A bisphosphonate monoamide analogue of DOTA: A potential agent for bone targeting

Article abstract of DOI:10.1021/ja054905u

A new macrocyclic DOTA-like ligand (BPAMD) for bone imaging and therapy containing a monoamide bis(phosphonic acid) bone-seeking group was designed and synthesized. Its lanthanide(III) complexes were prepared and characterized by ¹H and ³¹P NMR spectroscopy. The Gd(III)-BPAMD complex was investigated in detail by ¹H and ¹⁷O relaxometric studies to inspect parameters relevant for its potential application as an MRI contrast agent. Sorption experiments were conducted with Gd(III) and Tb(III) complexes using hydroxyapatite (HA) as a model of bone surface. Very effective uptake of the Gd-BPAMD complex by the HA surface was observed in NMR experiments. Radiochemical studies with the (¹⁶⁰Tb-BPAMD)-HA system proved the sorption to be remarkably fast and strong on one hand and fully reversible on the other hand. The strong (Gd-BPAMD)-HA interaction was also supported by ¹H NMRD measurements in the presence of a hydroxyapatite slurry, which showed an increase of the rotational correlation time upon adsorption of the complex on the HA surface, resulting in a significant relaxivity enhancement. The amide-bis(phosphonate) moiety is the only factor responsible for the binding of the complex to HA.

Full text of DOI:10.1021/ja054905u

Published on Web 11/04/2005
A Bisphosphonate Monoamide Analogue of DOTA: A
Potential Agent for Bone Targeting
Vojteˇch Kub´ıcˇek,†,‡ Jakub Rudovsky´,‡ Jan Kotek,‡ Petr Hermann,‡
Luce Vander Elst,§ Robert N. Muller,§ Zvonimir I. Kolar,| Hubert Th. Wolterbeek,|
Joop A. Peters,*,† and Ivan Lukesˇ*,‡
Contribution from the Laboratory of Organic Chemistry and Catalysis, Delft UniVersity of
Technology, Julianalaan 136, 2628 BL Delft, The Netherlands, Department of Inorganic
Chemistry, Charles UniVersity, HlaVoVa 8, 128 40 Prague, Czech Republic, Department of
Organic and Biomedical Chemistry, NMR and Molecular Imaging Laboratory, UniVersity of
Mons-Hainaut, B-7000, Mons, Belgium, and Department of Radiation, Radionuclides &
Reactors, Faculty of Applied Sciences, Delft UniVersity of Technology,
2629 JB Delft, The Netherlands
Received July 21, 2005; E-mail: J.A.Peters@tnw.tudelft.nl; Lukes@natur.cuni.cz
Abstract: A new macrocyclic DOTA-like ligand (BPAMD) for bone imaging and therapy containing a
monoamide bis(phosphonic acid) bone-seeking group was designed and synthesized. Its lanthanide(III)
complexes were prepared and characterized by 1H and 31P NMR spectroscopy. The Gd(III)-BPAMD
1
complex was investigated in detail by H and 17O relaxometric studies to inspect parameters relevant for
its potential application as an MRI contrast agent. Sorption experiments were conducted with Gd(III) and
Tb(III) complexes using hydroxyapatite (HA) as a model of bone surface. Very effective uptake of the Gd-
BPAMD complex by the HA surface was observed in NMR experiments. Radiochemical studies with the
(
160Tb-BPAMD)-HA system proved the sorption to be remarkably fast and strong on one hand and fully
reversible on the other hand. The strong (Gd-BPAMD)-HA interaction was also supported by 1H NMRD
measurements in the presence of a hydroxyapatite slurry, which showed an increase of the rotational
correlation time upon adsorption of the complex on the HA surface, resulting in a significant relaxivity
enhancement. The amide-bis(phosphonate) moiety is the only factor responsible for the binding of the
complex to HA.
Chart 1. Structure of Ligands Discussed
Introduction
Magnetic resonance imaging (MRI) is one of the most
powerful noninvasive diagnostic methods in medicine. It relies
on variations of the water content in different tissues and on
differences in the composition of body liquids causing variations
of water proton relaxation rates at the clinically used magnetic
fields (0.5-2 T). Paramagnetic compounds can affect relaxation
rates, which have been exploited extensively to increase the
contrast of MRI images. Among the MRI contrast agents (CAs)
routinely used, the most widespread are based on the Gd(III)
ion, which is favorable because of its high magnetic moment
and long electron spin relaxation time. Due to the toxicity of
unbound lanthanide(III) ions, Gd(III) cannot be used as the aqua-
ion, but it has to be encapsulated in a thermodynamically and
kinetically stable complex. Polyaminocarboxylate ligands de-
rived from DOTA and DTPA (Chart 1) are commonly used for
medical applications. However, sequestering of the Gd(III) ion
reduces its ability to enhance the water 1H relaxation rate
(usually expressed as the relaxivity r1,1 which is the relaxation
rate enhancement of a millimolar solution of the Gd(III) complex
under study). As the contrast gained is usually limited, injection
of relatively high doses of CA (up to 5 g of CA for each
examination) is necessary so far.1-3 There are two major ways
to decrease the administrated dose of CA without losing
† Laboratory of Organic Chemistry and Catalysis, Delft University of
Technology.
‡ Charles University.
(1) Merbach, A. E., To´th, E., Eds. The Chemistry of Contrast Agents in Medical
Magnetic Resonance Imaging; John Wiley and Sons: New York, 2001.
