Imparting multivalency to a bifunctional chelator: A scaffold design for targeted PET imaging probes

Full text of DOI:10.1002/anie.200903556

Communications
DOI: 10.1002/anie.200903556
Imaging Probes
Imparting Multivalency to a Bifunctional Chelator: A Scaffold Design
for Targeted PET Imaging Probes**
Wei Liu, Guiyang Hao, Michael A. Long, Tiffani Anthony, Jer-Tsong Hsieh, and Xiankai Sun*
Positron emission tomography (PET) has become a
standard clinical practice in the diagnostic or prog-
nostic imaging of cancer, mainly owing to the great
success of [18F]FDG (2-deoxy-2-18F-fluoro-d-glu-
cose) for non-invasive detection of glucose uptake
in tumors.[1] Currently 11C and 18F are the most
commonly used PET nuclides for the development
of PET imaging probes. However, the short half-lives
of these two radioisotopes (11C t1/2 = 20.3 m, 18F t1/2
=
109 m) limit their applications to biomolecules with
relatively fast in vivo biodistribution kinetics, and
the chemical procedures to incorporate these iso-
topes must be carried out in the proximity of a
biomedical cyclotron. Among nonstandard PET
nuclides, 64Cu (t1/2 = 12.7 h; b+ 0.653 MeV, 17.4%)
has drawn considerable interest in PET research
owing to its low positron range, commercial avail-
ability, and reasonably long decay half-life. Such
characteristics could enable a variety of imaging
applications involving peptides, antibodies or their
fragments, and nanoparticles.[2,3]
It has been demonstrated that creation of multi-
mers of a targeting molecule on one scaffold can
efficiently improve cell-specific binding affinity by
several orders of magnitude.[3] As such, various
approaches have been reported to exploit multi-
valent scaffolds for the construction of molecular
imaging probes.[4–10] However, their chemistries are
often complicated and become even more so when a
bifunctional chelator (BFC) must be attached to a
separately multimerized construct in order to intro-
duce a metal radionuclide for nuclear imaging.
Scheme 1. Chemical structures of relevant compounds
[*] W. Liu, G. Hao, M. A. Long, T. Anthony, X. Sun
Herein we present an approach to take advantage of the
Department of Radiology and Advanced Imaging Research Center
University of Texas Southwestern Medical Center
Dallas, TX 75390-8542 (USA)
Fax: (+1)214-645-2885
E-mail: xiankai.sun@utsouthwestern.edu
pendent arms of the commonly used BFCs to build simplified
but potentially versatile multivalent scaffolds for multimeric
presentation of targeting molecules. This type of multivalent
scaffold features a chelator that forms a stable and neutral
complex with a radiometal and multiple functional groups for
the anchoring of targeting molecules. If required by in vivo
pharmacokinetics, poly(ethylene glycol) (PEG) chains can be
introduced between the chelate and targeting moieties.
To test the rationale of our design, we use a cyclic RGD
peptide (c(RGDyK),[5,11] a well-validated aub3 integrin ligand,
for the construction of a divalent PET imaging probe from a
chelate known to have a high affinity for 64Cu (CB-TE2A,
Scheme 1). The divalent probe is anticipated to have a
prolonged biological half-life and enhanced specific binding
and retention in tissues expressing the aub3 integrin.
J.-T. Hsieh
Department of Urology
University of Texas Southwestern Medical Center
Dallas, TX 75390-9110 (USA)
[**] This work was partially supported by a USAMRMC grant (W81XWH-
08-1-0305) and two NIH grants (P01 DK058398; U24 CA126608).
The authors acknowledge the generous support of a private donor
that allowed the purchase of the Inveon PET-CT system. The
c(RGDyK) peptide was provided by Dr. Xiaoyuan Chen at Stanford
University
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200903556.
7346
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7346 –7349

Angewandte
Chemie
The stability of the metal complex is of critical importance
in the design of a metal radiopharmaceutical. CB-TE2A
forms one of the most stable complexes with 64Cu,[12] and the
CuII-CB-TE2A complex is more resistant to reductive metal
loss than are other tetramacrocyclic complexes.[13] However,
its stability with 64Cu may be reduced when used as a BFC in
which one of the carboxylate groups is converted to an amide
for conjugation to a targeting molecule,[14] leading to a
positive charge on the 64Cu moiety. To avoid this potential
problem, we choose a-bromoglutaric acid-1-tert-butyl ester-5-
benzyl ester (3) for the alkylation of CB-cyclam, which allows
selective deprotection of the carboxylate groups by two
separate procedures to conjugate with c(RGDyK) through
the peripheral carboxylates to afford H21 and H22, respec-
tively. As such, the unique feature of CB-TE2A is preserved
in the conjugates, in which the two inner carboxylates can
form a neutral octahedral complex with 64CuII along with the
four nitrogen atoms of the macrocycle.
