18F-Glyco-RGD peptides for PET imaging of integrin expression: Efficient radiosynthesis by click chemistry and modulation of biodistribution by glycosylation

Article abstract of DOI:10.1021/mp4004817

Glycosylation frequently improves the biokinetics and clearance properties of macromolecules in vivo and could therefore be used for the design of radiopharmaceuticals for positron emission tomography (PET). Recently, we have developed a click chemistry method for ¹⁸F-fluoroglycosylation of alkyne-bearing RGD-peptides targeting the integrin receptor. To investigate whether this strategy could yield an ¹⁸F-labeled RGD glycopeptide with favorable biokinetics, we generated a series of new RGD glycopeptides, varying the 6-fluoroglycosyl residue from monosaccharide to disaccharide units, which provided the glucosyl ([¹⁹F]6Glc-RGD, 4b), galactosyl ([ ¹⁹F]Gal-RGD, 4c), maltosyl ([¹⁹F]Mlt-RGD, 4e), and cellobiosyl ([¹⁹F]Cel-RGD, 4f) conjugated peptides in high yields and purities of >97%. All of these RGD glycopeptides showed high affinity to α_vβ₃ (11-55 nM), α_vβ ₅ (6-14 nM), and to α_vβ₃-positive U87MG cells (90-395 nM). ¹⁸F-labeling of the various carbohydrate precursors (1a-f) using cryptate-assisted reaction conditions (CH₃CN, 85 C, 10 min) gave ¹⁸F-labeled glycosyl azides in radiochemical yields (RCYs) of up to 84% ([¹⁸F]2b). The deacetylation and subsequent click reaction with the alkyne-bearing cyclic RGD peptide proceeded in one-pot reactions with RCYs as high as 81% in 15-20 min at 60 C, using a minimal amount of peptide precursor (100 nmol). Optimization of the radiosynthesis strategy gave a decay-uncorrected RCY of 16-24% after 70-75 min (based on [¹⁸F]fluoride). Due to their high-yield radiosyntheses, the glycopeptides [¹⁸F]6Glc-RGD and [¹⁸F]Mlt-RGD were chosen for comparative biodistribution studies and dynamic small-animal PET imaging using U87MG tumor-bearing nude mice. [¹⁸F]6Glc-RGD and [ ¹⁸F]Mlt-RGD showed significantly decreased liver and kidney uptake by PET relative to the 2-[¹⁸F]fluoroglucosyl analog [ ¹⁸F]2Glc-RGD, and showed specific tumor uptake in vivo. Notably, [¹⁸F]Mlt-RGD revealed uptake and retention in the U87MG tumor comparable to that of [¹⁸F]Galacto-RGD. Both [¹⁸F]6Glc-RGD and [¹⁸F]Mlt-RGD were obtained by a reliable and easy click chemistry-based procedure, much more rapidly than was [¹⁸F]Galacto- RGD. Due to its favorable biodistribution and tissue clearance in vivo, [ ¹⁸F]Mlt-RGD represents a viable alternative radiotracer for imaging integrin expression in solid tumors by PET.

Full text of DOI:10.1021/mp4004817

Article
pubs.acs.org/molecularpharmaceutics
18F‑Glyco-RGD Peptides for PET Imaging of Integrin Expression:
Efficient Radiosynthesis by Click Chemistry and Modulation of
Biodistribution by Glycosylation
Simone Maschauer,† Roland Haubner,‡ Torsten Kuwert,† and Olaf Prante*,†
†Department of Nuclear Medicine, Molecular Imaging and Radiochemistry, Friedrich Alexander University, 91054 Erlangen,
Germany
‡Department of Nuclear Medicine, Innsbruck Medical University, 6020 Innsbruck, Austria
S
* Supporting Information
ABSTRACT: Glycosylation frequently improves the biokinetics
and clearance properties of macromolecules in vivo and could
therefore be used for the design of radiopharmaceuticals for
positron emission tomography (PET). Recently, we have
developed a click chemistry method for 18F-fluoroglycosylation
of alkyne-bearing RGD-peptides targeting the integrin receptor.
To investigate whether this strategy could yield an 18F-labeled
RGD glycopeptide with favorable biokinetics, we generated a
series of new RGD glycopeptides, varying the 6-fluoroglycosyl
residue from monosaccharide to disaccharide units, which
provided the glucosyl ([19F]6Glc-RGD, 4b), galactosyl
([19F]Gal-RGD, 4c), maltosyl ([19F]Mlt-RGD, 4e), and cellobiosyl ([19F]Cel-RGD, 4f) conjugated peptides in high yields
and purities of >97%. All of these RGD glycopeptides showed high affinity to αvβ3 (11−55 nM), αvβ5 (6−14 nM), and to αvβ3-
positive U87MG cells (90−395 nM). 18F-labeling of the various carbohydrate precursors (1a−f) using cryptate-assisted reaction
conditions (CH3CN, 85 °C, 10 min) gave 18F-labeled glycosyl azides in radiochemical yields (RCYs) of up to 84% ([18F]2b).
The deacetylation and subsequent click reaction with the alkyne-bearing cyclic RGD peptide proceeded in one-pot reactions with
RCYs as high as 81% in 15−20 min at 60 °C, using a minimal amount of peptide precursor (100 nmol). Optimization of the
radiosynthesis strategy gave a decay-uncorrected RCY of 16−24% after 70−75 min (based on [18F]fluoride). Due to their high-
yield radiosyntheses, the glycopeptides [18F]6Glc-RGD and [18F]Mlt-RGD were chosen for comparative biodistribution studies
and dynamic small-animal PET imaging using U87MG tumor-bearing nude mice. [18F]6Glc-RGD and [18F]Mlt-RGD showed
significantly decreased liver and kidney uptake by PET relative to the 2-[18F]fluoroglucosyl analog [18F]2Glc-RGD, and showed
specific tumor uptake in vivo. Notably, [18F]Mlt-RGD revealed uptake and retention in the U87MG tumor comparable to that of
[18F]Galacto-RGD. Both [18F]6Glc-RGD and [18F]Mlt-RGD were obtained by a reliable and easy click chemistry-based
procedure, much more rapidly than was [18F]Galacto-RGD. Due to its favorable biodistribution and tissue clearance in vivo,
[18F]Mlt-RGD represents a viable alternative radiotracer for imaging integrin expression in solid tumors by PET.
KEYWORDS: RGD peptide, fluorine-18, glycosylation, click chemistry, positron emission tomography
patients for antiangiogenic therapies and for monitoring effects
of treatment in individualized therapy.
INTRODUCTION
■
The molecular mechanisms underlying angiogenesis have been
the subject of intensive research in the last few decades.1 These
efforts have led to considerable knowledge of the role of
angiogenesis and its disturbances in the pathogenesis of
diseases such as macular degeneration, and notably in solid
tumors. These insights have led to the development and
validation of new therapeutic drugs interfering with angio-
genesis.2 Moreover, efforts have been expended in developing
radiopharmaceuticals for molecular imaging of processes
pertinent to the formation of new vessels.3 Angiogenesis
tracers for positron emission tomography (PET) have the
potential to diagnose foci of inflammation or tumor deposits
and could also be used clinically for rational selection of
One class of PET angiogenesis tracer is based on RGD
containing peptides which target the integrin αvβ3, a receptor,
which plays a major role during angiogenesis by mediating
adhesion between endothelial cells. One approach relies on the
pioneering work by Kessler et al., who successfully developed
cyclic pentameric RGD-containing peptide inhibitors with high
specificity and affinity for this integrin.4 These efforts have
yielded a number of radiolabeled RGD derivatives for imaging
Received: August 14, 2013
Revised: October 25, 2013
Accepted: December 10, 2013
Published: December 10, 2013
© 2013 American Chemical Society
505
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
Figure 1. Synthetic route toward a series of 18F-labeled RGD glycopeptides following a click chemistry-based strategy by the use of a small library of
glycosyl azide precursors.
αvβ3 expression in tumors and nontumor tissues.5 Very
recently, Al[18F]F-NOTA-PRGD2 ([18F]alfatide) has been
produced using fluoride-aluminum complex chemistry, and
has entered clinical studies.6 The glycoconjugate [18F]Galacto-
RGD has also been applied in various clinical PET studies,
including patients with malignant melanoma, glioblastoma,
head and neck cancer, breast cancer, sarcoma, nonsmall cell
lung cancer, and prostate cancer.7−9 These studies demon-
strated beneficial pharmacokinetics and in vivo clearance
properties of [18F]Galacto-RGD, that is, rapid, predominately
renal tracer elimination, resulting in low background radio-
activity in almost all organs, and consequently with tumor-to-
background ratios favoring the detection of lesions. Excellent
PET imaging properties have also been reported for other
glycopeptides and glycoconjugates, including octreotide ana-
logues and, more recently, nonpeptide tracers.8,10,11 However,
the accessibility of [18F]Galacto-RGD has been hitherto
hampered by its time-consuming multistep radiosynthesis; a
more reliable 18F-synthesis for a RGD-based glycopeptide is
still needed.
