Synthesis, characterization, and pharmacokinetic evaluation of a potential MRI contrast agent containing two paramagnetic centers with albumin binding affinity

Article abstract of DOI:10.1002/chem.200401207

A dinuclear gadolinium(III) complex of an amphiphilic chelating ligand, containing two diethylenetriamine-N,N,N′,N″,N″ -pentaacetate (DTPA) moieties bridged by a bisindole derivative with three methoxy groups, has been synthesized and evaluated as a poten

Full text of DOI:10.1002/chem.200401207

FULL PAPER
Synthesis, Characterization, and Pharmacokinetic Evaluation of a Potential
MRI Contrast Agent Containing Two Paramagnetic Centers with Albumin
Binding Affinity
Tatjana N. Parac-Vogt,[a] Kristof Kimpe,[a] Sophie Laurent,[b] Luce Vander Elst,[b]
Carmen Burtea,[b] Feng Chen,[c] Robert N. Muller,[b] Yicheng Ni,[c]
Alfons Verbruggen,[d] and Koen Binnemans*[a]
Abstract: A dinuclear gadolinium(iii)
complex of an amphiphilic chelating
ligand, containing two diethylenetri-
amine-N,N,N’,N’’,N’’-pentaacetate
the presence of human serum albumin
(HSA) the relaxivity value of the non-
covalently bound dinuclear complex in-
talation of the dinuclear gadolinium(iii)
complex by zinc(ii) has been investigat-
ed. Biodistribution studies suggest that
the complex is excreted by the renal
pathway, and possibly by the hepato-
biliary route. In vivo studies indicated
that half of the normal dose of the
gadolinium(iii) complex enhanced the
contrast in hepatic tissues around 40%
more effectively than [Gd(DTPA)].
The dinuclear gadolinium(iii) complex
was tested as a potential necrosis avid
contrast agent (NACA), but despite
the binding to HSA, it did not exhibit
necrosis avidity, implying that binding
to albumin is not a key parameter for
necrosis-targeting properties.
creases to 15.2 sꢀ1 per mmol of Gd3+
,
(DTPA) moieties bridged by a bisin-
dole derivative with three methoxy
groups, has been synthesized and eval-
uated as a potential magnetic reso-
nance imaging (MRI) contrast agent.
Nuclear magnetic relaxation dispersion
(NMRD) measurements indicate that
at 20 MHz and 378C the dinuclear
gadolinium(iii) complex has a much
higher relaxivity than [Gd(DTPA)] (6.8
vs 3.9 sꢀ1 mmolꢀ1). The higher relaxivity
of the dinuclear gadolinium(iii) com-
plex can be related to its reduced
motion and larger rotational correla-
tion time relative to [Gd(DTPA)]. In
due to its relatively strong interaction
with this protein. The fitted value of
the binding constant to HSA (Ka) was
found to be 104 mꢀ1. Because of its in-
teraction with HSA, the dinuclear com-
plex exhibits a longer elimination half-
life from the plasma, and a better con-
finement to the vascular space com-
pared to the commercially available
[Gd(DTPA)] contrast agent. Transme-
Keywords: gadolinium
agents lanthanides
spectroscopy · rare earths
·
imaging
· NMR
·
Introduction
ments showed that contrast agents might greatly increase its
diagnostic value. Consequently, the development of MRI
techniques has been accompanied by an increased interest
in developing contrast agents.[1–3] The first MRI contrast
agent approved for use in humans, the gadolinium(iii) com-
Magnetic resonance imaging (MRI) is routinely used in vari-
ous diagnostic imaging procedures. Although excellent soft-
tissue images can be obtained by this method, early experi-
Department of Radiology, University Hospitals, K.U. Leuven
Herestraat 49, 3000 Leuven (Belgium)
[a] Dr. T. N. Parac-Vogt, Dr. K. Kimpe, Prof. Dr. K. Binnemans
Department of Chemistry, Katholieke Universiteit Leuven
Celestijnenlaan 200F, 3001 Leuven (Belgium)
Fax : (+32)16-327-992
[d] Prof. Dr. A. Verbruggen
Laboratory of Radiopharmaceutical Chemistry
Department of Pharmaceutical Sciences
Van Evenstraat 4, 3000 Leuven (Belgium)
E-mail: koen.binnemans@chem.kuleuven.ac.be
[b] Dr. S. Laurent, Prof. Dr. L. Vander Elst, Dr. C. Burtea,
Prof. Dr. R. N. Muller
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
NMR and Molecular Imaging Laboratory
Department of Organic and Biomedical Chemistry
University of Mons-Hainaut, 7000 Mons (Belgium)
[c] Dr. F. Chen, Prof. Dr. Y. Ni
Biomedical Imaging
Interventional Therapy and Contrast Media Research
Chem. Eur. J. 2005, 11, 3077 – 3086
DOI: 10.1002/chem.200401207
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3077

plex of diethylenetriamine-N,N,N,’N’’,N’’-pentaacetate, [Gd-
(DTPA)], is used in clinical MRI under the name Magnev-
istꢁ. This prototype of the first generation of contrast agents
distributes into the vascular and interstitial space and is clas-
sified as an “unspecific agent” or “extracellular fluid agent”.
In recent years, efforts have been directed towards the de-
velopment of contrast agents that are tissue specific with
higher proton relaxivities.
weight, supramolecular complexes increases the rotational
correlation time, and the presence of multiple paramagnetic
centers results in an increased relaxivity. Complexes contain-
ing multiple gadolinium(iii) chelate units designed as blood-
pool contrast agents have been reported by Martin et al.[28]
More recently, heterocyclic complexones containing two pyr-
azole rings and two iminodiacetic units per molecule have
been shown to bind two gadolinium(iii) ions.[29] Although
the observed relaxivity was improved significantly relative
to [Gd(DTPA)] and [Gd(DOTA)], the thermodynamic sta-
bility of the complexes was low, probably because of only
tetradentate coordination around the gadolinium(iii) ion.
In the search for more efficient and specific contrast
agents, we designed a novel amphiphilic ligand bearing two
DTPA units capable of forming stable octacoordinate com-
plexes with two gadolinium(iii) ions. The two DTPA units
are linked by a 2,3-disubstituted bisindole derivative, and
we anticipated that this type of linker would noncovalently
bind to plasma proteins. By using this approach, relaxivity
was increased not only by the presence of multiple paramag-
netic centers, but also by the reduced rate of molecular tum-
bling.
Higher proton relaxivity can be achieved by slowing the
rotational motion of the contrast agent. Various routes have
been explored to increase correlation time, such as the in-
corporation of the contrast agent into liposomes.[4,5] Al-
though incorporation of paramagnetic compounds into lipo-
somes efficiently increases rotational correlation time, the
application of these particles in vivo has some disadvantag-
es, such as their rapid physiological removal by Kupffer cells
of the liver and spleen, and their relatively slow clearance
rate, which may be harmful to human tissue.[6] An alterna-
tive strategy is the synthesis of macromolecular gadolini-
um(iii) chelates, such as dendrimers,[7] linear polymers,[8–10]
or proteins.[2,11–12] This strategy, however, has proved to be
rather disappointing, as the relaxivity gained by increasing
the molecular size is often much less than expected, because
of the internal flexibility or nonrigid attachment of the che-
late to the macromolecule.[7–14] More recently, an increase in
rotational correlation time was achieved by either incorpo-
rating amphiphilic gadolinium(iii) chelates into mixed mi-
celles,[15,16] or by designing complexes that can self-assemble
to form micelles.[17,18] Due to their reduced particle size
(50 nm or less), these agents are less recognizable by the
Kupffer cells for physiological removal, with the result that
their residence time in blood increases. Consequently, they
could be very promising as potential MRI contrast
agents.[15,19]
The so-called noncovalent binding strategy offers an alter-
native to the macromolecular approach,[20,21] and involves
targeting of a small gadolinium(iii) complex to a particular
protein. Binding to a protein increases the concentration of
the complex around the protein receptor molecule, which
selectively enhances the target relative to the background.
