Copper-free click - A promising tool for pre-targeted PET imaging

Article abstract of DOI:10.1039/c1cc16220a

The copper-free click (CFC) reaction has been evaluated for its potential application to in vivo pre-targeting for PET imaging. A promising biodistribution profile is demonstrated when employing [¹⁸F]2- fluoroethylazide ([¹⁸F]1) and

Full text of DOI:10.1039/c1cc16220a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 991–993
www.rsc.org/chemcomm
COMMUNICATION
Copper-free click—a promising tool for pre-targeted PET imagingw
Helen L. Evans,a Rozanna L. Slade,a Laurence Carroll,*a Graham Smith,a
a
a
a
b
b
¨
Quang-De Nguyen, Lisa Iddon, Nazila Kamaly, Henning Stockmann, Finian J. Leeper,
Eric O. Aboagyea and Alan C. Spiveyc
Received 6th October 2011, Accepted 17th November 2011
DOI: 10.1039/c1cc16220a
The copper-free click (CFC) reaction has been evaluated for its
potential application to in vivo pre-targeting for PET imaging. A
promising biodistribution profile is demonstrated when employ-
ing [18F]2-fluoroethylazide ([18F]1) and optimisation of the CFC
reaction with a series of cyclooctynes shows that reactions
proceed efficiently with tantalizing opportunities for application-
specific tuning.
(i.e. drug, peptide or antibody) and allowed to localise in the
tissue to be imaged (e.g. tumour) before systemic administra-
tion of a ‘pull-down’ reagent (e.g. an azide having a fluorescent
label), which, after wash-out of the excess pull-down reagent,
allows the ‘pre-targeted’ tissue to be imaged. This is in contrast
to traditional ‘targeted’ radiotracers, which accumulate in the
desired area of interest with the radiolabelled tag already
attached to the targeting group—a protocol which is more
severely constrained by the limited half-life of PET radio-
nuclides. The CFC reaction was shown to have superior
reaction kinetics relative to other bio-orthogonal reactions
such as the Staudinger ligation, in this context.7,8 We therefore
envisaged that the CFC reaction had great potential for analo-
gous in vivo pre-targeting for PET imaging. Prompted by a
recent preliminary disclosure by Wuest et al. of an approach to
this in which a radiolabelled cyclooctyne is used in conjunction
with a series of functionalised azides,9 we describe here a
complementary approach10 in which a radioabelled azide is
used in conjunction with a series of functionalised cyclooctynes.
Specifically, we report the biodistribution profile of
[18F]2-fluoroethylazide ([18F]1) in BALB/c nude mice and the
reactivity of this radiotracer with a series of five structurally
diverse cyclooctynes as a function of temperature and solvent.
We chose [18F]2-fluoroethylazide ([18F]1) as our labelled
azide, due to its synthesis being well-described in the literature,
its small size and its expected rapid clearance from major
organs in vivo (Scheme 2).10–13
Since the discovery of fluorine-18 labelled 2-deoxy-2-fluoro-
glucose ([18F]FDG) approximately forty years ago, positron
emission tomography (PET) has expanded to become a crucial
technique in the fight against diseases such as Parkinson’s
disease and cancer.1,2 Throughout this period, there has been
great interest in developing new radiolabelling methods to
expand the range of possible tracers that can be synthesised.