(2) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem. ReV.
1999, 99, 2293-2352.
§ University of Mons-Hainaut.
| Department of Radiation, Radionuclides & Reactors, Delft University
of Technology.
9
10.1021/ja054905u CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 16477-16485
16477

A R T I C L E S
Kub´ıcˇek et al.
contrast: (i) by employing new Gd(III) complexes endowed
with much higher relaxivities; (ii) by improvement of the
biodistribution of CAs.4-6 The latter can be achieved by
conjugating the metal complex with a targeting group. Com-
monly, this is a biologically active compound, which exhibits
a high and specific affinity for a particular tissue. Hence, after
administration, the CA accumulates in the tissue concerned,
resulting in local contrast enhancement.
containing a bis(phosphonic acid) pendant arm. The properties
of its lanthanide(III) complexes with respect to its potential
application as a bone-targeted CA were evaluated. The sorption
and MRI contrast-enhancement abilities of the Ln(III) com-
plexes were tested using hydroxyapatite (HA) as a model of
bone surface. The sorption was studied using the 160Tb
radioisotope (half-life 72.1 days, â- 0.6 and 1.7 MeV, γ 879,
299, and 966 keV) as a model for its neighbor in the periodical
system, gadolinium. This isotope was effectively produced by
neutron activation of 159Tb (100% natural abundance) and could
be easily handled because of its long half-life.
In medical treatment of bone diseases, geminal bis(phospho-
nates) have been widely used during the last 20 years (Chart
1).7 Their success can be ascribed to the high affinity of the
bis(phosphonic acid) group, which is a non-hydrolyzable
analogue of natural pyrophosphate, to the bone surface. The
two other substituents in the bis(phosphonates) are attached to
the central carbon atom (Chart 1), and these groups are
responsible for the clinical effect of the drug (e.g., by affecting
osteoclasts and/or osteocystsscells responsible for bone forma-
tion and desorption).7 Bis(phosphonates) are adsorbed prefer-
entially to the active parts of bones (areas of growing and
pathological changes),7 which further enhances the desired
effect. More recently, CAs for MRI with phosphonates and bis-
(phosphonates) as targeting group have been developed.8-11
Introduction of the bis(phosphonate) group into the molecule
of a Gd(III) complex of a suitable polyaminocarboxylic ligand
results in a strong affinity of the resulting species for the desired
part of the bone surface, which may be reflected in an increase
of contrast of this tissue. Strong interaction of such species with
calcified tissues has been demonstrated for the bis(phosphonate)-
containing amides of DTPA.9,10 Unfortunately, these compounds
lack sufficient stability of their Ln(III) complexes for in vivo
applications.12 Furthermore, bis(phosphonates) attached to su-
perparamagnetic iron oxide (SPIO) particles have been reported
as promising bone-targeting CAs.11 Another approach to bone-
targeting CAs involves phosphonate analogues of polyami-
nocarboxylate ligands (DOTP, DTPP, EDTP; Chart 1), which
exhibit a strong adsorption on HA. However, upon binding of
the Gd(III) complex of DOTP, the relaxivity is practically silent
due to expulsion of second-sphere water molecules upon
adsorption.13,14 In the case of DTPP and EDTP, the strong
interaction of the phosphonate groups with the surface of HA
weakens the binding of the Gd(III) ion, resulting in a decrease
in stability of these complexes.13
Experimental Section
Materials. Commercially available chemicals had synthetic purity
and were used as received. Acetonitrile and toluene were dried by
distillation with P2O5. t-Bu3DO3A‚HBr was prepared according to a
published procedure.15 Hydroxyapatite (HA) was purchased from Fluka
(cat. number: 55496); the specific surface area is 53 m2/g as determined
by N2 adsorption by means of a Quantachrome Autosorb-6B apparatus.
Thin-layer chromatography was performed on TLC aluminum sheets
with silica gel 60 F254 (Merck KGaA, Germany). For the detection,
iodine vapors were used. Elemental analyses were done in the Institute
of Macromolecular Chemistry (Academy of Sciences of the Czech
Republic, Prague).
NMR Spectroscopy. 1H (300 MHz),13C (75.5 MHz), 17O (40.7
MHz), and 31P (121.5 MHz) NMR spectra were recorded on a Varian
Unity Inova-300 spectrometer using 5 mm sample tubes. Unless stated
otherwise, NMR experiments were performed at 298 K. Chemical shifts
are given in ppm (δ-scale). For the measurements in D2O, tert-butyl
alcohol was used as an internal standard with the methyl signal
referenced to 1.2 ppm (1H) or 31.2 ppm (13C). Water was used as an
external chemical shift reference for 17O resonances. The 31P chemical
shifts were measured with respect to 1% H3PO4 in D2O as an external
standard (substitution method). The pH values of the samples were
measured at ambient temperature using a Corning 125 pH meter with
a calibrated combined microelectrode purchased from Aldrich Chemical
Co. The pH values of the solutions were adjusted using dilute aqueous
solutions of NaOH and HCl.
Mass Spectrometry. Mass spectra were recorded on a Bruker
Esquire 3000 spectrometer equipped with an electrospray ion source
and an ion trap. All measurements were carried out in the positive mode.
Activity Measurements. The radioactivity of 160Tb in samples was
determined by measuring the intensity of emitted γ rays using a NaI-
(Tl)-scintillation detector based γ-ray counter (Wallac).