The first example of this type of contrast agent was [MS-
325], a compound designed to image the blood pool by ex-
ploiting its noncovalent binding to the plasma protein
human serum albumin (HSA).[22–24] This biophysical trick
limits the amount of free drug that can extravasate from the
blood pool into the nonvascular space, and provides selec-
tive enhancement of the vascular relaxation rate. In addi-
tion, this noncovalent binding causes a further slowing of
the rotational motion, resulting in a much higher relaxivi-
ty.[14,25] This binding mechanism has also been related to ne-
crosis avidity (i.e., detection of necrotic tissue).[26]
The dinuclear gadolinium(iii) complex was characterized,
preclinically evaluated, and tested in vivo. Its ability to bind
to albumin was also examined, and because it has been sug-
gested that the albumin binding mechanism may be the
reason for necrosis avidity,[26] the dinuclear gadolinium(iii)
complex was also tested for necrosis-targeting applications.
Results and Discussion
The ligand and complexes: The key difference between our
design of ligand 3 (Scheme 1) and the approach used for
other agents with high albumin affinity (such as [MS-
325]),[22–25] was the use of the lipophilic group, which can
bind two DTPA moieties, as a starting point. A bisindole de-
rivative of trimethoxybenzaldehyde, 1, was chosen as the
basis for the attachment of two DTPA moieties. Hydrazine
was used as a bridging molecule because of its ability to
form two amide bonds. The first amide bond is formed by
reaction of hydrazine with the ethylcarboxylate group of the
ethylindole-2-carboxylate moiety. Compound 2, formed by
this reaction, was further coupled in the presence of the cou-
pling agent o-benzotriazol-1-yl-N,N,N’,N’-tetramethyluroni-
um tetrafluoroborate (TBTU) to two DTPA molecules to
yield ligand 3.
The 1H NMR spectrum of this final product (Figure 1)
was recorded in D2O and its full assignment was achieved
by conducting two-dimensional COSY experiments (see
Supporting Information). The spectrum consists of 16 sig-
nals, which indicate the symmetry of the ligand molecule
(see Scheme 1 for labeling). Electrospray mass spectrometry
(ES-MS) was used to confirm the structure of ligand 3 in so-
lution. The ES-MS spectrum in the positive mode showed a
molecular peak [M+Na]+ at m/z 1301.4, consistent with the
structure shown in Scheme 1.
Recently, heteroditopic ligands featuring a DTPA- or
DOTA-like unit for the chelation of gadolinium(iii)
(DOTA=1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic
acid), and a phenanthroline or terpyridine unit able to form
highly stable complexes with iron(ii), have been designed.[27]
The resulting spontaneous formation of high molecular
3078
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3077 – 3086

MRI Contrast Agents
FULL PAPER
the region 1690–1625 cmꢀ1, cor-
responding to CO stretching
modes. A shift of approximately
50 cmꢀ1 to lower energy was
observed for the carbonyl
stretching frequencies after
complexation, indicating amide
oxygen coordination to the lan-
thanide ion. Positive mode
ES-MS of the lanthanum(iii)
and gadolinium(iii) complexes
showed molecular peaks at m/z
1595 and 1632, respectively, in-
dicating the binding of two
metal ions to the ligand. This
finding was also supported by
elemental analysis data. The so-
lution structure of the dinuclear
complex was investigated by re-
cording the proton NMR spec-
trum of the lanthanum(iii) com-
plex. Complexation of
3 to
lanthanum(iii) causes remark-
able changes in the proton
NMR spectra in the form of
shifts and broadening of the
resonances.
Although
the
proton resonances of the ligand
are sharp and well separated,
the resonances of the lan-
thanum(iii) complex are very
broad, and multiple peaks were
observed for the protons adja-
cent to the carboxylic groups.
Heating the solution to 368 K
resulted in the severe broaden-
ing of all the resonances (see
Supporting Information). The
Scheme 1. Synthesis of ligand 3. Conditions: i) reflux in EtOH/conc. HCl; ii) MeOH/pyridine, 1008C;
iii) TBTU/TEA, DMSO.
Under slightly alkaline conditions, the ligand readily
formed complexes with lanthanide(iii) ions (Ln=La, Gd).
IR spectral data of the ligand showed strong absorptions in
spectra are consistent with the occurrence of several inter-
converting isomers; a typical characteristic of lanthanide(iii)
complexes with DTPA ligands.[30] However, based on the
available NMR spectroscopy data, no speculation on the
structures of the isomers present in the solution is possible.
Unfortunately, we were unable to grow crystals of sufficient
quality to allow structure determination by X-ray single
crystal diffraction; however, the IR, ES-MS, and elemental
analysis data are consistent with the formation of a dinu-
clear complex in which two lanthanide(iii) ions are coordi-
nated to two DTPA moieties. From hereon, the dinuclear
gadolinium(iii) complex is denoted [Gd2–3].
Residence time of the coordinated water molecule in [Gd2–
3]: The residence time of the coordinated water molecule
(tM) was obtained by analysis of the temperature depen-
dence of the reduced transverse relaxation rate of the water
17O nucleus in solutions of the gadolinium complex
(Figure 2). The theoretical adjustment of these experimental
Figure 1. Proton NMR spectrum of ligand 3 in D2O at 298 K. For the res-
onance assignment, see Scheme 1.
Chem. Eur. J. 2005, 11, 3077 – 3086
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3079

K. Binnemans et al.
for other monoamide derivatives of [Gd(DTPA)], such as
[Gd(DTPA-N-MA)] (tM310 =346 ns),[1b] and with the previous-
ly reported dimeric complexes based on DOTA ligands
[Gd(pip(DO3A)2)] (t3M10 =372 ns)[31b] and [Gd(bisoxa-
(DO3A)2)] (t3M10 =394 ns)[31b] (DOTA=1,4,7,10-tetrakis(carb-
oxymethyl)-1,4,7,10-tetraazacyclododecane, DO3A=1,4,7,10-
tetraazacyclododecane-1,4,7-triacetic acid, pip(DO3A)2 =
bis(1,4-(1-(carboxymethyl)-1,4,7,10-tetraaza-4,7,10-tris(carb-
oxymethyl)-1-cyclododecyl-1,4-diazacyclohexane, and bis-
oxa(DO3A)2 =bis(1,4-(1-(carboxymethyl)-1,4,7,10-tetraaza-
4,7,10-tris(carboxymethyl)-1-cyclododecyl))-1,10-diaza-3,6-di-
oxadecane).
The scalar coupling constant (A/ꢀh) is in good agreement
with those values reported for other gadolinium(iii) poly-
aminocarboxylate complexes that have one inner-sphere
water molecule (typically ꢀ3.8ꢂ106 radsꢀ1).[31b] Inspection of
the data listed in Table 1 reveals that, except for an entropy
of activation (DS#) somewhat lower than for [Gd(DTPA)]
and [MS-325], all other parameters determined for [Gd2–3]
are within the normal range observed for gadolinium(iii)
chelates.