One of the most exciting recent developments has been the
deployment of the copper(I)-catalysed Huisgen [2, 3]-cyclo-
addition between alkynes and azides, which is widely referred
to as the ‘click reaction’, following independent development
work by Sharpless and Meldal in 2001.3,4 The use of click
chemistry for the synthesis of a number of novel radiotracers
has been documented, but to date, none have reached the
clinic, in part due to the difficulties associated with removing
traces of copper, which has been shown to be toxic to humans
even at low levels.5
The copper-free click (CFC) reaction, using strained cyclo-
octynes, was pioneered by Bertozzi, and eliminates any issues
of copper toxicity (Scheme 1).6
Using the method of Glaser et al., tosylate 2 was reacted
with a mixture of [18F]fluoride and Kryptofix[2.2.2] in aceto-
nitrile, giving the desired labelled azide [18F]1 in 30% RCY
n.d.c, after distillation.11
The reaction was developed specifically to enable in vivo pre-
targeting for fluorescent imaging, albeit for ex vivo tumour
analysis, due to the limitations of fluorescence detection tech-
nology. Pre-targeting refers to a strategy in which a reactive tag
(e.g. the cyclooctyne) is attached to a specific biomolecule
In order to determine the level of background radioactivity
that could be expected when employing the CFC reaction for
in vivo pre-targeting using this compound, biodistribution data
was collected for [18F]1 in BALB/c nude mice (n = 3) (Fig. 1).
a Comprehensive Cancer Imaging Centre, Department of Surgery &
Cancer, Hammersmith Campus, Imperial College, London,
United Kingdom. E-mail: l.carroll@imperial.ac.uk;
Tel: +44 (0)20 8383 3759
b Department of Chemistry, University of Cambridge, Lensfield Road,
Cambridge, CB2 1EW, UK
c Department of Chemistry, Imperial College London,
South Kensington Campus, London, SW7 2AZ, UK
w Electronic supplementary information (ESI) available: Synthetic
procedures for 3–7 and selected HPLC traces for click reactions. See
DOI: 10.1039/c1cc16220a
Scheme 1 The copper-free ‘click’ reaction.
Chem. Commun., 2012, 48, 991–993 991
c
This journal is The Royal Society of Chemistry 2012

View Article Online
Scheme 2 Radiosynthesis of [18F]2-fluoroethylazide ([18F]1).
Fig. 2 Structures of cyclooctynes used in the study.
Fig. 1 Representative small animal PET images of a BALB/c nude
mouse at t = 15 s and t = 60 min after intravenous injection in the tail
of 3.7 MBq of [18F]1.
of azide 1 to each cyclooctyne. After stirring until the reactions
reached completion, excess azide 1 was evaporated at reduced
pressure to afford the desired triazole products for characterisation
and HPLC analysis; for cyclooctynes 3–7, a mixture of regio-
isomeric 1, 4, 5-trisubstituted-1, 2, 3-triazoles could be identified
by 19F NMR and HPLC in ratios varying from 2 : 1 to 7 : 1.
Next, we examined the corresponding CFC reactions using
radiolabelled [18F]1 under different conditions of temperature,
solvent and time (Table 1).
Over the course of the experiment, approximately 9 % of
the total radioactivity was eliminated either via excretion or
from the lungs via exhalation. By the end of the scan a small
amount of activity was found in the bone, suggesting a minor
amount of [18F]fluoride anion generation, but this is not
anticipated to present a problem for future imaging, due to
its low level relative to the initial injection. The remaining
radioactivity was evenly spread throughout the other tissue
(see ESI for full biodistribution dataw). Azide [18F]1 is also
completely stable in phosphate buffer (pH = 7.4), acetonitrile
and dimethylsulfoxide for up to six hours at room tempera-
ture. This intrinsic stability, combined with the even distri-
bution within the body within 60 min, suggests that azide
[18F]1 could prove to be an advantageous labelled azide for use
following pre-targeting with a CFC reagent.
