Controlled axial coordination: Solid-phase synthesis and purification of metallo-radiopharmaceuticals

Article abstract of DOI:10.1002/anie.200801936

(Chemical Equation Presented) No sweat PET: Controlled axial coordination of pseudo-square-planar zinc(II) complexes to 4-(dimethylamino)pyridine- functionalized polystyrene resin facilitates the synthesis (by transmetalation) and purification (by selective ligation) of metallo-radiopharmaceuticals. These complexes have potential as positron emission tomography (PET) imaging agents.

Full text of DOI:10.1002/anie.200801936

Communications
DOI: 10.1002/anie.200801936
Radiochemical Synthesis
Controlled Axial Coordination: Solid-Phase Synthesis and Purification
of Metallo-Radiopharmaceuticals**
Helen M. Betts, Peter J. Barnard, Simon R. Bayly, Jonathan R. Dilworth,* Antony D. Gee, and
Jason P. Holland
Solid-supported reagents show great potential for improving
the synthesis of radiodiagnostic agents, in terms of radio-
chemical yield and purity, as well as convenience and safety.
The preparation of the positron emission tomography (PET)
imaging agent [18F]-2-fluoro-2-deoxy-d-glucose has recently
been demonstrated using a solid-bound substrate which is
selectively cleaved from the solid support upon reaction with
[18F]-fluoride ions.[1] Whilst this covalent approach is appro-
priate for the nucleophilic substitution chemistry of fluoride
ions, its utility with metallonuclides is limited and the only
examples to date have used 99mTc in conjunction with
proligands attached to resins[2,3] and gold surfaces.[4] Conven-
tional solid-phase synthesis is therefore limited by the
requirement for proligands possessing a donor group which
can both be covalently attached to the solid support and
cleaved from it upon coordination to the desired metal ion.
Herein, we report a novel strategy for solid-phase syn-
thesis based on selective axial coordination of ZnII substrates
to 4-(dimethylamino)pyridine(DMAP)-functionalized poly-
styrene resin[5] and exemplify its use in the preparation and
purification of known and potential 64Cu and 99mTc radio-
pharmaceuticals. This strategy is applicable to a wide range of
metallic radionuclides and is suitable for the macrocyclic
ligand systems that are favored in nuclear medicine because
of their high in vivo stability.
Jahn–Teller distortion disfavors the coordination of a fifth
donor atom in the axial site of analogous CuII complexes. A
ZnII precursor can be bound to polymer-supported DMAP by
axial coordination. Upon transmetalation of the ZnII complex
with CuII, the coordinate bond to the solid support is broken
and only the transmetalated complex is released from the
resin (Scheme 1a). Similarly, polystyrene-supported DMAP
can be used to selectively bind the ZnII complex from a
solution-phase reaction mixture leaving only the radiolabeled
complex in solution (Scheme 1b).
ZnII complexes of tetradentate ligands that are con-
strained in pseudo-square-planar conformations can poten-
tially bind a fifth donor atom in an axial coordination site.[6]
Scheme 1. a) Transmetalation of ZnII complex with Cu2+ ions on a
solid support and release of a CuII complex into solution. b) Selective
binding of a ZnII complex to DMAP-modified resin in the presence of a
CuII complex.
[*] H. M. Betts, Dr. P. J. Barnard, Dr. S. R. Bayly, Prof. J. R. Dilworth,
Dr. J. P. Holland
Chemistry Research Laboratory, University of Oxford
Mansfield Road, Oxford OX1 3TA (UK)
Fax: (+44)1865-272690
Initially a range of potential donor systems were assessed
for binding to the axial site of [ZnII(atsm)] (1; atsm =
E-mail: jon.dilworth@chem.ox.ac.uk
diacetyl-bis(N4-methyl-3-thiosemicarbazonato)
zinc(II))
Dr. S. R. Bayly, Prof. J. R. Dilworth
Siemens Oxford Molecular Imaging Laboratory, University of Oxford
South Parks Road, Oxford OX1 3QR (UK)
which is a precursor to the hypoxia-selective complex [CuII-
(atsm)] (2).[7] Phosphine, thiolate, and nitrogen-donor mole-
cules were tested by preparing a 1:1 mixture of the donor
molecule with 1 in THF, acetone, or chloroform, and the
mixture was analyzed by reverse-
Prof. A. D. Gee
PET and Radiotracer Development GSK Clinical Imaging Centre
Hammersmith Hospital, Du Cane Road, London, W12 ONN (UK)
[**] H.M.B. thanks GlaxoSmithKline for funding. P.J.B. thanks Glaxo-
SmithKline for a fellowship. S.R.B. thanks DTI and Siemens
Molecular Imaging Ltd for a fellowship. J.P.H. thanks Merton
College and the EPSRC for funding. We thank Prof. Jennifer Green
and Philip Waghorn for helpful discussions, Dr. Christoph Salzmann
for assistance with Raman spectroscopy, and the Oxford Super-
computing Centre.
phase
HPLC
(acetonitrile/water
mobile phase). Only nitrogen-donor
molecules bound under these condi-
tions.
The binding of DMAP to 1 in the
solution phase was investigated by
NMR spectroscopy. Aliquots of a
solution of DMAP were added to a
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801936.
8416
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8416 –8419