To evaluate the anticipated multivalent effect, a recently
reported CB-TE2A analogue[15] was also synthesized and
coupled with c(RGDyK) to form a monovalent CB-TE2A-
RGD conjugate (H21) as a control for comparison. Scheme 2
shows the synthetic routes to H21 and H22. The multiple-step
synthetic route involves three parts: 1) synthesis of orthogo-
nally protected compounds 5 and 7; 2) formation of NHS-
activated ester intermediates 6 and 9 after selective depro-
tection of the peripheral carboxylate groups (NHS = N-
hydroxysuccinimide); and 3) conjugation of c(RGDyK) to
the NHS esters followed by acid deprotection of the inner
carboxylate groups to form the products H21 and H22.
Alkylation of CB-cyclam by 3 was asynchronous at the two
nonbridged nitrogen atoms, with the monoalkylation product
4 predominating at room temperature (35–55% yield even in
the presence of excess of 3).[15] In comparison, the dialkylated
product 7 was only formed at elevated temperatures. At 508C
using two equivalents of 3 to CB-cyclam, only 20–45% of 7
was found, and it was always accompanied by the mono-
alkylation product. Although 7 was difficult to elute from
silica gel using common organic solvents, a good separation
was obtained by adding 5–10% isopropyl amine in ethyl
acetate. It is noteworthy that the debenzylation of the
peripheral carboxylate groups catalyzed by 10% Pd/C in a
hydrogen atmosphere always resulted in formation of the
corresponding esters if the reaction was carried out in an
alcohol solvent. This finding is likely due to the “proton
sponge” nature of the CB-cyclam core that induces transes-
terification during debenzylation. No clean debenzylization
could be accomplished in either methanol or ethanol, even in
the presence of formic acid as described in literature.[15]
However, THF/H2O (1:1) successful circumvented this prob-
lem and afforded debenzylation products in quantitative
yield. These were then activated by NHS and conjugated to
c(RGDyK) in the presence of
ten equivalents N,N-diisopro-
pylethylamine
(DIPEA).
After HPLC purification, H21
and H22 were obtained by
removal of the tert-butyl
groups using 95% trifluoro-
acetic acid (TFA; see the Sup-
porting
details).
Information
for
Both H21 and H22 were
efficiently labeled by 64Cu at
708C in 0.4 m NH4OAc buffer
within 30 min. The specific
activity of [64Cu]1 and [64Cu]2
was in the range of 8.4–
20.4 GBqmmolÀ1. The in vitro
stability of the 64Cu-labeled
peptide conjugates was evalu-
ated in rat serum by radio-
HPLC.
Chromatographic
results showed no release of
64Cu from the conjugates over
a period of 48 h. This high
stability is attributed to the
CB-TE2A moiety in the con-
jugates. The aub3 binding affin-
ities of H21 and H22 were
measured by
cell-binding
a
competitive
assay using
U87MG cells in which 125I-
echistation was employed as
aub3-specific radioligand for
Scheme 2. Synthesis of H21 and H22. Reagents: a-1,a-2) CH3CN, K2CO3; b) BrCH2COOtBu, CH3CN,
K2CO3; c,g) 10% Pd/C, THF/H2O; d,h) MeCN, NHS, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide
(EDC); f,i) c(RGDyK), DIPEA, DMF; e,j) 95% TFA.