Recently developments in so-called “click chemistry” have
provided improved strategies for 18F-labeling, taking advantage
of the speed, reliability, and selectivity of these reactions.12 The
most prominent example of a click chemistry reaction is the
copper(I)-catalyzed azide−alkyne cycloaddition (CuAAC),
which is a highly chemo- and regioselective and high-yielding
reaction.13,14 Only a limited number of click chemistry-based
radiosyntheses of 18F-labeled glycopeptides have been
described, including the oxime formation of aminooxy-
functionalized peptides with 4-[18F]fluorobenzaldehyde or 2-
deoxy-2-[18F]fluoroglucose ([18F]FDG)15−17 and the use of
thiol-reactive glycosyl donors derived from [18F]FDG.18,19
Recently, we developed a strategy for concomitant 18F-
labeling and glycosylation of alkyne-bearing molecules using
CuAAC click chemistry.20,21 Meanwhile, this 18F-click glyco-
sylation method has been successfully adapted by others to the
radiosyntheses of 18F-glycoproteins and nonpeptide glycocon-
jugates.11,22 We applied this strategy to the radiosynthesis of a
glucosylated cyclic RGD peptide ([18F]2Glc-RGD) that was
successfully used for imaging of αvβ3 integrin expression in a
mouse model.21 However, the biodistribution of this tracer
suffered from clearance properties in liver and kidney that were
inferior to the pharmacokinetics of [18F]galacto-RGD, which
had relatively rapid washout from nontumor tissues.
The aim of this study was to develop an 18F-labeled RGD
glycopeptide with improved pharmacokinetic properties by
varying the 18F-labeled glycosyl residue. We predicted that this
might be achieved by aiming at suitable glycosyl precursors for
the introduction of 6-deoxy-6-[18F]fluorogalactosyl, 6-deoxy-6-
[18F]fluoroglucosyl, 6′-deoxy-6′-[18F]fluoromaltosyl, and 6′-
deoxy-6′-[18F]fluorocellobiosyl residues to the RGD motif
(Figure 1). Clickable 18F-glycosyl azides should facilitate and
simplify the radiosynthesis of 18F-labeled glycopeptides. There-
fore, we synthesized a series of novel 6-O-sulfonylglycosyl
azides as 18F-labeling precursors (Figure 1): 2,3,4-tri-O-acetyl-
6-O-toluenesulfonyl-β-D-glucopyranosyl azide 1b, 2,3,4-tri-O-
acetyl-6-O-toluenesulfonyl-β-D-galactopyranosyl azide 1c, 2,3,4-
tri-O-acetyl-6-O-(2-nitrophenyl)sulfonyl-β-D-galactopyranosyl
azide 1d, 2′,3′,4′,2,3,6-hexa-O-acetyl-6′-O-toluenesulfonyl-β-D-
maltosyl azide 1e, and 2′,3′,4′,2,3,6-hexa-O-acetyl-6′-O-tolue-
nesulfonyl-β-D-cellobiosyl azide 1f and studied their applic-
ability as 18F-labeling precursors compared with the 2-triflate
precursor 1a previously developed in our laboratory.20 We then
tested our series of 18F-glycosyl azides for click labeling of the
alkyne-bearing cyclic RGD peptide c(RGDfPra) to provide the
series of RGD glycopeptides [18F]6Glc-RGD, [18F]Gal-RGD,
[18F]Mlt-RGD, and [18F]Cel-RGD (Figure 1). In addition, we
measured affinities of the resulting 18F-glyco-RGD peptides to
αvβ3 and αvβ5 integrin receptors and characterized the influence
of heteroglycosylation on the biodistribution and clearance
properties in vivo in small animal PET studies of tumor-bearing
nude mice.
EXPERIMENTAL SECTION
■
General. All chemicals and reagents were of analytical grade
and obtained from commercial sources unless stated otherwise.
Analytical high-performance liquid chromatography (HPLC)
was performed on an Agilent 1100 system with a quarternary
pump and variable wavelength detector and radio-HPLC
detector D505TR (Canberra Packard). Computer analysis of
the HPLC data was performed using FLO-One software
(Canberra Packard). Electron-spray ionization (ESI) mass
spectrometry analysis was performed using a Bruker Esquire
2000 instrument. The carbohydrate syntheses of the glycosyl
506
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
labeling precursors 1e and 1f, all intermediates and precursors
for the syntheses of the reference RGD peptides [19F]6Glc-
RGD (4b), [19F]Gal-RGD (4c), [19F]Mlt-RGD (4e), and
[19F]Cel-RGD (4f), and their analytical data are given in the
Supporting Information file (see associated content). The
syntheses of 1a, 2a, 3a, 4a and [18F]2a, [18F]3a, and [18F]4a
were performed as previously described.20,21
2,3,4-Tri-O-acetyl-6-O-p-toluenesulfonyl-β-D-galacto-
pyranosyl Azide (1c). The synthesis was performed according
to the procedure described in the literature.23 To a solution of
p-toluenesulfonyl chloride (465 mg, 2.44 mmol) in anhydrous
pyridine (7 mL) at 0 °C β-galactosyl azide (500 mg, 2.44
mmol; Sigma-Aldrich) was added, and the solution was stirred
overnight at 0 °C. Acetic anhydride (2 mL) was added
dropwise at 0 °C and the solution was stirred for 1 h. The
solution was diluted with dichloromethane, and the organic
phase was washed several times with saturated NaHCO3, 0.2 M
HCl, and water, dried with Na2SO4, and concentrated to afford
the crude product 1c which was crystallized from absolute
2,3,4-Tri-O-acetyl-6-O-p-toluenesulfonyl-β-D-gluco-
pyranosyl Azide (1b). To a solution of p-toluenesulfonyl
chloride (3.2 g, 16.7 mmol) in anhydrous pyridine (45 mL) at 0
°C D-glucose (3 g, 16.7 mmol) was added. The solution was
stirred overnight at 0 °C. Subsequently, acetic anhydride (13.2
mL) was added dropwise. The solution was allowed to warm to
room temperature, and after 1 h the solvent was removed in
vacuo. The residue was dissolved in dichloromethane, and the
solution was washed several times with saturated NaHCO3, 0.2
M HCl, and water, dried with Na2SO4, and concentrated again
to afford the crude 1,2,3,4-tetra-O-acetyl-6-O-p-toluenesulfonyl-
β-D-glucopyranose 7b which was crystallized from ethanol: 3 g,
1
EtOH: 350 mg, 730 μmol, 30% yield. H NMR (600 MHz,
CDCl3): δ ppm 1.97 (s, 3H), 2.07 (s, 3H), 2.08 (s, 3H), 2.46
(s, 3H), 4.03 (m, 2H, H-5, H-6a), 4.13 (dd, J = 9.42, 5.73 Hz,
1H, H-6b), 4.55 (d, J = 8.77 Hz, 1H, H-1), 5.00 (dd, J = 10.37,
3.37 Hz, 1H, H-3), 5.10 (dd, J = 10.37, 8.77 Hz, 1H, H-2), 5.40
(dd, J = 3.33, 0.71 Hz, 1H, H-4), 7.38−7.35 (m, 2H), 7.79−
7.76 (m, 2H). 13C NMR (151 MHz, CDCl3): δ ppm 169.83,
169.80, 169.32, 145.38, 132.20, 130.02, 128.07, 88.24, 72.73,
70.49, 67.84, 66.66, 65.98, 21.69, 20.65, 20.48.
1
6.0 mmol, 36% yield. H NMR (600 MHz, CHCl3): δ ppm
1.99 (s, 3H, OAc), 2.00 (s, 3H, OAc), 2.02 (s, 3H, OAc), 2.09
(s, 3H, OAc), 2.46 (s, 3H, CH3-Ts), 3.84 (ddd, J = 10.07, 4.55,
2.87 Hz, 1H, H-5), 4.15 (dd, J = 11.17, 2.89 Hz, 1H, H-6b),
4.11 (dd, J = 11.21, 4.59 Hz, 1H, H-6a), 5.04 (dd (t), J = 9.5
Hz, 1H, H-4), 5.05 (dd, J = 9.41, 8.23 Hz, 1H, H-2), 5.20 (t, J =
9.37 Hz, 1H, H-3), 5.65 (d, J = 8.19 Hz, 1H, H-1), 7.35 (d, J =
7.97 Hz, 2H), 7.77 (d, J = 8.29 Hz, 2H). 13C NMR (151 MHz,
CHCl3): δ ppm 170.09, 169.27, 169.12, 168.75 (4x CO,
OAc), 145.14 (Ts), 132.43 (Ts), 129.86 (2C, Ts), 128.17 (2C,
Ts), 91.53 (C-1), 72.60, 72.17, 70.05, 67.95, 66.74 (C-2 − C-
6), 21.68 (CH3, Ts), 20.74, 20.55, 20.52, 20.49 (4x CH3, OAc).
To a solution of 1,2,3,4-tetra-O-acetyl-6-O-p-toluenesulfonyl-
β-D-glucopyranose 7b (3 g, 6 mmol) in dichloromethane (20
mL) HBr (33% in acetic acid, 30 mL) was added dropwise at 0
°C, and the solution was stirred for 3 h at room temperature.