Angewandte
Chemie
solution of 1 in 10% [D6]DMSO/[D6]acetone, and the
chemical shift of the methyl protons on the bis(thiosemicar-
bazone) backbone was monitored. A logK value of (2.73 Æ
0.32) for the formation of a 1:1 adduct was calculated using
WinEQNMR software.[9] The binding constant is relatively
small;[10] however, this may be a result of DMSO (added to
ensure all species remained in solution) competing for the
binding site.
Similar DFT calculations were performed on a range of
nitrogen- and oxygen-donor ligands to assess their potential
to bind to the axial site. All coordination reactions were found
to be enthalpically favorable, but only [1–DMAP] complex
formation is calculated to be thermodynamically spontaneous
in the gas phase (DrG = À4.9 kJmolÀ1). Calculations including
solvation[15] using a water-polarizable continuum model show
the same trend as the gas phase, but all complexation
reactions become spontaneous. DFT calculations support
the selection of DMAP as the preferred nitrogen donor for 1.
Electronic absorption spectroscopy (UV/Vis) confirmed
the selective binding of 1 to DMAP-functionalized polystyr-
ene resin in the presence of 2. The spectrum of a 1:1 mixture
of the ZnII and CuII complexes was recorded in ethanol, and
the spectrum remeasured after agitation of the solution with
DMAP resin (20 mg) for 10 min (Figure 2a). After treatment
with the resin, only the spectrum of 2 was observed, indicating
The X-ray crystal structure of the [1–DMAP] adduct is
shown in Figure 1.[11] The ZnII ion lies in a pseudo-square-
pyramidal coordination geometry, and is displaced by 0.517 ꢀ
Figure 1. X-ray crystal structure of [1–DMAP]. Selected geometric
parameters of the X-ray [DFT-calculated] structures in ꢀ: Zn–
S2=2.3946(6) [2.421], Zn–N5=2.112(2) [2.169], Zn–N8=2.117(2)
[2.116], Zn–S11=2.3725(7) [2.422], Zn–N18=2.0688(19) [2.114], Zn–
N2S2 plane=0.517 [0.559]. Estimated standard deviations (esd) are
shown in parentheses, DFT calculations were carried out using the
B3LYP/631++G(d,p) water-phase model.
above the plane of best fit defined by the bis(thiosemicarb-
azonato) N2S2 donor atoms. Both gas- and solution- phase
density functional theory (DFT)[12] optimized structures of
the adduct are in excellent agreement with X-ray crystal
structure with a weighted root mean square deviation
(RMSD) calculated for all 26 heavy atoms of only 0.505 ꢀ
(see the Supporting Information). In the X-ray crystal
structure the Zn–N18 bond is shorter (2.0688(19) ꢀ) than
the other two bonds to N5 and N8, which is indicative of
relatively strong axial ligation. Natural bond orbital (NBO)
analysis[13] shows that the electron donation from the DMAP
nitrogen-donor-atom lone pair, N18, to the ZnII ion contrib-
utes approximately 11% (163 kJmolÀ1) of the total stabiliza-
tion (1453 kJmolÀ1) from the five donor atoms, and is
comparable to the donation from nitrogen atoms N5 and
N8. Donation from the sulfur atoms S2 and S11 is stronger
and contributes a total of 65% (947 kJmolÀ1) of the ligand-to-
Figure 2. a) UV/Vis spectrum of a 1:1 mixture of 1 (0.025 mm) and 2
(0.025 mm) (solid line), spectrum after addition of DMAP resin
(dashed line), and calculated difference spectrum (dotted line).[8]
b) Powder-phase Raman spectra recorded with an excitation wave-
length of 632 nm, of DMAP resin (top), loaded [1-DMAP] (middle) and
1 (bottom).
À
metal stabilization energy. Given the strength of these Zn S
bonds it is surprising that thiolate ligands showed no affinity
for coordination to the axial site.
Angew. Chem. Int. Ed. 2008, 47, 8416 –8419
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8417

Communications
that 1 was removed quantitatively by binding to the solid
phase. The calculated difference spectrum corresponds to that
of 1. The powder Raman spectrum of 1-loaded DMAP resin
was also recorded (Figure 2b). Comparison with the spectra
of 1 and the unloaded resin shows that characteristic bands of
1 are present in the loaded resin, and provides further
evidence of the binding of the ZnII complex to the solid
support.
Transmetalation of ZnII complexes is a convenient method
to prepare 64Cu complexes for biological testing and PET.[16]
However, as low concentrations (ca. 10À9 m) of 64Cu are used,
a large excess of the precursor of the ZnII complex (ca. 106-
fold) is usually present as a contaminant in the radiolabeled
solution. Rapid radiosynthesis of 64Cu complexes that are free
from the ZnII precursor were achieved by transmetalation of
the preloaded DMAP resin. The ZnII complex was loaded
onto the resin by stirring in a methanol/acetone (1:9) solution
at room temperature. A prototype cartridge was prepared by
packing 50 mg of ZnII-loaded resin into a syringe, and
64Cu(OAc)2 solution was added. The cartridge was eluted
with 200 mL ethanol/water (1:1) after 5 min. HPLC analysis
(radio and UV detection) of the eluted product showed that
[64Cu]-2 and glucose-derivatized bis(thiosemicarbazonato)
complex [64Cu]-4 (Figure 3) were prepared using this proce-
dure in greater than 97% radiochemical purity. The ZnII
Figure 3. Overlay of the radio-HPLC chromatograms of [64Cu]-4
(tR =7.05 and 7.12 min),[14] [64CuL1] (tR =5.90 min), and [99mTcOL1]
(tR =6.65 min). Chromatograms have been normalized and offset for
clarity. tR =Retention time.
to a monoxo-TcV complex of the macrocycle H2L1. Excess
[ZnL1] was removed quantitatively postradiolabeling, by the
addition of DMAP resin.
Routine preparation of [64Cu]-2 for in vivo studies uses
excess proligand, but a separation step is not performed.[21]
The effect of unlabeled proligand on the biodistribution of the
complex is unknown. For receptor-targeting complexes such
as those containing bioactive peptides or monoclonal anti-
bodies, unlabeled precursor in the solution can saturate the
target receptor sites and lead to reduced signals in imaging
experiments.[22] It is desirable that the labeling of such
biomolecules does not significantly perturb their physico-
chemical properties. However, this makes their separation
from unlabeled precursor difficult and time consuming even
with techniques such as preparative HPLC. The solid-phase-
synthesis strategy described above, by trapping unlabeled
substrate by axial coordination, has the potential to produce
solutions with highly specific activity in a single, rapid step.
The procedure is not restricted to radiosynthesis, and could
potentially be used in a range of applications that involve
pseudo-square-planar ZnII complexes. We are currently
investigating the use of this technology for different metal
ions and ligand systems, multistep solid-supported synthesis,
and purification of porphyrins.
precursor was absent according to the HPLC trace (UV
detection).[17] The ethanol/water eluate is biologically com-
patible and suitable for immediate formulation in saline
solution for in vivo studies, with no further purification steps
required.
A postradiolabeling procedure was developed for com-
plexes where transmetalation kinetics of the solid-bound ZnII
complex are slow. This method was used for complexes of the
macrocyclic
ligand
2,10-dioxo-1,4,8,11-tetraazabicyclo-
[11.4.0]1,12-heptadeca-1(12),14,16-triene (H2L1). [64CuL1] is
of interest as a potential radiopharmaceutical agent as the
CuII center is not reduced to CuI at biologically relevant
potentials.[18] Enzymatic reduction followed by loss of CuI
from macrocyclic chelators is thought to be a major pathway
of decomposition in vivo.[19,20] [64CuL1] was prepared in
greater than 98% radiochemical purity by transmetalation
of [ZnL1] with 64Cu(OAc)2 in methanol solution. Excess
[ZnL1] was subsequently removed from the mixture by
addition of DMAP resin (20 mg), as confirmed by HPLC.
In addition, the complex [99mTcOL1] was prepared by trans-
metalation from [ZnL1] with [99mTcO4]À in the presence of the
reductant tin(II) chloride. HPLC analysis of the resultant
solution showed that a single 99mTc species was formed in
90% radiochemical purity (Figure 3). This species is assigned
Received: April 24, 2008
Revised: July 26, 2008
Published online: September 4, 2008
Keywords: coordination modes · copper · radiochemistry ·
.
radiopharmaceuticals · solid-phase synthesis
[1] L. J. Brown, D. R. Bouvet, S. Champion, A. M. Gibson, Y. Hu,
A. Jackson, I. Khan, N. Ma, N. Millot, H. Wadsworth, R. C. D.
Brown, Angew. Chem. 2007, 119, 959; Angew. Chem. Int. Ed.
2007, 46, 941.
8418
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8416 –8419