Tetraethyl (N,N-Dibenzyl)aminomethyl-bis(phosphonate) (2).16
Triethyl orthoformate (10.6 g, 71 mmol), diethyl phosphite (25.6 g,
186 mmol), and dibenzylamine (11.8 g, 59.9 mmol) were mixed in a
100 mL three-necked round-bottom flask. The solution was refluxed
under argon atmosphere (150 °C in an oil bath). After 5 h, the cooler
was removed and the solution was heated for 24 h at 160 °C under
argon flux to remove the ethanol. After cooling to RT, CHCl3 (300
mL) was added and the solution was washed with 3 × 60 mL of 5%
aqueous NaOH and 2 × 75 mL of brine. The organic fraction was
dried with anhydrous Na2SO4, and then the solvents were removed by
rotary evaporation. The resulting oil was purified by column chroma-
tography on silica gel (hexane-EtOH, from 100:0 to 0:100). Yield:
12.4 g (43%) of a viscous, colorless oil. Rf (hexane-EtOAc, 1:1) )
Here, we report on a new DOTA-like ligand ((4-{[(bis-
(phosphonomethyl))carbamoyl]methyl}-7,10-bis(carboxymethyl)-
1,4,7,10-tetraazacyclododec-1-yl)acetic acid (BPAMD; Chart 1))
(3) Aime, S.; Botta, M.; Fasano, M.; Terreno, E. Chem. Soc. ReV. 1998, 27,
19-29.
(4) Weissleder, R.; Mahmood, U. Radiology 2001, 219, 316-333.
(5) Jacques, V.; Desreux, J. F. Top. Curr. Chem. 2002, 221, 123-164.
(6) Aime, S.; Cabella, C.; Colombatto, S.; Crich, S. G.; Gianolio, E.; Maggioni,
F. J. Magn. Reson. Imaging 2002, 16, 394-406.
(7) Fleisch, H. Bisphosphonates in Bone Disease, 4th ed.; Academic Press:
London, 2000.
(8) Adzamli, I. K.; Gries, H.; Johnson, D.; Blau, M. J. Med. Chem. 1989, 32,
139-144.
(9) Adzamli, I. K.; Johnson, D.; Blau, M. InVest. Radiol. 1991, 26, 143-148.
(10) Adzamli, I. K.; Blau, M.; Pfeffer, M. A.; Davis, M. A. Magn. Reson.
Med. 1993, 29, 505-511.
1
3
0.2-0.3. H NMR (CDCl3, 300 MHz): δ 1.32 (td, 12H, -CH3, JHH
4
2
(11) Greb, W.; Blum, H.; Roth, M. PCT WO2003/097074, 2003; Chem. Abstr.
2003, 931193.
) 7.2 Hz, JPH ) 2.0 Hz), 3.55 (t, 1H, P-CH-P, JPH ) 25.2 Hz),
4.07 (m, 4H, N-CH2-Ph), 4.15 (m, 8H, O-CH2-), 7.20-7.45 (m,
(12) Adzamli, I. K.; Blau, M. Magn. Reson. Med. 1991, 17, 141-148.
(13) Bligh, S. W. A.; Harding, C. T.; McEwen, A. B.; Sadler, P. J.; Kelly, J.
D.; Marriott, J. A Polyhedron 1994, 13, 1937-1943.
(14) Alves, F. C.; Donato, P.; Sherry, A. D.; Zaheer, A.; Zhang, S.; Lubag, A.
J. M.; Merritt, M. E.; Lenkiski, R. E.; Frangioni, J. V.; Neves, M.; Prata,
M. I. M.; Santos, A. C.; de Lima, J. J. P.; Geraldes, C. F. G. C. InVest.
Radiol. 2003, 38, 750-760.
10H, Ar-H). 31P NMR (CDCl3, 300 MHz): δ 20.7 (d, JHP ) 25.2
2
Hz). MS: calcd 506.5, obsd 506.7 (M + Na+).
(15) Dadabhoy, A.; Faulkner, S.; Sammes, P. G. J. Chem. Soc., Perkin Trans.
2 2002, 348-357.
9
16478 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

A Potential Agent for Bone Targeting
A R T I C L E S
Tetraethyl Aminomethyl-bis(phosphonate) (3). A 250 mL two-
necked round-bottom flask was flushed with argon and charged with
10% Pd/C (0.8 g), and then a solution of 2 (4.0 g, 8.3 mmol) in absolute
EtOH (150 mL) was added. The mixture was refluxed with vigorous
stirring under a hydrogen atmosphere for 24 h. After filtration and
evaporation of volatiles, the product was obtained as a colorless oil.
the pH value was readjusted to 9 with 30% NaOH in H2O and the
solution obtained was heated at 80 °C overnight. After cooling to RT,
the pH value and the concentration of the sample were adjusted to the
desired values. Sometimes, a precipitate was formed after mixing of
the ligand and the metal salt solutions; however, upon increase of the
pH value, this precipitate dissolved.