The mixture was diluted with dichloromethane (150 mL) and
washed with saturated NaHCO3 (3 × 50 mL), water (2 × 50
mL), dried with Na2SO4 and concentrated in vacuo. The crude
product was purified by column chromatography (EtOAc-n-
hexan 1:1) to afford 2,3,4-tri-O-acetyl-6-O-p-toluenesulfonyl-β-
D-glucopyranosyl bromide as a yellow oil (2.8 g, 5.3 mmol, 88%
yield) that was dissolved in in EtOAc (30 mL) and
tetrabutylammonium hydrogen sulfate (1.8 g, 5.3 mmol) and
sodium azide (1.4 g, 21.2 mmol), and saturated NaHCO3 (30
mL) was added. The two-phase mixture was vigorously stirred
for 2 h at room temperature; EtOAc (70 mL) was added, and
the organic phase was washed with saturated NaHCO3 (2 × 50
mL) and brine (50 mL). After drying with Na2SO4 and
evaporation of the solvent, the crude product 1b was
recrystallized twice from EtOH: 1.8 g, 3.8 mmol, yield 70%.
1H NMR (600 MHz, CDCl3): δ ppm 1.99 (s, 3H, OAc), 2.02
(s, 3H, OAc), 2.06 (s, 3H, OAc), 2.46 (s, 3H, CH3-Ts), 3.83
(ddd, J = 10.10, 5.70, 2.66 Hz, 1H, H-5), 4.09 (dd, J = 11.24,
5.72 Hz, 1H, H-6a), 4.15 (dd, J = 11.24, 2.65 Hz, 1H, H-6b),
4.56 (d, J = 8.87 Hz, 1H, H-1), 4.85 (t, J = 9.26 Hz, 1H, H-2),
4.95 (t, J = 9.77 Hz, 1H, H-4), 5.18 (t, J = 9.50 Hz, 1H, H-3),
7.37 (d, J = 8.08 Hz, 2H, Ts), 7.79 (d, J = 8.29 Hz, 2H, Ts). 13C
NMR (151 MHz, CDCl3): δ ppm 170.04, 169.32, 169.11 (3×
CO, OAc), 145.33 (Ts), 132.25 (Ts), 129.94 (2C, Ts),
128.13 (2C, Ts), 87.69 (C-1), 73.65, 72.34, 70.40, 68.14, 67.21
(C-2−C-6), 21.69 (CH3, Ts), 20.53, 20.51, 20.50 (3× CH3,
OAc). ESI-MS: m/z = 508.1 [M + Na]+.
2,3,4-Tri-O-acetyl-6-O-p-nitrobenzenesulfonyl-β-D-
galactopyranosyl Azide (1d). 1d was synthesized starting
from β-galactosyl azide (500 mg, 2.44 mmol) following the
protocol described for 1c but using p-nitrobenzenesulfonyl
chloride (540 mg, 2.44 mmol) instead of p-toluenesulfonyl
chloride. 1d was obtained after column chromatography
(toluene−EtOAc 1:1) as an off-white solid in a yield of 40%
1
(500 mg, 970 μmol). H NMR (360 MHz, CDCl3): δ ppm
1.98 (s, 3H), 2.08 (s, 3H), 2.12 (s, 3H), 4.07 (ddd, J = 6.77,
5.40, 1.20 Hz, 1H, H-5), 4.20 (dd, J = 10.53, 5.39 Hz, 1H, H-6),
4.22 (dd, J = 10.52, 6.88 Hz, 1H, H-6), 4.56 (d, J = 8.68 Hz,
1H, H-1), 5.01 (dd, J = 10.37, 3.32 Hz, 1H, H-3), 5.11 (dd, J =
10.46, 8.63 Hz, 1H, H-2), 5.41 (dd, J = 3.31, 1.11 Hz, 1H, H-4),
8.44−8.40 (m, 2H), 8.13−8.09 (m, 2H). 13C NMR (90.56
MHz, CDCl3): δ ppm 169.91, 169.84, 169.26, 151.08, 141.08,
129.40, 124.58, 88.26, 72.67, 70.42, 67.72, 67.24, 66.68, 20.62,
20.53, 20.46. ESI-MS (m/z): = 539.0 [M + Na]+, m/z = 555.0
[M + K]+.
Syntheses of Reference Compounds 4b, 4c, 4e, or 4f.
To a solution of 6-deoxy-6-fluoro-D-glucopyranosyl azide (3b)
(1.86 mg, 2.4 μmol), 6-deoxy-6-fluoro-D-galactopyranosyl azide
(3c) (1.86 mg, 2.4 μmol), 6′-deoxy-6′-fluoro-β-maltosyl azide
(3e) (2.25 mg, 2.4 μmol), or 6′-deoxy-6′-fluoro- β-cellobiosyl
azide (3f) (2.25 mg, 2.4 μmol) and c(RGDfPra)21 (1.37 mg,
2.4 μmol) in a solution of phosphate-buffered saline (PBS; 10%
ethanol) (0.5 mL), a solution of copper(II)sulfate pentahydrate
(0.2 M, 10 μL) and a solution of sodium ascorbate (0.6 M, 10
μL) were added. The mixture was stirred at room temperature
for 30 min, then diluted with water/acetonitrile (9:1, 0.5 mL),
and subjected to semipreparative RP-HPLC for purification of
the desired product (Kromasil C8, 125 × 8 mm, 10−40%
acetonitrile (0.1% trifluoroacetic acid, TFA) in water (0.1%
TFA) in a linear gradient over 30 min, 4 mL/min, tR(4b) =
11.0 min, tR(4c) = 9.3 min, tR(4e) = 7.5 min, tR(4f) = 6.4 min).
After lyophilization of the product fraction, the glycopeptides
were obtained as white solids in yields of 70−80%. ESI-MS: 4b:
m/z = 778.4 [M + H]+; 4c: m/z = 778.4 [M + H]+; 4e: m/z =
940.6 [M + H]+; 4f: m/z = 940.4 [M + H]+, 481.7 [M + H +
Na]2+.
Synthesis of c(RGDfPra). The peptide was synthesized
using Fmoc-protocols as described previously.21
507
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
Production of [18F]fluoride. No-carrier-added (n.c.a.)
[18F]fluoride was produced by the 18O(p,n)18F reaction on
18O-enriched water (>98%, Rotem Ind. LTD) using a proton
beam of 11 MeV generated by a RDS 111 cyclotron (Siemens/
CTI; Pet Net GmbH, Erlangen).
uncorrected radiochemical yield) of [18F]4b or [18F]4e in a
total synthesis time of 70−75 min.
Stability in Serum in Vitro. An aliquot of the 18F-labeled
glycopeptide [18F]4b or [18F]4e in PBS (35 μL) was added to
human serum (200 μL) and incubated at 37 °C. Aliquots (25
μL) were taken at various time intervals (5−60 min) and
quenched in methanol/water (1:1, 100 μL). The samples were
centrifuged, and the supernatants were analyzed by radio-
HPLC.
18F-Labeling under “Mild” Reaction Conditions. A
QMA-cartridge with [18F]fluoride was eluted with a solution of
10 mg of Kryptofix 2.2.2, 0.1 M K2CO3 (17.5 μL), and 0.1 M
KH2PO4 (17.5 μL) in 1 mL of acetonitrile/water (8:2). The
solution was evaporated using a stream of nitrogen at 85 °C
and coevaporated to dryness with CH3CN (2 × 500 μL). The
respective labeling precursor (1a, 1b, 1c, 1d, 1e, or 1f, 15
μmol) in anhydrous acetonitrile (450 μL) was added, and the
solution was stirred for 10 min at 85 °C. Samples were
withdrawn from the reaction mixture for the analysis of the
radiochemical yields of the corresponding 18F-labeled glycosyl
azides [18F]2b, [18F]2c, [18F]2e, and [18F]2f by radio-HPLC
(Figure S1). The 18F-labeled products were identified by
retention time (tR) on the radio-HPLC system and by
coinjection of the corresponding reference compounds
(Kromasil C8, 250 × 4.6 mm, 40−100% acetonitrile (0.1%
TFA) in water (0.1% TFA) in a linear gradient over 50 min, 1.5
mL/min, tR (2b) = 9.2 min, tR (2c) = 9.4 min, tR (2e) = 14.3
min, tR (2f) = 13.4 min).
Determination of the Partition Coefficients (log D).
The lipophilicity of the 18F-labeled radioligands was assessed by
determination of the water−octanol partition coefficient. 1-
Octanol (0.5 mL) was added to a solution of approximately 25
kBq of the 18F-labeled compound in PBS (0.5 mL, pH 7.4) ,and
the layers were vigorously mixed for 3 min at room
temperature. The tubes were centrifuged (14000 rpm, 1 min)
and three samples of each layer (each 100 μL) were counted in
a γ-counter (Wallac Wizard). The partition coefficient was
determined by calculating the ratio cpm (octanol)/cpm (PBS)
and expressed as log D7.4 = log(cpmoctanol/cpmPBS). The
experiments were performed in triplicate and data reported as
mean ( SD).
Cell Culture. The human αvβ3-expressing glioblastoma cell
line U87MG (ECACC no. 89081402) was grown in DMEM
supplemented with nonessential amino acids (1%), sodium
pyruvate (1 mM), and FBS (10%) at 37 °C in a humidified
atmosphere of 5% CO2 as described previously.18 The cells
were routinely subcultured every 4 days. Routine test of the
U87MG cells for contamination with mycoplasma were always
negative.
