Scavenging strategy for specific activity improvement: Application to a new CXCR4-specific cyclopentapeptide positron emission tomography tracer

Article abstract of DOI:10.1002/jlcr.3095

Huisgen cycloaddition is attractive to label peptide because of its rapidity and bioorthogonality. However, for larger tracers, the physico-chemical differences between the precursor and the tracer are usually insufficient to allow their separation by HPLC, reducing the specific activity. This is of importance for peptidic tracers because the combination of their high-affinity receptor with low specific activity results in the precursor saturating the receptors, causing non-specific tracer binding. Here, we report a fast, one-pot, general strategy to circumvent this issue, yielding a tracer of improved specific activity. It consists in adding a lipophilic azide after the labeling step to scavenge unreacted precursor into a more lipophilic species that does not co-elute with the tracer. We applied this strategy to a new fluorinated cyclopentapeptidic CXCR4 antagonist for the PET imaging of cancer, CCIC15, for which we managed to reduce the apparent peptide concentration by a factor of 34 in 10 min. This tracer was radiolabeled by click chemistry with 2-[ ¹⁸F]fluoroethylazide, yielding the tracer in 18 ± 6% (n = 5) end-of-synthesis radiochemical yields (EOS-RCY) in 1.5 h from [ ¹⁸F]fluoride with a specific activity of 19.4 GBq μmol ^-1. Preliminary biological evaluation of the probe confirmed potency and specificity for CXCR4; further biological evaluation is underway. Copyright

Full text of DOI:10.1002/jlcr.3095

Research Article
Received 17 April 2013,
Revised 29 May 2013,
Accepted 18 June 2013
Published online 16 July 2013 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.3095
Scavenging strategy for specific activity
improvement: application to a new CXCR4-
specific cyclopentapeptide positron emission
tomography tracer†
a
a
Guillaume P. C. George,a,b Federica Pisaneschi, Elizabeth Stevens,
*
Quang-Dé Nguyen,a Ola Åberg,a,c Alan C. Spivey,b and Eric O. Aboagyea
Huisgen cycloaddition is attractive to label peptide because of its rapidity and bioorthogonality. However, for larger tracers,
the physico-chemical differences between the precursor and the tracer are usually insufficient to allow their separation by
HPLC, reducing the specific activity. This is of importance for peptidic tracers because the combination of their
high-affinity receptor with low specific activity results in the precursor saturating the receptors, causing non-specific tracer
binding. Here, we report a fast, one-pot, general strategy to circumvent this issue, yielding a tracer of improved specific
activity. It consists in adding a lipophilic azide after the labeling step to scavenge unreacted precursor into a more lipophilic
species that does not co-elute with the tracer. We applied this strategy to a new fluorinated cyclopentapeptidic CXCR4
antagonist for the PET imaging of cancer, CCIC15, for which we managed to reduce the apparent peptide concentration
by a factor of 34 in 10 min. This tracer was radiolabeled by click chemistry with 2-[18F]fluoroethylazide, yielding the tracer
in 18 6% (n = 5) end-of-synthesis radiochemical yields (EOS-RCY) in 1.5 h from [18F]fluoride with a specific activity of
19.4 GBq μmolÀ1. Preliminary biological evaluation of the probe confirmed potency and specificity for CXCR4; further
biological evaluation is underway.
Keywords: scavenging; specific activity; CXCR4; cancer; FC131
In general, peptides tend to show favorable properties for
Introduction
cancer imaging, such as their rapid pharmacokinetics, their specific
tumor-targeting properties, and their efficient tumor penetration.7
The chemokine receptor CXCR4 is a heptahelical transmembrane
G protein-coupled receptor. With its native ligand SDF-1 (stromal
Furthermore, the precursors for the labeling step are also easily
cell-derived factor-1), it is a key part of a signal transduction
available via solid-phase peptide synthesis, which can even be
pathway that is exploited for instance during embryo
automated and therefore allow a molecule of high binding
development and inflammatory responses.1 Over-expression of
this receptor is associated with various diseases, including
particularly aggressive types of cancer and/or metastatic tumors.
As a consequence, CXCR4 appears to be a particularly attractive
affinities to be engineered without much effort. However, a low
IC50 value increases the tracer’s vulnerability to its specific activity:
unlabeled species, for example, the unreacted precursor, are likely
to saturate the receptors causing non-specific binding of the tracer
and thus reducing the signal-to-noise ratio.8
marker to specifically image invasive and metastatic phenotypes.
Most of the CXCR4 inhibitors reported in the literature are
based on peptide (T140, FC131) or cyclam (AMD3100) scaffolds,2
although the number and classes of inhibitors are expected to
increase following the recent disclosure of the receptor’s crystal
structure.3 A selection of these inhibitors has been derivatized
to allow CXCR4-specific labeling for different imaging modalities,
mostly fluorescence, PET, and single-photon emission computed
tomography.4
aComprehensive Cancer Imaging Centre, Faculty of Medicine, Hammersmith
Hospital Campus, Imperial College London, Du Cane Road, London W12 0NN,
United Kingdom
bDepartment of Chemistry, Imperial College London, South Kensington Campus,
London SW7 2AZ, United Kingdom
cDepartment of Medicinal Chemistry, Preclinical PET Platform, Uppsala
University, SE-752 23 Uppsala, Sweden
Positron emission tomography is a non-invasive and highly
sensitive imaging technique that requires the incorporation of
a
positron-emitting radioisotope into administered probe
*Correspondence to: Guillaume P. C. George, Comprehensive Cancer Imaging
Centre, Faculty of Medicine, Hammersmith Hospital Campus, Imperial College
London, Du Cane Road, London W12 0NN, United Kingdom.