Figure 2. Evolution of the reduced 17O transverse relaxation rate
(Rp2ara*55.55/GdIII concentration) versus the reciprocal of the temperature
for [Gd2–3]. Concentration of [Gd2–3]=12.42 mm.
data was performed as previously described, assuming the
presence of one water molecule in the first coordination
sphere.[31] This procedure allows for the determination of:
1. The hyperfine coupling constant (A/ꢀh) between the
oxygen nucleus of bound water molecules and the
gadolinium(iii) ion.
2. Parameters describing the electronic relaxation times of
Gd3+, that is, the correlation time modulating the elec-
tronic relaxation (tV), the activation energy (EV) related
to tV, and the mean-square of the zero-field splitting
energy D2.
3. Parameters related to the water exchange, that is, the en-
thalpy (DH°) and entropy (DS°) of activation of the
process.
Nuclear magnetic relaxation dispersion (NMRD) measure-
ments of [Gd2–3]: The [Gd2–3] dinuclear complex was inves-
1
tigated by recording H longitudinal relaxation time meas-
urements in water at 310 K and at magnetic field strengths
ranging from 0.24 mT to 1.4 T. The proton NMRD profile
recorded at 310 K (Figure 3) shows that at high magnetic
fields the proton relaxivity of [Gd2–3], expressed in sꢀ1 per
mmol of Gd3+ per liter, is much larger than that of the
parent [Gd(DTPA)] compound. The data were analyzed by
using the classical inner-sphere[32,33] and outer-sphere[34] the-
ories. Some parameters were fixed during the fitting proce-
dure: the distance (d) of closest approach for the outer-
sphere contribution was set at 0.36 nm, the tM value was ob-
tained by performing 17O NMR spectroscopy as described
above, and the number of water molecules in the first coor-
dination sphere of Gd3+ was set to one, in agreement with
At 310 K, the water residence time of [Gd2–3] is equal to
599ꢁ32 ns (Table 1). This value is greater than that for both
the parent compound [Gd(DTPA)] (t3M10 =132 ns[31b] and
Table 1. Parameters for [Gd2–3] obtained from the theoretical fitting of
the 17O NMR data. Parameters for [Gd(DTPA)] and [MS-325] are shown
for comparison.
Parameter
[Gd2–3]
[Gd(DTPA)]
[MS-325]
t3M10 [ns]
599ꢁ32
143ꢁ25
83ꢁ13[a]
51.3ꢁ0.3[a]
53.0ꢁ1.8[b]
55.8ꢁ0.4[a]
ꢀ4.1ꢁ0.06[a]
ꢀ3.8[b]
DH° [kJmolꢀ1
]
49.4ꢁ0.1
51.5ꢁ0.3
DS° [Jmolꢀ1 Kꢀ1
A/ꢀh [106 radsꢀ1
]
33.5ꢁ0.2
ꢀ4.0ꢁ0.1
52.1ꢁ0.6
ꢀ3.4ꢁ0.1
]
D2 [1020 sꢀ2
]
0.71ꢁ0.06
10.0ꢁ1.0
1.08ꢁ0.03
12.3ꢁ0.3
2.18ꢁ0.10[a]
0.25ꢁ0.06[b]
18.7ꢁ0.9[a]
1.7ꢁ0.1[b]
tV [298 K, ps]
[a] From ref. [25a]. [b] From ref. [25c].
143 ns[31a]
)
and the albumin-binding [MS-325] (tM310
=
83 ns),[25b] but smaller than that reported for the DTPA–
bisamide derivative [Gd(DTPA–BMA)] (t3M10 =1050 ns)[31b]
(DTPA–BMA=1,7-bis[(N-methylcarbamoyl)methyl]-1,4,7-
tris(carboxymethyl)-1,4,7-triazaheptane). The water resi-
dence time is in good agreement with the values reported
*
Figure 3. Proton NMRD profile of [Gd2–3] in water at 310 K ( ). The
[31]
&
NMRD profile of [Gd(DTPA)] is added for comparison ( ). Concen-
tration of [Gd2–3]=2.38 mm.
3080
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3077 – 3086

MRI Contrast Agents
FULL PAPER
the value reported for other related complexes.[1b] The corre-
lation time modulating the electronic relaxation (tV) and the
electronic relaxation time at zero field (tSO =(12D2tV)ꢀ1)
were optimized for the outer-sphere and inner-sphere contri-
butions, simultaneously. The distance (r) between the pro-
tons of the coordinated water molecule and the Gd3+ ion
was set to 0.31 nm. At 20 MHz and 378C the relaxivity of
[Gd2–3], expressed in 6.8 sꢀ1 per mmol of Gd3+, is higher
than that of [Gd(DTPA)] and [MS-325] (3.9 and
5.5 sꢀ1 mmolꢀ1, respectively). Moreover, if the relaxivity of
[Gd2–3] is expressed in sꢀ1 per mmol of the complex, its
value is equal to 13.6 at 378C and 0.47 T.
Inspection of the parameters shown in Table 2 reveals
that the higher relaxivity of [Gd2–3] can be related to its
larger rotational correlation time (tR), that is, to its reduced
Figure 4. Proton longitudinal paramagnetic relaxation rate in a solution
containing 4% HSA and increasing amounts of [Gd2–3] (&), at 20 MHz
and 310 K. The relaxation rates in the absence of HSA are represented
Table 2. Parameters obtained from fitting of the proton NMRD data, ob-
tained in water at 310 K, for [Gd2–3] compared to the data for [Gd-
(DTPA)] and [MS-325].
*
by . The diamagnetic contribution of the albumin solutions was mea-
310
SO
sured by using the solution of 4% HSA.
Compound
t3M10
t3R10
t
t3V10
[ns]
[ps]
[ps]
[ps]
[Gd2–3]
599
143
83
188
59
108
82
82
93
34
23
32
Fitting of the data by means of Equation (1) provides, on
the one hand, an estimation of the association constant Ka,
which characterizes the interaction between HSA and [Gd2–
3], and on the other hand, a value for the relaxivity of the
noncovalently bound complex, rc1. The relaxivity of the free
contrast agent, rf1, was determined for the dimeric [Gd2–3]
to be 13.6 sꢀ1 per mmol of [Gd2–3] (or 6.8 sꢀ1 per mmol of
Gd3+), which is larger than the rf1 value for [MS-325]
(5.5 sꢀ1 mmolꢀ1) at the same field (Bo =0.4 T) and tempera-
ture (T=310 K).[25a]
[Gd(DTPA)][a]
[MS-325][b]
[a] From ref. [25a] [b] From ref. [25c] with a distance r=0.31 nm.
motion. The rate of molecular tumbling observed for [Gd2–
3] is a third of that of the parent [Gd(DTPA)] complex,
demonstrating the significant effect of increasing the molec-
ular size of the dinuclear complex. The other parameters
(tSO and tV) are in good agreement with those reported for
both [Gd(DTPA)],[31a,b] and [MS-325].[25a]
ꢀ
ꢁ
obs
Rp1 ¼ 1000 ꢂ ðr1f ꢂ soÞ þ 0:5ðr1cꢀr1f Þ ðN ꢂ p0Þ þ s0þ
ð1Þ
ꢂꢃ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Binding of [Gd2–3] to serum albumin: Contrast agents can
interact by noncovalently binding to proteins present in
blood plasma. The most important protein in human plasma
is serum albumin (HSA), which constitutes 4 to 4.5% of
plasma (0.67 mm). This large (67 kDa) globular protein
binds to a variety of molecules, such as drugs, metabolites,
and fatty acids.[35] Noncovalent interactions of MRI contrast
agents with HSA result in the reduction of the molecular
tumbling rate of the contrast agent (i.e., the value of tR in-
creases), and hence in an increase in its relaxivity. Figure 4
displays the longitudinal proton relaxation rate in a solution
containing 4% HSA, as a function of [Gd2–3] concentration.
