAnGA was rapidly cleaved by both
Escherichia coli ꢀ-GUS and bovine liver ꢀ-GUS. Radio-TLC
analysis confirmed that the cleavage of [18F]-FEAnGA by both
enzymes is a two-step process (33). The consumption of [18F]-
FEAnGA leads to the formation of an intermediate which we
hypothesized being the 4-(hydroxymethyl)-2-nitrophenol self-
immolative spacer attached to [18F]-FEA ([18F]-FEAn). The
intermediate ([18F]-FEAn) is subsequently converted to [18F]-
FEA resulting in almost complete cleavage of the tracer into
In addition to the uptake in C6 cells, [18F]-FEAnGA and [18F]-
FEA were evaluated in genetically engineered CT26mꢀGUS
murine colon adenocarcinoma cells that expressed membrane-
anchored ꢀ-GUS on the outer cell membrane. Therefore, CT26
cells and CT26mꢀGUS were incubated with [18F]-FEAnGA or
[18F]-FEA for 60 min at 37 °C. Incubation with [18F]-FEA
resulted in a 14-fold higher cell-associated radioactivity in
Figure 6. MicroPET image (left), microPET/CT fusion image (middle), and microCT image (right) of a mouse bearing a CT26mꢀGUS tumor (a)
and a CT26 tumor (b) injected with either [18F]-FEAnGA (upper row) or [18F]-FEA (lower row).

[18F]-FEAnGA for ꢀ-Glucuronidase Activity Imaging
Bioconjugate Chem., Vol. 21, No. 5, 2010 919
control CT26 cells than incubation with [18F]-FEAnGA. There
is no significant difference between CT26 and CT26mꢀGUS
associated radioactivity when incubated with [18F]-FEA, while
incubation with [18F]-FEAnGA caused 3-fold higher uptake in
ꢀ-glucuronidase expressing cells (CT26mꢀGUS) than control
CT26 cells (Figure 4). Radio-TLC of the medium showed that
[18F]-FEAnGA was converted into [18F]-FEA (20%) by the
enzyme expressing cells only (Supporting Information).
For a good PET tracer, not only tracer uptake, but also its
retention is important. To evaluate whether the tracer after
cleavage of ꢀ-GUS is retained by C6 cells, the efflux of [18F]-
FEA was investigated as shown in Figure 5, the efflux analysis
seems to indicate that 50% of the cell-associated radioactivity
is bound loosely to the extracellular domain. Nevertheless, it
was observed that, in the absence of radioactivity in the medium,
50% of the radioactivity still remains trapped in the cells.
To test the potential of [18F]-FEAnGA as a PET tracer probe
for detecting ꢀ-GUS expression in vivo, a preliminary PET study
was performed in two mice bearing CT26 and CT26mꢀGUS
tumors. Mice were injected with either [18F]-FEAnGA or [18F]-
FEA, and a microPET scan was acquired for 60 min followed
by a 15 min microCT scan.
This material is available free of charge via the Internet at http://
pubs.acs.org.
LITERATURE CITED
(1) de Graaf, M., Boven, E., Scheeren, H. W., Haisma, H. J., and
Pinedo, H. M. (2002) Beta-glucuronidase-mediated drug release.
Curr. Pharm. Des. 8, 1391–1403.
(2) George, J. (2008) Elevated serum beta-glucuronidase reflects
hepatic lysosomal fragility following toxic liver injury in rats.
Biochem. Cell Biol. 86, 235–243.
(3) Boyer, M. J., and Tannock, I. F. (1993) Lysosomes, lysosomal
enzymes, and cancer. AdV. Cancer Res. 60, 269–291.
(4) Chen, X., Wu, B., and Wang, P. G. (2003) Glucuronides in
anti-cancer therapy. Curr. Med. Chem. Anticancer Agents 3, 139–
150.
(5) de Groot, F. M., Damen, E. W., and Scheeren, H. W. (2001)
Anticancer prodrugs for application in monotherapy: targeting
hypoxia, tumor-associated enzymes, and receptors. Curr. Med.
Chem. 8, 1093–1122.
(6) Denny, W. A. (2003) Prodrugs for gene-directed enzyme-
prodrug therapy (suicide gene therapy). J. Biomed. Biotechnol.
48–70.
(7) Papot, S., Tranoy, I., Tillequin, F., Florent, J. C., and Gesson,
J. P. (2002) Design of selectively activated anticancer prodrugs:
elimination and cyclization strategies. Curr. Med. Chem. Anti-
cancer Agents 2, 155–185.
(8) Rivault, F., Tranoy-Opalinski, I., and Gesson, J. P. (2004) A
new linker for glucuronylated anticancer prodrugs. Bioorg. Med.
Chem. 12, 675–682.