Preparation of Solutions of 160Tb-BPAMD-Containing Tb-
BPAMD (160Tb-containing materials and their solutions in following
text are denoted as “labeled”; other Tb-containing materials are denoted
as “nonlabeled”). Labeled Tb(NO3)3‚5H2O was prepared by thermal
neutron irradiation of Tb(NO3)3‚5H2O target (20 mg, 46 mmol) in a
nuclear reactor (thermal neutron flux ) 4.49 × 1012 cm-2 s-1, irradiation
time ) 1.5 h) (Reactor Institute Delft, Faculty of Applied Sciences,
Delft University of Technology, Delft, The Netherlands). This material
(specific activity A ) 65 GBq mol-1) was used without further
purification.
The hydrate of BPAMD (180 mg, 252 mmol) was suspended in
H2O (3 mL). A solution of NaOH (20% in H2O) was slowly added
until all ligand was dissolved (pH ) 8-9). This mixture was slowly
dropped into a solution of Tb(NO3)3‚5H2O (10 mg, 23 mmol of labeled;
99 mg, 228 mmol of nonlabeled salts) in H2O (2 mL). After the addition
was completed, the pH was adjusted to 9 (with 20% NaOH in H2O),
after which the solution was heated at 80 °C for 12 h. The solution of
the complex obtained was transferred into a 25 mL volumetric flask
and diluted to a final concentration of 0.01 mol L-1 (A ) 60 MBq
L-1), and the pH was readjusted to the desired value (pH ) 7.5).
Estimation of the Time Required To Establish the Thermody-
namic Equilibrium. In a 10 mL vial, HA (50 mg) was suspended in
a TRIS buffer solution (0.10 mol L-1, pH ) 7.5). To this solution was
added the solution of the labeled Tb(III)-BPAMD complex (the final
concentration of labeled Tb(III) in samples was 1.00 mmol L-1, the
total volume of the solution was 3 mL), and the resulting mixture was
gently shaken. After incubation (1, 3, 5, 7, 24, 48, 72 h; one vial for
each point), the suspension was filtered through a Millipore filter (0.22
µm) and 1 mL of the filtrate was counted.
Reversibility of Tb-BPAMD Adsorption. In a 10 mL vial, HA
(50 mg) was suspended in a TRIS buffer solution (0.10 mol L-1, pH
) 7.5). The solution of labeled Tb(III)-BPAMD complex was added,
and the mixture was gently shaken for 3 days. Then, a solution of
nonlabeled complex was added (final concentration of labeled Tb(III)
in the samples was 0.33 mmol L-1, the final concentration of nonlabeled
Tb(III) in the samples was 1.50 mmol L-1, the total volume of the
solution was 3 mL), and the mixture obtained was shaken for another
3 days. The suspension was then filtered through a Millipore filter (0.22
µm), and 1 mL of the filtrate was counted.
1
Yield: 2.4 g (96%). H NMR (CDCl3, 300 MHz): δ 1.36 (t, 12H,
3
2
-CH3, JHH ) 6.8 Hz), 3.43 (t, 1H, P-CH-P, JPH ) 20.4 Hz), 4.23
(m, 8H, O-CH2-). 31P NMR (CDCl3, 121.5 MHz): δ 20.8 (d, 2JHP
20.4 Hz). MS: calcd 326.2, obsd 326.2 (M + Na+).
)
Chloroacetamide (4). A solution of 3 (2.4 g, 7.9 mmol) in dry
acetonitrile (50 mL) was slowly dropped into a suspension of Na2CO3
(5.2 g, 49 mmol) in a solution of chloroacetyl chloride (2.9 g, 25.6
mmol) in dry acetonitrile (50 mL), which was cooled at -40 °C. After
that, the mixture was allowed to warm to room temperature and stirred
overnight. After treatment with charcoal and filtration, the excess of
chloroacetyl chloride was removed by repeated evaporation with
toluene. Traces of chloroacetic acid were removed by stirring with
K2CO3 (1.0 g, 9.4 mmol) in dry toluene (50 mL) overnight, followed
by filtration and evaporation of the volatiles. The product was obtained
as a colorless oil in a yield of 3.1 g (96%). Rf (concentrated aqueous
ammonia-MeOH-EtOAc, 1:4:14) ) 0.8-0.9. 1H NMR (CDCl3, 300
3
4
MHz): δ 1.34 (td, 12H, -CH3, JHH ) 7.2 Hz, JPH ) 4.8 Hz), 4.09
(s, 2H, Cl-CH2-CO), 4.21 (m, 8H, O-CH2-), 5.02 (td, 1H, P-CH-
P, 2JPH ) 21.6 Hz, 3JHH ) 10.0 Hz), 7.50 (bd, 1H, NH, 3JHH ) 10 Hz).
31P NMR (CDCl3, 121.5 MHz): δ 16.1 (d, 2JHP ) 21.6 Hz). MS: calcd
402.7, obsd 402.4 (M + Na+).
Ligand BPAMD (1). A solution of 4 (3.1 g, 7.6 mmol) in dry
acetonitrile (50 mL) was slowly dropped into a solution of t-Bu3DO3A‚
HBr (2.6 g, 4.4 mmol) in a suspension of K2CO3 (3.5 g, 25 mmol) in
dry acetonitrile (50 mL), and then the mixture was stirred for 24 h.
