A benzene-core trinuclear GdIII complex: Towards the optimization of relaxivity for MRI contrast agent applications at high magnetic field

Article abstract of DOI:10.1039/b717390c

A novel ligand, H₁₂L, based on a trimethylbenzene core bearing three methylenediethylenetriamine-N,N,N″,N″-tetraacetate moieties (-CH₂DTTA^4-) for Gd³⁺ chelation has been synthesized, and its trinuclear Gd³⁺ complex [Gd₃L(H ₂O)₆]^3- investigated with respect to MRI contrast agent applications. A multiple-field, variable-temperature ¹⁷O NMR and proton relaxivity study on [Gd₃L(H ₂O)₆]^3- yielded the parameters characterizing water exchange and rotational dynamics. On the basis of the ¹⁷O chemical shifts, bishydration of Gd³⁺ could be evidenced. The water exchange rate, k_ex²⁹⁸ = 9.0 ± 3.0 s^-1 is around twice as high as k_ex²⁹⁸ of the commercial [Gd(DTPA)(H₂O)]^2- and comparable to those on analogous Gd³⁺-DTTA chelates. Despite the relatively small size of the complex, the rotational dynamics had to be described with the Lipari-Szabo approach, by separating global and local motions. The difference between the local and global rotational correlation times, τ_lO²⁹⁸ = 170 ± 10 ps and τ_gO²⁹⁸ = 540 ± 100 ps respectively, shows that [Gd₃L(H₂O)₆]^3- is not fully rigid; its flexibility originates from the CH₂ linker between the benzene core and the poly(amino carboxylate) moiety. As a consequence of the two inner-sphere water molecules per Gd³⁺, their close to optimal exchange rate and the appropriate size and limited flexibility of the molecule, [Gd₃L(H₂O)₆]^3- has remarkable proton relaxivities when compared with commercial contrast agents, particularly at high magnetic fields (r₁ = 21.6, 17.0 and 10.7 mM^-1s ^-1 at 60, 200 and 400 MHz respectively, at 25 °C; r₁ is the paramagnetic enhancement of the longitudinal water proton relaxation rate, referred to 1 mM concentration of Gd³⁺). The Royal Society of Chemistry.

Full text of DOI:10.1039/b717390c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
A benzene-core trinuclear GdIII complex: towards the optimization of
relaxivity for MRI contrast agent applications at high magnetic field†
a
a
a
b
a
a,c
´
Joa˜o Bruno Livramento, Lothar Helm, Ange´lique Sour, Conlin O’Neil, Andre´ E. Merbach and Eva To´th*
Received 9th November 2007, Accepted 20th November 2007
First published as an Advance Article on the web 20th December 2007
DOI: 10.1039/b717390c
A novel ligand, H12L, based on a trimethylbenzene core bearing three methylenediethylenetri-
amine-N,N,Nꢀꢀ,Nꢀꢀ-tetraacetate moieties (–CH2DTTA4−) for Gd3+ chelation has been synthesized, and
its trinuclear Gd3+ complex [Gd3L(H2O)6]3− investigated with respect to MRI contrast agent
applications. A multiple-field, variable-temperature 17O NMR and proton relaxivity study on
[Gd3L(H2O)6]3− yielded the parameters characterizing water exchange and rotational dynamics. On the
basis of the 17O chemical shifts, bishydration of Gd3+ could be evidenced. The water exchange rate,
kex = 9.0 3.0 s−1 is around twice as high as kex of the commercial [Gd(DTPA)(H2O)]2− and
comparable to those on analogous Gd3+-DTTA chelates. Despite the relatively small size of the complex,
the rotational dynamics had to be described with the Lipari–Szabo approach, by separating global and
298
298
298
local motions. The difference between the local and global rotational correlation times, slO = 170
10 ps and sgO = 540 100 ps respectively, shows that [Gd3L(H2O)6]3− is not fully rigid; its flexibility
298
originates from the CH2 linker between the benzene core and the poly(amino carboxylate) moiety. As a
consequence of the two inner-sphere water molecules per Gd3+, their close to optimal exchange rate and
the appropriate size and limited flexibility of the molecule, [Gd3L(H2O)6]3− has remarkable proton
relaxivities when compared with commercial contrast agents, particularly at high magnetic fields (r1 =
21.6, 17.0 and 10.7 mM−1s−1 at 60, 200 and 400 MHz respectively, at 25 ◦C; r1 is the paramagnetic
enhancement of the longitudinal water proton relaxation rate, referred to 1 mM concentration of Gd3+).