Radiosynthesis of [18F]6Glc-RGD ([18F]4b), [18F]Gal-
RGD ([18F]4c), or [18F]Mlt-RGD ([18F]4e). Acetylated glycosyl
azides [18F]2b, [18F]2c, or [18F]2e were isolated by evaporating
the solvent in a stream of nitrogen, dissolving the residue with
acetonitrile/water (30:70, 500 μL), and separation by semi-
preparative HPLC (Kromasil C8, 125 × 8 mm, 4 mL/min,
acetonitrile (0.1% TFA)/water (0.1% TFA) 30−70% in 30
min). The product fraction was diluted with water (1:10),
transferred to a C18-cartridge (Merck, 100 mg), dried in a
stream of nitrogen, and eluted with ethanol (1 mL). [18F]2b,
[18F]2c, or [18F]2e was dried at 60 °C in a stream of nitrogen,
and subsequently NaOH (250 μL, 60 mM, 10% ethanol) was
added. The deacetylation was complete after stirring for 5 min
at 60 °C, and the crude product [18F]3b, [18F]3c, or [18F]3e
was used for the click reaction with the alkyne in a one-pot
procedure. This was accomplished by neutralizing the solution
with HCl (1 N, 13.5 μL), followed by addition of c(RGDfPra)
(100 μg, dissolved in ethanol/water (260 μL, 1:1)), CuSO4
(0.2 M, 10 μL), and sodium ascorbate (0.6 M, 10 μL). The
reaction mixture was stirred for 15−20 min at 60 °C. The
radiochemical yield was 81% for [18F]4b, 36% for [18F]4c, and
80% for [18F]4e as determined by analytical HPLC from a
sample withdrawn from the reaction mixture. tR ([18F]4b) =
1.57 min, tR ([18F]4c) = 1.37 min, tR ([18F]4e) = 1.10 min
(radio-HPLC: Chromolith C8, 10 × 4.6 mm, 10−50%
acetonitrile (0.1% TFA) in water (0.1% TFA) in a linear
gradient over 5 min, 4 mL/min). 18F-glycopeptides [18F]4b,
[18F]4c, or [18F]4e were isolated by semipreparative HPLC
(Kromasil C8, 125 × 8 mm, 4 mL/min tR ([18F]4b) = 11.0
min, tR ([18F]4c) = 9.3 min, tR ([18F]4e) = 7.5 min (10−40%
acetonitrile (0.1% TFA) in water (0.1% TFA) in a linear
gradient over 30 min). The product fraction was diluted in
water and passed through a RP-18 cartridge (Lichrolut 100 mg,
Merck), and the product was eluted with a solution of ethanol/
PBS (1:1, 800 μL). For in vitro and in vivo experiments the
solvent was evaporated in vacuo, and the 18F-glycopeptides were
formulated with PBS (pH 7.4). Starting from [18F]fluoride (500
MBq), this procedure yielded 80−120 MBq (16−24% decay-
αvβ3 and αvβ5 Solid-Phase Receptor Binding Assay.
The receptor binding assay was performed following the
procedure described previously.18 In brief, commercially
available purified αvβ3 or αvβ5 (Triton-X solution, Chemicon
International) was diluted at 250 ng/mL or 500 ng/mL,
respectively, in coating buffer (25 mM Tris-HCl, 150 mM
NaCl, 1 mM CaCl2, 0.5 mM MgCl2, 1 mM MnCl2, pH 7.4),
and an aliquot of 100 μL/well was added to a 96-well microtiter
plate (Maxisorb, Nunc, Wiesbaden, Germany) and incubated at
4 °C overnight. The wells were washed once with blocking
buffer (200 μL, coating buffer containing 1% (w/v) BSA) and
incubated for two hours with blocking buffer (200 μL) at room
temperature. The plate was rinsed twice with binding buffer
(coating buffer with 0.1% (w/v) BSA) and incubated in the
presence of the competing ligand (90 μL, 0.5 pM to 500 nM
echistatin; 0.05 nM to 50 μM RGD peptide in binding buffer)
with [125I]echistatin (0.37 kBq/well, 10 μL, 0.05 nM;
PerkinElmer, Germany). After incubation for 3 h, the wells
were washed three times with binding buffer, and bound
[
125I]echistatin was solubilized with warm NaOH (2 M). The
radioactivity in the resulting samples was measured by a γ-
counter (Wallac Wizard). Each data point represents the mean
from three wells. All measurements were repeated at least two
times. Ki values were calculated by the use of the software
program GraphPad PRISM (GraphPad Software Inc., CA,
USA), employing the algorithm for “heterologous competitive
binding with ligand depletion”.
Cell Binding Assay. U87MG cells were harvested,
suspended in growth media, and seeded in 96-well plates at a
concentration of 500 000 cells per well (200 μL). The 96-well
plates were centrifuged at 150 g for 2 min; the supernatant was
removed by aspiration, and cells were suspended in blocking
buffer (25 mM Tris-HCl, 150 mM NaCl, 1 mM CaCl2, 0.5 mM
508
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
MgCl2, 1 mM MnCl2, pH 7.4, 1% BSA). The plates were
centrifuged again at 150 g for 2 min, and the supernatant was
removed. After incubation of the cell suspension with blocking
buffer for 15 min at room temperature, centrifugation was
repeated, and the cell pellets were washed twice with binding
buffer. Subsequently, the cells were incubated on a shaker in the
presence of varied amounts of competing ligand (90 μL, 0.05
nM to 50 μM RGD peptide) with 370 Bq/Eppendorf tube
RESULTS
■
Chemistry. Based on our previous work on the
comprehensive carbohydrate chemistry of glycosyl precursors,
including 1a,20 we envisaged the synthesis of an extended series
of glycosyl azides 1b−1f bearing a tosylate leaving group for
nucleophilic 18F-substitution in 6-position of the carbohydrate
(Figure 1). Employing classical synthetic methods, such as
acetylation and deacetylation steps, fluorination by diethylami-
nosulfur trifluoride (DAST), and anomeric azido-for-bromo
substitution reactions,23−27 we achieved the synthesis of the
labeling precursors 1b−1f, following the route of the synthesis
shown in Figure 2.
[
125I]echistatin (10 μL, 0.5 nM) in a total volume of 100 μL of
binding buffer (25 mM Tris-HCl, 150 mM NaCl, 1 mM CaCl2,
0.5 mM MgCl2, 1 mM MnCl2, pH 7.4, 0.1% BSA). After the
incubation time of 1 h cells were rinsed twice with binding
buffer, and the supernatant was removed; the cell pellets with
bound [125I]echistatin were resuspended in 2 N NaOH (200
μL) and transferred in counting tubes for measuring the
radioactivity with a γ-counter (Wallac Wizard). Each data point
is the mean of three determinations. All measurements were
repeated at least twice. Ki values were calculated as above.
Animal Model. All animal experiments were performed in
compliance with the protocols approved by the local Animal
Protection Authorities (Regierung Mittelfranken, Germany, no.
54-2532.1-15/08). Athymic nude mice (nu/nu) were obtained
from Harlan Winkelmann GmbH (Borchen, Germany) at 9−
11 weeks of age and were kept under standard conditions (12 h
light/dark) with food and water available ad libitum. U87MG
cells were harvested and suspended in sterile PBS at a
concentration of 2 × 107 cells/mL. Viable cells (2 × 106) in
PBS (100 μL) were injected subcutaneously in the back. One to
two weeks after inoculation the mice, now weighing about 40 g
and bearing tumors of 400−800 mg, were used for
biodistribution and small-animal PET studies.
Biodistribution Studies. U87MG xenografted mice were
injected with [18F]6Glc-RGD ([18F]4b) or 18F-Mlt-RGD ([18F]
4e) to a tail vein (4−10 MBq/mouse). The animals were killed
by cervical dislocation 30, 60, or 120 min postinjection (p.i.).
The tumors, other tissues (lung, liver, kidneys, heart, brain,
muscle, and intestine), and blood were removed and weighed.
Radioactivity of the samples was measured using a γ-counter
and expressed as percentage of injected dose per gram of tissue
(%ID/g), from which we calculated tumor-to-organ ratios. The
blockade experiment was carried out by coinjecting randomly
chosen mice with 500 μg (c(RGDfK)) (12.5 mg/kg body
weight) together with the radiotracer. These mice were killed
by cervical dislocation at 65 min p.i., and organs and tissue were
removed, weighed, and counted as described above.
Small-Animal PET Imaging. PET scans and image analysis
were performed using a small-animal PET scanner (Inveon,
Siemens Medical Solutions). About 3−15 MBq of [18F]6Glc-
RGD ([18F]4b) or [18F]Mlt-RGD ([18F]4e) were intravenously
injected into each mouse (n = 3−5) under isoflurane anesthesia
(4%). Thereupon, a 60-min dynamic emission scan (12 × 10 s,
3 × 1 min, 5 × 5 min, 3 × 10 min, total of 23 frames) was
initiated. After iterative maximum a posteriori (MAP) image
reconstruction of the decay and attenuation corrected images,
regions of interest (ROIs) were drawn over the tumor and
other targets. The radioactivity concentration within the tumor
region was obtained from the mean value within the multiple
ROIs and then converted to percent injected dose per gram (%
ID/g) relative to the injected dose. For receptor-blocking
experiments U87MG tumor-bearing mice were scanned for the
period 0−60 min p.i. after coinjection with radiotracer
[18F]6Glc-RGD or [18F]Mlt-RGD (4−8 MBq) containing
c(RGDfK) (12.5 mg/kg).