E-mail: g.george09@imperial.ac.uk
molecules.5 Fluorine-18 is an attractive radionuclide for PET as
its 110-min half-life allows multistep radiolabeling strategies
and enables the transfer of radiotracers to hospitals without an
on-site cyclotron.6
†This supporting information can be found in the electronic version of the article.
J. Label Compd. Radiopharm 2013, 56 679–685
Copyright © 2013 John Wiley & Sons, Ltd.

G. P. C. George et al.
1
In labeling experiments of low-molecular-weight compounds spectrometer. Only selected absorbances are reported. H, 13C,
utilizing a fluorine-18-labeled prosthetic group, a large excess and 19F nuclear magnetic resonance (NMR) spectra were
of precursor is often used to increase the reaction rate in order obtained at 300 K on Bruker (Billerica, MA, U.S.A.) AV-400, DRX-
to be compatible with the radionuclide’s short half-life. The 400, or AV-500 instruments. Chemical shifts (δ) are given in parts
starting material and the product have distinctive physico- per million (ppm) as referenced to the appropriate residual
chemical properties (e.g., charge and lipophilicity) permitting solvent peak (for Me2SO-d6 1H 2.50 ppm and 13C 39.5 ppm, for
the isolation by reverse-phase HPLC (RP-HPLC) of the radioactive MeOH-d4 1H 3.31 ppm and 13C 49.0 ppm) or to fluorobenzene
product with high specific activity. In the case of high-molecular- (19F À113.5 ppm). The 13C chemical shifts are assigned as s, d, t,
weight precursors, such as peptides, differences in physico- and q for C, CH, CH2, and CH3, respectively. Coupling constant
chemical properties can be insufficient to prevail over the (J) are given in Hertz (Hz). Mass spectra were obtained in positive
unmodified parts of the molecule, leading to the precursor and ESI mode on a Micromass LCT Premier, Atlantis, 600 pump,
the product to co-elute and resulting in low apparent specific Empower Pro software etc.: Waters (Elstree, U.K.) equipped with
activities.9
a Waters Atlantis C18 3 μ column 2.1 × 30 mm. Mobile phase (A)
For instance, the use of the 2-[18F]fluoroethylazide ([18F]FEA) water (0.1% formic acid) and (B) acetonitrile. High resolution mass
prosthetic group to label an alkyne-containing peptide by click spectrometry values are valid up to 5 ppm. A Mettler Toledo
chemistry is attractive because of the bioorthogonality of this (Leicester, U.K.) pH meter was used to prepare the acetate buffer.
reaction. Nevertheless, the relatively high amounts of precursor The [18F]fluoride was acquired from PETNet Solutions
needed for the reaction to occur reduce the apparent specific (Manchester, UK). Conventional heating was carried out with a
activity of the tracer thus produced, limiting the use of such heating block, and the temperature was measured in the heating
methods for the production of radiotracers.
block by using a digital thermometer. Preparative ultraviolet
In this study, we report the design and synthesis of an FC131 HPLC (UV-HPLC) was carried out using a Waters 600 pump and
analogue as a potential CXCR4-specific tracer. Modifications EMPOWER PRO software (Waters) equipped with a linear Water
to FC131 were designed following the structure-activity 2487 UV detector (λ = 254 and 270 nm) and a Phenomenex
relationship analysis of the structures reported in the literature (Macclesfield, U.K.) Luna C18(2) 5μ 75 × 30.00 mm HPLC column.
as described in the succeeding texts. We also present the Injection loop 5000 μL. Mobile phase (A) water [0.1% trifluoroacetic
development of a method, unprecedented to the best of our acid (TFA)] and (B) MeCN (0.1% TFA). Gradient 5–95% B over
knowledge, that allows scavenging of the unreacted precursor, 10 min. Flow rate 20 mLminÀ1. Semi-preparative UV-radio-HPLC
thus affording the tracer in much higher apparent specific was carried out using a Beckman Pump 127 (Beckman Coulter, High
activity.