Gd3+ complexes are widely used to increase image contrast in
MRI.2,3 Their efficiency is expressed by their proton relaxivity,
r1, which is the paramagnetic enhancement of the longitudinal
water proton relaxation rate, in reference to a 1 mM concentration
of Gd3+. Proton relaxivity is directly related to the microscopic
parameters of the chelate, such as water exchange, rotational
dynamics or electronic relaxation. Much effort has been devoted
to the optimization of these factors on Gd3+ complexes with
the objective of increasing relaxivity. In particular, slowing down
the rotation by using macromolecules resulted in a remarkable
relaxivity improvement at intermediate frequencies (20–60 MHz)
as compared to commercial agents.2–4 For macromolecular agents,
however, above 60 MHz r1 drops strongly with increasing mag-
netic field and, at high frequencies (above 100 MHz), they
are hardly superior to small chelates. The optimization of the
relaxivity at high magnetic field requires the fine-tuning of the
microscopic parameters of the Gd3+ complex to optimal values
which are different from those at intermediate fields. Namely,
as the Solomon–Bloembergen–Morgan theory predicts, at proton
Larmor frequencies above 200 MHz r1 increases with the inverse
rotational correlation time 1/sR, in contrast to lower frequencies
where it is proportional to sR (Fig. 1). Thus at high frequencies,
rigid molecules of intermediate size are favoured over large ones,
with an optimal sR of ∼400 ps at 400 MHz (the exact value of
the optimal rotational correlation time will also be dependent on
the other influencing parameters). One peculiarity of high field
optimization is that the optimal rotational correlation time is very
sensitive to the magnetic field but remains, nevertheless, in the
range of 400–1000 ps. The optimal value of the water exchange
Introduction
In the last two decades, magnetic resonance imaging (MRI) has
become one of the most prominent techniques in clinical diagnos-
tics and biomedical research. In MRI, the amount of available
signal is inextricably associated with the static magnetic field
strength. Higher field strength offers considerable improvement
7/4
in the signal-to-noise ratio (proportional to B0 at mid-fields,
with a more linear dependence at high fields), which translates
to an increased spatial and temporal resolution. In addition, T1
relaxation times of grey and white matter also increase with field
strength, thus at high field the uptake of a contrast agent will result
in a more significant shortening of T1. Until recently, common
clinical scanners operated at ≤1.5 T. Due to recent improvements
in magnet design, 3 T scanners have become widely available in
the clinics, and for experimental animal studies much higher fields
(≥9.4 T; 400 MHz) have entered the everyday practice.1
a
´
Institut des Sciences et Inge´nierie Chimiques, Ecole Polytechnique Fe´de´rale
de Lausanne, ISIC, BCH, CH-1015, Lausanne, Switzerland; Fax: +41
(0)21 693 9875; Tel: +41 (0)21 693 98 71
b
´
Laboratoire de me´decine re´ge´ne´rative et de pharmacobiologie, Ecole
Polytechnique Fe´de´rale de Lausanne, EPFL-AAB, CH-1015, Lausanne,
Switzerland
cCentre de Biophysique Mole´culaire, CNRS, rue Charles Sadron,
45071, Orle´ans, France. E-mail: eva.jakabtoth@cnrs-orleans.fr; Fax: +33
(0)2 3863 1517; Tel: +33 (0)2 3825 7625
† Electronic supplementary information (ESI) available: Equations used in
the analysis of 17O NMR and NMRD data; experimental data of proton
relaxivities, 17O relaxation rates and chemical shifts, details of HPLC
separation. See DOI: 10.1039/b717390c
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1195–1202 | 1195
©

View Article Online
MR imaging experiments at 4.7 T confirmed the good contrast
agent efficiency of the metallostar under in vivo conditions.9
Nevertheless, one drawback of this system is the limited stabil-
ity of the Fe2+-tris(bipyridine) core, which might lead to slow
decomposition of the metallostar in biological conditions,8 and
hence raise free iron toxicity concerns. In order to overcome
this problem, we have designed a novel chelate, H12L, involving
covalent linking of three DTTA units to a central trimethylbenzene
core (–CH2DTTA4−) = methylenediethylenetriamine-N,N,Nꢀꢀ,Nꢀꢀ-
tetraacetate). H12L affords a trinuclear Gd3+ complex (Scheme 1)
of medium size with a rotational correlation time expected to
be in the appropriate range to attain maximum relaxivities at
4.7–9.4 T. The DTTA4− chelator has been previously proved to
possess several positive features with regard to MRI contrast agent
applications. It guarantees a relatively fast water exchange on the
Gd3+ complex, as evidenced for analogue chelates.8,10 The complex
has two inner sphere water molecules to double the inner sphere
relaxivity contribution. The animal imaging results obtained with
the metallostar9 showed that the efficiency of the chelate is not
reduced in vivo, which proves that the two inner sphere water
molecules in GdDTTA− are not replaced by endogenous anions
or other potential donors from proteins etc. in the biological
medium. This is an important feature in favour of the DTTA-
chelates in comparison with the macrocyclic bishydrated LnDO3A
complexes, known for their ability to form ternary complexes with
various endogenous carboxylate donors.11
For in vivo applications, the Gd3+ chelate has to be sufficiently
stable so that no release of free Gd3+ occurs before total excretion
of the contrast agent from the body. Thermodynamic stability
was assessed for several GdDTTA−-type complexes and showed a
limited decrease compared to [Gd(DTPA)(H2O)]2−. The stability
constants determined for the Gd3+ complexes formed with DTTA-
chelators attached to a benzene or a bipyridine core were all in
the range of log KGdL = 17–19, with corresponding pGd values of
∼15–16 ([L]total = 10 lM; [Gd]total = 1 lM; pH = 7.4).8,10 Evidently,
these stabilities would not be acceptable for human applications,
nevertheless, they can be sufficient for animal studies.12 In fact,
high field imaging is essentially intended for animal experiments.