Figure 2. Course of the synthesis of 18F-labeling precursors 1b−1f. a:
1. TsCl, pyridine, 0 °C, 20 h, 2. Ac2O, 1 h, 0 °C to rt; b: 1. HBr/
AcOH, 0 °C to rt, 3 h, 2. NaN3, tetrabutylammonium hydrogen
sulfate, NaHCO3, ethyl acetate, rt, 2 h; c: 1. NsCl, pyridine, 0 °C, 20 h,
2. Ac2O, 1 h, 0 °C to rt. Please see the Supporting Information for
details (Schemes S5 and S6).
Starting from the peracetylated glycosyl azides 2a−2f and
Zemplen deacetylation (Figure 3A), the corresponding free
́
glycosyl azides 3a−3f were reacted with the alkyne-bearing
peptide c(RGDfPra) (Pra = propargyl glycine) by applying the
chemoselective 1,3-dipolar cycloaddition under Cu(I)-catalyzed
reaction conditions (click chemistry) as described previously.21
Following this chemical strategy, the series of four RGD
glycopeptides 19F-6Glc-RGD (4b), 19F-Gal-RGD (4c), 19F-Mlt-
RGD (4e), and 19F-Cel-RGD (4f) were obtained in high yields
of about 80% after preparative HPLC purification; their identity
and purity (>98%) were proven by LC-ESI-MS.
Radiochemistry. The applicability of our new series of
labeling precursors 1b, 1c, 1d, 1e, and 1f for concomitant 18F-
labeling and glycosylation was investigated by applying a three-
step two-pot labeling procedure, as depicted in Figure 4,
following the strategy recently developed by our group for
precursor 1a.20 The glucosyl 6-tosylate precursor 1b gave the
18F-labeled product [18F]2b in 84
7% radiochemical yield
(RCY) after 10 min, whereas the galactosyl 6-nosylate
509
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
Figure 3. (A) Synthesis of the series of glycosylated RGD peptides 4a (ref 21) and 4b−4f. a: NaOMe, MeOH, rt, 1 h; b: c(RGDfPra), CuSO4,
sodium ascorbate, PBS/10% ethanol, rt, 30 min. (B) Displacement of [125I]echistatin by RGD peptides using immobilized αvβ3. (C) Displacement of
[
125I]echistatin by RGD peptides using U87MG cells. Each data point is the mean value SD from three independent experiments, each performed
in triplicate.
the classical [18F]FDG synthesis.28,29 To investigate the
potential of [18F]3b, [18F]3c or [18F]3c as click labeling agents,
the two-step one-pot approach was completed by neutralization
and direct addition of the alkyne-functionalized c(RGDfPra) in
the presence of copper(II)sulfate and sodium ascorbate. The
RCY of this step was determined by analytical radio-HPLC
being 81% for [18F]6Glc-RGD ([18F]4b), 36% for [18F]Gal-
RGD ([18F]4c), and 80% for [18F]Mlt-RGD ([18F]4e) after
15−20 min at 60 °C (Figure S1, Table S1). After semi-
preparative HPLC separation, the identity of 18F-labeled
compounds was verified by coinjection of the radiolabeled
compounds together with the authentic 19F-substituted
compounds on an analytical HPLC system (Figure S2, Figure
S3). [18F]6Glc-RGD and [18F]Mlt-RGD were obtained in an
overall RCY of 16−24% (relative to [18F]fluoride) in a total
synthesis time of 70−75 min and in specific activities of 50−
200 GBq/μmol. Due to their high-yielding and reliable
radiosynthesis, we selected the glucosyl and the maltosyl
RGD peptide ([18F]6Glc-RGD and [18F]Mlt-RGD) to proceed
with a comparative in vivo study on the biodistribution and
small-animal PET imaging.
In Vitro Studies. In Vitro Binding Assay for Glyco-RGD
Peptides. Integrin binding assays for [19F]2Glc-RGD (4a)
(only αvβ5; data on αvβ3 binding: see ref 21), [19F]6Glc-RGD
(4b), [19F]Gal-RGD (4c), [19F]Mlt-RGD (4e), and [19F]Cel-
RGD (4f) were performed as competition binding experiments
using [125I]echistatin as a radioligand and immobilized αvβ3,
αvβ5, or αvβ3−positive U87MG cells, respectively, as described
previously.18 Notably, the αvβ3 integrin expression of U87MG
cells was confirmed by immunofluorescence staining indicating
αvβ3-positive and αvβ5-negative staining results (data not
shown). The results of integrin binding affinities are presented
in Table 1 and Figure 3B,C. As expected, all our RGD peptides
inhibited the binding of [125I]echistatin to αvβ3, αvβ5 and to
U87MG cells in a concentration-dependent manner. The
affinities of the glycosylated RGD peptides ranged from 11 to
55 nM for αvβ3, 6−14 nM for αvβ5 and 90−395 nM for
U87MG cells. Interestingly, the glycosylation of c(RGDfPra)
enhanced the binding affinity determined in U87MG cells by a
Figure 4. Radiosynthesis of 18F-glyco-RGD peptides [18F]6Glc-RGD,
[18F]Gal-RGD, and [18F]Mlt-RGD. a: [K+ ⊂ K222]18F−, MeCN, 85
°C, 10 min; b: 60 mM NaOH, 60 °C, 5 min; c: c(RGDfPra), CuSO4,
sodium ascorbate, ethanol/water 1:1, 60 °C, 15−20 min.
precursor 1d gave [18F]2c in only 38 8% RCY (Figure S1,
Table S1). Notably, in case of the alternative galactosyl 6-tosyl-
precursor 1c an unidentified product occurred in radiochemical
yields of 80%, whereas the desired product [18F]2c had RCY of
only 6 3%. Alternative reaction parameters, such as the use of
K2CO3 instead of K2CO3/KH2PO4, did not result in higher
RCY (Table S1). The disaccharide precursors 1e and 1f were
18F-labeled in RCYs of 61% and 6%, giving [18F]2e and [18F]2f,
respectively. The low RCY of the cellobiosyl azide [18F]2f
prevented its use for the subsequent reactions. After semi-
preparative HPLC isolation of the 18F-labeled products [18F]2b,
[18F]2c, or [18F]2e, deacetylation was accomplished by the
addition of sodium hydroxide (60 mM) to yield pure [18F]3b,
[18F]3c, or [18F]3e, respectively, after a reaction time of 5 min
at 60 °C. These reaction conditions are very similar to those of
510
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
Table 1. In Vitro Binding Affinities of a Series of RGD
Glycopeptides and c(RGDfPra) and c(RGDfK) to αvβ3, αvβ5,
and Human U87MG Cells (n = 3)
Table 3. Biodistribution Data for [18F]6Glc-RGD ([18F]4b)
Measured in U87MG Bearing Nude Mice
a
a
65 min
b
organ
blood
30 min
65 min
blocking
120 min
compound
c(RGDfPra)
Ki (αvβ3)
Ki (αvβ5)
Ki (U87MG)
b
b
b
0.50 0.32
0.80 0.17
3.65 0.45
4.53 1.63
0.32 0.10
0.63 0.08
1.33 0.43
0.09 0.04
0.43 0.05
2.46 0.29
1.69 0.62
0.26 0.22
0.49 0.26
0.79 0.32
0.04 0.01
0.12 0.02
2.19 0.36
2.23 0.49
0.05 0.02
0.10 0.06
0.11 0.03
0.04 0.01
0.23 0.05
1.33 0.23
1.07 0.12
0.10 0.04
0.27 0.10
0.33 0 07
26 6 nM
11 4 nM
31 5 nM
22 4 nM
13 9 nM
1396 365 nM
b
4a: [19F]2Glc-RGD
4b: [19F]6Glc-RGD
4c: [19F]-Gal-RGD
4e: [19F]Mlt-RGD
4f: [19F]Cel-RGD
c(RGDfK)
8
7
2 nM
3 nM
395 152 nM
lung
liver
287 44 nM
270 150 nM
90 45 nM
228 33 nM
n.d.
kidney
heart
spleen
14 9 nM
55 14 nM
30 6 nM
32 8 nM
6
2 nM
n.d.
U87MG
tumor
12 7 nM
a
Data are expressed as mean values standard deviation from three
brain
0.07 0.02
1.06 0.70
1.25 0.74
0.51 0.40
0.07 0.04
0.68 0.32
0.60 0.36
0.23 0.23
0.06 0.05
1.41 1.29
2.29 2.68
0.13 0.12
0.02 0 01
0.29 0.34
0.96 0.99
0.15 0.06
independent experiments, each determined in triplicate; n.d. = not
muscle
intestine
femur
b
determined. Values from ref 21 for comparison.
factor of 15 in the case of the maltosyl peptide [19F]Mlt-RGD
relative to the nonglycosylated parent compound c(RGDfPra)
(Table 1).
a
Data represent mean values
Coinjection of 12.5 mg/kg c(RGDfK).
standard deviation (n = 3−4).
b
Lipophilicity and Stability in Serum of the RGD
Glycopeptides in Vitro. The log D7.4 values of the glycosylated
RGD peptides are listed in Table 2. The disaccharyl peptide
kidney and tumor-to-liver values were higher for [18F]Mlt-RGD
by about a factor of 2 (Figure S4).