Wycombe, U.K.) and Laura 3 software (Lablogic) equipped with a
FC131 is a cyclo(-Gly1-D-Tyr2-Arg3-Arg4-Nal5-) pentapeptide, linear Hewlett Packard (Wokingham, U.K.) Series 1100 G1314A UV
which was first synthesized by Fujii et al.10 in 2003 and displayed detector (λ = 254 nm), a Bioscan (Washington, D.C., U.S.A.)
good inhibition of [125I]SDF-1 binding to CXCR4-transfected Flowcount FC-3400 PIN diode detector (Lablogic) and
a
Chinese hamster ovary cells. Over 100 analogues of FC131 have Phenomenex Luna C18(2) 5 μ 100 × 10.00 mm HPLC column.
been published in the literature along with their inhibitory Injection loop 2000μL. Mobile phase (A) water (0.1% TFA) and (B)
potencies,11 providing valuable insight into its structure-activity MeCN (0.1% TFA). Gradient 10% B for 1 min, 10–40% over 9 min,
relationship. Notably, substitution of Arg3 seems to have a 40–80% over 5 min, and 80% for 1 min. Flow rate 3.15 mLminÀ1
.
relatively moderate effect on the binding provided the replaced Analytical UV-radio-HPLC was carried out as aforementioned but
residue is not a negatively charged residue such as glutamate. This by using a Hewlett Packard Series 1100 G1312A pump and a
position is therefore attractive for the introduction of the required Phenomenex Luna C18(2) 5μ 150 × 4.6mm HPLC column.
terminal alkyne and subsequent triazole following Huisgen alkyne- Injection loop 20μL. Flow rate 1 mLminÀ1
.
azide cycloaddition.
Chemistry
Materials and methods
Preparation of 2-([(9H-fluoren-9-ylmethoxy)carbonyl]amino)-3-[1-
(2-fluoroethyl)-1H-1,2,3-triazol-4-yl]propanoic acid (Fmoc-Fta-OH)
General
All reagents and solvents were purchased from Sigma-Aldrich The CuI (5.78 g, 30.4 mmol, 2 equiv.) and N,N-diisopropylethylamine
(Gillingham, U.K.), Fluka, or Fluorochem (Hadfield, U.K.) and used (DIPEA) (10.4 mL, 60.8 mmol, 4 equiv.) were added to a solution
without further purification. Flash column chromatography was of Fmoc-propargyl-Gly-OH (5.08 g, 15.2mmol, 1 equiv.) and 2-[18F]
carried out on silica gel (Fluka 230–400 mesh, for flash FEA (91.2mL at 0.5 M in DMF, 45.6mmol, 3 equiv.) and stirred at
chromatography). Thin layer chromatography was performed room temperature (r.t.) under N2 for 2 h. The reaction mixture
on aluminum plates pre-coated with silica (200 μm, 60 F254), was diluted with 0.5N HCl aqueous solution (500 mL) and
which were visualized either by quenching of ultraviolet extracted with EtOAc (100 mL). The organic layer was washed with
fluorescence (λmax = 254 nm) or by charring with a KMnO4 0.5 N HCl aqueous solution (4× 500 mL). The organic layer was
dip. Peptides were shaken on an IKA (Staufen, Germany) KS dried over MgSO4 and concentrated under reduced pressure.
130-basic device. The m.p.’s were determined on a Khofler The residue was purified by flash column chromatography (AcOH
hot stage and are uncorrected. The [α]2D0 were recorded on a – MeOH – CH2Cl2 1:10:90, Rf 0.33) to afford 6.01 g (14.2 mmol, 93%)
Perkin-Elmer (Waltham, MA, U.S.A.) 241 polarimeter at 589 nm of amorphous dark yellow powder. The m.p. is 78–79 °C; [α]2D0 –5.0
1
(Na D-line) with a path length of 1 dm. Concentrations (c.) are (c. 0.08, MeOH); infrared (IR): 1713 (s), 1220 (s), 739 (vs) cmÀ1; H
quoted in g 100 mLÀ1, and specific rotations are quoted in NMR (400 MHz, MeOH-d4): δ 7.79 (3H, d, J = 7.4Hz, H-16, H-20,
units of 10À1 deg cm2 gÀ1 at 20 °C. Infrared spectra were recorded and H-25), 7.63 (2H, m, H-19, and H-23), 7.39 (2H, t, J = 7.4 Hz, H-
neat on Perkin-Elmer Paragon 1000 Fourier transform 17, and H-21), 7.31 (2H, t, J = 7.4Hz, H-18, and H-22), 4.72 (4H, tt,
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 679–685

G. P. C. George et al.
J = 4.5, 27.1Hz, H-29, and H-30), 4.54 (1H, dd, J = 4.8, 9.1 Hz, H-2), cyclo(– Gly – D-Tyr – Prg – Arg – Nal –). TFA, (1)
4.30 (2H, d, J = 7.1Hz, H-9), 4.20 (1H, t, J = 7.0Hz, H-11), 3.23
TOF MS ES + m/z (rel. int.) 669 [M + H]+ (100), 670 (40), 671 (10);
(3H, m, H-4); 13C NMR (101 MHz, MeOH-d4): δ 173.1 (s, C-3), 157.0
(s, C-7), 143.9 (s, C-12, and C-15), 143.6 (s, C-24), 141.2 (s, C-13,
and C-14), 127.4 (d, C-17, and C-21), 126.8 (d, C-18, and C-22),
124.9 (d, C-19, and C-23), 123.6 (d, C-25), 119.5 (d, C-16 and C-20),
81.4 [t (d, J = 170.6 Hz), C-30], 66.7 (t, C-9), 53.7 (d, C-2), 50.4 [t (d,
J = 20.3 Hz), C-29], 46.9 (d, C-11), 27.4 (t, C-4); 19F NMR (471 MHz,
MeOH-d4): δ -223.7 (tt, J = 47.1 and 27.2Hz, -CH2F); Time of Flight
(TOF) Mass spectroscopy (MS) Electrospray (ES)+ m/z [elative
intensity(rel. int.)] 426 (25), 425 [M+ H]+ (100); High resolution mass
spectrometry (HRMS) ES+ [M+ H]+ m/z calcd for C22H21FN4O4
425.1625, found 425.1616, Δ = –2.1ppm.