We have previously reported the synthesis and physico-chemical
characterization of dinuclear, bishydrated Gd3+ complexes
[Gd2(m-X(DTTA)2)(H2O)4]2− and [Gd2(p-X(DTTA)2)(H2O)4]2−,
based on a xylene core substituted in para and meta positions
by the same DTTA chelators as in ligand L (X = xylene).10
These complexes have elevated proton relaxivities at high field.
On introducing a third DTTA unit and replacing the benzene core
with 1,3,5-trimethylbenzene, we expected a further rigidification
of the molecule, a slightly longer rotational correlation time and a
concomitant increase in the high field relaxivities.
Fig. 1 Inner sphere proton relaxivities calculated using the Solomon–
Bloembergen–Morgan theory for various values of the rotational corre-
lation time, sR, as a function of the magnetic field (upper x-axis) or the
proton Larmor frequency (lower x-axis). q = 2, kex = 9 × 106 s−1, sv =
38 ps, D2 = 0.1 × 1020 s−2
.
rate will also be different (and considerably higher) from that
for current clinical fields.5 The development of contrast agents
specifically designed for high field applications is an emerging
domain, and apart from some preliminary studies,6–8 no dedicated
agents have been reported. We recently published a self-assembled,
4−
metallostar-structured system, {Fe[Gd2bpy(DTTA)2(H2O)4]3} ,
with a remarkable relaxivity at high magnetic fields (r1
=
15.8 mM−1 s−1 at 4.7 T and 37 ◦C; Scheme 1).
Here we report the synthesis of ligand H12L and the
physico-chemical characterization of its trinuclear Gd3+ complex,
[Gd3L(H2O)6]3−. We have performed a combined 17O NMR and
1H NMRD (nuclear magnetic relaxation dispersion) study which
allowed us to assess all significant parameters influencing proton
relaxivity. The promising in vitro relaxivities obtained at high
magnetic fields prompted us to pursue the assessment of this
trinuclear Gd3+ chelate in in vivo animal imaging experiments
at 9.4 T. These results, to be published elsewhere,13 confirmed
the considerably higher efficacy of the complex with respect to
commercial contrast agents in high field applications.
Scheme 1 Structure of potential high field MRI contrast agents.
1196 | Dalton Trans., 2008, 1195–1202
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Molecular modelling
Results and discussion
The spatial demand of the tris-DTTA substitution on the
trimethylbenzene core was assessed by molecular modelling
studies. As Fig. 2 shows, the molecule bearing three DTTA-
units in 1,3,5-positions of the central benzene is not sterically
overcrowded. The exact Gd-Gd distances are determined at any
given time by the conformation of the methylene bridges linking
the DTTA unit on the central nitrogen to the benzene ring.
Synthesis
The synthesis of ligand H12L is presented in Scheme 2. The
synthetic route is composed of two major steps: synthesis of the
tetraester amine compound 4 and its conjugation onto the alkyl
halide 1,3,5-tris(bromomethyl)-2,4,6-trimethylbenzene, which is
subsequently hydrolyzed to give the desired ligand 6. Compound
4 was prepared in a succession of protection/deprotection steps.
Two consecutive protections are done, first on the terminal
primary amines of the diethylenetriamine (to form compound
1), and second on the central secondary amine to yield com-
pound 2. Deprotections of both amines are done selectively. The
terminal amines are first deprotected and then alkylated with t-
butylbromoacetate to give compound 3. The central amine is then
deprotected with Pd over charcoal to yield the building block 4.
Reaction of 1,3,5-tris(bromomethyl)-2,4,6-trimethylbenzene with
compound 4 in the presence of K2CO3 gave the ester 5, which
was hydrolyzed to the product 6. This desired trisubstituted
ligand H12L was obtained as a mixture with the corresponding
disubstituted molecule where the third arm previously bearing a
bromine has been converted into an alcohol.
The separation of these two similar compounds proved to be
hard. It was attempted on a cationic exchange resin, which is
a conventional separation method for poly(amino carboxylates).
However, the separation was revealed to be unsuccessful even when
repeated. Therefore we have performed an HPLC separation by
using ESI-MS and UV-Vis spectroscopy as combined detection
methods. The purity of the desired product was further confirmed
by 1H NMR spectroscopy.
Fig. 2 Molecular modelling representation of [Gd3L(H2O)6]3−
.
Scheme 2 Synthesis of H12L.
The Royal Society of Chemistry 2008
This journal is
Dalton Trans., 2008, 1195–1202 | 1197
©

View Article Online
However a rough estimation of the average distance could be
˚
to preclude important dipolar interactions between Gd3+ ions that
could accelerate electronic relaxation and cut back relaxivity at
high magnetic fields.14
the longitudinal 17O relaxation, which is governed by dipole–
dipole (influenced by the Gd3+–water oxygen distance, rGdO) and
quadrupolar mechanisms (influenced by the quadrupolar coupling
constant, v(1 + g2/3)1/2).
made: d(Gd–Gd) = 9.8
0.6 A. This distance is long enough
The 17O relaxation rates and chemical shifts and the proton
relaxivity data were analyzed simultaneously on the basis of the
Solomon–Bloembergen–Morgan theory.16 The rotational dynam-
ics has been described by using the Lipari–Szabo approach which
allows the separation of the local motion of the Gd3+ chelating
subunits, characterized by a local rotational correlation time, sl,
from the motion of the overall molecule, described by an overall
rotational correlation time, sg.17,18 Despite the relatively small size
of [Gd3L(H2O)6]3−, we could not achieve a satisfactory fit of the
NMRD curves by using a single rotational correlation time and
the Lipari–Szabo treatment was necessary.
