Maximizing the relaxivity of HSA-bound gadolinium complexes by simultaneous optimization of rotation and water exchange

Article abstract of DOI:10.1039/b714438e

Two new GdEGTA (EGTA = ethylene glycol-bis(2-aminoethylether)-N,N,N′, N′-tetraacetic acid) derivatives incorporating aromatic moieties into the oxoethylenic bridge have been prepared and characterised, their conjugates to HSA investigated and an unprecede

Full text of DOI:10.1039/b714438e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Maximizing the relaxivity of HSA-bound gadolinium complexes by
simultaneous optimization of rotation and water exchange{
Stefano Avedano,a Lorenzo Tei,a Alberto Lombardi,b Giovanni B. Giovenzana,b Silvio Aime,*c Dario Longoc
and Mauro Botta*a
Received (in Cambridge, UK) 20th September 2007, Accepted 16th October 2007
First published as an Advance Article on the web 26th October 2007
DOI: 10.1039/b714438e
ca. 80 and 120 mM21 s21 (per bound water molecule and per Gd,
Two new GdEGTA (EGTA = ethylene glycol-bis(2-aminoethy-
at 20 MHz and 298 K), depending on the electronic relaxation
parameters.1–3 At this frequency the optimal residence lifetime can
be set around 30 ns, while most of the complexes based on DTPA
and DOTA ligands present tM values between ca. 70 and 1000 ns.6
A slow rate of water exchange corresponds to an inefficient
transfer of the paramagnetic effect from the GdIII centre to the
bulk water and such an effect can severely limit the relaxivity in
the case of macromolecular systems. This is clearly exemplified by
the nearly linear dependence of the relaxivity of several GdIII
complexes non-covalently bound to human serum albumin (HSA)
and kex.7 As a consequence, the relaxivity of these adducts is
generally limited to the range 20 to 50 mM21 s21, i.e. much lower
than expected on the basis of the increase in molecular size (slower
tumbling rate, longer tR).
lether)-N,N,N9,N9-tetraacetic acid) derivatives incorporating
aromatic moieties into the oxoethylenic bridge have been
prepared and characterised, their conjugates to HSA investi-
gated and an unprecedented high relaxivity, close to that
predicted by theory, interpreted in terms of the combined effect
of restricted local rotation and fast rate of water exchange.
The advent and the successive rapid development of MRI into a
prominent diagnostic modality have seen a parallel and intense
activity in chemical research aimed at the design of effective,
specific and safe contrast enhancing agents (CA’s).1,2 These are
often required in order to improve the diagnostic content of the
tomographic image and shorten its acquisition time, intrinsically
long because of both the poor sensitivity of NMR, particularly at
the clinically used magnetic field strengths, and the long values of
However, more recently it has become evident that another
important factor opposes the attainment of high relaxivities in
macromolecular conjugates: the presence of a relatively fast
internal rotation superimposed on a global reorientation of the
system. The GdIII-based CA’s are typically conjugated to proteins,
dendrimers, micelles and virus capsids through a targeting group
connected to the coordination cage with a linker. The possible
internal rotation about the linker introduces a degree of flexibility
that results in an effective reorientational correlation time of the
GdIII–water–proton vector shorter than that associated with the
global rotation of the system.4,8–11 As a consequence any increase
in the molecular size does not translate into a proportional
decrease of the tumbling motion and then into a relaxivity
enhancement. Whereas a few possible strategies to overcome this
problem have been discussed,4,10,12 a clear and definitive example is
still lacking of a macromolecular complex that combines a fast
exchange of the inner-sphere water with an effective coupling
between the Gd–water vector and the tumbling motion of the
whole system.
1
the nuclear magnetic relaxation times, T1,2, of the H nuclei of
tissues. The T1’s for the water protons in tissues being much longer
than T2’s, T1-specific CA’s were soon established as the most
efficient systems and were the first to be clinically employed.
The gadolinium complexes of polyaminocarboxylate ligands are
the most common of these agents since they satisfy most of the
requirements necessary for safe in vivo use.3 Although the control
of parameters such as solubility, stability, inertness, biodistribution
and excretion pathways by suitable modifications of the chelate
basic structures has seen significant improvements over the years,
the efficacy of the GdIII chelates to catalyze the water protons’
relaxation times has not increased in a comparable manner.4 This
efficacy depends on a number of parameters that describe the
modulation of the dipolar interaction between the metal ion and
the proton nuclei of the water molecules belonging to its inner-,
second- and outer-spheres of hydration.2,4,5 Two of the most
important are the rate of exchange, kex = 1/tM, of the coordinated
water molecule(s) and the rate of reorientation, 1/tR, of the system.