In small-animal PET studies with U87MG tumor-bearing
mice, [18F]6Glc-RGD and [18F]Mlt-RGD were compared to
[18F]2Glc-RGD. The two novel tracers had similar time-
activity-curves SUV units (Figure 6). PET recordings for
[18F]6Glc-RGD and [18F]Mlt-RGD mirrored the biodistribu-
tion data at 60 min p.i., with almost equal tumor uptake (Figure
S5), which gave adequate tumor visualization. In animals
coinjected with radiotracer and c(RGDfK), tumor uptake was
negligible, demonstrating the specificity of both [18F]6Glc-
RGD and [18F]Mlt-RGD for imaging of integrin expression in
vivo by PET (Figure 6A−F).
Table 2. Log D7.4 Values of the Various Glycosylated RGD
Peptides
a
compound
log D7.4
b
[18F]2Glc-RGD
[18F]6Glc-RGD
[18F]Gal-RGD
[18F]Mlt-RGD
−3.77 0.22
−3.47 0.13
−3.56 0.07
−4.10 0.12
a
Data are expressed as mean values
standard deviation from 3
b
independent experiments. Values from ref 21 for comparison.
DISCUSSION
[18F]Mlt-RGD was slighty more hydrophilic than the RGD
peptides coupled to the monosaccharides glucose ([18F]2Glc-
RGD, [18F]6Glc-RGD) and galactose ([18F]Gal-RGD). The
high stability of cyclic RGD peptides in serum is well-known.
We confirmed the stability of [18F]6Glc-RGD and [18F]Mlt-
RGD in human serum in vitro at 37 °C (Figures S6 and S7).
Both radiotracers did not show any degradation within 60 min
as determined by analytical radio-HPLC.
■
We investigated the influence of different glycosyl moieties on
the biodistribution and clearance pattern of cyclic RGD
pentapeptides. To this end we synthesized five tosyl or nosyl
containing glycosyl-azides as 18F-labeling precursors and
examined the binding of the corresponding series of
fluoroglycosyl peptides in vitro and the biodistribution
properties in vivo. The synthesis of the labeling precursors
were easily achieved by reaction of glucose, galactosylazide,
maltosyl azide, or cellobiosyl azide, respectively, with tosyl or
nosyl chloride in pyridine and subsequent acetylation using
acetic anhydride.23 After column chromatography, the labeling
precursors were obtained as pure β-anomers. The 18F-labeling
was performed using mild reaction conditions with a lower
amount of K2CO3 in the presence of KH2PO4 compared to
commonly used 18F-labeling conditions.28,30 This modification
proved to be advantageous for the subsequent separation of the
18F-labeled products by semipreparative HPLC,20 in that the
less basic reaction medium resulted in less elimination reactions
and degradation of the labeling precursors.
In our initial proof-of-principle of the present concept,
[18F]2Glc-RGD displayed relatively high uptake in liver (4.8%
ID/g) and kidney (6.6%ID/g) at 120 min postinjection.21
Moving the 18F-label in the glucosyl residue from position 2 to
position 6 in the present study led to significantly reduced
kidney and liver uptake of [18F]6Glc-RGD (1.1%ID/g and
1.3%ID/g, respectively). Since the log D7.4 values for [18F]6Glc-
RGD and [18F]2Glc-RGD as well as their in vitro binding
properties are very similar, this favorable effect is most likely
In Vivo Studies. Biodistribution and Small-Animal PET
Imaging in Tumor-Bearing Nude Mice. The biodistribution
studies were performed with [18F]6Glc-RGD and [18F]Mlt-
RGD using U87MG tumor bearing nude mice, and results are
shown in Table 3 and Table 4 (see also Figure S4). Both
radiotracers showed a similar tumor uptake of 0.8%ID/g after
60 min p.i. (Figure 5A). At 120 min p.i. [18F]Mlt-RGD showed
a higher retention in the tumor than did [18F]6Glc-RGD (0.8
vs 0.3%ID/g, respectively). Both radiotracers displayed specific
tumor uptake as determined by coinjection of radiotracer with
the blocker c(RGDfK) (Table 3, Table 4, Figure S5). The
radioactivity concentrations in normal organs 60 min p.i. were
similarly expected for kidney and liver; [18F]Mlt-RGD had 40%
lower renal uptake at 60 min post injection and 61% lower
hepatic uptake. [18F]6Glc-RGD showed fast clearance from the
blood and most examined organs, whereas [18F]Mlt-RGD
showed a significantly decreased clearance from blood and
higher retention in most organs compared to [18F]6Glc-RGD
(see Figure S4). This led to higher tumor-to-blood values for
[18F]6Glc-RGD than for [18F]Mlt-RGD (Figure 5B), whereas
tumor-to-muscle ratios were similar (Figure 5B) and tumor-to-
511
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
a
Table 4. Biodistribution Data for [18F]Mlt-RGD ([18F]4e) Measured in U87MG Bearing Nude Mice
b
organ
blood
30 min
60 min
60 min blocking
120 min
0.36 0.10
0.85 0.10
1.00 0.06
1.76 0.43
0.33 0.03
0.56 0.04
1.26 0.14
0.12 0.01
0.24 0.04
0.69 0.11
0.32 0.09
0.29 0.06
0.92 0.05
0.27 0.01
0.50 0.05
0.95 0.06
1.01 0.05
0.25 0.02
0.45 0.02
0.85 0.11
0.14 0.01
0.30 0.13
0.57 0.11
0.28 0.07
0.19 0.03
0.72 0.17
0.08 0.02
0.17 0.02
0.39 0.08
0.59 0.16
0.05 0.01
0.12 0.09
0.09 0.02
0.02 0.01
0.41 0.38
0.25 0.27
0.23 0.17
0.17 0.16
0.18 0.07
0.29 0.04
0.45 0.06
0.83 0.15
0.83 0.16
0.27 0.03
0.43 0.06
0.76 0.15
0.22 0.04
0.22 0.05
0.44 0.02
0.24 0.03
0.20 0.04
0.61 0.13
lung
liver
kidney
heart
spleen
U87MG tumor
brain
muscle
intestine
femur
pancreas
duodenum
a
b
Data represent mean values standard deviation (n = 3−4). Coinjection of 12.5 mg/kg c(RGDfK).
Figure 5. Biodistribution data (A) and tumor-to-organ ratios (B) of [18F]2Glc-RGD, [18F]6Glc-RGD, and [18F]Mlt-RGD in nude mice bearing
U87MG tumors 60 min p.i. Data are expressed as mean values SD from 3 to 5 animals. *Data of [18F]2Glc-RGD are taken from ref 21 and
depicted for comparison.
due to differences in the metabolic fate of the tracers in vivo. We
can speculate that the formation of the glucuronic acid by
oxidation in the 6-position in the liver could result in high liver
uptake of radioactivity from [18F]2Glc-RGD, whereas the
oxidative trapping would not be possible in the case of
[18F]6Glc-RGD.
[18F]Mlt-RGD to [18F]Galacto-RGD in the U87MG nude mice
model, the tumor-to-kidney-ratio, the tumor-to-liver-ratio, and
the total tumor uptake at 120 min p.i. were almost equal (0.9%
ID/g).31 However, [18F]Mlt-RGD reached the peak value of
0.9%ID/g much earlier, at only 60 min post injection. The
uptake for [18F]Galacto-RGD in the M21 tumor model is
reportedly twice as high as that of [18F]Mlt-RGD in U87MG
tumors.9 This difference could be most likely ascribed to
differences in the overall αvβ3 or αvβ5 integrin expression
between the M21 tumors and U87MG tumors, as [18F]Galacto-
RGD showed similar uptake in U87MG tumors when
compared with [18F]Mlt-RGD.31
A variety of 18F-labeled RGD peptides have been synthesized
during the past decade.3 One radioactive RGD peptide which
already finds frequent clinical use is [18F]Galacto-RGD, which
was developed by Haubner et al. in 2001.9 Its use is encouraged
by the favorable pharmacokinetics of this glycopeptide. In
contrast to the large variety of other radiolabeled RGD
derivatives containing bulky aromatic groups (e.g., [18F]SFB
or [18F]FBA), [18F]GalactoRGD is a glycoconjugate, making it
a hydrophilic tracer with fast blood clearance and predom-
inantly renal excretion in vivo. However, a major drawback of
[18F]Galacto-RGD is the time-consuming four-step radiosy-
thesis with an overall preparation time of 200 min, which
includes three HPLC separations.8 In contrast, the RGD
glycopeptides presented in this study are prepared by efficient
and simple click-chemistry. Our new radiosythesis consists of a
three-step (two-pot) procedure and includes two HPLC
separations, saving at least 120 min in the overall preparation
time, which equals one-half life of fluorine-18, and thus
resulting in a significantly increased radiochemical yield of the
18F-labeled glycopeptides.