17O NMR and 1H NMRD measurements
The parameters characterizing water exchange and rotation of
[Gd3L(H2O)6]3− were obtained in a variable-temperature, multiple-
field 17O NMR and 1H NMRD study (Fig. 3). The 17O longitudinal
and transverse relaxation rates and chemical shifts were measured
in an aqueous solution of [Gd3L(H2O)6]3− at 4.7 and 9.4 T. Relaxiv-
ity measurements were performed ◦between 0.01–400 MHz proton
Larmor frequency at 5, 25 and 37 C. The reduced 17O transverse
relaxation rates, 1/T2r, attain a maximum at ∼283 K. Above this
value, they decrease with increasing temperature indicating a fast
exchange region. Here 1/T2r is given by the transverse relaxation
rate of the bound water oxygen, 1/T2m, which is influenced by the
water exchange rate, kex, the scalar coupling constant, A/, and the
longitudinal electronic relaxation rate, 1/T1e.
In the fitting procedure several parameters were fixed to
common values. The distance between the Gd3+ electron spin
and the coordinated water 17O nucleus, rGdO, was considered to be
2.50 A, based on available crystal structures19 and ESEEM data.20
˚
The Gd3+–H distance, rGdH, was set to 3.10 A, and the distance of
˚
closest approach of an outer sphere water proton to Gd3+, aGdH
,
2
to 3.5 A. The quadrupolar coupling constant, v(1+g /3)1/2) was
˚
7.58 MHz, the value for pure water.
The scalar coupling constant calculated for [Gd3L(H2O)6]3− is
A/ = −(3.4 0.7) × 106 rad s−1. For the electron spin relaxation,
the following parameters have been obtained: Ev = 0.5 kJ mol−1;
298
sv = 38
5 ps, D2 = (0.1
0.02) × 1020 s−2. In addition to
the static zero field splitting contribution, a spin rotation term has
2
been also included, characterized by the parameter dgL = 0.1. The
diffusion constant, DGdH298, and its activation energy, EGdH, were
20 × 10−10 m2 s and 23 kJ mol−1 respectively.
The motions of the Gd-coordinated water oxygen and the
Gd-coordinated water hydrogen vectors are reflected in the
1
longitudinal 17O and H relaxation. The ratio between the local
rotational correlation time of the Gd–Hwater and Gd–Owater vectors,
sgH/sgO, was fixed to 0.65, a usual value for this kind of complex.21
All equations used in the analysis are given in the ESI† and the
fitted curves are shown in Fig. 3. Tables 1 and 2 summarize the pa-
rameters characterizing water exchange and rotation respectively.
Water exchange
DTTA4− is
a
heptacoordinate ligand for Gd3+, leaving
two coordination sites to be occupied by water. Crystallo-
graphic data published on the analogous chelate [(CNH2)3]3-
[Gd(TTAHA)]·9H2O proved the coordination of seven donor
atoms of the poly(amino carboxylate) (TTAHA6− = N-tris(2-
aminoethyl)amine-Nꢀ,Nꢀ,Nꢀꢀ,Nꢀꢀ,Nꢀꢀꢀ,Nꢀꢀꢀ-hexaacetate).22 In this
structure two binding sites were occupied by neighboring
carboxylates which are then replaced by water molecules in
aqueous solution. The bishydration of the lanthanide ion in
[Gd3L(H2O)6]3− has been confirmed by the experimental 17O
chemical shifts. UV-Vis measurements were previously performed
Fig. 3 (a) 1H nuclear magnetic relaxation dispersion profiles, recorded at
5 (᭺), 25 (ꢀ) and 37 ◦C (ꢁ); temperature dependence of (b) reduced
transverse 1/T2r [9.4 (ꢀ) and 4.7 T (ꢂ)] and longitudinal 1/T1r 17O
relaxation rates [9.4 T (᭺) and 4.7 T (᭹)]; and (c) reduced chemical
shifts Dxr [9.4 (D) and 4.7 T (ꢃ)]. The lines represent the least-squares
simultaneous fitting of all data points.
2−
on the dinuclear Eu3+ analogues m-X[Eu(DTTA)(H2O)2]2 and
p-X[Eu(DTTA)(H2O)2]22− in the 7F0–5D0 region which proved the
absence of hydration equilibrium on DTTA-chelates.10
The water exchange parameters obtained in the fit are sum-
marized in Table 1. As expected, the exchange rates of the
various complexes carrying a DTTA unit are similar, with the
The transverse 17O relaxation proceeds via a scalar mechanism
and holds no rotational information. This latter is contained in
1198 | Dalton Trans., 2008, 1195–1202
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 1 Water exchange rates, kex298, activation enthalpies, DH‡, and activation entropies, DS‡, of selected Gd3+ complexes
Complex
kex298/106 s−1
DH‡/kJ mol−1
DS‡/J mol−1 K−1
Ref.