HRMS ES + [M + H]+ m/z calcd for C35H40N8O6 669.3149, found
669.3127, Δ = –3.3 ppm.
cyclo(– Gly – D-Tyr – Fta – Arg – Nal –). TFA, (2)
The 19F NMR (376MHz, MeOH-d4-CDCl3 1:1): δ À76.2 (s, TFA),
À222.6 (tt, J = 45.8 and 25.6 Hz, -CH2F); TOF MS ES+ m/z (rel. int.)
820 [M + Cu]+ (60), 758 [M+ H]+ (100); HRMS ES+ [M+ H]+ m/z
calcd for C37H44FN11O6 758.3538, found 758.3530, Δ = –1.1ppm.
Radiochemistry
Preparation of 2-[18F]Fluoroethylazide12
Preparation of cyclopentapeptides
2-[18F]Fluoroethylazide12 was synthesized from 740 MBq (20 mCi)
of [18F]fluoride as reported in the literature in 47 7% decay-
corrected radiochemical yield (n = 6) affording a solution of the
product in 300 μL in 40 min with >99% radiochemical purity.
In a syringe with a plastic filter and a stopper at its bottom,
H-Gly-2ClTrt resin (0.46 mmol gÀ1 and 100 μmol) was swollen
in dimethylformamide (DMF) for 30 min in a shaker and
the solvent removed by filtration. Hydroxybenzotriazole
(HOBt) (40 mg), O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium
hexafluorophosphate (HBTU) (113 mg), Fmoc-protected amino
acid (300 μmol), and DIPEA (52 μL) were mixed in 3.2 mL DMF; this
coupling mixture was added to the syringe and the reaction
mixture was shaken for 2 h at r.t. The resin was then washed
with DMF (3× 30 s); the Fmoc protecting group was removed
by the treatment of the resin with 20% piperidine in DMF (4mL)
for 30 min and the resin was washed again with DMF (3× 30 s).
The sequence coupling–washing–deprotection–washing was
repeated with the appropriate Fmoc-protected amino acids
until the desired resin-bound pentapeptide was synthesized. The
pentapeptide was cleaved from the resin by treatment with a
1:1:3 mixture of AcOH – TFE – CH2Cl2 (1× 2 h and 3 × 2 min,
10 mL each), and the combined filtrates were concentrated under
reduced pressure. To a solution of the resulting unprotected linear
pentapeptide and NaHCO3 (60 mg) in DMF (40 mL) at –40 °C was
added diphenylphosphorylazide (100 μL) and the reaction mixture
was allowed to stir at r.t. for 4 days. DMF was removed under
reduced pressure; the residue was suspended in MeOH–CHCl3
(9:1) and filtered through a small column of basic alumina. The
filtrate was concentrated in vacuo; the residue was taken up in a
minimum of CH2Cl2 and precipitated with Et2O. The solid was
filtered and washed with Et2O, affording the protected
cyclopentapeptide in good to high yields (73% to >99%) as white
powders. It was dissolved in TFA – triisopropylsilane – H2O (190:5:5)
(10 mL) and the reaction mixture was stirred at r.t. for 6 h. It was
concentrated in vacuo and the residue was suspended in Et2O
and filtered affording a white powder. It was purified by HPLC on
a preparative column: Luna C18(2) 5 μ 75 × 30.00 mm, solvent A:
Preparation of cyclo(– Gly – D-Tyr – [18F]Fta – Arg – Nal –), ([18F]
CCIC15)
To a 1 mL Wheaton vial sealed with a PTFA/Silicon septum under
N2 were successively added CuSO4 (12.5 μL, 238 mM in water),
sodium ascorbate [12.5 μL, 262 mM in acetate buffer (pH 5,
250 mM)], bathophenanthrolinedisulfonic acid disodium salt
(12.5 μL, 298 mM in water), 2-[18F]fluoroethylazide (50 μL, in
MeCN, 178 MBq) and cyclo(– Gly – D-Tyr – Prg – Arg – Nal –) (1)
(50 μL, 15.0 mM in DMF) and the mixture was allowed to react
at r.t. for 10 min. A solution of BnN3 in MeOH (10 μL, 1.50 M)
was added to the Wheaton vial and the reaction carried out
10 min under the same conditions. The reaction mixture was
then diluted to 1.9 mL with water and purified by reverse phase
semi-preparative HPLC, column: Luna C18(2) 5 μ 100 × 10.0 mm,
solvent A: water + 0.1% TFA, solvent B: MeCN + 0.1% TFA, flow:
3.15 mL.minÀ1, gradient: 10% B for 1 min, 10–40% over 9 min,
40–80% over 5 min, and 80% for 1 min. The decay-corrected
yield was 63 16% (n = 5) in approximately 30 min, starting from
178 MBq [18F]FEA. The non-decay-corrected EOS-RCY were of
18 6% (n = 5) in approximately 1.5 h from aqueous [18F]fluoride.