Theory predicts that the gadolinium complexes endowed with
optimal values of tM and tR should have a relaxivity, r1p, between
A suitable and simple model should consist of a complex
characterized by small size, compactness, one inner-sphere water
molecule with a fast rate of exchange and by the presence of a rigid
targeting moiety capable of interacting with HSA. To this purpose
we designed two new complexes derived from the basic structure of
GdEGTA, since this compound was shown to present an optimal
aDipartimento di Scienze dell’Ambiente e della Vita, Universita` del
Piemonte Orientale ‘‘Amedeo Avogadro’’, Via Bellini 25/G, I-15100
Alessandria, Italy. E-mail: mauro.botta@mfn.unipmn.it
bDiSCAFF & DFB Center, Universita` degli Studi del Piemonte
Orientale ‘‘Amedeo Avogadro’’, Via Bovio 6, I-28100 Novara, Italy
cDipartimento di Chimica IFM, Universita` degli Studi di Torino, Via P.
Giuria 7, I-10125 Torino, Italy. E-mail: silvio.aime@unito.it
{ Electronic supplementary information (ESI) available: Reaction schemes
and characterization of ligands L1 and L2; VT 17O NMR data and
relaxometric titration curves of the Gd-complexes. See DOI: 10.1039/
b714438e
tM value of 30 ns (298 K) explained in terms of steric compression
at the binding site induced by the oxoethylenic bridge.13 In the new
L1 and L2 ligands (Scheme 1) the basic structure of EGTA was
modified in the central ethylenic moiety, rigidified by fusion with
an aromatic ring. The modified ligands were prepared from
1,2-arenediols. L1 was obtained from 2,3-naphthalenediol, whereas
4726 | Chem. Commun., 2007, 4726–4728
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Table 2 Selected parameters obtained from the analysis of the
relaxometric titrations (20 MHz; 298 K) and 1/T1 NMRD profiles
(298 K) of GdL1 and GdL2 with HSA
Parameter
GdL1
GdL2
KA/M21
880 ¡ 100
5.7 ¡ 0.2
68 ¡ 2
940 ¡ 90
7.0 ¡ 0.3
45 ¡ 3
298
r
1p
/mM21 s21 a
Scheme 1
298rb1p/mM21 s21 b
t
RL/ns
6.0 ¡ 1.7
41
0.60 ¡ 0.13
1.1 ¡ 0.2
41
0.31 ¡ 0.05
L2 was prepared from 4-nitrocatechol, in which the nitro group,
once reduced, represents a useful site for further functionalization.
The naphthalenic backbone or the pendant hydrophobic arm are
also useful to exploit conjugation with HSA. GdL1 represents our
best model and GdL2 a closely related complex only differing by
the presence of a flexible pendant group. It must be noted that the
kinetic and thermodynamic stabilities of these complexes are
probably not high enough to permit their safe in vivo use and thus
GdL1 and GdL2 should be considered as model compounds for a
study intended to be a proof of principle.
t
RG/nsc
S2
a
b
Relaxivity of the free complex (20 MHz). Relaxivity of the HSA-
bound complex (20 MHz). Fixed during the fitting.16
c
two complexes at the binding site. Whereas an rb1p value of
45 mM21 s21 is similar to that previously observed in the case of
DOTA derivatives featuring benzyloxymethylenic pendant groups,
the relaxivity of 68 mM21 s21 is the highest so far reported for a
Gd-complex conjugated to a protein. At 310 K the rb1p value is
65 mM21 s21, ca. 28% higher than for GdMS-325-HSA.12 The
field dependence of the water proton relaxivities was measured for
dilute solutions of the complexes (0.11 and 0.10 mM for GdL1 and
GdL2 respectively) in the absence and in the presence of HSA
(# 1.7 mM) at 298 K in the frequency range 2–70 MHz. These so-
called nuclear magnetic relaxation dispersion (NMRD) profiles
allow a detailed characterization of the paramagnetic solutes in
terms of a large set of structural and dynamic parameters.15 From
the values of KA, the relaxivities of the free complexes and the
diamagnetic contribution of the protein we calculated the r1bp
values at each frequency (Fig. 1). The shape of the profiles of the
two adducts are very similar, with a rather narrow peak at ca.