[Gd3L(H2O)6]3−
9
3
40
3
+31
+25
+24
+41
+53
5
This work
4−
{Fe[Gd2bpy(DTTA)2(H2O)4]3}
[Gd2(m-X(DTTA)2)(H2O)4]2−
[Gd2(p-X(DTTA)2)(H2O)4]2
[Gd(DTPA)(H2O)]2−
7.4
8.9
9.0
3.3
41.3
39.2
45.4
51.6
8
10
10
15
exception of {Fe[Gd(tpy-DTTA)(H2O)2]2} which has a lower kex,
related to the neutrality of the complex. On the other hand,
kex on [Gd(DTPA)(H2O)]2− is almost half of that displayed by
the DTTA-based complexes. The faster exchange on the DTTA-
chelates can be rationalized in terms of their flexible inner
sphere structure, induced by the presence of the two inner sphere
water molecules. With respect to [Gd(DTPA)(H2O)]2−, this faster
exchange could be beneficial in attaining higher relaxivities at any
frequency, once the rotational dynamics are optimized. The values
of activation entropy and enthalpy obtained for [Gd3L(H2O)6]3−
are in agreement with a dissociatively activated water exchange
mechanism expected for a nine-coordinate complex.
{Fe[Gd(tpy-DTTA)(H2O)2]2} in which the poly(amino carboxy-
late) is directly linked to the terpyridine unit.24
Although the CH2 linker brings a certain flexibility to the
molecule, its presence is necessary for stability reasons: the direct
attachment of the amine of the DTTA chelator to a benzene
ring is known to reduce the basicity of the nitrogen which then
translates to a reduced stability of the Gd3+ complex.24 In order to
limit flexibility as much as possible, a 2,4,6-trimethyl derivative of
benzene was used as the central core.
Proton relaxivity
[Gd3L(H2O)6]3− displays considerably high relaxivities, although
at medium field they are still limited by fast rotation as the
temperature dependence of the relaxivity hump evidences (r1
decreases with increasing temperature). At 20 MHz, the trinuclear
complex has a relaxivity about four times higher than the
commercial [Gd(DOTA)(H2O)]− or [Gd(DTPA)(H2O)]2− and it is
close to the relaxivities of macromolecular, dendrimeric complexes
such as Gadomer 17 (Table 2). Several factors account for the high
r1 value. [Gd3L(H2O)6]3− is a rigid structure possessing three Gd3+
centers tightly bound around the benzene core. Each of these metal
ions carries two inner sphere water molecules. Furthermore, the
six inner sphere water molecules exchange with the bulk with a
rate close to the optimal value.
Rotational dynamics
The rotational parameters calculated for [Gd3L(H2O)6]3− are
presented and compared to those of different size Gd3+ complexes
in Table 2. The global rotational correlation time, sgO298, is much
lower than that for the metallostar and reflects the difference in
their molecular weight. On the other hand, the local rotational
correlation time, slO298, and the S2 parameter characterizing the
spatial restriction of the local with regard to the global motion are
very similar for the two systems. This similarity can be expected,
since in both cases the only relevant source of flexibility is the
CH2 moiety connecting the DTTA unit to the aromatic core.
In accordance with this, no internal flexibility was observed for
Table 2 Rotational correlation times, sRO 298, proton relaxivities, r1, and densities of relaxivity for selected Gd3+ complexes (20 MHz; 37 ◦C)
Complex
sRO 298/ps
r1/mM−1 s−1
Density of relaxivity/s−1 g−1
29.0
L
Ref.
[Gd3L(H2O)6]3−
sgO = 540 100
15.7
This work
298
298
slO = 173 10a
S2 = 0.5 0.1
Small MW complexes
[Gd(DTPA)(H2O)]2−
[Gd(DOTA)(H2O)]−
58
77
4.0
3.8
7.1
6.7
15
15
Dinuclear complexes
[Gd2bpy(DTTA)2(H2O)4]2−
[Gd2(m-X(DTTA)2)(H2O)4]2−
[Gd2(p-X(DTTA)2)(H2O)4]2
240
278
289
12.4
5.8
4.9
20.2
14.3
15.8
8
10
10
Heterometallic complexes
{Fe[Gd(tpy-DTTA)(H2O)2]2}
410
15.7
20.1
20.0
32.2
24
8
4−
298
{Fe[Gd2bpy(DTTA)2(H2O)4]3}
sgO = 930
298
slO = 190
S2 = 0.6
Dendrimers
Gadomer 17
298
sgO = 3050
17.5
19.3
23
298
slO = 760
S2 = 0.5
a Eg = 23 3 kJ mol−1; El = 22 8 kJ mol−1
.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1195–1202 | 1199
©

View Article Online
Table 3 High field relaxivities of selected Gd3+ complexes
r1/mM−1 s−1
200 MHz
25 ◦
17.0
16.4
13.4
14.4
400 MHz
25 ◦
10.7
9.32
9.18
10.7
C
37 ◦
C
C
37 ◦
10.2
8.53
C
[Gd3L(H2O)6]3−
14.1
15.9
10.2
11.9
3.04
3.20
4−
a
{Fe[Gd2bpy(DTTA)2(H2O)4]3}
[Gd2(m-X(DTTA)2)(H2O)4]2−
8.12
9.59
2.95
3.13
b
[Gd2(p-X(DTTA)2)(H2O)4]2−
b
[Gd(DOTA)(H2O)]−
4.02
4.22
3.86
4.06
c
[Gd(DTPA)(H2O)]2−c
a Ref. 8. b Ref. 10. c Ref. 9.