The purified fraction (tR = 11:00–12:20) was diluted with water
(3 mL) and >99% of the radioactivity was immobilized on Waters
(Elstree, U.K.) SepPak tC18 Light cartridge (preconditioned with
5 mL MeOH and 10 mL water). The cartridge was washed with
water (5 mL) and dried with air (10 mL), and the title compound
was eluted with 25 mM HCl in EtOH in 8 × 100 μL fractions. Over
90% of the radioactivity was collected in two fractions. The
labeled peptide was stable in 10% EtOH in phosphate buffered
saline (PBS) for at least 4 h at r.t. Analytical HPLC showed a
radiochemical purity of >98%.
water + 0.1% TFA, solvent B: MeCN + 0.1% TFA, flow: 20 mL minÀ1
,
gradient 5% B to 95% over 10 min. The collected product
fraction was freeze-dried to afford the TFA salt of
the desired product as a white powder. Alkyne-containing
deprotected cyclopentapeptide 1 (5 mg) underwent Huisgen
Labeling optimization
alkyne-azide cyclization with CuSO4.5H2O (0.32 mg, 1.3 μmol, Optimization of the radiolabeling was carried out following the
0.2 equiv.), sodium ascorbate (5.6 mg, 32 μmol, 5 equiv.), 2- aforementioned detailed procedure except with one variable.
fluoroethylazide (250 μL, 1.84 M in DMF) and acetate buffer In particular, we followed the evolution through time of the
(250 μL, pH 5, 250 mM) at r.t. over a week. The reaction reaction advancement for various reaction temperatures and
mixture was subjected to the same HPLC purification various peptide concentrations. The advancement was calculated
conditions as described in the previous texts and freeze- as the ratio between the integration of the product’s radio-peak
dried to afford CCIC15 (2) a white powder in quantitative to the total integration of the chromatogram (i.e., the integration
yield.
of the product’s radio-peak+ the integration of [18F]FEA).
J. Label Compd. Radiopharm 2013, 56 679–685
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

G. P. C. George et al.
Biological evaluation
37 °C. After incubation, the unbound radioactivity was washed
from the cells twice with ice-cold PBS, scraped in PBS, and
collected by centrifugation at 5000 rpm for 3 min at 4 °C and
then counted by gamma counter. The decay-corrected counts
per minute was normalized to the protein content of the wells
using GraphPad Prism 5.
Binding assay
U87.CD4. CXCR4 cells were aspirated with EDTA and incubated
at 37 °C in complete media for 2 h to recover, then resuspended
in binding buffer (PBS containing 2 mg mLÀ1 bovine serum
albumin) at a concentration of 1.5 × 106 cells mLÀ1. To each tube,
100 μl of this cell mixture was added [1.5 × 105 cells], along
with 10 μl of antagonist prepared in dimethyl sulfoxide to the
Results and discussion
final concentration indicated, and
[
125I]SDF-1 radioligand
To test our hypothesis, we decided to synthesize an FC131
analogue by substituting its Arg3 residues for either 1-(2-
fluoroethyl)-1H-1,2,3-triazol-4-ylalanine (Fta) or propargylglycine
(Prg) in order to afford the reference compound and the
precursor for radiolabeling.
(PerkinElmer) is added to give a final concentration of 0.1 nM in
a 200 μl total assay volume. The tubes were then incubated on
ice for 60 min, with frequent agitation. Bound and unbound
radioactivities were separated by filtration using a vacuum
manifold, followed by three washes in assay buffer. The filters
were then punched out and counted in a gamma counter for
240 s. The bound radioactivity was expressed as a percentage
of the maximum bound radioactivity using GRAPHPAD (La Jolla,
CA, U.S.A.) Prism 5.
We built both the precursor and the cold isotopologue by
solid phase peptide synthesis deploying a 2-chlorotrityl/Fmoc
strategy. Fmoc-Fta-OH was synthesized from commercially
available Fmoc-protected propargyl glycine (Fmoc-Prg-OH) via
copper(I)-mediated alkyne-azide cyclisation with 2-FEA in good
yields (Scheme 1). Solid-phase peptide synthesis was carried
out with both of these non-proteogenic amino acids to
introduce the desired acetylenic and fluoroethyltriazolyl side-
chains (Scheme 2). A mixture of HOBt/HBTU/DIPEA was used to
perform the amide bond formation and treatment of the resin
with a mixture of AcOH – trifluoroethanol – CH2Cl2 (1:1:3) was
used to afford the protected linear pentapeptides. After
successful cyclisation with diphenylphosphorylazide, the
resulting protected cyclic pentapeptides underwent side-chain
deprotection in TFA – triisopropylsilane – H2O (190:5:5). This last
step was followed by RP-HPLC purification to afford the desired
alkyne-containing analogue 1 in 29% overall yield. Although
the deprotection conditions used are common for peptide
synthesis, they were too harsh for the triazole-containing
analogue, resulting in isolation of insufficient quantities of the
Cell-uptake assay
U87.CD4 and U87.CD4.CXCR4 cells were seeded at 48 h prior to
the experiment in complete Dulbecco’s Modified Eagle Medium
in triplicate wells of 6-well plates at a density of 3 × 105 cells/well.