Angew. Chem. Int. Ed. 2009, 48, 7346 –7349
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7347

Communications
competitive displacement.[5] The U87MG cell line was chosen
compared to [64Cu]1. The enhanced tumor uptake and
retention of [64Cu]2 may be partially attributed to the
difference of in vivo kinetics of [64Cu]1 and [64Cu]2, given
the higher molecular weight of [64Cu]2. Indeed, [64Cu]2 was
not cleared as efficiently as [64Cu]1 from kidneys ([64Cu]1
94.7 Æ 3.6%ID excreted at 24 h p.i.; [64Cu]2 88.2 Æ 4.9%ID
excreted at 24 h p.i.; p < 0.05). The aub3 binding specificity of
[64Cu]1 and [64Cu]2 was demonstrated by signal loss (Figure 1:
1 h blockade) in tumors after co-injection of c(RGDyK) at a
dose of 10 mgkgÀ1. The quantitative PET image data and
post-PET biodistribution data are presented in Table 1 and
Tables S1–S3 (see the Supporting Information for details).
because the aub3 integrin density on the cell surface was the
highest among the solid tumor cell lines that have been
assessed.[16] The IC50 values of c(RGDyK), H21, and H22,
which represent their concentrations required to displace
50% of the 125I-echistation bound on the U87MG cells, were
determined to be 110, 139, and 35 nm, respectively (n = 5).
The slightly decreased aub3 binding of H21 as compared to
c(RGDyK) indicates a minute impact of CB-TE2A on the
binding of c(RGDyK) to the aub3 integrin. As anticipated,
H22 exhibited a strong divalent effect measured by the
multivalent enhancement ratio (MVE) calculated by dividing
the IC50 value of H21 by that of H22 (MVE = 4).[6] The
distance between the two RGD motifs in H22 is greater than
25 bonds (including the lysine spacers), the minimum spacing
length required to realize multivalent binding of RGD motifs
to the aub3 integrin.[5] It is noteworthy that as a downstream
effect of multivalent binding, oligomerization of cell-surface
receptors could initiate cellular internalization events, which
might further enhance the specific accumulation in the target
tissues.[17]
Table 1: Tumor uptake of [64Cu]1 and [64Cu]2 as determined from PET
imaging quantitation and post-PET biodistribution. Data are presented
as %IDgÀ1 Æstandard deviation (n=3).[a]
1 h PET
4 h PET
24 h
Post-PET
1 h PET
(blockade)
PET
[64Cu]1 1.95Æ0.10 1.85Æ0.26 1.10Æ0.15 1.39Æ0.17 0.31Æ0.05
[64Cu]2 2.92Æ0.26 2.40Æ0.22 1.72Æ0.18 1.79Æ0.13 0.71Æ0.04
[a] For [64Cu]2, p<0.05 for all times.
In vivo small-animal imaging studies were performed on a
Siemens Inveon PET-CT multimodality system. Six SCID
mice (6–7 weeks old) bearing PC-3 human prostate cancer
xenografts in both front flanks (tumor size ca. 230 mg) were
randomized into two groups (n = 3) for the evaluation of
[64Cu]1 or [64Cu]2, which was injected into the tail vein. As
shown in Figure 1, both tumors were visualized by [64Cu]1 and
[64Cu]2 at 1 and 4 h post injection (p.i.), while [64Cu]2 showed
a significantly stronger PET signal than [64Cu]1 at all time
points. At 24 h p.i., the tumors were still clearly visible with
[64Cu]2 but were rather faint with [64Cu]1. Owing to the fact
that the aub3 integrin is also expressed in other tissues (e.g.
liver, kidneys, stomach, intestines) in young mice, but to a
lesser extent (personal communications[18]), an elevated
uptake was observed in those organs with [64Cu]2 as
The significantly greater uptake and prolonged signal
intensity of [64Cu]2 in tumors reflects the advantages of the
scaffolding design of H22, which affords optimal in vivo
kinetics in addition to the anticipated multivalent effects. It
should be pointed out that there are two chiral centers in the
pendent arms of Cu2, which should statistically give rise to
three diastereomers (R/R, S/S, and a meso R/S), even though
they could not be distinguished by the techniques used herein.
While the purpose of this work is to demonstrate the
feasibility of building multivalent imaging probes from a
bifunctional chelator, an enantiopure isomer of 3 should be
considered for future clinical applications of this type of
multivalent scaffold. Used as a sample targeting molecule
herein, c(RGDyK) can obviously be replaced with other
targeting peptides or small organic molecules for imaging of
various diseases or non-invasive cell-surface receptor map-
ping. In summary, we have demonstrated a divalent scaffold-
ing design for targeted imaging probe development. Obvi-
ously this concept can be applied to the design of other
multivalent scaffolds based on NOTA (1,4,7-triazacyclono-
nane-1,4,7-triacetic acid) or DOTA (1,4,7,10-tetraazacyclo-
dodecane-1,4,7,10-tetraacetic acid).