An important characteristic of the trinuclear complex is its
exceptionally broad relaxivity hump centered around 30–60 MHz.
In fact, the relaxivity is almost constant between 20 MHz
Synthesis of compound 5. To a stirred solution of 1,3,5-
tris(bromomethyl)-2,4,6-trimethylbenzene (200 mg, 0.5 mmol)
in 60 ml of dry acetonitrile was added compound 4 (1.03 g,
1.84 mmol) and K2CO3 (276 mg, 2.0 mmol). The resulting mixture
was refluxed for 7 h (disappearance of starting material was
monitored by ESI-MS). The solvent was removed in vacuo and the
residue was redissolved in dichloromethane. This organic phase
was washed three times with water, once with brine, then dried over
Na2SO4. The solvent was removed to afford 1.11 g of a product
which was used without further purification. 1H NMR (CDCl3 =
7.25 ppm, 400 MHz) d: 1.42 (s, 108H), 2.35 (s, 9H), 2.56 (t, J =
7.2 Hz, 12H), 2.78 (t, J = 7.2 Hz, 12H), 3.34 (s, 24H), 3.59 (s, 6H);
MS (ESI): m/z (%): 918 (100) [MH22+]; 612.85 (50) [MH33+].
◦
(15.7 mM−1 s−1, 37 C) and 200 MHz (14.1 mM−1 s−1, 37 ◦C).
As expected, the high field relaxivities are considerably higher
than those of [Gd(DOTA)(H2O)]− and [Gd(DTPA)(H2O)]2−
(Table 3). However, they do not attain the maximum values which
can be theoretically calculated on the basis of the Solomon–
Bloembergen–Morgan theory by using the water exchange and
electronic parameters as determined for [Gd3L(H2O)6]3−, and
assuming optimal rotational dynamics, as shown in Fig. 1. The
theoretically attainable relaxivities would be ∼25 and 18 mM−1 s−1
at 200 and 400 MHz respectively (considering ∼2 mM−1 s−1
for the outer sphere contribution). Instead of these values, for
[Gd3L(H2O)6]3− we measure 17.0 and 10.2 mM−1 s−1 at 200
and 400 MHz respectively. This difference clearly shows the
importance of the rotational motion and in particular of the
internal flexibility, responsible for cutting back the relaxivity.
It is interesting to note that at 200 and 400 MHz the
Synthesis of compound 6. A solution of compound 5 (900 mg,
0.5 mmol) in 70 ml of HCl (6 M) was refluxed for 10 h. After
evaporation to dryness, the compound was dissolved in 70 ml of
water and evaporated (this procedure was repeated 3 times). The
resulting solid was then dissolved in a minimum amount of H2O,
loaded onto a conditioned cation-exchange column (Bio-Rad AG
50W-X2, H+ form, 3.0 × 9.5 cm). The column was washed with
H2O till the pH of the eluate was neutral. The absence of Cl−
was evaluated by a typical AgNO3/HNO3 test. The column was
eluted with a gradient of HCl (from 0.2 to 5 M), rinsed with H2O
until neutral pH and finally eluted with a 1 M NH3 solution.
The product was concentrated from the appropriate fractions, re-
dissolved in a minimum amount of H2O and loaded a second time
onto a cation-exchange column (Bio-Rad AG 50W-X2, H+ form,
2.0 × 8.5 cm). The elution scheme used was the same as for the first
cation-exchange resin. This time, the product was collected from
the fractions 3–4 M of HCl and evaporated to dryness. Further
purification was done by HPLC (Waters Corp., Milford, MA,
USA). All details are given in the ESI.† Solvent evaporation of
the appropriate fractions yielded the desired product (0.05 g) as a
white solid. 1H NMR (D2O ≈ 4.80 ppm, pD = 2.0, 400 MHz) d:
2.45 (s, 9H), 3.16 (m, 12H), 3.52 (m, 12H), 3.72 (s, 24H), 4.1 (s,
6H); MS (ESI): m/z (%): 1162.4 (30) [MH+]; 581.8 (100) [MH22+].
2−
previously reported dinuclear m-X[Gd(DTTA)(H2O)2]2 and p-
2−
X[Gd(DTTA)(H2O)2]2 complexes have relaxivities very similar
to those measured for [Gd3L(H2O)6]3− (Table 3). A direct compar-
ison of these three systems shows practically no difference above
10 MHz in terms of molar relaxivity per Gd3+. Nevertheless, one
advantage of the trinuclear complex can be a more important
concentration of the relaxation effect into a restricted molecular
space. We have recently introduced the concept of “density of
relaxivity”,8 which we defined as the paramagnetic relaxation rate
enhancement per unit mass of the agent (s−1 g−1 L). The three
Gd3+ centers present in [Gd3L(H2O)6]3− each have a high molar
relaxivity. Given the low molecular mass of the complex, this
important relaxation effect is concentrated into a small molecular
volume. Table 2 lists densities of relaxivity at 20 MHz for a series
of Gd3+ complexes. For [Gd3L(H2O)6]3− a value of 29.0 s−1 g−1
L
is calculated, which is ∼4 times as higher than that for the
commercial [Gd(DTPA)(H2O)]2− and [Gd(DOTA)(H2O)]−, and
∼2 times more important than those for the macromolecular
contrast agent Gadomer 17.
