A practical synthesis of [18F]FtRGD: An angiogenesis biomarker for PET

Article abstract of DOI:10.1002/jlcr.3019

Integrins have become increasingly attractive targets for molecular imaging of angiogenesis with positron emission tomography or single-photon emission computed tomography, but the reliable production of radiopharmaceuticals remains challenging. A strategy for chemoselective labeling of the integrin ligand - c(RGDyK) peptide - has been developed on the basis of the Cu(I)-catalyzed conjugation reaction. Recently, we reported a nucleophilic detagging and fluorous solid-phase extraction method providing an easy way to implement an approach for obtaining 2-[¹⁸F]fluoroethyl azide. In this work, we report the practical use of this method for the preparation of the 2-[ ¹⁸F]fluoroethyl-triazolyl conjugated c(RGDyK) peptide: [ ¹⁸F]FtRGD. The two-step, two-pot synthesis, HPLC purification, and reformulation could be readily performed with a standard nucleophilic radiofluorination synthesizer (GE TRACERlab FX_FN), with minimal modifications. [¹⁸F]FtRGD was obtained in a solution for injection (>500 MBq/mL) in 10-30% nondecay-corrected radiochemical yield, excellent radiochemical purity (>98%), and 28 ± 13 GBq/μmol specific activity. [¹⁸F]FtRGD (K_i = 54 ± 14 nM for α_Vβ₃ and 1.7 ± 0.2 nM for α_Vβ₅) was evaluated in mice and showed good stability in vivo, good tumor-to-background ratio (1.6 ± 0.3 %ID/g at 1.5 h post-injection in U87-MG tumors), and rapid urinary excretion. Therefore, [¹⁸F]FtRGD proved valuable for preclinical positron emission tomography imaging of integrin expression. A two-step, two-pot, and chemoselective strategy was developed for the radiolabeling of the c(RGDyK) peptide. Via nucleophilic detagging and fluorous solid-phase extraction, we obtain 2-[¹⁸F]fluoroethyl azide that is further used in the CuAAC conjugation reaction to prepare the 2-[¹⁸F]fluoroethyl-triazolyl conjugated c(RGDyK) peptide: [¹⁸F]FtRGD. [¹⁸F]FtRGD displayed a good affinity and selectivity for integrins both in vitro and in vivo, and proved valuable for preclinical positron emission tomography imaging of integrin expression. Copyright

Full text of DOI:10.1002/jlcr.3019

Research Article
Received 6 August 2012,
Revised 5 October 2012,
Accepted 6 December 2012
Published online 30 January 2013 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.3019
A practical synthesis of [18F]FtRGD: an
angiogenesis biomarker for PET†
a
a,b
Romain Bejot, Julian Goggi, Shebbrin S. Moonshi,a and
*
Edward G. Robinsa
Integrins have become increasingly attractive targets for molecular imaging of angiogenesis with positron emission
tomography or single-photon emission computed tomography, but the reliable production of radiopharmaceuticals remains
challenging. A strategy for chemoselective labeling of the integrin ligand—c(RGDyK) peptide—has been developed on the
basis of the Cu(I)-catalyzed conjugation reaction. Recently, we reported a nucleophilic detagging and fluorous solid-phase
extraction method providing an easy way to implement an approach for obtaining 2-[18F]fluoroethyl azide. In this work,
we report the practical use of this method for the preparation of the 2-[18F]fluoroethyl-triazolyl conjugated c(RGDyK)
peptide: [18F]FtRGD. The two-step, two-pot synthesis, HPLC purification, and reformulation could be readily performed with
a standard nucleophilic radiofluorination synthesizer (GE TRACERlab FXFN), with minimal modifications. [18F]FtRGD was
obtained in a solution for injection (>500 MBq/mL) in 10–30% nondecay-corrected radiochemical yield, excellent radioche-
mical purity (>98%), and 28 Æ 13 GBq/mmol specific activity. [18F]FtRGD (Ki = 54 Æ 14 nM for aVb3 and 1.7 Æ 0.2 nM for aVb5)
was evaluated in mice and showed good stability in vivo, good tumor-to-background ratio (1.6 Æ 0.3 %ID/g at 1.5 h post-
injection in U87-MG tumors), and rapid urinary excretion. Therefore, [18F]FtRGD proved valuable for preclinical positron
emission tomography imaging of integrin expression.
Keywords: PET; F-18; angiogenesis; integrins; RGD; fluorous; CuAAC
sequence so as to optimize of the pharmacokinetics and the
Introduction
binding affinity and selectivity.
The capability to induce the formation of new blood supply from
Nowadays, most of the radiolabeled ligands are based on the
optimized c(RGDyK) and c(RGDfK) sequences.11 Various radiola-
preexisting vasculature—angiogenesis—is widely accepted as
one of the hallmarks of cancer.1 The development of capillaries
beled derivatives of the cyclic pentapeptides have actually been
is essential for the supply of nutrients and oxygen to growing
reported to selectively bind the endothelial integrins in vivo,
including the 18F-, 68Ga-, and 64Cu-labeled ones for PET and
tumors and metastases. Multiple factors are known to stimulate
endothelial cells and induce angiogenesis.2 All of those growth
factors promote endothelial cell adhesion and migration through
integrin ligation. Integrins are cell surface receptors that bind to
extracellular matrix proteins. Amongst the integrin family, the
aVb3-, aVb5-, and a5b1-glycoproteins are overexpressed on
activated endothelial cells versus quiescent cells and play a key
role in the induction of blood vessel formation. Therefore, the
expression of such endothelial integrins has been shown to
correlate with the angiogenic activity.3,4
Integrins are attractive targets for angiogenesis imaging, and
significant effort has been made toward the development of
ligands for positron emission tomography (PET) and single-
photon emission computed tomography (SPECT) molecular
imaging.5–8 Such biomarkers can prove useful for early detection
of integrin-positive tumors and for monitoring the efficacy of
antiangiogenic therapies. Those integrins share a common
characteristic; the recognition of the amino acid sequence
Arg-Gly-Asp (RGD).3 Therefore, numerous RGD-containing ligands
have been developed as radioactive probes for nuclear imaging
the 99mTc- and 111In-labeled ones for SPECT imaging.12 These
radiopharmaceuticals all exhibit excellent tumor-targeting effi-
cacy, good metabolic stability, and a favorable pharmacokinetic
profile. Noteworthy multimeric peptides, such as [18F]FPP(RGD)2, have
demonstrated a higher binding affinity than monomeric ones.7,12–14
It is also worth noting that a GE Healthcare (Little Chalfont, UK) pro-
prietary compound—fluciclatide, an 18F-labeled polyethylene glycol
conjugate of an RGD-containing cyclic peptide—and a c(RGDfK)
derivative, developed by Siemens Healthcare (Erlangen, Germany)
(18F-RGD-K5), are gaining interest and have entered clinical trials.15–17
As part of our ongoing PET oncology programs with rodents,
we expressed some interest for such molecular imaging probes
aSingapore Bioimaging Consortium (A*STAR), Helios, 02-02, 11 Biopolis Way,
Singapore 138667, Singapore
bDepartment of Physiology, Yong Loo Lin School of Medicine, National University
of Singapore, Singapore 117456, Singapore
*Correspondence to: Romain Bejot, Singapore Bioimaging Consortium (A-Star),
since the structure–activity relationship study with RGD-containing Helios 02-02, 11 Biopolis way, 138667, Singapore, Singapore.