Received: June 30, 2009
Revised: July 18, 2009
Published online: September 1, 2009
Keywords: chelates · imaging agents · multivalent effect ·
.
radiochemistry
Figure 1. Representative microPET-CT images of PC-3 tumor bearing
mice (n=3) at 1, 4, 24 h after intravenous injection of [64Cu]1 (upper
panel) and [64Cu]2 (lower panel). Images obtained with co-injection of
10 mgkgÀ1 of c(RGDyK) are only shown for 1 h blockade (right).
Arrows indicate tumors.
[1] M. Allen-Auerbach, W. A. Weber, Oncologist 2009, 14, 369.
[2] R. Rossin, S. Muro, M. J. Welch, V. R. Muzykantov, D. P.
Schuster, J. Nucl. Med. 2008, 49, 103.
[3] R. Haubner, H. J. Wester, Curr. Pharm. Des. 2004, 10, 1439.
7348
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7346 –7349

Angewandte
Chemie
[4] A. Almutairi, R. Rossin, M. Shokeen, A. Hagooly, A. Ananth, B.
Capoccia, S. Guillaudeu, D. Abendschein, C. J. Anderson, M. J.
Welch, J. M. Frechet, Proc. Natl. Acad. Sci. USA 2009, 106, 685.
[5] Z. B. Li, W. Cai, Q. Cao, K. Chen, Z. Wu, L. He, X. Chen, J. Nucl.
Med. 2007, 48, 1162.
[6] X. Montet, M. Funovics, K. Montet-Abou, R. Weissleder, L.
Josephson, J. Med. Chem. 2006, 49, 6087.
[7] I. Dijkgraaf, J. A. Kruijtzer, S. Liu, A. C. Soede, W. J. Oyen, F. H.
Corstens, R. M. Liskamp, O. C. Boerman, Eur. J. Nucl. Med.
Mol. Imaging 2007, 34, 267.
[12] X. Sun, M. Wuest, G. R. Weisman, E. H. Wong, D. P. Reed, C. A.
Boswell, R. Motekaitis, A. E. Martell, M. J. Welch, C. J. Ander-
son, J. Med. Chem. 2002, 45, 469.
[13] K. S. Woodin, K. J. Heroux, C. A. Boswell, E. H. Wong, G. R.
Weisman, W. Niu, S. A. Tomellini, C. J. Anderson, L. N.
Zakharov, A. L. Rheingold, Eur. J. Inorg. Chem. 2005, 4829.
[14] M. Woods, P. Caravan, C. F. Geraldes, M. T. Greenfield, G. E.
Kiefer, M. Lin, K. McMillan, M. I. Prata, A. C. Santos, X. Sun, J.
Wang, S. Zhang, P. Zhao, A. D. Sherry, Invest. Radiol. 2008, 43,
861.
[8] I. Dijkgraaf, A. Y. Rijnders, A. Soede, A. C. Dechesne, G. W.
van Esse, A. J. Brouwer, F. H. Corstens, O. C. Boerman, D. T.
Rijkers, R. M. Liskamp, Org. Biomol. Chem. 2007, 5, 935.
[9] M. Janssen, W. J. Oyen, L. F. Massuger, C. Frielink, I. Dijkgraaf,
D. S. Edwards, M. Radjopadhye, F. H. Corstens, O. C. Boerman,
Cancer Biother. Radiopharm. 2002, 17, 641.
[15] C. A. Boswell, C. A. Regino, K. E. Baidoo, K. J. Wong, A.
Bumb, H. Xu, D. E. Milenic, J. A. Kelley, C. C. Lai, M. W.
Brechbiel, Bioconjugate Chem. 2008, 19, 1476.
[16] X. Zhang, Z. Xiong, Y. Wu, W. Cai, J. R. Tseng, S. S. Gambhir, X.
Chen, J. Nucl. Med. 2006, 47, 113.
[17] J. E. Gestwicki, C. W. Cairo, L. E. Strong, K. A. Oetjen, L. L.
Kiessling, J. Am. Chem. Soc. 2002, 124, 14922.
[10] S. Liu, Mol. Pharm. 2006, 3, 472.
[11] W. Cai, S. S. Gambhir, X. Chen, Methods Enzymol. 2008, 445,
141.
[18] S. Liu, unpublished results.
Angew. Chem. Int. Ed. 2009, 48, 7346 –7349
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7349