30 MHz, typical of slowly tumbling systems. The highest relaxivity
values are 78 and 50 mM21 s21 for GdL1–HSA and GdL2–HSA
respectively. Clearly, the relaxivity gain in physiological conditions
is expected to be sensibly lower due to the relatively weak protein
binding. The data were fitted to the Lipari–Szabo model-free
approach that takes into account the presence of an internal
The water exchange rate at the Gd centre was assessed by the
well established VT 17O NMR procedure,14 measuring the
temperature dependence of R2 at 2.12 T. For both complexes
rather similar exchange rates were obtained (tM $ 17 ¡ 3 ns at
298 K) nearly twice as fast as with GdEGTA (Table 1). It is worth
noting that, in spite of their similar exchange rates the two new
derivatives are endowed with negative values of the activation
entropy. This suggests the interesting possibility of a change in the
exchange mechanism that requires further investigation by variable
pressure 17O NMR. In any case, the presence of the aromatic
group in the ligand backbone does not modify the hydration
number (q = 1) of the complexes. This is confirmed by the
relaxivities of 5.7 and 7.0 mM21 s21 (298 K, pH 7.2, 20 MHz)
measured for GdL1 and GdL2 respectively, typical of monohy-
drated Gd-complexes and strictly proportional to their molecular
weight.5 Unlike the parent GdEGTA complex, the presence of
aromatic moieties allows the new complexes to set up non-covalent
interactions with the hydrophobic sites of HSA. The affinity
constants, KA, and relaxivities of the adducts with the protein, rb1p,
were determined, at 20 MHz and 298 K, from the analysis of the
data obtained from titration experiments of dilute solutions of the
complexes with HSA (Table 2).7
rotation, characterized by a correlation time tRL, superimposed on
16
a global motion described in terms of the correlation time tRG
.
The degree of correlation between the two motions is described by
the parameter S2 whose value is zero when the two motions are
completely independent; in the absence of local fluctuations the
complex is immobilized and S2 = 1. The value of the parameter
tRG was fixed to 40 ns in order to account for the global
Both complexes present similar and weak affinity constants as
expected on the basis of the chemical nature of the binding
moieties, but they significantly differ in the relaxivities of their
adducts, estimated to be 68 ¡ 2 and 45 ¡ 3 mM21 s21 for GdL1
and GdL2 respectively. The different relaxivities can be easily
attributed to the expected different degree of local flexibility of the
Table 1 Selected best-fit parameters obtained from analysis of the
17O NMR data (2.12 T)
Parameter
GdEGTAa
GdL1
5.3 ¡ 0.3
17.3 ¡ 2.9
239 ¡ 2
23.6 ¡ 0.2
4.5 ¡ 0.1
17 ¡ 2
GdL2
6.1 ¡ 0.4
16.0 ¡ 2.3
242 ¡ 3
23.5 ¡ 0.1
4.6 ¡ 0.3
18 ¡ 1
298
k
/107 s21
3.1 ¡ 0.2
42.7 ¡ 3.1
+42 ¡ 3
23.2 ¡ 0.1
3.4 ¡ 0.2
24 ¡ 1
ex
DH#/kJ mol21 b
DS#/J mol21 K21 c
(A/)/106 rad s21 d
D2/1019 s22 e
tV/psf
a
b
Data from ref. 13. Activation enthalpy of the exchange process.
c
d
Activation entropy of the exchange process. Scalar (hyperfine)
e
coupling constant. Trace of the squared zero field splitting (ZFS)
Fig. 1 1/T1 NMRD profiles and fits for GdL1–HSA (squares) and
GdL2–HSA (circles) at 298 K.
f
tensor. Correlation time of the modulation of the transient ZFS.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 4726–4728 | 4727

View Article Online
coordinated water molecule) for a model complex, unprecedented
for macromolecular conjugates of similar or even larger molecular
weight (Fig. 2). Finally, these results suggest that a further
refinement of the electronic relaxation should allow attainment of
relaxivities close to or above 100 mM21 s21
.
We thank MIUR (PRIN 2005: 2005039914) for financial
support. This work was performed within the frame of EU COST
Action D-38.
Notes and references
1 (a) S. Aime, M. Botta, M. Fasano and E. Terreno, Chem. Soc. Rev.,
1998, 27, 19–29; (b) P. Caravan, J. J. Ellison, T. J. McMurry and
R. B. Lauffer, Chem. Rev., 1999, 99, 2293–2352; (c) D. Parker,
R. S. Dickins, H. Puschmann, C. Crossland and J. A. K. Howard,
Chem. Rev., 2002, 102, 1977–2010.
2 The Chemistry of Contrast Agents in Medical Magnetic Resonance
Imaging, ed. E. Toth and A. E. Merbach, Wiley, New York, 2001.