cyclic pentapeptides in 1995 and the first 125I-radiolabeling of
E-mail: romain_bejot@sbic.a-star.edu.sg
c(RGDyV) in 1999.9,10 Many studies have led to an increase of
† Additional Supporting Information may be found in the on-line version of
the hydrophilicity and the conformational restriction of the RGD this article.
J. Label Compd. Radiopharm 2013, 56 42–49
Copyright © 2013 John Wiley & Sons, Ltd.

R. Bejot et al.
that could be used to determine the therapy efficacy of • The radiopharmaceutical shall target angiogenesis in vivo.
antiangiogenic agents in experimental tumor models. However, • The radiopharmaceutical shall display a favorable metabolic
we noticed that most of the reported 18F-radiopharmaceuticals
rely on very specific and proprietary procedures, reactants, and
radiosynthesis systems: for instance, various radiopharmaceuti-
cals production requires custom-made precursors and multistep,
multipot synthesis modules with several HPLC purification steps.
Hence, the procedures do not allow for the quick implementa-
tion of the on-demand preparation radiopharmaceuticals, in a
typical setup equipped with a standard nucleophilic radiofluori-
nation synthesizer (e.g., GE TRACERlab FXFN). Such synthesis
units are available worldwide and allow for the full automation
of the process but are limited to (i) aqueous fluoride processing;
(ii) nucleophilic fluorination and additional reaction steps in the
same reaction vessel; (iii) HPLC or solid-phase extraction (SPE)
purification; (iv) and reformulation. The growing distribution of
68Ge/68Ga generators should allow for the relatively easy
preparation of 68Ga-labeled RGD biomarkers, but the shorter
half-life of Ga-68 (68 vs. 110 min. for F-18) limits the number of
animals to be imaged with a single radiopharmaceutical
preparation.
profile in rodents: fast clearance and optimal tumor contrast
1–2 h post-injection (p.i.).
This manuscript describes the procedure for the reliable and repro-
ducible preparation of 18F-labeled RGD-containing peptide for angio-
genesis imaging in rodents. This procedure relies on the synthesis
and purification of a radiolabeled synthon 2-[18F]fluoroethyl azide,
its subsequent conjugation with the pentynoic derivative of c
(RGDyK), and finally the purification and reformulation of the
radiopharmaceutical. The conjugation reaction was based on the
1,3-dipolar cycloaddition providing the 1,2,3-triazole linker, an
amide bond isostere, not susceptible to cleavage. The radiolabeled
derivative of the cyclic pentapeptide c(RGDyK)—[18F]FtRGD
(Scheme 1)—was further evaluated in vivo: its tumor targeting
efficacy and pharmacokinetics were assessed using PET.
Materials and methods
General
18F-radiolabeling of peptides relies heavily on the radio-
synthesis of a prosthetic group and then its conjugation to
the functionalized peptide. Various conjugation methods have
been employed for the synthesis of radiolabeled RGD derivatives.
For instance, the chemoselective oxime ligation can be employed
between an 18F-labeled aldehyde (e.g., 4-[18F]fluorobenzaldehyde,
All chemicals obtained commercially were of analytical grade and
used without further purification. 2-Azidoethyl 1H,1H,2H,2H-
perfluorodecane-1-sulfonate and 2-fluoroethyl azide were prepared
as described previously.23 Solvents and chemicals were purchased
from Sigma-Aldrich (Singapore) unless stated otherwise. c(RGDyK),
pentynoic c(RGDyK), and c(RGDyK)-Bn-NOTA-Ga were purchased
from FutureChem Co Ltd (Seoul, S. Korea). Axitinib was purchased
from Selleck Chemicals (Houston, TX, USA). The solution tetrabuty-
lammonium bicarbonate solution (TBA•HCO3, 1 M in MeOH) was
prepared by purging CO2 into a solution of tetrabutylammonium
hydroxide (1 M in MeOH, Sigma Aldrich) for 3 h. The solution could
be stored for several months at room temperature without notice-
able decrease of efficiency.
Nonradioactive analytical and semi-preparative reversed-phase
HPLC was performed with a Shimadzu Prominence HPLC system
(Kyoto, Japan) equipped with a diode array detector. Plasma meta-
bolites HPLC analysis was performed with the same system
equipped with two high-pressure flow-line selection valves
(two-positions/six-ports valves). Radioactive analytical HPLC was
performed with a PerkinElmer Series 200 HPLC (Shelton, CT, USA)
system equipped single-wavelength ultraviolet (UV) detector and
a NaI/PMT-radiodetector (Flow-RAM, LabLogic, Sheffield, UK).
Radioactive semi-preparative HPLC was performed with a Sykam
S-1122 pump (Eresing, Germany) and a Knauer K-2001 single-
wavelength UV detector (Berlin, Germany). Following the elution
and collection of [18F]FtRGD, the semi-preparative HPLC column
was washed and stored in 70% ethanol.