The relaxivity of [Gd2–3] is much greater in a solution of
HSA than in pure water, indicating a noncovalent interac-
tion. Binding to albumin reduces the amount of free drug
that can extravasate from the blood pool into the nonvascu-
lar space and provides the selective enhancement of the vas-
cular relaxation rate. Binding also reduces the fraction of
free chelate available for glomerular filtration by the kid-
neys. This slows the renal excretion rate, which extends the
half-life in the blood and increases the time available for
imaging.
ꢀ1
ꢀ1
2
Ka
ꢀ
ððN ꢂ p0Þ þ s0 þ Ka Þ ꢀ4 ꢂ N ꢂ s0 ꢂ p0
in which p0 and s0 are the initial concentrations of protein
and [Gd2–3], respectively.
To fit the data, the number of equivalent and independent
interactions sites (N) was set to 1. The fitted Ka and r1c
values are 10100mꢀ1 and 15.2 sꢀ1 per mmol of Gd3+ (30.4 sꢀ1
per mmol of [Gd2–3]), respectively, which implies that the
relaxivity is much higher than the values for the free [Gd2–
3].
This value of rc1 is lower than those reported for other Gd-
complexes, such as [MS-325] (rc1 ꢃ49 sꢀ1 mmolꢀ1 in water at
310 K,[25a] r1c ꢃ42–43 sꢀ1 mmolꢀ1 at pH 7.5 in Hepes buffer at
310 K,[25c]) or [Gd–DTPA(BOM)3] (BOM=benzyloxymeth-
yl) (rc1 ꢃ52 sꢀ1 mmolꢀ1 at pH 7.5 in Hepes buffer at
310 K[25c]). This difference can be related to a quenching of
the relaxivity of the bound [Gd2–3] by the long water resi-
dence time. In addition, for both [MS-325] and [Gd–DTPA-
(BOM)3], only one Gd3+ ion is located within the complex,
whereas in [Gd2–3], the two paramagnetic centers are locat-
Chem. Eur. J. 2005, 11, 3077 – 3086
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3081

K. Binnemans et al.
ed quite far from the area of the complex likely to interact
with the protein. Consequently, either both paramagnetic
centers will experience the same slowing of their rotational
motion, or, if one center is protein-bound, this will experi-
ence a larger effect than the other center. However, because
both paramagnetic centers of [Gd2–3] could have different
rotational mobilities, the relaxometric measurement cannot
distinguish between the gadolinium centers. The value calcu-
lated by the fitting procedure is, therefore, an average of
two different rc1 values.
49% after 5000 min), and much more favorable than anoth-
er commercially available [Gd(DTPA–BMA)] contrast
agent, for which the Rp1 decreased to 9% after 5000 min.
This indicates that [Gd2–3] has a stability comparable to
[Gd(DTPA)], but appears to be much more stable than the
[Gd(DTPA–BMA)] contrast agent towards transmetalation
by zinc(ii).
Pharmacokinetic profile of [Gd2–3]: The pharmacokinetic
profile provides information about both the residence time
of [Gd2–3] in the bloodstream, and its clearance after intra-
venous injection. The profiles of [Gd2–3] in rats with respect
to [Gd(DTPA)] are presented in Figure 6. Pharmacokinetic
Transmetalation of [Gd2–3] by zinc(ii): In general, there is
no direct correlation between the thermodynamic stability
constant and the acute toxicity of contrast agents. The con-
ditional stability constants at physiological pH (7.4) are
better indicators of the in vivo degree of chelation of the
cation. Cacheris et al. have argued that, due to possible
transmetalation reactions in the body, the in vivo toxicity for
DTPA derivatives is determined by the selectivity or differ-
ence in stability between a complex of ligand with Gd3+
,
and a complex of that same ligand with Zn2+, thermody-
namically expressed as Ksel. These authors have shown that
logKsel at physiological pH correlates with toxicity.[36] Trans-
metalation of the complex by Zn2+ results in the release of
Gd3+ ions, which form an insoluble phosphate complex in
the presence of phosphate ions, and no longer contribute to
the proton paramagnetic relaxation rate of the solution. The
consequent decrease in this rate during the transmetalation
process can be used to monitor quantitatively the evolution
of the system.[37] Figure 5 displays the transmetalation data
for [Gd2–3]. After 5000 min, Rp1 is decreased to 40% of its
initial value, showing that significant transmetalation takes
place. These data are less favorable than those for [MS-
325], in which the Rp1 decreased to about 75% of its initial
value after 5000 min.[25a] However, the data are comparable
with those of the parent [Gd(DTPA)] complex (decrease to
*
Figure 6. Pharmacokinetic profiles of [Gd2–3] ( , n=4) with respect to
*
[Gd(DTPA)] ( , n=3) in rats. Data are presented as absolute values of
plasma concentration (top) and percentages of C0 (bottom). The solid
lines represent the fit of the data to a biexponential model. Data points
are presented as averagesꢁSEM.
parameters were calculated by fitting the curves of blood
concentration as a function of time following a single bolus
in vivo injection, and are presented in Table 3. Except for
the volume of distribution (VDb), which is comparable to
that of [Gd(DTPA)], the other parameters for [Gd2–3] indi-
cate a significantly prolonged vascular residence time. The
delayed blood clearance is probably attributable to the bind-
ing of [Gd2–3] to albumin in the blood. Notably, although
the VDb for [Gd2–3] in rats (0.189 Lkgꢀ1) is similar to the
value for [Gd(DTPA)], it is even smaller than that of other
Figure 5. Time course of the normalized water proton paramagnetic re-
laxation rate at 310 K and 20 MHz in a solution containing 2.5 mm of
Zn2+ and 2.5 mm of [Gd2–3] ( ), [Gd(DTPA–BMA)] ( ), [Gd(DTPA)]
&
*
~
!
(
), or [MS-325] ( ).
3082
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3077 – 3086

MRI Contrast Agents
FULL PAPER
albumin-binding blood-pool contrast agents, such as [MS-
325] (VDb =0.236 Lkgꢀ1).[23] The longer elimination half-life
indicates that the agent has a long vascular residence, and
could have potential as a blood-pool contrast agent.
6). At only half of the dose of [Gd(DTPA)], [Gd2–3] caused
a signal intensity (SI) of the normal liver that was signifi-
cantly higher over the first 60 min than that measured with
[Gd(DTPA)]. Although the liver SI for [Gd(DTPA)] leveled
off rapidly within 40 min, in the case of [Gd2–3], it only
slowly decreased over 24 h (Figure 8). Such distinctive con-
Table 3. Pharmacokinetic parameters of [Gd2–3] and [Gd(DTPA)] in
rats. Td1/2 and Te1/2 are the distribution and elimination half-lives in
plasma, respectively; Cltot is the total clearance; VDb is the distribution
volume in the elimination phase.
Parameter
[Gd2–3]
(n=4)
[Gd(DTPA)]
(n=3)
Td1/2 [min]
1.56ꢁ0.025[b]
75.93ꢁ12.80[a]
1.10ꢁ0.15[b]
0.70ꢁ0.039
14.94ꢁ1.25
8.66ꢁ1.18
Te1/2 [min]
Cltot [mLkgꢀ1 minꢀ1
]
VDb [Lkgꢀ1
]
0.189ꢁ0.016
0.180ꢁ0.005
[a] p<0.05. [b] p<0.01.