The tumor-to-blood ratio for [18F]Galacto-RGD at 120 min
p.i. is reported to be 30,9 while [18F]6Glc-RGD gave a ratio of 8
and [18F]Mlt-RGD only 3 (Figure 5B). The lower ratio of
[18F]Mlt-RGD is due to its relatively slow blood clearance
between 30 and 120 min p.i. as compared with the glucosyl
analogue (Table 4; Figure S4). The hydrophilicities of
[18F]Mlt-RGD and [18F]6Glc-RGD are comparable (log D7.4
of −3.5 vs. −4.1), with [18F]Mlt-RGD having the highest
hydrophilicity in the series of glycopeptides under study. Thus,
hydrophobicity gives no direct explanation for the apparently
slow blood clearance of [18F]Mlt-RGD. However, a similar
effect of glycosylation on blood clearance was found by
Schottelius et al. for 125I-iodinated glycosylated octreotide
derivatives, where elongation of the glucose moiety from a
monosaccharide to a trisaccharide chain decreased the tumor-
to-blood ratio from 6 to 2.32
Our results of the biodistribution of [18F]Mlt-RGD are very
similar to that reported for [18F]Galacto-RGD.9,31 Comparing
512
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
Figure 6. Small-animal PET using [18F]2Glc-RGD (A), [18F]6Glc-RGD (B), and [18F]Mlt-RGD (C) at 60 min p.i.: Transaxial (top) and coronal
(bottom) images from U87MG-bearing mice injected with the tracer alone (control, left) and coinjection of c(RGDfK) (12.5 mg/kg, blocking,
right). (D−F) Time−activity curves of dynamic small-animal PET acquisitions. Data represent mean values SD from 3 to 5 animals.
Multimeric RGD peptides have the principle advantage of
higher tumor uptake and high tumor retention and sufficient
clearance properties for imaging purposes, as recently been
demonstrated by various studies.33−35 Among these, Guo et al.
recently reported the comparison of three dimeric 18F-AlF-
NOTA-RGD tracers, demonstrating tumor uptake values of
2.1−2.8%ID/g in U87MG xenografts, liver uptake of 1.6−2.0%
ID/g, and kidney uptake of 2.0−2.6%ID/g at 120 min
postinjection.35 Furthermore, the recently reported 18F-labeled
improved renal clearance and sufficient tumor uptake in
combination with a relatively high tumor retention of 60%
(120 min p.i. vs 30 min p.i., Table 4), resulting in tumor/
background ratios similar to those of dimeric peptides.
Taken together, in this study of fluoroglycosylation of a cyclic
RGD peptide we indentified two glycopeptide tracers, namely,
[18F]6Glc-RGD and [18F]Mlt-RGD, which were prepared by
the same radiochemical procedure, but with greatly different
pharmacokinetic properties in vivo. Both radiotracers showed a
dramatic decrease in kidney and liver uptake when compared
with the previously published [18F]2Glc-RGD.21 While
[18F]6Glc-RGD has favorable tumor-to-blood ratios in the
U87MG tumor model, but only low tumor retention, [18F]Mlt-
RGD proved to have slower blood clearance but even less
uptake in the kidney and liver. These properties, in conjunction
with high tumor retention, make [18F]Mlt-RGD a promising
candidate for molecular studies of the integrin expression.
−
dimeric RGD peptide [c(RGDfK)]2E-18F-ArBF3 showed
rather low tumor uptake of 0.5%ID/g in U87MG tumor-
bearing nude mice.36 The authors ascribed this result to the
marked renal clearance of the tracer.
With [18F]Mlt-RGD we found a tumor uptake of
approximately 1% ID/g at 120 min p.i., being lower than that
of most dimeric RGD tracers, but significantly higher than the
tumor uptake of the dimeric RGD peptide [c(RGDfK)]2E-18F-
36
ArBF3 . Moreover, the uptake values of [18F]Mlt-RGD in
liver and kidney were desirably low (both 0.8%ID/g) compared
to the dimeric tracers reported by Guo et al. (1.6−2.6%ID/g;35
see above). Notably, the low kidney uptake of [18F]Mlt-RGD
resulted in a tumor/kidney ratio of 1, very closely reflecting the
tumor/kidney ratio of the dimeric RGD peptides reported by
Guo and co-workers (1.02−1.06).35 Therefore, dimeric RGD
peptides show higher tumor uptake which is consistent with the
general theory of improved avidities of multimeric ligands to
the αvβ3 integrin,33 whereas the fluoroglycosylation method, in
the case of [18F]Mlt-RGD, induced a suitable balance between
−
CONCLUSION
■
The aim of this study was to develop an alternative
radiopharmaceutical for the clinically used tracer [18F]-
Galacto-RGD for the noninvasive determination of the αvβ3
expression. To this end we extended our previously designed
radiosynthesis method based on the CuAAC click chemistry,
using different 18F-glycosyl azides and an alkyne-bearing RGD
peptide. Our approach generalized to a reliable and high-
yielding 18F-synthetic strategy. The series of new 18F-
fluoroglycosylated RGD peptides displayed high affinity toward
513
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
(5) Schottelius, M.; Laufer, B.; Kessler, H.; Wester, H.-J. Ligands for
Mapping αvβ3-Integrin Expression in Vivo. Acc. Chem. Res. 2009, 42,
969−980.
(6) Wan, W.; Guo, N.; Pan, D.; Yu, C.; Weng, Y.; Luo, S.; Ding, H.;
Xu, Y.; Wang, L.; Lang, L.; Xie, Q.; Yang, M.; Chen, X. First
Experience of 18F-Alfatide in Lung Cancer Patients Using a New
Lyophilized Kit for Rapid Radiofluorination. J. Nucl. Med. 2013, 54,
691−698.
(7) Beer, A. J.; Schwaiger, M. PET imaging of avb3 expression in
cancer patients. Methods Mol. Biol. 2011, 680, 183−200.
(8) Haubner, R.; Kuhnast, B.; Mang, C.; Weber, W. A.; Kessler, H.;
Wester, H. J.; Schwaiger, M. [18F]Galacto-RGD: synthesis, radio-
labeling, metabolic stability, and radiation dose estimates. Bioconjugate
Chem. 2004, 15, 61−69.
(9) Haubner, R.; Wester, H. J.; Weber, W. A.; Mang, C.; Ziegler, S. I.;
Goodman, S. L.; Senekowitsch-Schmidtke, R.; Kessler, H.; Schwaiger,
M. Noninvasive imaging of αvβ3 integrin expression using 18F-labeled
RGD-containing glycopeptide and positron emission tomography.
Cancer Res. 2001, 61, 1781−1785.
the integrins αvβ3 and αvβ5 and showed distinct in vivo
properties. Compared to [18F]Galacto-RGD, [18F]Mlt-RGD
has similar tumor-to-background values, resulting from a
predominantly renal clearance together with high tumor
retention. Thus, [18F]Mlt-RGD is distinguished by a favorable
in vivo biodistribution, while presenting benefits over [18F]-
Galacto-RGD imparted by more rapid and simplified radiosyn-
thesis, thus representing a very promising alternative radiotracer
for imaging integrin expression by PET in further (pre)clinical
studies.
ASSOCIATED CONTENT
■
S
* Supporting Information
General information on Materials and Methods. Syntheses of
1e, 1f, 2b, 2c, 2e, 2f, 3c, 3e, 3f, 5e, 7e, 8e, 9e, 10e, 11e;
Schemes S1−S6. Radiochemical yields (RCY) of the two-step
glycosylation reaction; Table S1. Radiochemical yields of the
18F-labeled glycosyl azides [18F]2a−f; Figure S1A. Time
dependency of the RCY of the RGD glycopeptides [18F]4a−
c and [18F]4e for the glycosylation reaction step; Figure S1B.
HPLC analysis of [18F]6Glc-RGD (radiochemical identity);
Figure S2. HPLC analysis of [18F]Mlt-RGD (radiochemical
identity); Figure S3. Biodistribution data of [18F]6Glc-RGD
and [18F]Mlt-RGD in U87MG tumor-bearing nude mice 30,
60, and 120 min p.i.; Figure S4. Tumor uptake of [18F]2Glc-
RGD ([18F]4a21) in comparison with [18F]6Glc-RGD ([18F]
4b) and [18F]Mlt-RGD ([18F]4e) in U87MG tumor-bearing
nude mice at 60 min postinjection; Figure S5. Stability of
[18F]6Glc-RGD and [18F]Mlt-RGD in human serum in vitro;
Figures S6 and S7. This material is available free of charge via
the Internet at http://pubs.acs.org.
(10) Schottelius, M.; Rau, F.; Reubi, J. C.; Schwaiger, M.; Wester, H.
A. Modulation of pharmacokinetics of radioiodinated sugar-conjugated
somatostatin analogues by variation of peptide net charge and
carbohydration chemistry. Bioconjugate Chem. 2005, 16, 429−437.
(11) Fischer, C. R.; Muller, C.; Reber, J.; Muller, A.; Kramer, S. D.;
̈
̈
Ametamey, S. M.; Schibli, R. [18F]Fluoro-Deoxy-Glucose Folate: A
Novel PET Radiotracer with Improved in Vivo Properties for Folate
Receptor Targeting. Bioconjugate Chem. 2012, 23, 805−813.
(12) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Click Chemistry:
Diverse Chemical Function from a Few Good Reactions. Angew.
Chem., Int. Ed. Engl. 2001, 40, 2004−2021.
(13) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
stepwise Huisgen cycloaddition process: Copper(I)-catalyzed regiose-
lective “ligation” of azides and terminal alkynes. Angew. Chem., Int. Ed.