After treatment with charcoal, filtration, and rotary evaporation of the
filtrate, the resulting oil was purified by column chromatography on
silica gel (concentrated aqueous ammonia-MeOH-EtOAc, from 1:8:
14 to 1:8:0). The yellow oil obtained after the chromatography was
dissolved in 30% HBr in dry AcOH (50 mL) and stirred for 24 h at
room temperature. After removal of volatiles on a rotary evaporator,
the resulting solid was suspended in dry THF (100 mL). The white
solid was filtered off, dissolved in water (20 mL), and purified on a
strongly acidic cation-exchange resin (Dowex 50×4, elution with water
followed by 10% aqueous pyridine). After rotary evaporation, the
remaining pyridine was removed on a weakly acidic cation-exchange
resin (Amberlite CG50, elution with water). The product-containing
fractions were combined and treated with charcoal, solids were filtered
off, and the filtrate was concentrated on a rotary evaporator to a final
volume of 25 mL. The product was obtained as a white fine powder
on standing for 5 days. The precipitate was filtered, washed with EtOH,
and dried over P2O5 at RT. The yield was 1.2 g of 1 as BPAMD‚7.5H2O
(39%). Rf (concentrated aqueous ammonia-MeOH-EtOAc, 1:4:14)
) 0.2-0.3. Anal. Calcd for C17H48N5O20.5P2: C, 28.66; H, 6.79; N,
Adsorption of Tb-BPAMD on HA. In a 10 mL vial HA (50 mg)
was suspended in a TRIS buffer solution (0.10 mol L-1, pH ) 7.5).
The solution of labeled Tb(III)-BPAMD complex was added (final
concentration of labeled Tb(III) in samples was 0.17-2.00 mmol L-1
,
the total volume of the solution was 3 mL), and the mixture obtained
was gently shaken for 3 days. The suspension was then filtered through
a Millipore filter (0.22 µm), and 1 mL of the filtrate was counted.
Desorption of Tb-BPAMD from HA by Methylenediphosphonic
Acid. In a 10 mL vial HA (50 mg) was suspended in a TRIS buffer
solution (0.10 mol L-1, pH ) 7.5). The solution of labeled Tb(III)-
BPAMD complex was added, and the mixture was gently shaken for
3 days. Then, a solution of methylenediphosphonic acid (MDP) was
added (the final concentration of labeled Tb(III) in the samples was
0.33 mmol L-1, the final concentration of MDP in the samples was
0-33.33 mmol L-1, the total volume of the solution was 3 mL), and
the mixture obtained was shaken for another 3 days. The suspension
was then filtered through a Millipore filter (0.22 µm), and 1 mL of the
filtrate was counted.
1
9.83. Found: C, 28.75; H, 6.84; N, 9.72. H NMR (D2O, 300 MHz):
δ 3.78 (bs, 8H, cyclen-CH2-N), 3.87 (bs, 8H, cyclen-CH2-N), 4.31
(s, 4H, N-CH2-CO), 4.44 (s, 2H, N-CH2-CO), 4.47 (s, 2H,
N-CH2-CO), 5.14 (t, 1H, P-CH-P, 2JPH ) 19.8 Hz). 13C{1H} NMR
(D2O, 75.5 MHz): δ 45.75 (1C, t, P-CH-P, 1JPC ) 130.2 Hz), 47.40
(2C, cyclen-CH2-N), 47.71 (2C, cyclen-CH2-N), 48.32 (4C, cy-
clen-CH2-N), 52.61 (1C, N-CH2-pendant), 52.77 (1C, N-CH2-
pendant), 53.32 (2C, N-CH2-pendant), 166.18 (1C, -CON-), 169.17
(2C, -COO-), 170.24 (1C, -COO-). 31P NMR (D2O, 121.5 MHz):
2
δ 13.8 (d, JHP ) 19.8 Hz). MS: calcd 578.4, obsd 578.2 (M + H+).
Preparation of the Lanthanide(III) Complexes. General Proce-
dure. The hydrate of BPAMD (100 mg, 0.140 mmol) was suspended
in H2O (0.4 mL). A solution of NaOH (30% in H2O) was slowly added
until all ligand was dissolved. The pH of the resulting solution was
8-9. This mixture was slowly dropped into a solution of hydrated
lanthanide(III) chloride (0.9 equiv, 0.127 mmol) in H2O (0.2 mL). Then,
Competitive Sorption of Ligand BPAMD and Tb-BPAMD
Complex on HA. In a 10 mL vial HA (50 mg) was suspended in a
TRIS buffer solution (0.10 mol L-1, pH ) 7.5). A solution of labeled
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16479

A R T I C L E S
Kub´ıcˇek et al.
Scheme 1. Synthesis of the Ligand BPAMDa
Tb(III)-BPAMD complex and a solution of ligand BPAMD were
added (final concentration of labeled Tb-BPAMD complex in samples
was 0.55 mmol L-1 (experiment A), 1.10 mmol L-1 (experiment B),
final concentration of free ligand in samples was 0-1.67 mmol L-1
(experiment A), 0-3.33 mmol L-1 (experiment B), the total volume
of the solution was 3 mL), and the mixture obtained was gently shaken
for 3 days. The suspension was then filtered through a Millipore filter
(0.22 µm), and 1 mL of the filtrate was counted.
Preparation of the Slurry of HA with Adsorbed Gd-BPAMD.