Biodistribution of [Gd2–3]: The biodistribution studies show
whether the contrast agent accumulates in specific organs.
In good agreement with its slower elimination, the biodistri-
bution of [Gd2–3] in rats 60 min after intravenous adminis-
tration (Figure 7) indicates significantly higher concentra-
Figure 8. Comparison of the liver signal intensity of T1-weighted MR
images after intravenous injection of [Gd(DTPA)] at 0.1 mmolkgꢀ1 (gray
bars) and [Gd2–3] at 0.05 mmolkgꢀ1 (black bars). Over the first 60 min
postcontrast, liver contrast enhancement is significantly stronger with
[Gd2–3] than with the commercially available [Gd(DTPA)].
trasts can be well explained by the different structural prop-
erties of [Gd(DTPA)] and [Gd2–3]; whereas [Gd(DTPA)] is
a hydrophilic, extracellular fluid agent, [Gd2–3] is more lipo-
philic due to the presence of the trimethoxybenzene ring
and two indole moieties, and is, therefore, more likely to be
taken up by the liver.
To evaluate whether [Gd2–3] can act as a potential ne-
crosis avid contrast agent (NACA), we compared it to the
well-known NACA Gadophrin-2 in a rat model with re-per-
fused partial liver infarction. The same dose of [Gd2–3] and
Gadophrin-2 (0.05 mmolkgꢀ1) caused a prompt contrast en-
hancement (CE) in normal liver tissue, rendering the in-
farcted liver lobe easily distinguishable by MRI as a hypoin-
tense zone. Contrary to a reversed contrast between infarct-
ed and normal liver lobes that occurred with Gadophrin-2, a
continuing liver signal intensity attenuation was observed
with [Gd2–3]. The SI of normal and infarcted liver after
[Gd2–3] injection was normalized overnight with no appreci-
able necrosis/normal contrast (contrast ratio, CRꢃ1). The
SI of the infarcted lobe after Gadophrin-2 injection in-
creased further and persisted for up to 24 h, leading to a
positive contrast between necrotic and normal liver (CR=
1.4), consistent with results of previous studies.[38,39]
Despite its high albumin-binding affinity, [Gd2–3] does
not appear to be a necrosis avid contrast reagent, supporting
the suggestion that binding to albumin is probably not re-
sponsible for necrosis avidity. This is not unexpected, as al-
bumin binding is a general feature of many drugs, including
contrast agents, only a few of which have been shown to
have necrosis avidity.[40] However, it is notable that with half
Figure 7. Biodistribution of [Gd2–3] (black bars) and [Gd(DTPA)] (gray
bars) 60 min after administration. The results are presented as averagesꢁ
SEM; the Student t-test was calculated for [Gd2–3] versus [Gd(DTPA)]
(*: p<0.05; **: p<0.01).
tions than [Gd(DTPA)] in all of the organs analyzed. The
Gd3+ concentration measured in kidney suggests that [Gd2–
3] has a renal route of excretion. On the other hand, the rel-
atively high Gd3+ concentration detected in liver (ca.
3.5 times higher than for [Gd(DTPA)], p<0.01) could be re-
lated to its lipophilic properties, leading to a possible hepa-
tobiliary route of excretion. In fact, many albumin-binding
complexes are believed to interact with hepatocytes, and al-
though the mechanism of interaction is not completely un-
derstood, the dissociated chelate is removed from circula-
tion by the hepatobiliary route.
In vivo evaluation of contrast-enhancing properties and ne-
crosis avidity of [Gd2–3]: The liver contrast-enhancing effi-
cacy on T1-weighted MRI in vivo was evaluated by compar-
ing [Gd2–3] at an intravenous dosage of 0.05 mmolkgꢀ1 with
[Gd(DTPA)] at a dose of 0.1 mmolkgꢀ1 in normal rats (n=
Chem. Eur. J. 2005, 11, 3077 – 3086
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3083

K. Binnemans et al.
ed to pH 7 with 0.1m HCl, containing 0.5 mm EDTA; flow rate
1 mLminꢀ1). After removal of the solvent, the residue was dissolved in
200 mL of water and the solution was evaporated again. This process was
repeated until no free hydrazine could be detected in the solution by
using I2. The product was then suspended in acetonitrile and filtered.
Yield: 14.1 g (90%); 1H NMR (300 MHz, [D3]CD3OH): d=3.45 (s, 6H;
2ꢂCH3), 3.66 (s, 3H; 1ꢂCH3), 6.43 (s, 2H; Ph), 6.55 (d, 2H; Ph), 6.69
(m, 2H; Ph), 6.86 (s, 1H; CH-Ph), 7.28 (m, 2H; Ph), 7.45 (m, 2H; Ph),
11.70 ppm (s, 2H; 2ꢂNH); ES-MS (positive mode): m/z: 529.2 [M+H]+.
of the normal dose, [Gd2–3] enhances the contrast in liver
tissues much more effectively than does [Gd(DTPA)].
Conclusion
A new type of contrast agent bearing two DTPA units able
to bind two paramagnetic centers has been synthesized,
characterized, preclinically evaluated, and tested for necrosis
avidity. Data from NMRD analysis proved that the dimeric
complex has a much higher relaxivity than [Gd(DTPA)],
and that the relaxivity increases significantly in a solution of
human serum albumin because of a relatively strong interac-
tion with this blood protein. The interaction with HSA also
results in longer elimination half-lives and a better confine-
ment to the vascular space, as shown by pharmacokinetic
evaluation. The fact that two gadolinium(iii) ions are present
in one molecule of the contrast agent permits the use of
smaller injection volumes. With half the normal dose, the di-
meric complex enhanced the contrast in tissues more effec-
tively than [Gd(DTPA)].
The synthetic approach described here offers the potential
to design high molecular weight contrast agents with defined
structure. The increase in relaxivity achieved is due not only
to the longer rotational correlation times, but also to the in-
teraction of the linker with HSA. Application of this strat-
egy should facilitate the use of many different types of link-
ers with high affinities towards blood proteins.
Preparation of compound 3: A dispersion of DTPA (61 g, 155 mmol) in a
mixture of DMSO (1800 mL) and triethylamine (TEA, 50 mL; 36 g,
0.36 mol) was sonicated for 15 min. TBTU (25 g, 78 mmol) was added
and the mixture was sonicated again for 15 min. Product 2 (10.44 g,
20 mmol) was added and the mixture was stirred for 1 h. After removal
of the solvent, the residue was dissolved in a concentrated solution of
NaHCO3. The pH was adjusted to 7 with concentrated hydrochloric acid,
followed by sonication for 30 min. The resulting mixture was evaporated
and placed on a C18 reversed-phase column (10 mmꢂ80 cm), which was
then eluted successively with 5 L of a solution of ammonium acetate in
distilled water (0.1m) containing 2.5% methanol, 3 L of a solution con-
taining 5% methanol, and finally with 3 L of a solution containing 10%
methanol. The product was collected in 200 mL fractions and the purity
was checked by conducting HPLC (conditions as for the preparation of
2, above) and monitoring the peak eluting at 4.4 min. Yield: 13 g (51%);
1H NMR (300 MHz, D2O, pD =6.0): d=3.11 (m, 8H; 2ꢂN-CH2), 3.27
(m, 4H; 2ꢂN-CH2), 3.37 (m, 4H; 2ꢂN-CH2), 3.43 (s, 4H; 2ꢂCH2-
COOH), 3.53 (s, 8H; 4ꢂCH2-COOH), 3.65 (s, 9H; 3ꢂOCH3), 3.69 (b,
8H; 2ꢂCH2-COOH, 2ꢂCH2-CO-NH), 6.66 (s, 2H; Ph), 6.75 (m, 2H;
Ph), 6.79 (m, 2H; Ph), 6.86 (s, 1H; CH-Ph), 7.19 (m, 2H; Ph), 7.45 ppm
(m, 2H; Ph); IR (KBr): n˜ =3406 (N-H), 3008 (C-H, phenyl), 2917 and
2852 (C-H alkyl), 1687 (CO, amide I), 1624 (CO acid), 1591 (COOꢀ
assym. stretch), 1404 cmꢀ1 (COOꢀ sym. stretch); elemental analysis calcd
(%) for NaC56N12O23H65(NH4)4: C 48.66, H 6.39, N 16.33; found: C 49.1,
H 6.0, N 16.36; ES-MS (positive mode): m/z: 1301.4 [M+Na]+, 662.3
[M+2Na]2+
.