Sample preparation
[Gd3L(H2O)6]3− was prepared by mixing the solid ligand with a
GdCl3 stock solution (solution prepared from Gd2O3 of 99.9%
purity from Fluka by dissolution in excess HCl, which was evap-
orated off). The concentration of the metal ion was determined
by complexometric titration with standardized Na2H2EDTA so-
lution. A slight excess of ligand (2%) was used to ensure complete
coordination of Gd3+, and the pH, measured with a combined
glass electrode calibrated with Metrohm buffers, was adjusted to
Experimental
Synthesis
All reagents were purchased from commercial sources and used
without further purification. Compounds 2, 3 and 4 were synthe-
sized as described in ref. 25.
1200 | Dalton Trans., 2008, 1195–1202
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
6.0 with addition of known amounts of NaOH (0.8 M) and HCl
(0.1 M). The absence of free Gd3+ was controlled with the xylenol
orange test.
dynamics could only be described by separating global and local
motions. The flexibility is related to the CH2 linker between the
benzene core and the chelating units, which is, however, a necessary
compromise in the aim of maintaining thermodynamic stability of
the complex. The two inner sphere water molecules, their fast
exchange, and the relatively limited flexibility of the molecule are
all important factors that contribute to the remarkable proton
relaxivities, particularly at high magnetic fields (r1 = 21.6, 17.0
and 10.7 mM−1 s−1 at 60, 200 and 400 MHz respectively, 25 ◦C).
17O NMR spectroscopy
Variable-temperature 17O NMR measurements were performed
on Bruker DPX-400 (9.4 T, 54.2 MHz) and Bruker Avance-
200 (4.7 T, 27.1 MHz) spectrometers and a Bruker VT-3000
temperature control unit was used to stabilize the temperature,
which was measured by a substitution technique.26 Transverse
and longitudinal 17O relaxation rates and chemical shifts were
measured between 277.1 and 371.1 K. The samples were sealed in
glass spheres adapted to 10 mm NMR tubes, to avoid susceptibility
corrections to the chemical shifts. 17O-enriched water (Irakli
Gverdtsiteli Research and Technology Center on High Technolo-
gies of Isotopes and Super Pure Materials. 17O: 10.5%) was added
to the gadolinium-containing samples to improve the sensitivity.
Longitudinal relaxation rates 1/T1 were obtained by the inversion
recovery method and transverse relaxation rates 1/T2 by the
Carr–Purcell–Meiboom–Gill spin-echo technique. Acidified water
(HClO4, pH = 3.71) was used as the external reference. The Gd3+
concentration of the sample was 0.0305 mol kg−1.
Acknowledgements
We acknowledge financial support of the Swiss National Sci-
ence Foundation, the Swiss State Secretariat for Education and
Research (SER) and the Centre National pour la Recherche
Scientifique (CNRS, France). This work was performed in the
frame of the EU COST Action D38 “Metal Based Systems for
Molecular Imaging Applications”.
Notes and references
1 R. G. Pautler, Physiology, 2004, 19, 168.
2 P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev.,
1999, 99, 2293.
1H NMRD
3 The Chemistry of Contrast Agents in Medical Magnetic Resonance
´
Imaging, ed. A. E. Merbach and E. To´th, John Wiley & Sons,
The 1/T1 NMRD profiles were obtained at 278.0, 298.0 and
310.0 K, on a Stelar Spinmaster Fast Field-Cycling relaxometer
(covering a continuous magnetic field from B = 2.35 × 10−4–0.47
T, proton Larmor frequencies of 0.01–20 MHz) equipped with
a VTC90 temperature control unit (temperature was fixed by a
gas flow and monitored by a substitution technique), on Bruker
Minispecs (30, 40 and 60 MHz); and on Bruker spectrometers (50,
100, 200 and 400 MHz). The Gd3+ concentration of the sample
was 2.00 mM.
Chichester, 2001.
4 P. Caravan, Chem. Soc. Rev., 2006, 35, 512.
´
5 E. Balogh, M. Mato-Iglesias, C. Platas-Iglesias, E. To´th, K.
Djanashvili, J. A. Peters, A. de Blas and T. Rodrıguez-Blas, Inorg.
´
Chem., 2006, 45, 8719.
6 V. C. Pierre, M. Botta and K. N. Raymond, J. Am. Chem. Soc., 2005,
127, 504.
´
7 J. B. Livramento, E. To´th, A. Sour, A. Borel, A. E. Merbach and R.
Ruloff, Angew. Chem., Int. Ed., 2005, 44, 1504.
´
8 J. B. Livramento, A. Sour, A. Borel, A. E. Merbach and E. To´th,
Chem.–Eur. J., 2006, 12, 989.