[
18F]FDG) and the aminooxy-functionalized peptide.18,19N-
succinimidyl-4-[18F]fluorobenzoate can be coupled with a free
amino group of the peptide to provide the 4-[18F]fluorobenzoyl-
functionalized radiopharmaceutical.20 The Cu(I)-catalyzed (CuAAC)
or strain-promoted Huisgen 1,3-dipolar cycloaddition between an
18F-labeled alkyne and the azide-functionalized peptide or
between an 18F-labeled azide and the alkyne-functionalized
peptide provides an alternative way to access 18F-labeled RGD
peptides.21,22 Those radiolabeled prosthetic groups all require an
HPLC purification step before the conjugation takes place,
resulting in incompatibilities with the general nucleophilic
fluorination units and lengthy preparation procedures (and lower
nondecay-corrected yields).
Hence, we developed a new 18F-labeled radiopharmaceutical
for angiogenesis imaging, satisfying the following criteria:
• The radiosynthesis, purification, and reformulation of the
radiopharmaceutical shall be performed with a standard
nucleophilic radiofluorination synthesizer, with minimal and
reversible modifications.
• The radiosynthesis shall rely on commercially available or easy
to synthesize reactants.
Levels of residual solvents were analyzed using a Varian 430-
GC (Palo Alto, CA, USA) with a flame ionization detector and a
Varian FactourFourTM VF-200 ms column (30 m, 0.32 mm, 1 mm).
• The formulation for in vivo injection shall be compatible with
rodents imaging, that is, >500 MBq/mL (<10% ethanol).
Scheme 1. Radiosynthesis of [18F]FtRGD.
J. Label Compd. Radiopharm 2013, 56 42–49
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

R. Bejot et al.
Mass spectrometry analysis was performed on an Agilent To this solution were added a solution of bathophenanthroline
6224 system (electrospray time-of-flight, Santa Clara, CA, USA), disulfonate (16 mL, 0.1 M) and a solution of sodium ascorbate
with a 1200 capillary HPLC, column Vydac C18 0.5 x 150 mm (12 mL, 1.0 M). The mixture was purged with a gentle stream of
(Grace Davison Discovery Sciences, Deerfield, IL, USA), 20mL/min, helium for 5 min, before adding a solution of copper(II) sulfate
gradient in 5 min from 10% acetonitrile, 0.1% formic acid, water (16 mL, 0.1 M).
to 80% acetonitrile, 0.1% formic acid, water, injection 5mL, column
temperature 45ꢀC). Data are recalibrated using the internal trapping on a Chromafix 30-PS-HCOÀ3 cartridge (Macherey-
standards.
Nagel) and further elution of [18F]fluoride, into the glassy carbon
Isolation of the [18F]fluoride from [18O]H2O was achieved by
The fluorous SPE (FSPE) cartridges were prepared in-house reaction vessel, with a solution of TBA•HCO3 (10 mmol) in 1.0-mL
with 100-mg fluorous silica gel (40 mm, Fluorous Technologies, methanol : water (9:1) (vial 1). Simultaneously, 1.0-mL acetonitrile
Pittsburgh, PA, USA) loaded into an empty polypropylene tube (vial 2) was added to the glassy carbon reaction vessel, and
fitted with a 20-mm polyethylene frit (empty Chromafix 30-PS- anhydrous [18F]fluoride was obtained after azeotropic evapora-
HCO3 tube, Macherey-Nagel, Düren, Germany), washed with 2-mL tion for 7 min at 60ꢀC under a flow of helium and 7 min at
ethanol, 2-mL acetone, and conditioned with 3-mL water.
100ꢀC under vacuum.
A 0.8-mL solution of 2-azidoethyl 1H,1H,2H,2H-perfluorodecane-
1-sulfonate (2mg) in acetonitrile (vial 3) was added to the reaction
vessel, and fluorination was performed at 90ꢀC for 10 min. The
crude solution was cooled down to 40ꢀC, diluted with 0.35-mL
water (vial 4), and eluted through a neutral alumina cartridge
(Sep-Pak light, Waters, Milford, MA, USA) and an FSPE cartridge
connected in series between the valves V14 and V12. The reactor
was washed with 0.35-mL water (vial 5) and eluted through the
same cartridges into the second reaction vessel containing the
peptide and the copper catalytic system. The 1,3-dipolar cycloaddi-
tion was performed in the second reaction vessel for 5 min at 60ꢀC.
Acetonitrile was further evaporated for 5 min at 60ꢀC with a stream
of helium, and the reaction mixture was diluted with water (4.5mL,
0.1% TFA, vial 6).
The resulting solution was purified by HPLC (Phenomenex
Synergy Polar-RP 250 Â 10 mm, 4 mm, 80 Å, guard column; eluent
water : acetonitrile : TFA (850:150:1), 5 mL/min; l = 280 nm). [18F]
FtRGD was collected at 18–19 min into the round-bottomed
dilution flask, containing 40-mL water (0.1% TFA). The resulting
solution was eluted through an hydrophilic–lipophilic balance
(HLB) cartridge (Oasis Plus Light, Waters); the cartridge was then
washed with water (4 mL, vial 9), and [18F]FtRGD was finally
eluted with 50% ethanol (0.4 mL, vial 8) followed by 1.6-mL phos-
phate buffered saline (pH 7.0, vial 7). The radiopharmaceutical
solution was then sterile filtered into a sterile vial through a
0.2-mm mixed cellulose esters membrane (Millex-OR, Millipore,
Bedford, MA, USA).