Synthesis of the lanthanide(iii) complexes
The lanthanide(iii) complexes of ligand 3 were synthesized according to a
Experimental Section
general procedure as follows:
a solution of hydrated LnCl3 salt
(1.1 mmol) in H2O (1 mL) was added to ligand 3 (0.5 mmol) dissolved in
pyridine (30 mL), and the mixture was heated at 708C for 3 h. The sol-
vents were evaporated under reduced pressure and the crude product
was then refluxed in ethanol for 1 h. After cooling to room temperature,
the complex was filtered off and dried in a vacuum. The absence of free
gadolinium was verified by using xylenol orange indicator.[41]
Materials: Reagents were obtained from Aldrich Chemical (Bornem,
Belgium), Acros Organics (Beerse, Belgium), and Fluka (Bornem, Bel-
gium), and used without further purification. Gadolinium(iii) chloride
hexahydrate was obtained from GFS chemicals (Powell, Ohio, USA).
Non-defatted human albumin (fraction V) was obtained from Sigma
(Bornem, Belgium).
Lanthanum(iii) complex [La2–3]: Yield: 95% (1.51 g); IR (KBr): n˜ =3401
(N-H), 3016 (C-H, phenyl), 2922 and 2850 (C-H alkyl), 1630 (CO, amide
I), 1590 (COOꢀ assym. stretch), 1400 cmꢀ1 (COOꢀ sym. stretch); elemen-
tal analysis calcd (%) for La2C56N12O23H62Na2(H2O)4: C 40.51, H 4.75, N
10.04; found: C 40.30, H 4.23, N 10.08; ES-MS (positive mode): m/z:
1595 [M+H]+, 1613 [M+H+H2O]+, 1635 [M+Na+H2O]+, 798
Syntheses
Preparation of compound 1: Some 3,4,5-trimethoxybenzaldehyde (10.8 g,
55 mmol) was added to a solution of ethylindole-2-carboxylate (18.9 g,
100 mmol) in ethanol (200 mL) under a nitrogen atmosphere, and the
mixture was heated to reflux temperature. Concentrated hydrochloric
acid (3.7 mL) was added, and TLC on silica gel plates with the chloro-
form/hexane (1:1) solvent system was used to monitor the reaction by the
disappearance of the ethylindole-2-carboxylate (Rf =0.7) and the appear-
ance of the reaction product spot (Rf =0.1). After about 1 h, the solution
was left to cool to room temperature and the white product was filtered
off and washed thoroughly with cold ethanol. Yield: 24 g (88%);
1H NMR (300 MHz, [D6]DMSO): d=1.19 (t, 6H; 2ꢂCH3), 3.46 (s, 6H;
2ꢂCH3), 3.66 (s, 3H; 1ꢂCH3), 4.23 (q, 4H; 2ꢂCH2), 6.43 (s, 2H; Ph),
6.55 (m, 2H; Ph), 6.69 (m, 2H; 2ꢂPh), 6.86 (s, 1H; CH-Ph), 7.12 (m,
2H; Ph), 7.46 (m, 2H; Ph), 11.69 ppm (s, 2H; 2ꢂNH); ES-MS (positive
mode): m/z: 557.2 [M+H]+.
[M+2H]2+
.
Gadolinium(iii) complex [Gd2–3]: Yield: 96% (1.56 g); IR (KBr) n˜ =
3406 (N-H), 2915 and 2840 (C-H alkyl), 1637 (CO amide I), 1585 (COOꢀ
assym. stretch), 1404 cmꢀ1 (COOꢀ sym. stretch); elemental analysis calcd
(%) for Gd2C56N12O23H62Na2(H2O)8: C 37.65, H 4.91, N 9.90; found: C
37.87, H 4.43, N 9.47; ES-MS (positive mode): m/z: 1632 [M+H]+, 1610
[MꢀNa+2H]+, 795 [Mꢀ2Na+4H]2+
.
Instruments: Elemental analysis was performed by using a CE Instru-
ments EA-1110 elemental analyzer. 1H NMR spectra were recorded by
using a Bruker Avance 300 (Bruker, Karlsruhe, Germany), operating at
300 MHz. IR spectra were measured by using an FTIR spectrometer
Bruker IFS66, using the KBr pellet method. Mass spectra were obtained
by using a Q-tof 2 (Micromass, Manchester, UK). Samples for the mass
spectrometry were prepared by dissolving the complex (2 mg) in metha-
nol (1 mL), then adding 200 mL of this solution to a water/methanol mix-
ture (50:50, 800 mL). The resulting solution was injected at a flow rate of
Preparation of compound 2: Hydrazine monohydrate (50 mL) in metha-
nol (50 mL) was added to a solution of product 1 (16.2 g, 30 mmol) in
pyridine (150 mL), and the mixture was heated at 1008C for 36 h. The re-
action was monitored by conducting reversed phase HPLC; a peak was
detected at 7.3 min (Hypersil BDS 5mm column 4.6 mmꢂ250 mm (All-
tech, Laarne, Belgium), detection wavelength 254 nm, linear gradient
from 5 to 95% CH3CN over 20 min, mobile phase 0.05m NaHCO3 adjust-
5 mLminꢀ1
.
3084
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3077 – 3086

MRI Contrast Agents
FULL PAPER
17O NMR measurements: Samples of solutions (2 mL) contained in NMR
tubes (OD 10 mm) were analyzed by using an AMX-300 spectrometer
(Bruker, Karlsruhe, Germany). The temperature was regulated by air or
nitrogen flow controlled by a BVT 2000 unit (Bruker, Karlsruhe, Germa-
ny). No field frequency lock was used. All 17O NMR spectra were proton
decoupled. The 17O transverse relaxation times of distilled water were
measured by using a CPMG (Car–Purcell Meiboom Gill) sequence and a
subsequent two-parameters-fit of the data points. The 90 and 1808 pulse
lengths were 25 and 50 ms, respectively. The 17O T2 of the complex in
aqueous solution was obtained from line width measurement. The con-
centration of the samples was less than 25 mm.