2002, 41, 2596−2599.
(14) Tornoe, C. W.; Christensen, C.; Meldal, M. Peptidotriazoles on
solid phase: [1,2,3]-triazoles by regiospecific copper(I)-catalyzed 1,3-
dipolar cycloadditions of terminal alkynes to azides. J. Org. Chem.
2002, 67, 3057−3064.
(15) Namavari, M.; Cheng, Z.; Zhang, R.; De, A.; Levi, J.; Hoerner, J.
K.; Yaghoubi, S. S.; Syud, F. A.; Gambhir, S. S. A Novel Method for
Direct Site-Specific Radiolabeling of Peptides Using [18F]FDG.
Bioconjugate Chem. 2009, 20, 432−436.
(16) Hultsch, C.; Schottelius, M.; Auernheimer, J.; Alke, A.; Wester,
H. J. 18F-Fluoroglucosylation of peptides, exemplified on cyclo-
(RGDfK). Eur. J. Nucl. Med. Mol. Imaging 2009, 36, 653−658.
(17) Schottelius, M.; Poethko, T.; Herz, M.; Reubi, J. C.; Kessler, H.;
Schwaiger, M.; Wester, H. J. First 18F-labeled tracer suitable for routine
clinical imaging of sst receptor-expressing tumors using positron
emission tomography. Clin. Cancer Res. 2004, 10, 3593−3606.
(18) Prante, O.; Einsiedel, J.; Haubner, R.; Gmeiner, P.; Wester, H.
J.; Kuwert, T.; Maschauer, S. 3,4,6-tri-O-acetyl-2-deoxy-2-[18F]-
fluoroglucopyranosyl phenylthiosulfonate: A thiol-reactive agent for
the chemoselective 18F-glycosylation of peptides. Bioconjugate Chem.
2007, 18, 254−262.
AUTHOR INFORMATION
■
Corresponding Author
*Molecular Imaging and Radiochemistry, Nuclear Medicine
Clinic, Friedrich-Alexander University, Schwabachanlage 6, D-
91054 Erlangen, Germany. Tel.: +49-9131-8544440. Fax: +49-
9131-8534440. E-mail: olaf.prante@uk-erlangen.de.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Bianca Weigel, Susanne Knauf, and Dr.
Carsten Hocke for expert technical support and thank Dr. Paul
Cumming for critical revisions to the manuscript. This work
was supported by the Deutsche Forschungsgemeinschaft
(DFG, grants MA 4295/1-2 and PR 677/5-1).
(19) Wuest, F.; Berndt, M.; Bergmann, R.; van den Hoff, J.; Pietzsch,
J. Synthesis and application of [18F]FDG-maleimidehexyloxime
([18F]FDG-MHO): A [18F]FDG-based prosthetic group for the
chemoselective 18F-labeling of peptides and proteins. Bioconjugate
Chem. 2008, 19, 1202−1210.
(20) Maschauer, S.; Prante, O. A series of 2-O-trifluoromethylsul-
fonyl-D-mannopyranosides as precursors for concomitant 18F-labeling
and glycosylation by click chemistry. Carbohydr. Res. 2009, 344, 753−
761.
REFERENCES
■
(1) Folkman, J. Angiogenesis: an organizing principle for drug
discovery? Nat. Rev. Drug Discovery 2007, 6, 273−286.
(2) Al-Husein, B.; Abdalla, M.; Trepte, M.; Deremer, D. L.;
Somanath, P. R. Antiangiogenic therapy for cancer: an update.
Pharmacotherapy 2012, 32, 1095−1111.
(3) Haubner, R.; Beer, A. J.; Wang, H.; Chen, X. Positron emission
tomography tracers for imaging angiogenesis. Eur. J. Nucl. Med. Mol.
Imaging 2010, 37, S86−S103.
(21) Maschauer, S.; Einsiedel, J.; Haubner, R.; Hocke, C.; Ocker, M.;
Hubner, H.; Kuwert, T.; Gmeiner, P.; Prante, O. Labeling and
̈
(4) Aumailley, M.; Gurrath, M.; Muller, G.; Calvete, J.; Timpl, R.;
Kessler, H. Arg-Gly-Asp constrained within cyclic pentapeptides.
Strong and selective inhibitors of cell adhesion to vitronectin and
laminin fragment P1. FEBS Lett. 1991, 291, 50−54.
Glycosylation of Peptides Using Click Chemistry: A General Approach
to 18F-Glycopeptides as Effective Imaging Probes for Positron
Emission Tomography. Angew. Chem., Int. Ed. 2010, 49, 976−979.
514
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515

Molecular Pharmaceutics
Article
(22) Boutureira, O.; Bernardes, G. J. L.; D’Hooge, F.; Davis, B. G.
Direct radiolabelling of proteins at cysteine using [18F]-fluorosugars.
Chem. Commun. 2011, 47, 10010−10012.
(23) Gyoergydeak, Z.; Szilagyi, L. Facile Syntheses of galactopyr-
anosyl and glucopyranosyl azides substituted at C-6. Liebigs Ann.
Chem. 1987, 235−241.
(24) Lopez, M.; Trajkovic, J.; Bornaghi, L. F.; Innocenti, A.; Vullo,
D.; Supuran, C. T.; Poulsen, S. A. Design, Synthesis, and Biological
Evaluation of Novel Carbohydrate-Based Sulfamates as Carbonic
Anhydrase Inhibitors. J. Med. Chem. 2011, 54, 1481−1489.
(25) Porwanski, S.; Marsura, A. Tandem Staudinger-Aza-Wittig
Templated Reaction: De Novo Synthesis of Sugar-Ureido Cryptands.
Eur. J. Org. Chem. 2009, 2047−2050.
(26) Schengrund, C. L.; Kovac, P. UDP-6-deoxy-6-fluoro-a-D-
galactose binds to two different galactosyltransferases, but neither
can effectively catalyze transfer of the modified galactose to the
appropriate acceptor. Carbohydr. Res. 1999, 319, 24−28.
(27) Withers, S. G.; MacLennan, D. J.; Street, I. P. The synthesis and
hydrolysis of a series of deoxyfluoro-D-glucopyranosyl phosphates.
Carbohydr. Res. 1986, 154, 127−144.
(28) Hamacher, K.; Coenen, H. H.; Stocklin, G. Efficient
̈
stereospecific synthesis of no-carrier-added 2-[18F]-fluoro-2-deoxy-D-
glucose using aminopolyether supported nucleophilic substitution. J.
Nucl. Med. 1986, 27, 235−238.
(29) Fuchtner, F.; Steinbach, J.; Mading, P.; Johannsen, B. Basic
̈
̈
hydrolysis of 2-[18F]fluoro-1,3,4,6-tetra-O-acetyl-d-glucose in the
preparation of 2-[18F]fluoro-2-deoxy-d-glucose. Appl. Radiat. Isot.
1996, 47, 61−66.
(30) Coenen, H. H.; Schuller, M.; Stocklin, G.; Klatte, B.; Knochel,
̈
̈
̈
A. Preparation of N.C.A. [17-18F]-fluoroheptadecanoic acid in high
yields via aminopolyether supported, nucleophilic fluorination. J. Label.
Compds. Radiopharm. 1986, 23, 455−466.
(31) Liu, S.; Liu, Z.; Chen, K.; Yan, Y.; Watzlowik, P.; Wester, H. J.;
Chin, F. T.; Chen, X. 18F-labeled galacto and PEGylated RGD dimers
for PET imaging of αvβ3 integrin expression. Mol. Imaging Biol. 2010,
12, 530−538.
(32) Schottelius, M.; Wester, H. J.; Reubi, J. C.; Senekowitsch-
Schmidtke, R.; Schwaiger, M. Improvement of pharmacokinetics of
radioiodinated Tyr3-octreotide by conjugation with carbohydrates.
Bioconjugate Chem. 2002, 13, 1021−1030.
(33) Liu, S. Radiolabeled cyclic RGD peptides as integrin αvβ3-
targeted radiotracers: maximizing binding affinity via bivalency.
Bioconjugate Chem. 2009, 20, 2199−2213.
(34) Wangler, C.; Maschauer, S.; Prante, O.; Schafer, M.;
̈
Schirrmacher, R.; Bartenstein, P.; Eisenhut, M.; Wangler, B. Multi-
̈
merization of cRGD peptides by click chemistry: synthetic strategies,
chemical limitations, and influence on biological properties. Chem-
BioChem 2010, 11, 2168−2181.
(35) Guo, J.; Lang, L.; Hu, S.; Guo, N.; Zhu, L.; Sun, Z.; Ma, Y.;
Kiesewetter, D. O.; Niu, G.; Xie, Q.; Chen, X. Comparison of Three
Dimeric F-AlF-NOTA-RGD Tracers. Mol. Imaging Biol. 2013,
10.1007/s11307-013-0668-1.
(36) Li, Y.; Liu, Z.; Lozada, J.; Wong, M. Q.; Lin, K. S.; Yapp, D.;
Perrin, D. M. Single step F-labeling of dimeric cycloRGD for
functional PET imaging of tumors in mice. Nucl. Med. Biol. 2013,
40, 959−966.
515
dx.doi.org/10.1021/mp4004817 | Mol. Pharmaceutics 2014, 11, 505−515