Chemistry
Preparation of 3-(1-(2-fluoroethyl)-1H-1,2,3-triazol-4-yl)propanoic
c(RGDyK): FtRGD
To pentynoic c(RGDyK) (5mg, FutureChem) in N,N-dimethylforma-
mide (100 mL) were added a 0.1-M aqueous solution of copper sul-
fate (200mL), a 0.1-M buffered sodium ascorbate solution (200 mL,
in 0.1-M potassium carbonate buffer, pH 8.0), and a 0.3-M solu-
tion of 2-fluoroethylazide in THF (200 mL).24 The reaction mixture
was stirred overnight at room temperature, diluted with 1.2-mL
water, and purified by semi-preparative HPLC (Phenomenex Luna
C18(2), 150 x 10mm, 5 mm, 100 Å, Torrance, CA, USA; eluent:
1 mL/min acetonitrile: water : trifluoroacetic acid (TFA): 50:950:1
for 5 min, 50:950:1 to 130:870:1 linear gradient over 5 min,
130:870:1 to 180:820:1 linear gradient over 50 min). The collected
fractions were combined, analyzed using electrospray time of flight
mass spectrometry, and lyophilized to give FtRGD as a white fluffy
powder (est. 3 mg).
1H NMR (600 MHz, D2O) d (ppm) 7.85 (s, 1H), 7.11 (d, J = 8.5 Hz,
2H), 6.82 (d, J = 8.5 Hz, 2H), 4.83 (dt, J1 = 47 Hz, J2 = 4.6 Hz, 2H), 4.73
(dt, J1 = 28Hz, J2 = 4.6 Hz, 2H), 4.56 (dd, J1 = 10.3 Hz, J2 = 6.0Hz, 1H),
4.37 (dd, J1 = 9.1Hz, J2 = 5.7 Hz, 1H), 4.22 (d, J = 14.9 Hz, 1H), 3.81
(dd, J1 = 10.7 Hz, J2 = 4.2Hz, 1H), 3.50 (d, J = 14.9 Hz, 1H), 3.24–3.13
(m, 2H), 3.08–2.98 (m, 3H), 3.03 (t, J = 7.2 Hz, 2H), 2.93
(dd, J1 = 16.7 Hz, J2 = 7.9 Hz, 1H), 2.87 (dd, J1 = 13.3 Hz,
J2 = 10.5 Hz, 1H), 2.75 (dd, J1 = 16.7 Hz, J2 = 6.5 Hz, 1H), 2.61
(t, J = 7.2Hz, 2H), 1.90–1.83 (m, 1H), 1.69–1.57 (m, 2H), 1.55–1.28
(m, 2H), 1.47–1.40 (m, 1H), 1.26–1.20 (m, 2H), 0.88–0.82
(m, 2H).19F NMR (376MHz, D2O) d (ppm) À223.2 (tt, J1 = 47 Hz,
J2 = 28 Hz). HRMS (ESI+) m/z calcd 789.3802; found 789.3802
C34H50FN12O9 [M+ H]+.
Typical yields of [18F]FtRGD were 10–30% (not decay-corrected),
and the total synthesis and reformulation time was 70–75 min.
The radiopharmaceutical identity was confirmed by HPLC
(Phenomenex Gemini NX C18 150 Â 4.6 mm, 5 mm, guard column;
eluent: acetonitrile: water : TFA 100:900:1 to 300:700:1 gradient
over 10 min, 1 mL/min) and indicated >98% radiochemical purity.
The specific activity was 28 Æ 13 GBq/mmol at end of synthesis
[18F]FtRGD
No-carrier-added [18F]fluoride was produced by the 18O(p,n)18F (EOS) (n = 11). Contaminants were less than 5 mg/mL, estimated
nuclear reaction with a 16-MeV proton beam generated by a GE from their absorbance at 276nm and the molar extinction
PETtrace cyclotron in a silver target using [18O]H2O (Singapore coefficient of pentynoic c(RGDyK). The final product was a clear
Radiopharmaceuticals Pte Ltd). A TRACERlab FXFN radiochemistry and colorless solution (1.5Æ 0.8 GBq/mL at EOS) at pH 6.5–7.5.
module (GE Healthcare) was used for the 2-[18F]fluoroethyl-
Solid-phase receptor binding assay
triazole-c(RGDyK) two-step synthesis and purification. A 5-mL V vial
(Wheaton, Millville, NJ, USA) was used as a second reaction vessel The saturation binding assay was performed as described
in place of the HPLC loading vessel, and a remotely controlled previously.25 Purified aVb3- and aVb5-integrins (Millipore)
Elite dry bath incubator fitted with a 20-mm-well heating block were diluted at 0.1 mg/mL in coating buffer (20-mM Tris pH 7.4,
(Major Science, Saratoga, CA, USA) was used as the second reaction 150-mM NaCl, 2-mM CaCl2, 1-mM MgCl2, 1-mM MnCl2). A 100 ml/well
vessel heater.
was added to 96-well plates (high binding polystyrene, Corning
Prior to the radiopharmaceutical synthesis, in the 5-mL V vial, Stripwell) and incubated overnight at 4ꢀC. The coating solution
pentynoic c(RGDyK) (5 mg) was dissolved in a solution of gentisic was removed, and 200 mL of blocking/binding solution (50-mM Tris
acid (4 mg) in potassium carbonate buffer (100 mL, pH 8.0, 1.0 M). pH 7.4, 100-mM NaCl, 2-mM CaCl2, 1-mM MgCl2, 1-mM MnCl2, 1%
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 42–49

R. Bejot et al.
bovine serum albumin) was added to the wells and incubated for an In vivo imaging studies
additional 2 h at room temperature. The plates were rinsed with
The U87-MG tumor-bearing mice (n = 4 per group) were imaged
200 mL of blocking/binding solution and incubated for 3 h at
room temperature with increasing concentration (30 pM–2nM)
125I-echistatin (81.4 GBq/mmol, PerkinElmer, Boston, MA, USA) in
blocking/binding solution. After incubation for 3 h at room tempera-
ture, the plates were washed twice with blocking/binding buffer. The
wells were separated and counted with an automatic gamma
counter (2470 Wizard2, Perkin Elmer). Bound radioactivity was
plotted versus the radioligand concentration. The equilibrium disso-
ciation constant Kd and the density of receptors Bmax were
obtained by nonlinear regression analysis of the total binding,
accounting for ligand depletion (Prism 5.0c, GraphPad Software,
La Jolla, CA, USA).26
The competition-binding assay was performed as described
previously.25,27 Purified aVb3- and aVb5-integrins (Millipore) were
diluted at 0.5 and 1.0 mg/mL, respectively, in coating buffer. A
100 mL/well was added to 96-well plates (high binding polystyr-
ene, Corning Stripwell, Corning, NY, USA) and incubated over-
night at 4ꢀC. The coating solution was removed, and 200 mL of
blocking/binding solution was added to the wells and incubated
for an additional 2 h at room temperature. The plates were
rinsed with 200 mL of blocking/binding solution and incubated
for 3 h at room temperature with 0.05-nM 125I-echistatin
(PerkinElmer) and increasing concentrations (1 pM–10 mM) of
the c(RGDyK) derivatives in blocking/binding solution. After
incubation for 3 h at room temperature, the plates were
washed twice with blocking/binding buffer. The wells were
separated and counted with an automatic gamma counter
(2470 Wizard2, Perkin Elmer). Bound radioactivity was plotted
versus the logarithm of the competitor concentration. The
inhibition constant Ki was determined by nonlinear regression
analysis with hillslope, accounting for ligand depletion (Prism
5.0c, GraphPad Software).