(9) Sperker, B., Murdter, T. E., Schick, M., Eckhardt, K., Bosslet,
K., and Kroemer, H. K. (1997) Interindividual variability in
expression and activity of human beta-glucuronidase in liver and
kidney: consequences for drug metabolism. J. Pharmacol. Exp.
Ther. 281, 914–920.
(10) Sperker, B., Werner, U., Murdter, T. E., Tekkaya, C., Fritz, P.,
Wacke, R., Adam, U., Gerken, M., Drewelow, B., and Kroemer,
H. K. (2000) Expression and function of beta-glucuronidase in
pancreatic cancer: potential role in drug targeting. Naunyn Schmiede-
bergs Arch. Pharmacol. 362, 110–115.
(11) Chen, K. C., Cheng, T. L., Leu, Y. L., Prijovich, Z. M.,
Chuang, C. H., Chen, B. M., and Roffler, S. R. (2007)
Membrane-localized activation of glucuronide prodrugs by beta-
glucuronidase enzymes. Cancer Gene Ther. 14, 187–200.
(12) Ho, N. H., Weissleder, R., and Tung, C. H. (2007) A self-
immolative reporter for beta-galactosidase sensing. ChemBio-
Chem 8, 560–566.
(13) Su, Y. C., Chuang, K. H., Wang, Y. M., Cheng, C. M., Lin,
S. R., Wang, J. Y., Hwang, J. J., Chen, B. M., Chen, K. C.,
Roffler, S., and Cheng, T. L. (2007) Gene expression imaging
by enzymatic catalysis of a fluorescent probe via membrane-
anchored beta-glucuronidase. Gene Ther. 14, 565–574.
(14) Biber, F. Z., Unak, P., Ertay, T., Medine, E. I., Zihnioglu, F.,
Tasci, C., and Durak, H. (2006) Synthesis of an estradiol
glucuronide derivative and investigation of its radiopharmaceu-
tical potential. Appl. Radiat. Isot. 64, 778–788.
(15) Celen, S., Deroose, C., de Groot, T., Chitneni, S. K., Gijsbers,
R., Debyser, Z., Mortelmans, L., Verbruggen, A., and Bormans,
G. (2008) Synthesis and evaluation of F-18- and C-11-labeled
phenyl-galactopyranosides as potential probes for in vivo visu-
alization of LacZ gene expression using positron emission
tomography. Bioconjugate Chem. 19, 441–449.
When injected with [18F]-FEAnGA, the CT26mꢀGUS tumor
had more retention of radioactivity than the control tumor (CT26).
The tracer uptake ratio between the CT26mꢀGUS tumor and the
control CT26 tumor was 2 at 1 h post injection. This enhanced
tumor uptake was ascribed to the selective production of [18F]-
FEA due to the cleavage of [18F]-FEAnGA by ꢀ-GUS that is
expressed extracellularly in the CT26mꢀGUS tumor only. The
uncleaved [18F]-FEAnGA was rapidly excreted out of the body
through the kidneys leading to a very clear image with a good
tumor-to-background ratio. On the other hand, the animal injected
with [18F]-FEA did not show specific accumulation in the
CT26mꢀGUS tumor (Figure 6). It is possible to observe an absence
of uptake in the center of the CT26mꢀGUS tumor due to the
advanced necrosis of the tumor.
These preliminary imaging results suggest that extracellular
ꢀ-GUS efficiently cleaves the tracer, resulting in the formation
of [18F]-FEA, which is retained in the tumor. The increased
retention of radioactivity in the tumor that expresses ꢀ-glucu-
ronidase extracellularly can be visualized in the image; ad-
ditional animal experiments are in progress to further evaluate
the properties of [18F]-FEAnGA in vivo.
CONCLUSION
[18F]-FEAnGA was efficiently labeled with [18F]-fluoride in
good yield. [18/19F]-FEAnGA showed high in vitro stability in
PBS and rat plasma. Kinetic studies indicate that [18/19F]-
FEAnGA is a good substrate for E. coli ꢀ-GUS and bovine liver
ꢀ-GUS. Cleavage of the tracer by ꢀ-GUS caused an increase
in cell-associated radioactivity in rodent tumor cells. Initial
animal studies showed that [18F]-FEAnGA is specially retained
in CT26mꢀGUS tumors with high tumor uptake and good
tumor-to-background ratio. Therefore, [18F]-FEAnGA has proven
to be a promising PET tracer for imaging of extracellular
ꢀ-glucuronidase activity.
ACKNOWLEDGMENT
We gratefully thank Annie van Dam and Margot Jeronimus-
Stratingh for performing all LRMS analysis, Katica Stojanov
for her help in the C6 glioma cell studies, Dr. Steve Roffler for
providing the CT26mꢀGUS cells, and Jurgen Sijbesma for the
help in the in vivo study.