Under our conditions in MeCN at 90 1C, the apparent
reactivities of cyclooctynes 3, 4 and 5, based on radiochemical
yield (RCY) after 15 min, follow the order 3 4 5 4 4
(cf. entries 3, 7 and 11) rather than the 5 4 3 4 4 that would
be expected based on published literature (vide supra). How-
ever, TMDIBO (5) appears to be unstable at higher tempera-
tures (see ESI file for example HPLC UV traces for reactions
at 90 1Cw) and so the lower amount of labelled product in this
case is probably due to its decomposition being faster than its
reaction with [18F]1. Interestingly, the amide derivative 6 of
cyclooctyne 4 and the carbamate derivative 7 of TMDIBO (5)
show significantly altered reactivities relative to their acid/
alcohol parents: 6 4 4 and 7 o 5 (cf. entries 12 and 3 vs. 14
and 7). Importantly, both TMDIBO (5) and DIFO (3) react
efficiently with [18F]1 in water at physiological temperature
within the timeframe of an in vivo PET scan (i.e. B15 min,
entries 5 and 9). This, coupled with azide [18F]1 having an even
distribution profile within the body (vide supra), suggests that
these cyclooctynes could be suitable for pre-targeting studies
by PET. It is evident however that after 60 min in water the
RCY of product when using TMDIBO (5) is approximately
double that when using DIFO (3); this contrasts strongly with
the situation after 15 min in MeCN where the RCY of product
when using TMDIBO (5) is over eight-fold lower than when
using DIFO (3) (cf. entries 6 and 10 vs. 4 and 8). In view of this
inversion of relative reactivity in water as compared to MeCN,
TMDIBO (5) appears to be the most promising cyclooctyne of
those that we have investigated for use in biological studies.
The synthesis of a wide range of cyclooctynes have been
reported for use in CFC reactions, each presenting a unique
blend of attributes relevant to this specialist application
vis-a-vis synthetic accessibility, stability/reactivity and
potential for ligation to biomolecular targets.14,15 We selected
three of these cyclooctynes (3–5), as well as further-functionalised
derivatives 6 and 7, for our investigation (Fig. 2).
Cyclooctynes 3–5 were synthesised using previously des-
cribed methods and derivative 6 was obtained by subsequent
HBTU coupling of cyclooctyne 4 with 4-methoxybenzylamine.14,15
Derivative 7 was synthesised by direct coupling of cyclooctyne
6 with N-(2-aminoethyl)maleimide. The literature suggests
that DIFO (3) has a faster reaction rate for the CFC reaction
than cyclooctyne 4, and similarly that TMDIBO (5) has a
faster reaction rate than DIFO (3).14
Non-labelled reference samples of the four triazole products
derived from CFC reaction of cyclooctynes 3–7 with azide 1
(cf. Scheme 1) were prepared by addition of a five-fold excess
c
992 Chem. Commun., 2012, 48, 991–993
This journal is The Royal Society of Chemistry 2012

View Article Online
Table 1 Summary of CFC reactions carried out using [18F]1
Substratea T/1C Time/Mins Solvent RCY/%b,c
(3) shows greater reactivity towards azide [18F]1 in MeCN than
TMDIBO (5), but in water over longer time periods TMDIBO
(5) displays more robust reactivity and benefits from a shorter
synthesis route for preparation. We are currently exploring the
combination of azide [18F]1 and TMDIBO (5) for in vivo PET
imaging of certain cancers and these studies will be published
in due course.
1
2
3
40
40
90
15
15
15
MeCN
H2O
MeCN
4.4
0.3
24.9
4
5
6
7
40
40
40
90
15
15
60
15
MeCN
H2O
H2O
7.1
33.3
36.2
31.5
This work was supported by a studentship from the Imperial
Cancer Research UK Centre grant (C37990/A12196), and by
CR-UK&EPSRC Cancer Imaging Centre at Imperial College
London, in association with the MRC and Department of
Health (England) grant C2536/A10337 and UK Medical
Research Council core funding grant U.1200.02.005.00001.01.
MeCN
8
9
10
11
40
40
40
90
15
15
60
15
MeCN
H2O
H2O
62.1
26.1
17.8
97.1
MeCN
12
90
15
MeCN
60.4
Notes and references
13
14
40
90
15
15
MeCN
MeCN
3.2
9.6
1 B. Beuthien-Baumann, K. Hamacher, F. Oberdorfer and
J. Steinbach, Carbohydr. Res., 2000, 327, 107–118.
2 S. M. Ametamey, M. Honer and P. A. Schubiger, Chem. Rev.,
2008, 108, 1501–1516.