on days 0 and 10 with [18F]FtRGD. The mice were injected with a
solution of [18F]FtRGD (~10 MBq in 0.2 mL) via the lateral tail vein,
and the animals imaged under isoflurane anesthesia (2% alveolar
concentration) and biological monitoring for respiration, and tem-
perature was performed using a BioVet system (m2m imaging, Cle-
veland, OH). Small-animal PET imaging was performed from 70 to
90min p.i., on an Inveon PET/CT system (Siemens Inc., Washington
DC). Images were generated from sinogram data, rebinned to two-
dimensional format by the Fourier rebinning algorithm, followed
by two-dimensional filtered back projection. Low-dose CT images
(40 kV, 500 mA; 4 Â 4 binning, 200-mm resolution) were acquired
for anatomical information. Images were reconstructed using the
image reconstruction, visualization, and analysis program supplied
by the manufacturer. PET and CT data were analyzed by using the
Amide software (Sourceforge 10.1, http://amide.sourceforge.net).
The PET and CT images were coregistered, to confirm anatomical
location of the tumors. Uptake of radioactivity in the tumor was
determined by placement of a region of interest around the tumor
border delineated using the CT images. The tissue concentrations
were measured using region of interest analysis and are presented
as percent injected dose/gram (%ID/g) for tumor, muscle (lower
left hind limb), liver, kidneys, and heart and brain.
Plasma metabolites analysis
After intravenous injection of the radiopharmaceutical into
tumor-bearing mice, blood samples (50–200 mL) from the tail
vein were manually collected into 0.5-mL heparinized polypropy-
lene centrifuge tubes, at 2, 30, 60, and 90 min p.i.. Blood samples
were centrifuged at 2500 g for 10 min. The supernatant plasma
samples were diluted with 1-mL water (0.1% TFA) and analyzed
using on-line HLB SPE and reversed-phase HPLC, as described
by Chitneni et al.28 The diluted plasma samples were (i) while
bypassing the HPLC column: injected onto a preconditioned
OasisW HLB column (20 Â 3.9 mm, 5 mm, Waters) and rinsed with
water for 3 min; (ii) with the HLB column connected to the HPLC
column: eluted with the mobile phase (acetonitrile : water : TFA
0:1000:1 to 300:700:1 gradient over 10 min, 1 mL/min). All eluates
were analyzed with an online NaI/PMT radioactivity detector (IN/
US).
Animals
Outbred male nude CD1 mice were purchased from the Biological
Resources Centre (Singapore). Animal experiments were per-
formed in accordance with the Institutional Animal Care and Use
Committee guidelines under ethics number IACUC 090437.
U87-MG cancer cells
Results and discussion
Human U87-MG glioma cancer cells were purchased from the
American Type Culture Collection (ATCC). Cells were cultured in
Dulbecco’s modified Eagle medium (Invitrogen, Carlsbad, CA,
USA), supplemented with 10% fetal bovine serum, 100 U/mL peni-
cillin, and 100 mg/mL streptomycin, at 37ꢀC in a humidified atmo-
sphere, with 5% of CO2.
Radiochemistry
We recently developed an FSPE technique for the separation of
the radiolabeled synthon from its precursor: the fluorous tagged
precursor is detagged upon nucleophilic 18F-fluorination, and on
the basis of the unique separation properties of highly polyfluori-
nated molecules (fluorous), the resulting 18F-labeled synthon is
separated by simple elution although a fluorous stationary
phase.23,24 This technique allowed us to obtain 18F-labeled
Treatment studies
At 3-week post-inoculation of U87-MG cells, the tumors reached synthons reliably and with a minimum of purification effort. For
a volume of 200–250 mm3, and the mice were randomly divided instance, 2-[18F]fluoroethyl azide was synthesized by nucleophilic
into two treatment groups. One group received daily administra- 18F-fluorination of the polyfluorinated sulfonate precursor and
tions of Axitinib (25 mg/kg) by intraperitoneal injection, whereas purified rapidly by simple elution through alumina and fluorous
the second group received administrations of the polyethylene silica gel: alumina traps excess [18F]fluoride, whereas fluorous silica
glycol/H2O (30:70) vehicle alone. Animals were imaged on the gel traps excess precursor and polyfluorinated by-products.
day prior to initiation of the dosing regimen (day 0), and the
same animals were imaged after 10 days of therapy (day 10).