3 R. B. Lauffer, Chem. Rev., 1987, 87, 901–927.
Fig. 2 Plot of the proton relaxivity r1p for selected macromolecular Gd-
complexes at 20 MHz and 310 K. Data taken from ref. 1b.(1) GdDTPA–
PEG I–polylysine (linear synthetic polymer); (2) Gadomer 17 (dendrimer)
(3) G4([NCS]N-bz-GdDO3A)38 (dendrimer); (4) MS-325-HSA (non-
covalently bound to HSA adduct); (5) Albumin–GdDTPA (covalently
bound to HSA macromolecular system); (6) MPEG–PL–GdDTPA (linear
synthetic polymer).
4 P. Caravan, Chem. Soc. Rev., 2006, 35, 512–523.
5 S. Aime, M. Botta and E. Terreno, Gd(III)-based Contrast Agents for
MRI, in Advances in Inorganic Chemistry, ed. R. van Eldik and
I. Bertini, Elsevier, San Diego, 2005, vol. 57, pp. 173–237.
6 (a) S. Aime, M. Botta, M. Fasano and E. Terreno, Acc. Chem. Res.,
1999, 32, 941–949; (b) L. Helm and A. E. Merbach, Chem. Rev., 2005,
105, 1923–1959.
7 S. Aime, M. Botta, M. Fasano and E. Terreno, in The Chemistry of
Contrast Agents in Medical Magnetic Resonance Imaging, ed. E. Toth
and A. E. Merbach, Wiley, New York, 2001, ch. 5, pp. 193–242.
reorientation of the protein,17 whereas tRL and S2 were used as
variable parameters. The results are given in Table 2 and are in full
agreement with the expected higher degree of rigidity of GdL1 at
the binding site, as indicated by the significantly longer values of
tRL (6.0 vs. 1.1 ns) and S2 (0.60 vs. 0.31).
´
8 S. Laus, A. Sour, R. Ruloff, E. To´th and A. E. Merbach, Chem.–Eur. J.,
2005, 11, 3064–3076.
9 S. Torres, J. A. Martins, J. P. Andre´, C. F. G. C. Geraldes,
Preliminary molecular modelling studies based on docking of
GdL1 and GdL2 to HSA are in agreement with relaxometric
experimental results. GdL1 is held fixed inside the entrance of the
binding pocket, allowing very small rotational movements. On
the other hand, GdL2 is more flexible and the interaction with
the HSA binding pocket leaves the cage free to rotate outside the
binding site, explaining the shorter tRL for this complex. Thus, the
different rotational dynamics arising from the different flexibility
of the targeting groups fully accounts for the different observed
relaxivities of GdL1 and GdL2 bound to HSA.
´
A. E. Merbach and E. To´th, Chem.–Eur. J., 2006, 12, 940–948.
10 J. Rudovsky´, M. Botta, P. Hermann, K. I. Hardcastle, I. Lukeˇs and
S. Aime, Bioconjugate Chem., 2006, 17, 975–987.
11 D. E. Prasuhn, Jr., R. M. Yeh, A. Obenaus, M. Manchester and
M. G. Finn, Chem. Commun., 2007, 1269–1271.
12 Z. Zhang, M. T. Greenfield, M. Spiller, T. J. McMurry, R. B. Lauffer
and P. Caravan, Angew. Chem., Int. Ed., 2005, 44, 6766–6769.
13 S. Aime, A. Barge, A. Borel, M. Botta, S. Chemerisov, A. E. Merbach,
U. Mu¨ller and D. Pubanz, Inorg. Chem., 1997, 36, 5104–5112.
14 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, 9333–9346.
This work shows for the first time that it is possible to achieve
relaxivity values very close to those predicted by theory by
controlling not only the rate of water exchange but also the
rotational dynamics of the system. The relevant result is the
remarkable gain in relaxivity observed (per Gd and per
15 S. H. Koenig and D. R. Brown, III, Prog. Nucl. Magn. Reson.
Spectrosc., 1990, 22, 487–567.
16 (a) G. Lipari and A. Szabo, J. Am. Chem. Soc., 1982, 104, 4546–4559;
(b) G. Lipari and A. Szabo, J. Am. Chem. Soc., 1982, 104, 4559–4570.
17 J. D. Dattelbaum, O. O. Abugo and J. R. Lakowicz, Bioconjugate
Chem., 2000, 11, 533.
4728 | Chem. Commun., 2007, 4726–4728
This journal is ß The Royal Society of Chemistry 2007