In a 5 mL plastic vial, HA (0.3 g) was suspended in a TRIS buffer
solution (1.5 mL, 0.1 mol L-1, pH ) 7.5). A solution of Gd-BPAMD
or Gd-DTPA (0.08 mL, 20 mmol L-1) was added, and the mixture
was shaken for 24 h. The suspension was then transferred into an NMR
tube. After standing for 4 h, the ratio clear solution/slurry remained
unchanged. The solution (∼0.75 mL) was decanted and filtered into a
clean NMR tube, and then relaxometric measurements with both sample
tubes were performed.
a (i) Argon atmosphere, 160 °C; (ii) H2, Pd/C, EtOH, reflux; (iii)
ClCH2C(O)Cl, acetonitrile, Na2CO3, -40 °C, (iv) t-Bu3DO3A, acetonitrile,
K2CO3, RT; (v) 30% HBr in dry AcOH.
Variable-Temperature 17O NMR Study of Gd-BPAMD. An 80
mM solution of Gd-BPAMD was prepared as described above, and
the pH value of this solution was adjusted to 7.5. Then, 10 µL of 17O-
enriched water (10% H217O) was added. The exact concentration of
Gd(III) was determined by the bulk magnetic susceptibility shift
measurement (BMS) as described previously.17 All NMR spectra were
acquired without frequency lock. To correct the 17O NMR shift for a
contribution of the BMS, the difference between chemical shifts of
proton signals of tert-butyl alcohol in the paramagnetic sample and in
pure water was measured according to the procedure described
previously.18 Transversal (1/T2) relaxation rates were obtained by the
Carr-Purcell-Meiboom-Gill pulse sequence.19
Measurement of 1/T1 1H NMRD Profiles. A 2 mM sample solution
at pH 7.5 was prepared as described above. The exact concentration of
Gd(III) was determined using the BMS method.17 The 1/T1 NMRD
profiles were measured at 298 K at magnetic field strengths between
4.7 × 10-4 and 0.35 T using a Stelar SpinMaster FFC-2000 relaxometer.
Measurements at 0.47 and 1.42 T were performed on a Bruker Minispec
PC-20 and Bruker Minispec mq60, respectively. The experimental data
were fitted simultaneously with 17O NMR data by means of a least-
squares fitting procedure using the Micromath Scientist program version
2.0 (Salt Lake City, UT) as described previously.20
ligand 1 was obtained by hydrolysis of ester 5 in a solution of
HBr in dry AcOH.
Preparation and Characterization of Lanthanide(III)
Complexes of BPAMD. The complexation process was moni-
tored in the case of the Eu(III) complex by 31P NMR.
Immediately after mixing of solutions of Eu(III) and the ligand
BPAMD, the resulting pH was 2-4 and the resonances in the
31P spectrum were so broad that they disappeared in the baseline.
At this stage, sometimes a precipitate was formed that dissolved
after the solution was rendered alkaline to pH 9. Then, the
spectrum displayed a broad resonance (δ ) 60.2 ppm, ∆ν1/2
)
7310 Hz) in addition to the resonance for the excess of free
ligand. Upon heating the solution at 80 °C overnight, two sharp
resonances belonging to Eu-BPAMD complex of about equal
intensity were observed at 19.1 and 15.1 ppm (see Supporting
Information, Figure S1).
Most likely, these phenomena can be ascribed to a complex
formation mechanism involving several intermediates. When
the components are mixed (at pH 2-4), only bis(phosphonate)
oxygen atoms will be coordinated to the lanthanide(III) ion,
whereas the carboxylate pendant arms and macrocyclic amine
groups remain protonated.21,22 The incidental precipitation may
be due to formation of poorly soluble lanthanide salts typical
for bis(phosphonate) complexes with divalent and trivalent metal
ions.23 Once the pH value is set to 9, most likely, the pathway
that has been established for the parent compound DOTA24 is
followed. Thus initially, an “out-of-cage” complex is formed,
in which both the carboxylate and phosphonate oxygen atoms
bind the Ln(III) ion and the macrocyclic nitrogen atoms are
still protonated. Then in a last and rate-determining step the
nitrogen atoms are deprotonated, while the Ln(III) ion moves
into the cage. This rearrangement is strongly pH dependent, as
reflected in the complex formation rate. At pH 6.5 the complex
formation takes several days, whereas at pH 8.5, it is completed
overnight.
Results and Discussion
Synthesis of the Ligand. Ligand BPAMD (1) was synthe-
sized from the tris(tert-butyl ester) of DO3A (t-Bu3DO3A) and
bis(phosphonate) 4 (Scheme 1). The latter was prepared by
reaction of dibenzylamine, triethylorthoformate, and diethyl
phosphite to give intermediate 2, followed by debenzylation with
H2 and Pd/C as the catalyst.16 The yield of the condensation
was relatively low (43%) due to the formation of many
byproducts (monophosphonate derivatives, formic acid amides
and esters). The deprotected aminomethyl-bis(phosphonate) (3)
was converted quantitatively into compound 4 by means of a
substitution reaction with an excess of chloroacetyl chloride.
The chloroacetylamide (4) was used to alkylate the tris(tert-
butyl ester) of DO3A. The conversion was almost quantitative;
however the yield after purification was only 65% due to some
decomposition of the product during chromatography. Finally,
1
The H NMR spectra of the Ln-BPAMD complexes show
close similarity with those of the corresponding Ln-DOTA
complexes,25,26 although they are much more complex due to
(16) Kantoci, D.; Denike, J. K.; Wechter W. J. Synth. Commun. 1996, 26, 2037-
2043.
(17) Corsi, D. M.; Platas-Iglesias, C.; van Bekkum, H.; Peters, J. A. Magn.
Reson. Chem. 2001, 39, 723-726.