commercially available; [Gd(DTPA)], and the well-known NACA bis-
Gd-mesoporphyrin or Gadophrin-2 (supplied by the Institut fꢃr Diagnos-
tikforschung Berlin, Germany).[47]
MR imaging and quantification: The rat under anesthesia was placed su-
pinely into a plastic holder. A tail vein of the rat was cannulated with a
G27 infusion set connected to a 1 mL tuberculin syringe loaded with a
contrast agent. MR imaging was performed using a 1.5 T Siemens Sonata
scanner (Erlangen, Germany) with a commercially available four-channel
phased-array wrist coil. A T1-weighted (TR/TE=600 ms/15 ms) spin-echo
2D-imaging sequence, with 2 mm slice thickness (without gap), a field of
view of 4.6 cmꢂ8.0 cm, a 240ꢂ512 matrix, and four averaged acquisi-
tions, resulting in about 3 min of measurement, was recorded to study the
contrast agents for in vivo contrast enhancement. The SI was measured
for a region of interest on five consecutive liver slices on MR images and
averaged for precontrast and serial postcontrast phases. The conspicuity
of the necrotic lobe was expressed as the contrast ratio (CR) between
T1 measurements: Proton nuclear magnetic relaxation dispersion
(NMRD) profiles were recorded between 0.24 mT and 0.24 T by using
Field Cycling Relaxometers (Field Cycling Systems, Oradell, New Jersey,
USA and Stelar Spinmaster FFC-2000, Stelar SRL, Mede, Italy) with
0.6 mL solutions contained in 10 mm OD tubes. Proton relaxation rates
were also measured at 0.47 T and 1.4 T by using Minispec PC-120 and
mq-60, respectively (Bruker, Karlsruhe, Germany). The temperature was
maintained at 310 K. 1H NMRD data were fitted according to the theo-
retical inner-sphere model described by Solomon[32] and Bloembergen,[33]
and to the outer-sphere contribution described by Freed.[34] Calculations
were performed by using a previously described software package.[42–43]
the necrotic and the normal liver, and was calculated as CR=SInecrosis
SInormal
/
.
Acknowledgements
Transmetalation: Transmetalation by zinc(ii) ions was evaluated by the
decrease in the water longitudinal relaxation rate at 310 K and 20 MHz
(Bruker Minispec PC-120) of buffered phosphate solutions (pH 7,
T.N.P.V. and K.B. acknowledge F.W.O. Flanders (Belgium) for a Postdoc-
toral Fellowship. T.N.P.V., K.K., and K.B. also thank the K.U. Leuven for
financial support (VIS/01/006.01/20002–06/2004 and GOA 03/03). CHN
microanalyses were performed by Mrs. Petra Bloemen. ES-MS measure-
ments were made by Ms. Leen Vannerum. Mrs Corinne Piꢄrart and Mrs
Virginie Henrotte are acknowledged for their help in measuring the
proton relaxation rates. S.L., L.V.E., C.B., and R.N.M. thank the ARC
Program 00/05–258 of the French Community of Belgium and kindly ac-
knowledge the support and sponsorship provided by COST Action D18
“Lanthanide Chemistry for Diagnosis and Therapy”.
[KH2PO4]=26 mm, [Na2HPO4]=41 mm) containing 2.5 mm of the
[37]
gadolinium(iii) complex and 2.5 mm of Zn2+
.
Interaction with HSA: The binding constant and relaxivity value of
[Gd2–3] in a 4% solution of HSA was determined by measuring the
proton longitudinal relaxation rate at 20 MHz and 310 K as a function of
the [Gd2—3] concentration.
In vivo characterization: All of the animal experiments were performed
according to recommendations of the ethical commission of K.U.Leuven
and the University of Mons-Hainaut.
Pharamacokinetic characterization: The animals were anesthetized by
the intraperitoneal injection of pentobarbital (Nembutalꢁ, Sanofi Sante
Animale, Brussels, Belgium) at a dose of 60 mgkgꢀ1. Wistar rats (n=3 or
4/group, 269 gꢁ6 g, Harlan, The Netherlands) were tracheotomized and
the left carotide was catheterized (Becton Dickinson Angiocath, 0.7ꢂ
19 mm) for blood collection. The agent was administered as a bolus in
the femoral vein at a dose of 0.025 mmolkgꢀ1 (0.05 mmol of gadolinium
per kg). [Gd(DTPA)], used as a control, was injected at a dose of
0.1 mmolkgꢀ1. Blood samples (~0.3 mL) were collected 1, 2.5, 5, 15, 30,
45, 60, 90, and 120 min after injection. The concentration of gadolinium
in the blood samples was determined by conducting relaxometry at 378C
and 60 MHz with a Bruker Minispec mq60 (Bruker, Karlsruhe, Germa-
ny).[45] A two-compartment distribution model was used to calculate the
distribution (Td1/2) and elimination (Te1/2) half-lives, the apparent volume
of distribution (VDb), the total clearance (Cltot), and the initial blood
concentration C0.[23] The gadolinium concentrations in blood were con-
verted to plasma concentrations by assuming a hematocrit value of
0.53.[44]
[1] a) R. B. Lauffer, Chem. Rev. 1987, 87, 901–927; b) P. Caravan, J. J.
Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999, 99, 2293–
2352.
[2] S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev. 1998, 27,
19–29.
[3] H. J. Weinmann, W. Ebert, B. Misslewitz, H. Schmitt-Willich, Eur. J.
Radiol. 2003, 46, 33–44.
[4] D. J. Hnatowich, B. Friedman, B. Clancy, M. Novak, J. Nucl. Med.
1981, 22, 810–814.
[5] G. W. Kabalka, M. A. Davis, T. H. Moss, E. Buonocore, K. Hubner,
E. Holmberg, K. Maruyama, L. Huang, Magn. Reson. Med. 1991,
19, 406–415.
[6] E. C. Unger, D. K. Shen, T. A. Fritz, J. Magn. Reson. Imaging 1993,
3, 195–198.
[7] ꢅ. Tꢆth, D. Pubanz, S. Vauthey, L. Helm, A. E. Merbach, Chem.
Eur. J. 1996, 2, 1607–1615.
[8] M. Spanoghe, D. Lanens, R. Dommisse, A. van der Linden, F. Alder-
weireldt, Magn. Reson. Imaging 1992, 10, 913–917.
[9] V. S. Vexler, O. Clement, H. Schmitt-Willich, R. C. Brasch, J. Magn.
Reson. Imaging 1994, 4, 381–388.
[10] T. S. Desser, D. L. Rubin, H. H. Muller, F. Qing, S. Khodor, G. Za-
nazzi, S. U. Young, D. L. Ladd, J. A. Wellons, K. E. , Kellar, J. L.
Toner, R. A. Snow, J. Magn. Reson. Imaging 1994, 4, 467–472.
[11] S. Aime, M. Botta, M. Fasano, S. G. Crich, E. Terreno, J. Biol. Inorg.
Chem. 1996, 1, 312–319.
[12] ꢅ. Tꢆth, F. Connac, L. Helm, K. Adzamli, A. E. Merbach, J. Biol.
Inorg. Chem. 1998, 3, 606–613.
[13] S. W. A. Bligh, A. H. M. S. Chowdhury, M. McPartlin, I. J. Scowen,
R. A. Bulman, Polyhedron 1995, 14, 567–569.
Biodistribution: Adult Wistar rats (n=3/group, 273 gꢁ9 g, Harlan, The
Netherlands) were anesthetized and injected with the contrast agents, as
described above. Sixty minutes after injection, the animals were sacrificed
and samples of liver, kidneys, heart, lungs, and spleen were collected for
evaluation of the gadolinium content. The samples were weighed, dried
overnight at 608C, and subsequently digested (up to 0.4 g each sample)
in acidic conditions (3 mL HNO3 65%, 1 mL H2O2 33%) by microwaves
(Milestone MSL-1200, Sorisole, Italy). The gadolinium content was deter-
mined by performing inductively coupled plasma atomic emission spec-
troscopy (ICP-AES, Jobin Yvon JY70+, Longjumeau, France). The ga-
dolinium concentration was expressed as a percentage of the injected
dose per gram of tissue.