9 J. B. Livramento, C. Weidensteiner, M. I. M. Prata, P. R. Allegrini,
C. F. G. C. Geraldes, L. Helm, R. Kneuer, A. E. Merbach, A. C.
Santos, P. Schmidt and E. To´th, Contrast Media Mol. Imaging, 2006,
Data analysis: 17O NMR and 1H NMRD
´
The simultaneous least-squares fit of the 17O NMR and 1H NMRD
data were performed by the programs Visualiseur/Optimiseur on
a Matlab platform, version 6.5.27
1, 30.
´
10 J. Costa, E. To´th, L. Helm and Andre´ E. Merbach, Inorg. Chem., 2005,
44, 4747.
11 S. Aime, M. Botta, J. I. Bruce, V. Mainero, D. Parker and E. Terreno,
Chem. Commun., 2001, 115.
12 In the in vivo imaging experiments performed with the metallostar and
with the trinuclear complex [Gd3L(H2O)6]3−, the contrast agent was
well tolerated by the mice. No gross side effects were observed during
injection, immediately or days after the experiment.
Molecular modeling
To assess the volume of [Gd3L(H2O)6]3− and the distance between
the Gd3+ ions, a molecular model was built using the CAChe
program.28 The structure of a Sr2+, instead of the Gd3+ complex
was partially optimized using molecular mechanics (MM2 force
field) and semi-empirical quantum calculations (PM5 method).
13 P. Loureiro de Sousa, J. B. Livramento, L. Helm, A. E. Merbach,
W. M eˆme, B.-T. Doan, J.-C. Beloeil, M. I. M. Prata, A. C. Santos,
´
C. F. G. C. Geraldes and E. To´th, to be submitted.
14 G. M. Nicolle, L. Helm and A. E. Merbach, Magn. Reson. Chem., 2003,
41, 794.
15 D. H. Powell, O. M. Ni Dhubhghaill, D. Pubanz, L. Helm, Y. S.
Lebedev, W. Schlaepfer and A. E. Merbach, J. Am. Chem. Soc., 1996,
118, 9333.
Conclusions
´
With the objective of maximizing proton relaxivity at high
magnetic field, we have synthesized a trinuclear Gd3+ complex
based on a trimethylbenzene core bearing three diethylenetri-
aminetetraacetate units for lanthanide complexation. The water
16 E. T
o´th, L. Helm and A. E. Merbach, Relaxivity of Gadolinium(III)
Complexes: Theory and Mechanism, in The Chemistry of Contrast
´
Agents in Medical Magnetic Resonance Imaging, ed. E. To´th and
A. E. Merbach, Wiley, Chichester, 2001, p. 45.
17 (a) G. Lipari and S. Szabo, J. Am. Chem. Soc., 1992, 10, 4546; (b) G.
298
exchange rate of [Gd3L(H2O)6]3−, kex = 9.0 3.0 s−1, is twice as
Lipari and S. Szabo, J. Am. Chem. Soc., 1992, 10, 4559.
high as that of [Gd(DTPA)(H2O)]2− which can be rationalized
in terms of a more flexible inner coordination sphere, related
to the bishydration of the complex. Despite the relatively small
molecular weight, the complex is not fully rigid and the rotational
´
18 E. To´th, L. Helm, K. E. Kellar and A. E. Merbach, Chem.–Eur. J.,
1999, 5, 1202.
19 (a) M.-R. Spirlet, J. Rebizant, J. F. Desreux and M.-F. Loncin, Inorg.
Chem., 1984, 23, 359; (b) J. J. Stezowski and J. L. Hoard, Isr. J. Chem.,
1984, 24, 323.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1195–1202 | 1201
©

View Article Online
20 A. M. Raitsimring, A. V. Astashkin, D. Baute, D. Goldfarb and P.
Caravan, J. Phys. Chem. A, 2004, 108, 7318.
21 F. A. Dunand, A. Borel and A. E. Merbach, J. Am. Chem. Soc., 2002,
124, 710.
24 J. Costa, R. Ruloff, L. Burai, L. Helm and A. E. Merbach, J. Am. Chem.
Soc., 2005, 127, 5147.
25 J. Platzek, U. Niedballa, B. Radu¨chel, DE 19508058, 1996, p. 13.
26 C. Amman, P. Meyer and A. E. Merbach, J. Magn. Reson., 1982, 46,
319.
27 (a) F. Yerly, VISUALISEUR 2.3.5, Lausanne, Switzerland, 1999; (b) F.
Yerly, OPTIMISEUR 2.3.5, Lausanne, Switzerland, 1999.
28 CAChe Worksystem Pro 6.1, Fujitsu, 2000–2003.
22 R. Ruloff, T. Gelbrich, J. Sieler, E. Hoyer and L. Beyer, Z. Naturforsch.,
B: Chem. Sci., 1997, 52, 805.
´
23 G. M. Nicolle, E. To´th, H. Schmitt-Willich, B. Radu¨chel and A. E.
Merbach, Chem.–Eur. J., 2002, 8, 1040.
1202 | Dalton Trans., 2008, 1195–1202
This journal is
The Royal Society of Chemistry 2008
©