At the commencement of the experiment, the wells were
incubated with 0.74 MBq (20 μCi) of [18F]CCIC15 for 60 min at
Scheme 1. Synthesis of Fmoc-Fta-OH.
Scheme 2. Synthesis of an FC131 analogue, precursor, and reference compound.
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 679–685

G. P. C. George et al.
desired triazole-containing peptide. Instead, reaction of its low precursor concentration is desirable as it impacts the apparent
alkyne counterpart 1 by Huisgen alkyne-azide cycloaddition with specific activity. Although the optimal peptide concentration was
FEA followed by RP-HPLC purification efficiently afforded the found to be 5.4mM, reaching full conversion, the reaction rate
desired product in 89% isolated yield.
associated to a peptide concentration of 2.2mM was satisfactory.
Binding activity of the TFA salt of compound 2 was assessed A peptide concentration of 1.1mM caused reaction progress to
by competitive displacement of [125I]SDF-1 on U87.CD4.CXCR4 plateau at about 20% conversion. The identity of the radiolabeled
cells. The concentration of compound 2 that reduced the product was confirmed by co-elution with the cold reference
amount of radioactivity bound to the cells by 50% compared compound. On the other hand, temperature had a lesser effect
with controls (IC50 value) was determined after 60 min of on the reaction rate. Four temperatures were tested (0 °C,
incubation over ice. CCIC15 (2) showed an IC50 of 1.23 μM which 20 °C, 50 °C, and 80 °C) at 5.4 mM. With the exception of the
is, as expected, relatively close to what we found for FC131 coldest experiment, the reaction rates were relatively similar;
(0.55 μM). It was therefore decided to evaluate the biological however, higher reaction temperatures tended to deform
properties of compound CCIC15 (2).
the HPLC-radio peak. This behaviour, which likely stems from
The [18F]FEA was synthesized according to literature epimerization of the peptide’s stereogenic centers,
procedure13 affording a solution of the product in approximately encouraged by heat and the presence of the Lewis acidic
300 μL of MeCN in 47 7% decay-corrected radiochemical yield copper, was not observed at 20 °C. These observations made
(n = 6) in about 40 min.
20 °C the temperature of choice for further experiments. The
The radiochemical yields of the subsequent Huisgen alkyne-
azide cycloadditions were strongly dependent on reaction
conditions: temperature, duration, concentration, pH, and the
order of addition of reagents. Previous work in our group12 also
showed that both the addition of a copper(I)-stabilizing bidentate
ligand such as bathophenanthrolinedisulfonic acid disodium (4)
and the use of acetate buffer substantially increased the reaction
rate of this type of cycloaddition and that the optimal order of
addition of reagents is CuSO4, sodium ascorbate, the ligand, [18F]
FEA and the precursor, sequentially. Taking these data into
account, focussed screening was performed to determine the
influence of peptide concentration, temperature, and the duration
of the reaction (Scheme 3, Table 1).
Table 1. Radiolabeling of FC131 analogue by Huisgen
alkyne-azide cycloaddition
Precursor concentration (mM)
Reaction temperature (oC)
1.1
2.2
91
5.4 10.9
35
0
25
25
35
20
40
50
60
80
96
100
100
95
41
48
As expected, the cycloaddition reaction rate was highly
dependent on the peptide precursor concentration. However, a
Scheme 3. Radiolabeling and scavenging of FC131 analogue by Huisgen alkyne-azide cycloaddition.
J. Label Compd. Radiopharm 2013, 56 679–685
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

G. P. C. George et al.
optimal labeling conditions were found to be a 2.2mM peptide
concentration at room temperature for ~5 min.
Cellular uptake of [18F]CCIC15 was assessed in U87.CD4 cells
transfected with CXCR4 (U87.CD4.CXCR4) (Figure 1). Sixty-minute
Unsurprisingly, the unreacted precursor and the radiotracer uptake of [18F]CCIC15 showed a two-fold higher uptake in the
co-eluted on the HPLC, although it was not possible to reduce CXCR4 over-expressing cells compared with isogenic CXCR4
the precursor concentration of peptide 1 any further to increase negative cells, confirming binding of the tracer to CXCR4.
the apparent specific activity. To overcome this challenge, we Furthermore, addition of 20 μM of FC131 showed a decrease of
developed a strategy wherein the unreacted peptide alkyne 1 the cellular uptake to the level of the CXCR4 negative cells
was scavenged by a lipophilic azide. The retention time of the control, demonstrating specificity of the tracer for CXCR4. These
corresponding triazole was anticipated to differ significantly biological data, along with the IC50 value measured for our
from the radioactive product, allowing an easy HPLC purification. tracer, validated our structure-activity relationship analysis.
Additionally, we envisaged that such a scavenging reaction
could be performed under the same conditions as the labeling,
Conclusions
generating a practical one-pot labeling-scavenging procedure.