3 H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem., Int.
Ed., 2001, 40, 2004–2021.
4 C. W. Tornøe, C. Christensen and M. Meldal, J. Org. Chem., 2002,
67, 3057–3064.
a
Each reaction carried out using 50 mL of [18F]1 (typical activity of
b
B5 MBq); n = 3 in all cases. Radiochemical yield as determined via
radio-HPLC showing conversion of [18F]1 into the summation of the
c
regioisomers. For further details relating to these data see the ESI file.w
5 G. J. Brewer, Clin. Neurophysiol., 2010, 121, 459–460; A. Ala,
A. P. Walker, K. Ashkan, J. S. Dooley and M. L. Schilsky, Lancet,
2007, 369, 397–408.
6 J. A. Codelli, J. M. Baskin, N. J. Agard and C. R. Bertozzi, J. Am.
Chem. Soc., 2008, 130, 11486–11493.
In addition, a cell viability assay was carried out in HCT116
human colon cancer cells, and no toxicity of TMDIBO (5) was
observed at concentrations up to 100 mM (see ESI for further
detailsw).
7 L. Carroll, S. Boldon, R. Bejot, J. E. Moore, J. Declerck and
V. Gouverneur, Org. Biomol. Chem., 2011, 9, 136–140.
8 J. M. Baskin, J. A. Prescher, S. T. Laughlin, N. J. Agard,
P. V. Chang, I. A. Miller, A. Lo, J. A. Codelli and C. R. Bertozzi,
Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 16793–16797.
9 V. Bouvet, M. Wuest and F. Wuest, Org. Biomol. Chem., 2011,
9, 7393.
10 While this manuscript was in preparation Feringa et al. published
an approach to labelling a cyclooctyne-tagged Lys [3]-bombesin via
a strain-promoted click reaction using three [18F]-labelled azides:
L. S. Campbell-Verduyn, L. Mirfeizi, A. K. Schoonen, R. A.
Dierckx, P. H. Elsinga and B. L. Feringa, Angew. Chem., Int.
Ed., 2011, 50, 11117.
11 M. Glaser and E. Arstad, Bioconjugate Chem., 2007, 18, 989–993.
12 R. Bejot, T. Fowler, L. Carroll, S. Boldon, J. E. Moore, J. Declerck
and V. Gouverneur, Angew. Chem., Int. Ed., 2009, 48, 586–589.
13 G. Smith, M. Glaser, M. Perumal, Q.-D. Nguyen, B. Shan,
E. Arstad and E. O. Aboagye, J. Med. Chem., 2008, 51, 8057–8067.
14 H. Stockmann, A. A. Neves, S. Stairs, H. Ireland-Zecchini,
K. M. Brindle and F. J. Leeper, Chem. Sci., 2011, 2, 932–936.
15 A. Bernardin, A. Cazet, L. Guyon, P. Delannoy, F. Vinet,
D. Bonnaffe and I. Texier, Bioconj. Chem., 2010, 21, 583–588.
We have demonstrated that CFC reactions can be carried
out between a short-lived 18F-labelled azide and various
cyclooctynes for the first time and that these reactions can
be achieved under conditions which appear suitable for in vivo
pre-targeting for PET imaging. [18F]2-Fluoroethylazide
([18F]1) is shown to be a promising initial choice for systemic
administration as the pull-down reagent, partnered with a pre-
localised cyclooctyne(s)-tagged biomolecule. Both the ease of
radiosynthesis of [18F]1 and its favourable mouse biodistri-
bution profile indicate that this mode of pre-targeting may
constitute a more promising strategy than one in which the
systemic radiolabelled pull-down reagent is the cyclooctyne.9
The cyclooctyne partner of choice for biomolecule tagging
using this strategy appears to be TMDIBO (5), although
further studies will be required to confirm this for specific
applications. A delicate balance needs to be struck between
reactivity of the tag towards azide [18F]1, its in vivo stability
and its ease of synthesis and biomolecule attachment. DIFO
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 991–993 993