In vivo evaluation for liver contrast enhancement: Six normal rats and a
rat with re-perfused hepatic infarction induced under laparotomy with
temporary obstruction (3 h) of the blood supply to the right liver lobes
[14] S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev. 1998, 27,
19–29.
[15] K. Kimpe, T. N. Parac-Vogt, S. Laurent, C. Piꢄrart, L. Vander Elst,
R. N. Muller, K. Binnemans, Eur. J. Inorg. Chem. 2003, 3021–3027.
[38,39]
were included in MRI studies. Two reference contrast agents were
Chem. Eur. J. 2005, 11, 3077 – 3086
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3085

K. Binnemans et al.
[16] P. L. Anelli, L. Lattuada, V. Lorusso, M. Schneider, H. Tournier, F.
Uggeri, MAGMA 2001, 12, 114–120.
[17] J. P. Andrꢄ, ꢅ. Tꢆth, H. Fischer, A. Seelig, H. R. Mꢇcke, A. E. Mer-
bach, Chem. Eur. J. 1999, 5, 2977–2983.
[31] a) S. Laurent, L. Vander Elst, S. Houzꢄ, N. Guꢄrit, R. N. Muller,
Helv. Chim. Acta 2000, 83, 394–406; b) D. H. Powell, O. M. NiD-
hubhghaill, D. Pubanz, L. Helm, Y. S. Lebedev, W. Schlaeffer, A.
Merbach, J. Am. Chem. Soc. 1996, 118, 9333–9346.
[18] K. Binnemans, C. Gçrller-Walrand, Chem. Rev. 2002, 102, 2303–
2345.
[19] H. Tournier, B. Lamy, R. Hyacinthe, US-5.833.948, 1998.
[20] B. G. Jenkins, E. Armstrong, R. B. Lauffer, Magn. Reson. Med. 1991,
17, 164–178.
[32] I. Solomon, Phys. Rev. 1955, 99, 559–565.
[33] N. Bloembergen, J. Chem. Phys. 1957, 27, 572–573.
[34] J. H. Freed, J. Chem. Phys. 1978, 68, 4034–4037.
[35] T. J. Peeters, All about Albumin: Biochemistry, Genetics and Medical
Applications, Academic Press, San Diego, 1996.
[21] R. B. Lauffer, Magn. Reson. Med. 1991, 22, 339–342.
[22] R. B. Lauffer, D. J. Parmelee, H. S. Ouellet, R. P. Dolan, H. Sajiki,
D. M. Scott, P. J. Bernard, E. M. Buchanan, K. Y. Ong, Z. Tyeklar,
K. S. Midelfort, T. J. McMurry, R. C. Walovitch, Acta Radiol. 1996,
3, S356–S358.
[23] D. J. Parmelee, R. C. Walovitch, H. S. Ouellet, R. B. Lauffer, Invest.
Radiol. 1997, 32, 741–747.
[24] R. B. Lauffer, D. J. Parmelee, S. U. Dunham, H. S. Ouellet, R. P.
Dolan, S. Witte, T. J. McMurry, R. C. Walovitch, Radiology 1998,
207, 529–538.
[25] a) R. N. Muller, B. Raduchel, S. Laurent, J. Platzek, C. Piꢄrart, P.
Mareski, L. Vander Elst, Eur. J. Inorg. Chem. 1999, 1949–1955; b) P.
Caravan, N. J. Cloutier, M. T. Greenfield, S. A. McDermid, S. U.
Dunham, J. W. M. Bulte, J. C. Amedio, R. J. Looby, R. M. Supowski,
W. D. Horrocks, T. J. McMurry, R. B. Lauffer, J. Am. Chem. Soc.
2002, 124, 3152–3162; c) S. Aime, M. Chiaussa, G. Digilio, E. Gia-
nolio, E. Terreno, J. Biol. Inorg. Chem. 1999, 4, 766–774.
[26] B. Hofmann, A. Bogdanov, E. Marecos, W. Ebert, W. Semmler, R.
Weissleder, J. Magn. Reson. Imaging 1999, 9, 336–341.
[27] a) J. F. Desreux, J. Vincent, V. Humblet, M. Hermann, V. Comblin-
Tholet, M. F. Tweedle, US-6.056.939, 2000; b) R. Ruloff, G. van Ko-
ten, A. E. Merbach, Chem. Commun. 2004, 842–843.
[28] V. V. Martin, W. H. Ralston, M. R. Hynes, J. F. W. Keana, Bioconju-
gate Chem. 1995, 6, 616–623.
[36] W. P. Cacheris, S. C. Quay, S. M. Rocklage, Magn. Reson. Imaging
1990, 8, 467–481.
[37] S. Laurent, L. Vander Elst, F. Copoix, R. N. Muller, Invest. Radiol.
2001, 36, 115–122.
[38] Y. C. Ni, K. Adzamli, Y. Miao, E. Cresens, J. Yu, M.-P. Periasamy,
M. D. Adams, G. Marchal, Invest. Radiol. 2001, 36, 97–103.
[39] Y. Ni, C. Petrꢄ, Y. Miao, J. Yu, E. Cresens, P. Adriaens, H. Bosmans,
W. Semmler, A. L. Baert, G. Marchal, Invest. Radiol. 1997, 32, 770–
779.
[40] Y. Ni, E. Cresens, P. Adriaens, Y. Miao, K. Verbeke, S. Dymarkoski,
A. Verbruggen, G. Marchal, Acta Radiol. 2002, 9, S98–S101.
[41] G. Brunisholz, M. Randin, Helv. Chim. Acta 1959, 42, 1927.
[42] R. N. Muller, D. Declercq, P. Vallet, F. Giberto, B. Daminet, H. W.
Fischer, F. Maton, Y. Van Haverbeke, in Proceedings, ESMRMB, 7th
Annual Congress, Strasbourg (France), 1990, 394.
[43] P. Vallet, Ph.D. thesis, University of Mons-Hainaut (Belgium), 1992.
[44] B. Misselwitz, H. Schmitt-Willich, W. Ebert, T. Frenzel, H. J. Wein-
mann, MAGMA 2001, 12, 128–134.
[45] C. Burtea, S. Laurent, J M. Colet, L. Vander Elst, R. N. Muller,
Invest. Radiol. 2003, 38, 320–333.
[46] R. Knorr, A. Trzeciak, W. Bannwarth, D. Gillessen, Tetrahedron.
Lett. 1989, 1927–1930.
[47] Y. Ni, G. Marchal, J. Yu, G. Lukito, C. Petrꢄ, M. Wevers, A. L.
Baert, W. Ebert, C. S. Hilger, F. K. Maier, W. Semmler, Acad
Radiol. 1995, 2, 687–699.
[29] E. P. Mayoral, M. Garcꢈa-Amo, P. Lꢆpez, E. Soriano, S. Cerdꢉn, P.
Ballesteros, Bioorg. Med. Chem. 2003, 11, 5555–5567.
[30] C. F. G. C. Geraldes, A. M. Urbano, M. A. Hoefnagel, J. A. Peters,
Inorg. Chem. 1993, 32, 2426–2432.
Received: November 25, 2004
Published online: March 17, 2005
3086
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3077 – 3086