Benzyl azide14 was chosen as a potential scavenging agent. A
We devised an easy, rapid, one-pot method to circumvent the
common issue met when labeling large molecules, namely the
poor apparent specific activity afforded. This was achieved by
treating the unreacted precursor with an excess of scavenging
reagent at the end of the labeling process. We successfully
test reaction demonstrated that this azide was fit for purpose,
leading to a satisfactory 2-min difference in retention times
between the radioactive product [18F]CCIC15 and the benzyl
triazole derivative 3; thus, unreacted benzyl azide was easily
removed by preparatory HPLC. More detailed monitoring of
the scavenging reaction showed that the apparent peptide
radiosynthesized by Huisgen alkyne-azide cycloaddition with
concentration, obtained from the integration of the UV-HPLC
applied it to a fluoroethyl triazole analogue of FC131 of our
own design, of which the 18F-isotopologue, [18F]CCIC15, was
[18F]FEA in a copper(II) – ascorbic acid system. Although, as
anticipated, the radioactive product and its unlabeled precursor
traces and a calibration curve, had decreased from 3.48 mM to
103 μM in 10 min and to 71.2 μM in 20 min (n = 2), that is, a factor
co-eluted on the HPLC, resulting in a low apparent specific
of 34 in 10 min and of 49 in 20 min (n = 2).
activity, scavenging of the unreacted precursor with benzyl azide
The click labeling – scavenging step afforded the tracer in
allowed easy purification, reducing the peptide concentration
63 16% (n = 5) (decay corrected) yields in approximately
30 min starting from 178 MBq of [18F]FEA but overall, the desired
from 3.48 mM to 103 μM, that is, by a factor of 34, in 10 min.
Overall, the labeling – scavenging step was achieved in
radiolabeled peptide was obtained in EOS-RCY of 18 6% (n = 5)
from aqueous [18F]fluoride in approximately 1.5 h, with an
63 16% (n = 5) yield (decay corrected) in approximately
apparent specific activity of 19.4 GBq.μmolÀ1. The tracer was
30 min, whereas our new tracer was obtained in an overall
18 6% (n = 5) EOS-RCY in 1.5 h from aqueous [18F]fluoride with
stable in 10% EtOH/PBS for 4 h as assessed by analytical HPLC.
a specific activity of 19.4 GBq μmolÀ1. In vitro biological assay
showed specificity for CXCR4 and further biological evaluation
of CCIC15 is underway.
The distribution coefficient of the tracer [18F]CCIC15 was then
evaluated. At the physiological pH of blood serum, the product is
expected to be exclusively present in its monoprotonated form
as the pKa of guanidine is ~12. The distribution ratio at
physiological pH was assessed by measuring the distribution of
radioactivity between n-octanol and PBS (pH 7.4) using the
‘shake flask’ method.15 The logarithm of the ratio was calculated
as logDph 7.4 = –0.11 0.02 (n = 5). This balanced lipophilicity was
assumed to be appropriate for a short-lived radiotracer as it
Acknowledgements
This work was funded by Cancer Research UK–Engineering and
Physical Sciences Research Council grant [in association with the
Medical Research Council and Department of Health (England)]
grant C2536/A10337. Professor Eric O. Aboagye (E.O.A.’s)
laboratory receives core funding from the UK Medical Research
Council (MC_A652_5PY80).
should account for
a
optimal renal elimination,16 faster
pharmacokinetics, and increased contrast.17
Conflict of Interest
The authors did not report any conflict of interest.
References
[1] F. Balkwill, Nat. Rev. Cancer 2004, 4, 540, http://dx.doi.org/10.1038/
nrc1388.
[2] a) X. Liang, Chem. Biol. Drug Des. 2008, 72, 97, http://dx.doi.org/
10.1111/j.1747-0285.2008.00681.x;
b) V. Vinader, K. Afarinkia, Future Med. Chem. 2012, 4, 853, http://dx.
doi.org/10.4155/fmc.12.48.
[3] B. Wu, E. Y. T. Chien, C. D. Mol, G. Fenalti, W. Liu, V. Katritch, R.
Abagyan, A. Brooun, P. Wells, F. C. Bi, D. J. Hamel, P. Kuhn, T. M.
Handel, V. Cherezov, R. C. Stevens, Science 2010, 330, 1066, http://
dx.doi.org/10.1126/science.1194396.
[4] a) L. E. Woodard, S. Nimmagadda, J. Nucl. Med. 2011, 52, 1665, http://
dx.doi.org/10.2967/jnumed.111.097733;
b) J. C. Knight, F. R. Wuest, Med. Chem. Commun. 2012, 3, 1039,
http://dx.doi.org/10.1039/C2MD20117H;
c) J. Kuil, T. Buckle, F. W. B. Leeuwen, Chem. Soc. Rev. 2012, 41, 5239,
Figure 1. Cell uptake profile at 1 h of
[
18F]CCIC15 in U87.CD4.CXCR4 cells
compared with U87.CD4 cells.
http://dx.doi.org/10.1039/C2CS35085H;
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 679–685

G. P. C. George et al.
d) O. Åberg, F. Pisaneschi, G. Smith, Q.-D. Nguyen, E. Stevens, E. O.
Aboagye, J. Fluorine Chem. 2012, 135, 200, http://dx.doi.org/
10.1016/j.jfluchem.2011.11.003.
d) H. Tamamura, H. Tsutsumi, W. Nomura, N. Fujii, Perspect. Med.