It is worth noting that tetrabutylammonium was used as a
phase-transfer catalyst for the nucleophilic fluorination.29,30 It
J. Label Compd. Radiopharm 2013, 56 42–49
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

R. Bejot et al.
proved superior to the classical mixture of potassium carbonate acid described by Gill et al.37 Excess ascorbate was used to
and KryptofixW 2.2.2 (K222) cryptand, providing less by-products reduce Cu(II) to Cu(I), and the addition of an equimolar amount
and the expected 18F-labeled synthon with a moderate specific of the bathophenantroline disulfonate ligand stabilized the
activity. It should actually be noted that the fluorous purification oxidation state and promotes the solubilization of Cu(I). The
technique usually suffers from a decrease of the specific activity reaction mixture improved both the conjugation reaction
possibly from leaching of [19F]fluoride ion from the precursor (increased efficiency, reduced amount of peptide, and reduced
tag.31 Solvent evaporation and [18F]fluoride drying could then reaction time) and the purification step (better solubility and less
be performed with the addition of acetonitrile forming the tern- clogging). As recommended by Glaser et al., the reaction was
ary azeotrope methanol–acetonitrile–water. The resulting tetra- performed at pH 8.0, in a buffered carbonate solution.18 It is
butylammonium [18F]fluoride was then used to provide 2-[18F] noteworthy that the pentynoic derivative of c(RGDyK) is water
fluoroethyl azide with high radiochemical yield and higher speci- soluble; hence, the use of dimethylformamide or dimethyl sulf-
fic activity. Furthermore, the reaction mixture after nucleophilic oxide was avoided.
fluorination was less troublesome: the obtained solution was
Reaction monitoring indicated that the reaction was almost
colorless and fully soluble, whereas brown-colored solution with complete (>90% conversion) in 5 min at 60ꢀC. Higher tem-
insoluble residues could be obtained with K222
.
perature was not investigated because of the volatility of the
The Cu(I)-catalyzed Huisgen 1,3-dipolar cycloaddition between 18F-labeled synthon. Prior to HPLC purification, acetonitrile is
an alkyne and an azide is being used increasingly in drug discovery evaporated with a gentle stream of helium at 60ꢀC, and the
and radiopharmaceutical development.32–34 The efficient, water- diluted crude reaction mixture is purified by semi-preparative
compatible and orthogonal reaction is actually one of the HPLC. The collected fraction is reformulated into a volume as low
main technique for the preparation of RGD-containing peptide as possible: using an HLB SPE cartridge, the radiopharmaceutical
conjugates.35
is separated from the diluted HPLC eluate and further eluted
Once purified, 2-[18F]fluoroethyl azide could be conjugated into the final vial with only 0.4mL 50% aqueous ethanol. It is
to the alkyne-functionalized peptide, such as the pentynoic noteworthy that only 80–90% of the radiopharmaceutical was
derivative of c(RGDyK), commercially available from FutureChem. eluted from the HLB cartridge, but attempts for a higher recovery
The cyclic pentapepeptide c(RGDyK) was chosen over its popular rate led to excessive dilution of the final solution. The solution is
analogue c(RGDfK) because of its increase hydrophilicity: the further made isotonic and buffered with the addition of 1.6-mL
former was actually reported with lower liver uptake.36 Dimeric saline and phosphate buffer solution (pH 7.0, 10 mM). A 0.5mL
or tetrameric ligands based on the pentapeptides could also be was used for quality control, affording 1.5-mL solution for
used to improve the binding affinity of the radiopharmaceutical; injection (10% ethanol).
however, their commercial availability is still limited.
The total synthesis, purification, and formulation time for 2-
The Cu(I)-catalyzed Huisgen 1,3-dipolar cycloaddition was [18F]fluoroethyl-triazole c(RGDyK) was only 70–75 min. The
performed following an optimized procedure based on those overall nondecay-corrected radiochemical yield was 10–30%
reported in the literature. In details, the copper(I)-catalyzed (n = 11), providing 1.5 Æ 0.8 GBq/mL radiopharmaceutical. The
reaction was performed using the mixture of copper sulfate, specific activity was moderate: 28 Æ 13 GBq/mmol at the EOS.
bathophenantroline disulfonate, sodium ascorbate, and gentisic As mentioned previously, the moderate specific activity was
due to the leaching of fluorine from the fluorous tag of the
precursor. Interestingly, the specific activity obtained with
TBA•HCO3 was one order of magnitude higher than that obtained
Table 1. Dissociation constants of c(RGDyK) derivatives
with K222/K2CO3. Finally, it is worth mentioning that the final
radiopharmaceutical solution contained less than 5mg/mL of UV-
absorbing contaminant. Residual organic solvents in the final
product were analyzed using gas chromatography, and the high-
est concentrations measured were well below the specifications
for radiopharmaceuticals (acetone: 55 < 5000 ppm; acetonitrile:
6 < 410 ppm).
Ki (aVb3) (nM)
Ki (aVb5) (nM)
c(RGDyK)
Ga-NOTA-Bn-c(RGDyK)
FtRGD
46 Æ 11
69 Æ 19
54 Æ 14
3.7 Æ 0.4
5.4 Æ 0.5
1.7 Æ 0.2
Figure 1. Competition binding assay against [125I]echistatin with aVb3 (left) and aVb5 (right).
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 42–49

R. Bejot et al.
In vitro validation
derivative. As expected, the fluoroethyl-triazole scaffold has
minimal effect on the binding affinity of the cyclic pentapeptide.
The competitor concentrations required for half-maximal bind-
ing of to aVb3- and aVb5-integrins were determined in solid-
phase receptor binding assays.25,27 The dissociation constant
In vivo study
(Ki) of the inhibitors was further determined from the binding The biology studies described here were performed to investigate
affinity (Kd) of 125I-echistatin (Table 1, Figure 1). For comparison, the performance of 18F-labeled peptide for imaging angiogenesis
the competition assay was performed with the fluoroethyl conju- in an U87-MG xenograft model in vivo using PET. Previously, other
gate of c(RGDyK)—FtRGD—with the reference peptide—c 18F-labeled peptides have been used successfully to image
(RGDyK)—and the gallium complex derivative—Ga-NOTA-Bn-c angiogenesis and early response to antiangiogenic therapy using
(RGDyK).38 We could observe that the binding affinity of our fluori- the U87-MG model.39 U87-MG cells form solid tumors when
nated conjugate is similar to that of the c(RGDyK) and its Ga-NOTA- implanted subcutaneously in nude mice and have been shown
Figure 2. Time-activity curves of [18F]FtRGD (%ID/g Æ standard error of the mean) in the tumor, muscle, brain, liver, kidneys, and the heart, quantified using region of
interest analysis of positron emission tomography images (left) and coronal image of a mice at 70–90 min post-injection (right).