(21) Moreau, J.; Guillon, E.; Pierrard, J.-C.; Rimbault, J.; Port, M.; Aplincourt,
M. Chem. Eur. J. 2004, 10, 5218-5232.
(18) Zitha-Bovens, E.; Vander Elst, L.; Muller, R. N.; van Bekkum, H.; Peters,
J. A. Eur. J. Inorg. Chem. 2001, 3101-3105.
(22) Bollinger, J. E.; Roundhill, D. M. Inorg. Chem. 1993, 32, 2821-2826.
(23) Zeevaart, J. R.; Jarvis, N. V.; Louw, W. K. A.; Jackson, G. E.; Cukrowski,
I.; Mouton, C. J. J. Inorg. Biochem. 1999, 73, 265-272.
(24) To´th, EÄ .; Bru¨cher, E.; Lazar, I.; To´th, I. Inorg. Chem. 1994, 33, 4070-
4076.
(19) Meiboom, S.; Gill, D. ReV. Sci. Instrum. 1958, 29, 688-691.
(20) Kotek, J.; Lebdusˇkova´, P.; Hermann, P.; Vander Elst, L.; Muller, R. N.;
Maschmeyer, T.; Lukesˇ, I.; Peters, J. A. Chem. Eur. J. 2003, 9, 5899-
5915.
(25) Aime, S.; Botta M.; Ermondi, G. Inorg. Chem. 1992, 31, 4291-4299.
9
16480 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

A Potential Agent for Bone Targeting
A R T I C L E S
Figure 1. 1H NMR spectra (400 MHz, pH ) 7, 25 °C) of (A) Eu-BPAMD, (B) Eu-DOTA, (C) Yb-BPAMD, and (D) Yb-DOTA (see also Supporting
Information). The signals of solvents are marked with asterisks.
the lower symmetry of the complexes (C1 rather than C4).
Typically, two sets of spectra were observed, in which the
lanthanide-induced 1H shifts for one of these sets were
significantly larger than for the other one. This can be explained
by the presence of two diastereomeric structures as has been
well established for Ln(III) complexes of DOTA derivatives.
From a number of solid state and solution studies on the
lanthanide(III) complexes of DOTA-like ligands it is known
that these complexes occur in solution as a mixture of two
diastereomeric forms which differ in mutual orientation of the
“nitrogen” and “oxygen” planes of the coordination polyhe-
dron: the square-antiprismatic (usually labeled M) and twisted
square-antiprismatic (usually labeled m).25 The ratio of m/M
commonly changes along the lanthanide series; the smaller ions
at the end of the series prefer the M form. From 1H NMR spectra
of Ln-BPAMD, it can be concluded that the m/M ratios in
solution parallel those reported for Ln-DOTA complexes.25,27
with a ratio of intensities of 0.18:0.82, which is in accord with
the m/M ratio of 0.21:0.79 reported for the Eu-DOTA
complex.27
Interestingly, two separate 31P NMR resonances with ap-
proximate intensities in a ratio 1:1 were observed for the
BPAMD complexes for Eu(III), Dy(III), and Yb(III). Since
the1H NMR spectra of Dy3+ and Yb3+ complexes showed that
both occur exclusively as the M isomer, the occurrence of two
31P resonances with equal intensities must be ascribed to a
nonequivalent spatial orientation of the two phosphorus atoms,
which are in slow interconversion on the NMR time scale (see
Supporting Information, Figure S3). This was nicely demon-
strated with 31P spectra of the Eu(III) complex when the two
31P signals coalesced when the sample was heated at 80 °C
(Figure S4). A similar behavior was previously reported for other
phosphonate-containing DOTA monoamides, where it was
accounted for in terms of hindered rotation around the amidic
bond.28 Alternatively, this phenomenon can be explained with
a single rotamer and hindered rotation of the bis(phosphonate)
moiety due to the formation of a hydrogen bond between a
phosphonate oxygen and the Ln(III)-bound water molecule.
Molecular models indicate that one of the phosphonate groups
can be close enough to the coordinated water to allow such a
1
In the H spectrum of the Yb-BPAMD complex a single set
of signals corresponding to the M isomer was found (Figure
1C). However, the spectra of the Eu-BPAMD complex (Figure
1A and Figure S2) displayed two sets of signals for m and M
(26) Howard, J. A. K.; Kenwright, A. M.; Moloney, J. M.; Parker, D.; Port,
M.; Navet, M.; Rousseau, O.; Woods, M. Chem. Commun. 1998, 13, 1381-
1382.
(27) Aime, S.; Botta, M.; Fasano, M.; Marques, M. P. M.; Geraldes, C. F. G.
C.; Pubanz, D.; Merbach, A. E. Inorg. Chem. 1997, 36, 2059-2068.
(28) Aime, S.; Botta, M.; Garino, E.; Geninatti Crich, S.; Giovenzana, G.;
Pagliarin, R.; Palmisano, G.; Sisti, M. Chem. Eur. J. 2000, 6, 2609-2617.
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16481

A R T I C L E S
Kub´ıcˇek et al.
hydrogen bond. An inspection of molecular models also
indicates that the P atoms are about 6 Å from the Ln(III) ion.