(16) Ertay, T., Unak, P., Biber, F. Z., Tasci, C., Zihnioglu, F., and
Durak, H. (2007) Scintigraphic imaging with a peptide glucu-
ronide in rabbits: Tc-99m-exorphin C glucuronide. Appl. Radiat.
Isot. 65, 170–175.
(17) Kumar, P., Wiebe, L. I., Mannan, R. H., Zhang, Z., Xia, H. Y.,
and Mcewan, A. J. (2002) [Tc-99m]Technetium labelled PnAo-
azomycin glucuronides: a novel class of imaging markers of
tissue hypoxia. Appl. Radiat. Isot. 57, 719–728.
1
Supporting Information Available: H NMR, 19F NMR
spectra, and the raw LRMS of compounds 2, 3, and 4. Radio-
TLC of the medium of C6 glioma cells and CT26/CT26mꢀGUS.

920 Bioconjugate Chem., Vol. 21, No. 5, 2010
Antunes et al.
(18) Brady, F., Luthra, S. K., Brown, G. D., Osman, S., Aboagye,
E., Saleem, A., and Price, P. M. (2001) Radiolabelled tracers
and anticancer drugs for assessment of therapeutic efficacy using
PET. Curr. Pharm. Des. 7, 1863–1892.
(19) Rukhman, I., and Gutman, A. L. (2000) Synthesis of morphine
6-alpha-D-glucuronide. Tetrahedron Lett. 41, 6889–6892.
(20) Antunes, I. F., Haisma, H. J., and de Vries, E. F. (2008)
Tumor-specific activation of prodrugs: is there a role for nuclear
medicine? Nucl. Med. Commun. 29, 845–846.
(27) Houba, P. H., Leenders, R. G., Boven, E., Scheeren, J. W.,
Pinedo, H. M., and Haisma, H. J. (1996) Characterization of
novel anthracycline prodrugs activated by human beta-glucu-
ronidase for use in antibody-directed enzyme prodrug therapy.
Biochem. Pharmacol. 52, 455–463.
(28) Huwyler, J., Drewe, J., Klusemann, C., and Fricker, G. (1996)
Evidence for P-glycoprotein-modulated penetration of morphine-
6-glucuronide into brain capillary endothelium. Br. J. Pharmacol.
118, 1879–1885.
(21) Duimstra, J. A., Femia, F. J., and Meade, T. J. (2005) A
gadolinium chelate for detection of beta-glucuronidase: A self-
immolative approach. J. Am. Chem. Soc. 127, 12847–12855.
(22) Tewson, T. J. (1997) Synthesis of [F-18]fluoroetanidazole: A
potential new tracer for imaging hypoxia. Nucl. Med. Biol. 24,
755–760.
(23) Combaud, D., Thomas, M., Papot, S., and Gesson, J. P. (2005)
Synthesis and evaluation of p-nitrophenyl beta-D-glucopyrano-
siduronic analogues as new triggers for beta-glucuronidase
mediated prodrug monotherapy. Lett. Drug Des. DiscoVery 2,
631–637.
(24) Tomasic, J., and Keglevic, D. (1973) Kinetics of hydrolysis
of synthetic glucuronic esters and glucuronic ethers by bovine
liver and Escherichia-coli beta-glucuronidase. Biochem. J. 133,
789–795.
(25) Chittaboina, S., Hodges, B., and Wang, Q. (2006) A facile
route for the regioselective deacetylation of peracetylated
carbohydrates at anomeric position. Lett. Org. Chem. 3, 35–38.
(26) Stachulski, A. V., and Jenkins, G. N. (1998) The synthesis of
O-glucuronides. Nat. Prod. Rep. 15, 173–186.
(29) Huwyler, J., Drewe, J., Gutmann, H., Thole, M., and Fricker,
G. (1998) Modulation of morphine-6-glucuronide penetration into
the brain by P-glycoprotein. Int. J. Clin. Pharmacol. Ther. 36,
69–70.
(30) Leu, Y. L., Chen, C. S., Wu, Y. J., and Chern, J. W. (2008)
Benzyl ether-linked glucuronide derivative of 10-hydroxycamp-
tothecin designed for selective camptothecin-based anticancer
therapy. J. Med. Chem. 51, 1740–1746.
(31) Bosslet, K., Czech, J., and Hoffmann, D. (1994) Tumor-
selective prodrug activation by fusion protein-mediated catalysis.
Cancer Res. 54, 2151–2159.
(32) Wei, G. P., Loktionova, N. A., Pegg, A. E., and Moschel,
R. C. (2005) beta-glucuronidase-cleavable prodrugs of O-6-
benzylguanine and O-6-benzyl-2 ′-deoxyguanosine. J. Med.
Chem. 48, 256–261.
(33) Bakina, E., Wu, Z., Rosenblum, M., and Farquhar, D. (1997)
Intensely cytotoxic anthracycline prodrugs: glucuronides. J. Med.
Chem. 40, 4013–4018.
BC9004602