Chem. 2008, 2, 1;
e) H. Tsutsumi, T. Tanaka, N. Ohashi, H. Masuno, H. Tamamura,
K. Hiramatsu, T. Araki, S. Ueda, S. Oishi, N. Fujii, Pept. Sci. 2007, 88,
279, http://dx.doi.org/10.1002/bip.20653;
[5] S. M. Ametamey, M. Honer, P. A. Schubiger, Chem. Rev. 2008, 108,
1501, http://dx.doi.org/10.1021/cr0782426.
[6] V. Tolmachev, S. Stone-Elander, Biochim. Biophys. Acta, Gen. Subj.
2010, 1800, 487, http://dx.doi.org/10.1016/j.bbagen.2010.02.002.
[7] M. Fani, H. R. Maecke, S. M. Okarvi, Theranostics 2012, 2, 481, http://
dx.doi.org/10.7150/thno.4024.
[8] D. A. Mankoff, J. M. Link, H. M. Linden, L. Sundararajan, K. A. Krohn, J. Nucl.
Med. 2008, 49, 149S, http://dx.doi.org/10.2967/jnumed.107.045963.
[9] M. R. Kilbourne, J. R. Huizenga, Fluorine-18 Labeling of
f) H. Tamamura, T. Araki, S. Ueda, Z. Wang, S. Oishi, A. Esaka, J. O.
Trent, H. Nakashima, N. Yamamoto, S. C. Peiper, A. Otaka, N. Fujii,
J. Med. Chem. 2005, 48, 3280, http://dx.doi.org/10.1021/jm050009h;
g) H. Tamamura, K. Hiramatsu, S. Ueda, Z. Wang, S. Kusano, S.
Terakubo, J. O. Trent, S. C. Peiper, N. Yamamoto, H. Nakashima, A.
Otaka, N. Fujii, J. Med. Chem. 2005, 48, 380, http://dx.doi.org/
10.1021/jm049429h.
Radiopharmaceuticals, National Academy Press, Washington, D.C., [12] L. Iddon, J. Leyton, B. Indrevoll, M. Glaser, E. G. Robins, A. J. T. George,
1990, Chapter 2, pp. 17.
[10] N. Fujii, S. Oishi, K. Hiramatsu, T. Araki, S. Ueda, H. Tamamura, A.
A. Cuthbertson, S. K. Luthra, E. O. Aboagye, Bioorg. Med. Chem. Lett.
2011, 21, 3122, http://dx.doi.org/10.1016/j.bmcl.2011.03.016.
Otaka, S. Kusano, S. Terakubo, H. Nakashima, J. A. Broach, J. O. Trent, [13] a) A. Haslop, A. Gee, C. Plisson, N. Long, J. Label. Compd. Radiopharm.
Z.-X. Wang, S. C. Peiper, Angew. Chem. Int. Ed. 2003, 42, 3251, http://
dx.doi.org/10.1002/anie.200351024.
2013, 56, 313, http://dx.doi.org/10.1002/jlcr.3024;
b) M. Glaser, E. G. Robins, J. Label. Compd. Radiopharm. 2009, 52, 407,
http://dx.doi.org/10.1021/bc060301j.
[11] a) S. Ueda, S. Oishi, Z.-X. Wang, T. Araki, H. Tamamura, J. Cluzeau,
H. Ohno, S. Kusano, H. Nakashima, J. O. Trent, S. C. Peiper, N. Fujii,
J. Med. Chem. 2007, 50, 192, http://dx.doi.org/10.1021/jm0607350;
b) T. Tanaka, W. Nomura, T. Narumi, A. Esaka, S. Oishi, N. Ohashi,
K. Itotani, B. J. Evans, Z.-X. Wang, S. C. Peiper, F. Nobutaka,
H. Tamamura, Org. Biomol. Chem. 2009, 7, 3805, http://dx.doi.org/
10.1039/B908286G;
[14] K. Asano, S. Matsubara, Org. Lett. 2010, 12, 4988, http://dx.doi.org/
10.1021/ol101990d.
[15] A. A. Wilson, L. Jin, A. Garcia, J. N. DaSilva, S. Houle, Appl.
Radiat. Isot. 2001, 54, 203, http://dx.doi.org/10.1016/s0969-
8043(00)00269-4.
[16] M. V. S. Varma, B. Feng, R. S. Obach, M. D. Troutman, J. Chupka, H. R.
Miller, A. El-Kattan, J. Med. Chem. 2009, 52, 4844, http://dx.doi.org/
10.1021/jm900403j.
c) T. Tanaka, H. Tsutsumi, W. Noruma, Y. Tanabe, N. Ohashi, A. Esaka,
C. Ochiai, J. Sato, K. Itotani, T. Murakami, K. Ohba, N. Yamamoto, N.
Fujii, H. Tamamura, Org. Biomol. Chem. 2008, 6, 4374, http://dx.doi.
org/10.1039/B812029C;
[17] P. H. Elsinga, Methods 2002, 27, 208, http://dx.doi.org/10.1016/
S1046-2023(02)00076-2.
J. Label Compd. Radiopharm 2013, 56 679–685
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org