Figure 3. Tumor uptake of [18F]FtRGD following intravenous bolus injection. Images show static acquisition data from 70–90 min post-injection of [18F]FtRGD indicating
no change in tumor uptake in the Axitinib-treated U87-MG tumors, but a significant increase in tumor retention in the vehicle-treated tumors is observed. This figure is
available in colour online at www.interscience.wiley.com/journal/jlcr
J. Label Compd. Radiopharm 2013, 56 42–49
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

R. Bejot et al.
to express aVb3 and aVb5 on both the tumor cells and vascula-
ture.40 The acquired [18F]FtRGD PET data demonstrate uptake into
untreated U87-MG tumors of 1.6Æ 0.3 %ID/g and skeletal muscle
Acknowledgements
We gratefully acknowledge the assistance of Dr. M. Raida and
of 0.24 Æ 0.07 %ID/g at ~70–90 min p.i., displaying tumor uptake Ms. K. Ee (ETC, Mass Spectroscopy group).
values consistent with previously published 18F-labeled angiogen-
esis imaging peptides such as [18F]fluciclitide (~1.5Æ 0.5 %ID/g)
Conflict of Interest
and [18F]FPP(RGD)2 (~1.85 Æ 0.3 %ID/g).39,41 The excretion profile
of [18F]FtRGD was also similar to previously published results with
The authors did not report any conflict of interest.
other RGD-containing radiopharmaceuticals, displaying high initial
kidney accumulation and rapid urinary excretion (Figure 2). Liver
uptake was relatively low with some hepatobiliary excretion;
References
muscle retention was also low consistent with the low lipophilicity
[1] D. Hanahan, R. A. Weinberg, Cell 2011, 144, 646.
[2] P. Somanath, A. Ciocea, T. Byzova, Cell Biochem. Biophys. 2009,
53, 53.
of the RGD peptides. The tumor-to-muscle ratio was sufficiently
high to be able to determine therapy efficacy using angiogenesis
inhibitors such as Axitinib (Pfizer). Axitinib is an orally bioavailable,
receptor tyrosine kinase inhibitor that exhibits potent antiangio-
genic and antitumor activity at VEGFR-1, VEGFR-2, VEGFR-3, cKit,
and PDGFR.42 In this study, Axitinib was well tolerated, with no
significant effects on animal health observed. Previous data have
shown that Axitinib displays treatment-related tumor growth
inhibition within 10 days using calipers; therefore, animals were
imaged the day prior to initiation of the dosing regimen and after
10days.42 As can be seen in Figure 3, the tumor retention of [18F]
FtRGD stayed constant from day 0 (1.62 Æ 0.32 %ID/g) in the
Axitinib treated group reaching 1.65 Æ 0.25 %ID/g (p > 0.89) at
[3] K. Hodivala-Dilke, A. Reynolds, L. Reynolds, Cell Tissue Res. 2003,
314, 131.
[4] N. Laurens, M. Engelse, C. Jungerius, C. Löwik, V. van Hinsbergh,
P. Koolwijk, Angiogenesis 2009, 12, 275.
[5] M. R. Stacy, M. W. Maxfield, A. J. Sinusas, Yale J. Biol. Med. 2012,
85, 75.
[6] R. Haubner, A. Beer, H. Wang, X. Chen, Eur. J. Nucl. Med. Mol. Imaging
2010, 37, 86.
[7] F. Gaertner, H. Kessler, H. J. Wester, M. Schwaiger, A. Beer, Eur. J. Nucl.
Med. Mol. Imaging 2012, 39, 126.
[8] Y. Zhou, S. Chakraborty, S. Liu, Theranostics 2011, 1, 58.
[9] R. Haubner, R. Gratias, B. Diefenbach, S. L. Goodman, A. Jonczyk,
H. Kessler, J. Am. Chem. Soc. 1996, 118, 7461.
day 10 posttreatment. However, the tumor retention of [18F]FtRGD [10] R. Haubner, H.-J. Wester, U. Reuning, R. Senekowitsch-Schmidtke,
B. Diefenbach, H. Kessler, G. Stöcklin, M. Schwaiger, J. Nucl. Med.
significantly increased in the vehicle treated group from day 0
1999, 40, 1061.
(1.49 Æ 0.25 %ID/g) to day 10 (3.33 Æ 0.63 %ID/g (p < 0.001)).
[11] Molecular Imaging and Contrast Agent Database (MICAD) [database
online]. Bethesda (MD): National Library of Medicine (US), NCBI;
2004-2012. Available from: http://micad.nih.gov
[12] S. Liu, Bioconj. Chem. 2009, 20, 2199.
Plasma metabolites analysis
[13] W. Cai, S. S. Gambhir, X. Chen, Biotechniques 2005.
[14] F. Chin, B. Shen, S. Liu, R. Berganos, E. Chang, E. Mittra, X. Chen,
S. Gambhir, Mol. Imaging Biol. 2012, 14, 88.
[15] B. Indrevoll, G. M. Kindberg, M. Solbakken, E. Bjurgert, J. H. Johansen,
H. Karlsen, M. Mendizabal, A. Cuthbertson, Bioorg. Med. Chem. Lett.
2006, 16, 6190.
[16] L. M. Kenny, R. C. Coombes, I. Oulie, K. B. Contractor, M. Miller,
T. J. Spinks, B. McParland, P. S. Cohen, A.-M. Hui, C. Palmieri,
S. Osman, M. Glaser, D. Turton, A. Al-Nahhas, E. O. Aboagye, J. Nucl.
Med. 2008, 49, 879.
The time-course stability of the 18F-labeled peptide was examined
by reversed-phase HPLC. In detail, plasma samples, at 2, 30, 60, and
90min p.i., were analyzed by HPLC using the HLB SPE technique
described by Chitneni et al.28 No metabolites could be detected
in the plasma: the only radioactive signal that could be observed
was from the parent peptide: [18F]FtRGD.
[17] M. Doss, H. C. Kolb, J. J. Zhang, M. J. Bélanger, J. B. Stubbs, M. G.
Stabin, E. D. Hostetler, R. K. Alpaugh, M. von Mehren, J. C. Walsh,
M. Haka, V. P. Mocharla, J. Q. Yu, J. Nucl. Med. 2012, 53, 787.
[18] M. Glaser, M. Solbakken, D. Turton, R. Pettitt, J. Barnett, J. Arukwe,
H. Karlsen, A. Cuthbertson, S. Luthra, E. Ärstad, Amino Acids 2009,
37, 717.