This is in agreement with Ln(III)-induced 31P longitudinal
relaxation rate measurements on these complexes. For example,
for the Dy(III) complex the 1/T1 values measured were 9.95
and 12.93 s-1. With the use of an equation taking into account
both the dipolar and the Curie relaxation,20,29 the distances
between the Dy(III) ion and the two P atoms were estimated to
be 6.5 and 6.2 Å, respectively. Anyway, these data show that
the phosphonate groups are not bound to the Ln(III) ion.
The hydration number of the complex was determined by
measuring the 17O NMR chemical shifts of water induced by
the presence of the dysprosium complex. For its suitable
magnetic properties, Dy(III) is generally selected for this
purpose. Previously, it has been shown that the Dy(III)-induced
water 17O shift of a coordinated water molecule in a Dy(III)
complex is almost independent of the other ligands coordinated
to the Dy(III) ion.29-31 Consequently, the Dy(III)-induced shift
(DIS) can be used to determine the hydration numbers of the
Dy(III) complexes. The DIS measured for the Dy-BPAMD
system at 300 K (extrapolated to a molar ratio of Dy(III)/water
) 1) was -2253 ppm. The DIS determined for an aqueous
solution of DyCl3 under the same conditions amounts to
-19256 ppm. If it is assumed that in the absence of organic
ligands Dy(III) is coordinated to eight water ligands, it can be
concluded from these data that q ) 1 for Dy-BPAMD and
that the Ln(III) ions in the Ln-BPAMD complexes are nine-
coordinated, which is consistent with a general coordination
mode of this class of compounds.1,2 Therefore complexes of
the title ligand should be defined as [Ln(H2O)(HxBPAMD](4-x)-
(at pH ) 7, the bis(phosphonate) group is mono- or diproto-
nated32,33).
Relaxometric Studies of Gd(III) Complex. To evaluate
parameters governing the relaxivity of the Gd-BPAMD com-
plex, variable-temperature 17O transversal relaxation rates (T2)
and a 1H NMRD profile were measured (Figure 2). These data
were fitted simultaneously with a set of equations usually used
to evaluate these parameters (see Supporting Information).1,20
Since these equations contain a large number of parameters, it
was necessary to fix some of them either to independently
determined values or to values estimated for related complexes
of DOTA-like ligands.1,34 The number of Ln(III)-coordinated
water molecules (q), a crucial parameter, was determined by
the DIS experiment to be q ) 1 (see above) and fixed during
the calculation. The hyperfine coupling constant of Gd-O
interaction, A/p, was fixed at -3.6 × 106 rad s-1, which is the
value generally observed for complexes with q ) 1.1,29 The
diffusion coefficient DGdH, needed for the calculation of the
outer-sphere contribution to the relaxation, was taken to be equal
to 2.0 × 10-9 m2 s-1. The distances of gadolinium to the water
Figure 2. (A) 17O R2r temperature dependence (80 mM, 300 MHz); (B)
1H NMRD profile (2 mM, 298 K) of Gd-BPAMD in aqueous solution
(pH ) 7.5); the solid line represents the best simultaneous fit of the data
(see Table 1).
Table 1. Comparison of the Relaxometric Parameters (298 K)
Found for the Gd(III) Complex of the Title Ligand, DOTA, and
Related Monoamides (R1, R2 are groups attached to the amide
nitrogen atom)
r
1 (20 MHz)
1
ligand
∆
2 [s-2/1019
]
τv [ps]
τr [ps]
τM
[
µ
s]
[s-1 mM-
]
BPAMD
DOTAa
3.7 ( 0.2
1.6
1.8
1.6
3.3
17 ( 1
11
21
23
17
88 ( 3
77
97
91
78
1.18 ( 0.06
0.244
1.6
1.6
1.5
5.3
4.8
6.2
6.0
5.6
b
L1
L2
b
b
L3
a Ref 34. b Ref 28; L1: R1) -CH2PO3H2; R2) -CH2PO3H2; L2: R1)
-CH2COOH; R2) -CH2PO3H2; L3: R1) -H; R2) -CH2PO3H2.
O atom and H atom (RGdO and RGdH) were fixed at 2.5 and 3.1
Å, respectively, whereas the distance of closest approach of a
bulk water molecule to gadolinium was taken as 3.65 Å. The
most relevant parameters obtained are compared with parameters
reported for other Gd(III) complexes of DOTA-like ligands in
Table 1 (for complete fitting results, see Supporting Information,
Table S1).
As can be seen, the values of these parameters are fully
comparable with those reported on similar systems. The values
of parameters determining the electronic relaxation, the mean
square zero-field splitting energy (∆2), and the correlation time
(29) Peters, J. A.; Huskens, J.; Raber, D. J. Prog. NMR Spectrosc. 1996, 28,
283-350.
(30) Kieboom, A. P. G.; van der Toorn, J. M.; Peters, J. A.; Bove´e, W. M. M.
J.; Sinnema, A.; Vijverberg, C. A. M.; van Bekkum, H. Recl. TraV. Chim.
Pays-Bas 1978, 97, 247.
(31) Alpoim, M. C.; Urbano, A. M.; Geraldes, C. F. G. C.; Peters, J. A. J. Chem.
Soc., Dalton Trans. 1992, 463.
(32) Dyba, M.; Kozlowksi, H.; Tlalka, A.; Leroux, Y.; El Manouni, D. Pol. J.
Chem. 1998, 72, 1148-1153.
(33) Vanˇura, P.; Jedina´kova´-Krˇ´ızˇova´, V.; Hakenova´, L.; Munesawa, Y. J.
Radioanal. