Structure and stability of hexadentate complexes of ligands based on AAZTA for efficient PET labelling with gallium-68

Article abstract of DOI:10.1039/c2cc37544c

Pre-organised tricarboxylate ligands based on 6-amino-perhydro-1,4- diazepine bind ⁶⁸Ga rapidly and selectively in acetate buffer at pH 4 to 7, forming kinetically stable complexes suitable for use in PET imaging.

Full text of DOI:10.1039/c2cc37544c

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Structure and stability of hexadentate complexes of
ligands based on AAZTA for efficient PET labelling
with gallium-68†
Bradley P. Waldron,a David Parker,*a Carsten Burchardt,b Dmitry S. Yufit,a
Melanie Zimnyb and Frank Roeschb
Cite this: Chem. Commun., 2013,
49, 579
Received 16th October 2012,
Accepted 13th November 2012
DOI: 10.1039/c2cc37544c
www.rsc.org/chemcomm
Pre-organised tricarboxylate ligands based on 6-amino-perhydro-
1,4-diazepine bind 68Ga rapidly and selectively in acetate buffer at
pH 4 to 7, forming kinetically stable complexes suitable for use in
PET imaging.
Scheme 1 Schematic diagram8 of solution conformers for [H2AAZTA]2À
We present a series of hexadentate tribasic ligands based on
perhydro-1,4-diazepine that bind 68Ga rapidly in the pH range
4–7, forming radiolabelled complexes suitable for imaging
studies using positron emission tomography.1
(equilibrium shifts right to reduce repulsion).
accessible by N-functionalisation. Indeed, the archetypal ligand
There are many examples of aza-carboxylate (e.g. NOTA)2 and AAZTA, 2, (pKa 11.23, 6.52)5c forms stable complexes in water
related aza-phosphinate ligands based on 1,4,7-triazacyclononane,3,4 with lanthanide (log K ca. 20) and calcium ions (log K 12.8, 298 K,
that form stable 1 : 1 complexes with small metal ions. An I = 0.1 KCl). The relative steric demand of the methyl and
octahedral coordination geometry is favoured, for example with N-substituents at the 6-position means that AAZTA, as a
Ga3+, Fe3+, Mn2+ and Zn2+. This has drawn the attention of such di-N-protonated ligand, is not pre-organised for metal binding
ligands for radiolabelling studies, e.g. with 67/68Ga, and 111In. (Scheme 1). The methyl or related primary alkyl groups (A value
Recently, the coordination chemistry of various ligands derived ca. 7 kJ molÀ1) have a lower steric demand than the protonated
from 6-amino-6-methyl-perhydro-1,4-diazepine, 1, has also been nitrogen substituent, (A value ca. 10 kJ molÀ1). Hence, it prefers
studied.5,6
to take up the pseudo-axial site (Scheme 1) so that the ligand
presents isolated EDDA (ethylenediamine diacetate) and imino-
diacetate binding units in the major conformer, that could lead
to kinetically trapped complexes, under conditions that do not
allow formation of the most stable complex by cooperative
donor ligation. Such conditions occur during radiolabelling of
ligands with the positron emitting isotope 68Ga, (t1/2 = 67.7 min)
as the eluate from a 68Ge generator is acidic (pH 0.2 to 1.3).7
Moreover, gallium aqua species tend to hydrolyse above pH 6.
With this background in mind, we have compared AAZTA
with ligands L1–4, examining the efficiency of radiolabelling as a
function of pH, as well as assessing the relative stability of 68Ga
complexes. The selection of the new ligands L3 and L4 was
made in order to introduce a bulkier phenyl substituent (A value
11.7 kJ molÀ1) at the 6-position, favouring population of the
desired metal-binding conformation (Fig. 1).
The parent triamine 1 is constitutionally isomeric with 1,4,7-
triazacyclononane and presents three N atoms that serve as
donors in a facial array. The presence of the primary amine
group means that heptadentate ligands are particularly
a Department of Chemistry, Durham University, South Road, Durham, DH1 3LE, UK.
E-mail: david.parker@dur.ac.uk; Fax: +44 (0)191-3844737; Tel: +44 (0)191-3342033
b Institute of Nuclear Chemistry, University of Mainz, Fritz-Strassmann-Weg 2,
55128 Mainz, Germany
Ligands L1 and L2 were made following established methods.5,6
In modifying the synthesis for L3 and L4, a key issue was the need
to remove the N-benzyl groups and reduce the nitro group in the
cyclic intermediate, 3, without ring opening, e.g. by concomitant
hydrogenolysis at the quaternary centre (Scheme 2). By using a
† Electronic supplementary information (ESI) available: Ligand and complex
synthesis and characterisation; pH/1H NMR titration data; radiochemical labelling
details; X-ray structures of the 3 gallium complexes as well as two intermediates in
the synthesis of L1 and L3. CCDC 906036–906040. For ESI and crystallographic data
in CIF or other electronic format see DOI: 10.1039/c2cc37544c
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 579--581 579

View Article Online
Communication
ChemComm
Fig. 1 Hexadentate ligands based on perhydro-1,4-diazepine, shown in their
likely metal ion binding conformation.
Scheme 2 Synthetic scheme for [GaL3].
2,4-dimethoxybenzyl group in place of a simple benzyl moiety,
debenzylation with TFA was achieved at room temperature;
subsequent reduction of the nitro group could be undertaken by
RANEYs nickel hydrogenation, averting solvolysis at the benzylic
site, with loss of the nitro group. The more bulky phenyl ring in
L3 suppresses the lactamisation reaction that occurs readily in
acidic aqueous media between an endocyclic N-acetate group
and the secondary amine site. This has been observed to occur
readily, even at pH 3 for the analogue of L3 (Me at the quaternary
centre), and is apparent in aged aqueous solution samples of L2.
The gallium(III) complexes of L1–L3 crystallised readily from
aqueous solution at pH 4. Their molecular structures were deter-
mined by X-ray crystallography at 120 K. Complexes [GaÁL1] and
[GaÁL2] crystallized as hemi- and monohydrates respectively;
crystals of complex [GaÁL3] contain two almost identical crystallo-
graphically independent molecules.‡ Each gallium ion is coordi-
nated by the N3O3 donors forming charge neutral complexes,
and the geometry around the Ga(III) ion is a slightly distorted
octahedron (Table 1 and Fig. 2).
Fig. 2 Molecular structures of gallium(III) complexes (120 K): upper: [GaÁL1];
centre: (SSS) [GaÁL2]; lower: [GaÁL3].
Table 2 Radiolabelling yields (%, s.d. Æ1)) and times (min, in parenthesis) for
labelling of AAZTA and L1–4 (295 K, [68Ga = 0.66 nM (t = 0); [L] = 10 mM;
1.4 mL) (ESI)
pH
AAZTA
L1
L2
L3
L4
4.0b,c
97a
(1)
98a
(1)
99
(3)
98
(3)
98
(3)
97
(3)
98
(3)
97
(3)
99
(3)
99
(3)
97
(3)
5.3
6.8
96a
(1)
97a
(1)
96a
(3)
97a
(1)
a
Formation of more than one species occurred, e.g. for AAZTA at each
pH, three species were observed and ratio was typically 9 : 9 : 1, and for
b
L1, 18 : 1. At pH 3.3 ([68Ga] = 0.2 nM, [L] = 10 mM] > 98% labelling was
achieved in every case within 10 min at 323 K, but multiple species were
observed notably with AAZTA (2 : 2 : 1) and L1 (9 : 1), of differing
c
relative stability. Ligands L1–4 also exhibited quantitative labelling
The radiolabelling performance of ligands L1–4 was assessed in
comparison to AAZTA at pH 4.0, 5.3 (acetate buffer, 0.2 M) and
6.8 (1 M HEPES), using 100 MBq 68Ga (0.66 nM) and a ligand con-
centration of 10 mM, (Table 2). With AAZTA, three radiolabelled
at pH 4 with less ligand (298 K, 5 mM [L], 0.66 nM [68Ga].
species were observed, whose relative ratio varied with pH; in con-
trast, with L2–4, only one species was formed at each pH examined.
The relative stability of the radiolabelled complexes was
assessed by diluting the labelled complexes in fresh PBS buffer
(1 M, pH 7.4, 310 K, [GaL] = 0.16 nM), comparing behaviour as a
function of time by radio-TLC. Similar protocols were used in
the presence of either excess DTPA, Fe(III) (each 5000 eqs), apo-
transferrin (130 eqs) or in foetal calf serum. In each case, the
complexes of ligands L2–4 prepared at pH 4 and above resisted
these challenges completely, over a 2 h period.
Table 1 Selected bond distances (Å) and geometric parameters for [GaÁL1–3
]
(120 K;11,12 for numbering system, see Fig. 2) (ESI)
[GaÁL1]
[GaÁL2]
[GaÁL3]
Ga–O(1)
Ga–O(3)
Ga–O(5)
Ga–N(1)
Ga–N(2)
Ga–N(3)
Sa
1.939(2)
1.904(2)
1.933(2)
2.140(2)
2.138(2)
2.111(2)
9.27
1.967(1)
1.895(2)
1.934(2)
2.158(2)
2.150(2)
2.075(2)
8.72
1.945(4)
1.912(4)
1.918(4)
2.128(4)
2.144(5)
2.091(4)
10.3
However, with labelled AAZTA, the minor species formed
were much less stable and did not resist the DTPA and serum
challenge. This behaviour was even more pronounced for pre-
parations made at pH 2.3 or 3.3. Comparing sets of complexes
labelled at pH 2.3, the phenyl-substituted pair of ligands, L3/4
Yb
10.7
16.9
13.8
a
Octahedral distortion parameter S = S(|90 À ji|)/12 [S = 01 for an ideal
octahedron; ji represents the 12 smallest L–M–L angles].9 b Average
trigonal distortion angle Y = S(|60 À i|)/24 [Y = 01 for an ideal octahedron;
i represents the trigonal angles of the eight faces of the octahedron].10
showed better stability profiles than their methyl analogues, L1/2
,
c
580 Chem. Commun., 2013, 49, 579--581
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
4501 unique (Rint = 0.0433) used in all calculations. Final R1 0.0322
by factors of 2 to 6, with 50% of the observed species unchanged.
The formation of multiple species with AAZTA (and to a lesser
extent with L1) of differing relative stability is consistent with the
formation of ‘kinetically trapped’, constitutionally isomeric com-
plexes (Scheme 1). These less stable complexes may involve weaker
binding to the EDDA moiety. For AAZTA, the adoption of N2O4 as
well as the favourable N3O3 coordination type may occur competi-
tively. Importantly, with L2–4, the formation of a single, major stable
gallium bound species is most probably assisted by adoption of a
favourable, ‘pre-organised’ conformer of the di-N-protonated
ligand, in which the exocyclic N-substituent adopts an axial site.
These gallium-labelling characteristics compare favourably
with the behaviour reported for related acyclic and macrocyclic
hexadentate ligands.3,4,13–15 Indeed, in separate challenge
experiments with L1–4 and NOTA (20 mM of each ligand, 0.66 nM
[68Ga], 298 K, pH 4.0, 0.2 M acetate), the acyclic ligands were each
bound in preference within 1 minute and the order of preferential
binding was L3 > L1 > L2 > L4. The ratio of the gallium labelled
L3/NOTA complexes was 3 : 1 under these conditions and did
not vary thereafter with time, reflecting the relative rates of
formation of these kinetically stable complexes.
(>2s(I)); wR2 0.0702 (all data). [GaÁL3]: C17H20GaN3O6, M = 432.08,
%
triclinic, space group P1,
a = 10.6961(8), b = 13.0778(12), c =
13.8414(12) Å, a = 116.076(9), b = 105.208(7), g = 92.002(7)1, V =
1652.8(2) Å3, Z = 4, m(Mo Ka) = 1.709 mmÀ1, Dcalc = 1.736 g mmÀ3
,
16 836 reflections measured (5.04 r 2Y r 55), 7571 were unique (Rint
=
0.0881) and used in all calculations. Final R1 0.0699 (>2s(I)); wR2 0.1864
(all data).
1 For recent examples describing positron emission tomography
using low MW complexes of 68Ga: (a) F. Roesch and P. J. Riss, Curr.
Top. Med. Chem., 2010, 10/16, 1633; (b) I. Velikyan, H. Maecke and
B. Langstrom, Bioconjug. Chem., 2008, 19, 569; (c) P. J. Riss,
C. Burchardt and F. Roesch, Contrast Media Mol. Imaging, 2011,
6, 492.
2 (a) K. Wieghardt, U. Bossek, P. Chaudhuri, W. Herrmann, B. C.
Menke and J. Weiss, Inorg. Chem., 1982, 21, 4308; (b) A. S. Craig,
D. Parker, H. Adams and N. A. Bailey, J. Chem. Soc., Chem. Commun.,
1989, 1793; (c) C. J. Broan, J. P. L. Cox, A. S. Craig, R. Kataky,
D. Parker, A. Harrison, A. M. Randall and G. Ferguson, J. Chem. Soc.,
Perkin Trans. 1, 1991, 87; (d) A. Harrison, C. A. Walker, K. A. Pereira,
L. Royle, R. C. Matthews, D. Parker and A. S. Craig, Nucl. Med.
Commun., 1992, 13, 667; (e) J. P. Andre, H. R. Maecke, M. Zehnder,
L. Macko and K. G. Akyel, Chem. Commun., 1998, 1301;
( f ) E. T. Clarke and A. E. Martell, Inorg. Chim. Acta, 1991, 181, 273.
3 E. Cole, R. C. B. Copley, J. A. K. Howard, D. Parker, G. Ferguson,
J. F. Gallagher, B. Kaitner, A. Harrison and L. Royle, J. Chem. Soc.,
Dalton Trans., 1994, 1619.
Finally, preparations of [68GaL1] and [68GaL4] were injected
into Sprague-Dawley rats to assess preliminary biodistribution
behaviour. The compounds appear to be biologically inert. The
only organs detectable by positron emission tomography were
the liver, kidneys and bladder, with no evidence for complex
retention in any other organ. Twenty-five minutes after admini-
stration of [68GaL1] via the tail vein, the signal was only
detectable by positron emission tomography in the kidney
and bladder, with no evidence for retention in any other organ
(ESI†). Such behaviour is consistent with the high kinetic
stability profile observed in vitro, and suggests that these
complexes offer scope as efficient and effective imaging probes,
and may be simply adapted structurally to allow the labelling of
biomolecules. Current work is exploring this behaviour.
In conclusion, a new series of hexadentate ligands has
been created, suitable for radiolabelling with 68Ga over the
pH range 4 to 7. Using a ligand with an aryl substituent at the
quaternary site, a preferred binding conformation is adopted
that suppresses the formation of less stable, kinetically trapped
complexes.
4 (a) J. Notni, P. Hermann, J. Havlickova, J. Kotek, V. Kubicek, J. Plutnar,
N. Loktionova, P. J. Riss, F. Rosch and I. Lukes, Chem.–Eur. J., 2010,
16, 7174; (b) J. Notni, J. Simecek, P. Hermann and H. J. Wester,
Chem.–Eur. J., 2011, 17, 14718.
5 (a) S. Aime, L. Calabi, C. Cavalloti, E. Gianolio, G. B. Giovenzana,
P. Losi, A. Maiocchi, G. Palmisano and M. Sisti, Inorg. Chem., 2004,
43, 7588; (b) S. Aime, G. Bombieri, C. Cavollotti, G. B. Giovenzana,
D. Imperio and N. Marchini, Inorg. Chim. Acta, 2008, 361, 1534;
(c) Z. Baranyai, F. Uggeri, G. B. Giovenzana, A. Benyei, E. Brucher
and S. Aime, Chem.–Eur. J., 2009, 15, 1696; (d) E. Elemento,
D. Parker, S. Aime, E. Gianolio and L. Lattuada, Org. Biomol. Chem.,
2009, 7, 1120.
6 During the course of this work, managanese(II) complexes of L1 and
L2 were described and evaluated as contrast agents in MRI: L. Tei,
G. Gugliotta, M. Fekete, F. K. Kalman and M. Botta, Dalton Trans.,
2011, 40, 2025.
7 (a) F. Roesch, Curr. Radiopharm., 2012, 5, 202; (b) W. A. Breeman and
A. M. Verbruggen, Eur. J. Nucl. Med. Mol. Imaging, 2007, 34, 978;
(c) K. P. Zhernosekov, D. V. Filosofov, R. P. Baum, P. Aschoff, H. Bihl,
A. A. Razbash, M. Jahn, M. Jennewein and F. Rosch, J. Nucl. Med.,
2007, 48, 1741.
8 Such substituted, heterocyclic seven membered rings adopt several
low energy conformers, of which the twist-chair is often the lowest in
energy: F. Freeman, J. H. Hwang, E. H. Junge, P. D. Parmar, Z. Renz
and J. Trinh, Int. J. Quantum Chem., 2008, 108, 339. Full details of the
conformational analysis of the ligands and the detailed structural
analysis of the gallium complexes will be reported elsewhere.
9 M. G. B. Drew, C. J. Harding, V. McKee, G. G. Morgan and J. Nelson,
J. Chem. Soc. Chem. Commun., 1995, 1035. The corresponding value
for [GaÁNOTA]2b is 6.261.
We thank the Association of Commonwealth Universities for
a scholarship (BPW) and ESF COST Action D38 for support.
˜
10 N. Ortega-Villar, A. L. Thompson, M. C. Munoz, V. M. Ugalde-
Notes and references
´
Saldıvar, A. E. Goeta, R. Moreno-Esparza and J. A. Real,
‡ The X-ray single crystal data were collected at 120 K on an Agilent
Gemini S-Ultra diffractometer (graphite monochromator, lMo Ka, l =
Chem.–Eur. J., 2005, 11, 5721. The corresponding value for [GaÁNOTA]
is 6.451 2b
.
0.70073 Å) equipped with Cryostream (Oxford Cryosystems) open-flow 11 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard and
nitrogen cryostat. The structures were solved by direct methods and
H. Puschmann, J. Appl. Cryst., 2009, 42, 339.
refined by full-matrix least squares on F2 for all data using Olex211 and 12 G. M. Sheldrick, Acta Crystallogr., 2008, A64, 112.
SHELXTL12 software. [GaÁL1]: C13H21GaN3O6.5, M = 393.05, orthorhom- 13 Y. Sun, C. J. Anderson, T. Pajeau, D. Reichert, R. D. Hancock,
bic, space group Fdd2, a = 28.8536(7), b = 27.2438(7), c = 7.57981(19) Å,
R. Motekaitis, A. E. Martell and M. Welch, J. Med. Chem., 1996,
39, 458.
V = 5958.3(2) Å3, Z = 16, m(Mo Ka) = 1.888 mmÀ1, Dcalc = 1.753 g mmÀ3
,
18 337 reflections measured (5.64 r 2Y r 59.98), 4319 unique (Rint
=
14 E. Boros, C. L. Ferreira, J. F. Cawthray, E. W. Price, B. O. Patrick,
D. W. Wester, M. J. Adam and C. Orvig, J. Am. Chem. Soc., 2010,
132, 15726.
0.0417) were used in all calculations. Final R1 0.0307 (>2s(I)); wR2
0.0674 (all data). [GaÁL2]: C15H26GaN3O7, M = 430.11, monoclinic, space
group P21, a = 7.58267(20), b = 12.6657(3), c = 9.2185(2) Å, b = 15 D. J. Berry, Y. Ma, J. R. Ballinger, R. Tavare, A. Koers, K. Sunassee,
100.735(3)1, V = 869.85(4) Å3, Z = 2, m(Mo Ka) = 1.627 mmÀ1, Dcalc
=
T. Zhou, S. Nawaz, G. E. D. Mullen, R. C. Hider and P. J. Blower,
Chem. Commun., 2011, 47, 7068.
1.642 g mmÀ3, 12 473 reflections measured (5.46 r 2Y r 57.98),
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 579--581 581