Angewandte
Chemie
[2] R. W. Riddoch, P. Schaffer, J. F. Valliant, Bioconjugate Chem.
2006, 17, 226.
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
[3] S. Mundwiler, L. Candreia, P. Hafliger, K. Ortner, R. Alberto,
Bioconjugate Chem. 2004, 15, 195.
[12] M. J. T. Frisch et al., Gaussian03 Revision C.02, see the Support-
ing Information.
[4] A. Pollak, D. G. Roe, C. M. Pollock, L. F. L. Lu, J. R. Thornback,
J. Am. Chem. Soc. 1999, 121, 11593.
[13] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899.
[14] The two peaks correspond to anomers of glucose and are
consistent with the results of previous solution-phase radio-
labeling experiments.
[15] J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105, 2999.
[16] J. P. Holland, F. I. Aigbirhio, H. M. Betts, P. D. Bonnitcha P.
Burke, M. Christlieb, G. C. Churchill, A. R. Cowley, J. R.
Dilworth, P. S. Donnelly, J. C. Green, J. M. Peach, S. R. Vasuve-
dan, J. E. Warren, Inorg. Chem. 2007, 46, 465.
[17] The lowest concentration of 1 and 3 observable by the UV
detector (254 nm) is 1.0 mm.
[18] A. Dey, F. E. Jenney, M. W. W. Adams, E. Babini, Y. Takahashi,
K. Fukuyama, K. O. Hodgson, B. Hedman, E. I. Solomon,
Science 2007, 318, 1464.
[5] Commercially available 4-(dimethylamino)pyridine on polystyr-
ene cross-linked with divinylbenzene (2%). The loading of
DMAP was approximately 3 mmolgÀ1
.
[6] Y. R. De Miguel, N. Bampos, K. M. Nalin de Silva, S. A.
Richard, J. K. M. Sanders, Chem. Commun. 1998, 2267.
[7] A. L. Vavere, J. S. Lewis, Dalton Trans. 2007, 4893.
[8] [ZnII(atsm)] (1): UV/Vis (EtOH): lmax (e/molÀ1 m3 cmÀ1) = 427
(10942), 309 nm (11466). [CuII(atsm)] 2: UV/Vis (EtOH): lmax
(e/molÀ1 m3 cmÀ1) = 525 sh (3569), 478 (6036), 355 sh (9846), 311
nm (19864).
[9] M. J. Hynes, J. Chem. Soc. Dalton Trans. 1993, 311. See the
Supporting Information for details.
[10] A. Satake, Y. Kobuke, Tetrahedron 2005, 61, 13.
[11] [1–DMAP] ꢀ C15H24N8S2Zn, Mr = 445.95. Monoclinic P21/n, a =
11.0951(3), b = 13.9392(4), c = 13.2449(4) ꢀ, b = 97.6854(13)8,
[19] 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.
[20] L. A. Bass, M. Wang, M. J. Welch, C. J. Anderson, Bioconjugate
Chem. 2000, 11, 527.
V= 2030.01(10) ꢀ3,
1calcd
(Z = 4) = 1.459 MgmÀ3
,
m =
1.432 mmÀ1, specimen: 0.07 ꢁ 0.18 ꢁ 0.63 mm, 2qmax = 27.58 l =
0.71073 ꢀ, Nt = 14405, N = 4572 (Rint = 0.042), No = 3521, R
[21] F. Dehdashti, P. W. Grigsby, J. S. Lewis, R. Laforest, B. A. Siegel,
M. J. Welch, J. Nucl. Med. 2008, 49, 201.
(I > 2s(I)) = 0.0357, wR
(I > 2s(I)) = 0.0639, T= 150 K.
CCDC 686182 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
[22] E. M. Jagoda, J. J. Vaquero, J. Seidel, M. V. Green, W. C.
Eckelman, Nucl. Med. Biol. 2004, 31, 771.
Angew. Chem. Int. Ed. 2008, 47, 8416 –8419
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8419