[19] M. Namavari, Z. Cheng, R. Zhang, A. De, J. Levi, J. K. Hoerner,
S. S. Yaghoubi, F. A. Syud, S. S. Gambhir, Bioconj. Chem. 2009, 20, 432.
[20] Z. Wu, Z. B. Li, W. Cai, L. He, F. Chin, F. Li, X. Chen, Eur. J. Nucl. Med.
Mol. Imaging 2007, 34, 1823.
Conclusion
A novel 18F-labeled c(RGDyK) peptide derivative ([18F]FtRGD) was
developed on the basis of the fluorous methodology we
reported recently. The approach allowed for a quick and easy
purification of radiolabeled synthons by simple elution through
alumina and fluorous silica gel, and could be readily implemen-
ted with a standard nucleophilic radiosynthesizer to perform a
two-step, two-pot process: that is, nucleophilic 18F-fluorination
and CuAAC conjugation. The procedure including the radiolabel-
ing of 2-[18F]fluoroethyl azide, the FSPE purification, the CuAAC
conjugation with the alkyne-functionalized c(RGDyK) peptide,
the HPLC purification, and the reformulation of the radiopharma-
ceuticals provided a reliable method to obtain [18F]FtRGD for
preclinical PET. [18F]FtRGD could be obtained in less than 75 min
and in good radiochemical yield. The radiochemical purity, specific
activity, and reformulation solution proved suitable for preclinical
imaging. The evaluation of [18F]FtRGD in mice confirmed the
affinity of the radiopharmaceutical for integrins and exhibited a
favorable metabolic profile (high stability and fast renal clearance)
for angiogenesis PET imaging with 1–2 h p.i.. Further studies are
ongoing using [18F]FtRGD as a tool for preclinical PET monitoring
of antiangiogenic therapies.
[21] Z.-B. Li, Z. Wu, K. Chen, F. T. Chin, X. Chen, Bioconj. Chem. 2007, 18,
1987.
[22] H. L. Evans, R. L. Slade, L. Carroll, G. Smith, Q.-D. Nguyen, L. Iddon, N.
Kamaly, H. Stockmann, F. J. Leeper, E. O. Aboagye, A. C. Spivey,
Chem. Commun. 2012, 48, 991.
[23] R. Bejot, T. Fowler, L. Carroll, S. Boldon, J. E. Moore, J. Declerck, V.
Gouverneur, Angew. Chem. Int. Ed. 2009, 48, 586.
[24] R. Bejot, L. Carroll, K. Bhakoo, J. Declerck, V. Gouverneur, Bioorg. Med.
Chem. 2012, 20, 324.
[25] C. C. Kumar, H.-N. Rogers, Christine Prorock, M. Malkowski, E.,
Maxwell, J. J. Catino, L. Armstrong, J. Pharmacol. Exp. Ther.
1997, 283, 843.
[26] S. Swillens, Mol. Pharmacol. 1995, 47, 1197.
[27] A. Trabocchi, G. Menchi, E. Danieli, D. Potenza, N. Cini, A. Bottoncetti,
S. Raspanti, A. Pupi, A. Guarna, Amino Acids 2010, 38, 329.
[28] S. K. Chitneni, K. Serdons, N. Evens, H. Fonge, S. Celen, C. M. Deroose,
Z. Debyser, L. Mortelmans, A. M. Verbruggen, G. M. Bormans, J.
Chromatogr. A 2008, 1189, 323.
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 42–49

R. Bejot et al.
[29] S. J. Lee, S. J. Oh, W. Y. Moon, M. S. Choi, J. S. Kim, D. Y. Chi, D. H.
Moon, J. S. Ryu, Nucl. Med. Biol. 2011, 38, 593.
[30] S. J. Lee, S. J. Oh, D. Y. Chi, S. H. Kang, H. S. Kil, J. S. Kim, D. H. Moon,
Nucl. Med. Biol. 2007, 34, 345.
[38] A. N. Singh, W. Liu, G. Hao, A. Kumar, A. Gupta, O. K. Öz, J.-T. Hsieh,
X. Sun, Bioconj. Chem. 2011, 22, 1650.
[39] M. R. Battle, J. L. Goggi, L. Allen, J. Barnett, M. S. Morrison, J. Nucl.
Med. 2011, 52, 424.
[31] P. S. Johnström, S. Stone-Elander, Appl. Radiat. Isot. 1996, 47, 401.
[32] H. C. Kolb, J. C. Walsh, Chimia 2010, 64, 29.
[33] H. C. Kolb, K. B. Sharpless, Drug Discov. Today 2003, 8, 1128.
[34] J. E. Moses, A. D. Moorhouse, Chem. Soc. Rev. 2007, 36, 1249.
[35] M. Colombo, A. Bianchi, Molecules 2010, 15, 178.
[36] X. Chen, S. Liu, Y. Hou, M. Tohme, R. Park, J. R. Bading, P. S. Conti, Mol.
Imaging Biol. 2004, 6, 350.
[40] S. de Boüard, P. Herlin, J. G. Christensen, E. Lemoisson, P. Gauduchon,
E. Raymond, J.-S. Guillamo, Neuro Oncol. 2007, 9, 412.
[41] L. Lang, W. Li, N. Guo, Y. Ma, L. Zhu, D. O. Kiesewetter, B. Shen, G. Niu,
X. Chen, Bioconj. Chem. 2011, 22, 2415.
[42] L. J. Wilmes, M. G. Pallavicini, L. M. Fleming, J. Gibbs, D. Wang, K. L. Li,
S. C. Partridge, R. G. Henry, D. R. Shalinsky, D. Hu-Lowe, J. W. Park,
T. M. McShane, Y. Lu, R. C. Brasch, N. M. Hylton, J. Magn. Reson.
Imaging 2007, 25, 319.
[37] H. S. Gill, J. Marik, Nat. Protocols 2011, 6, 1718.
J. Label Compd. Radiopharm 2013, 56 42–49
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org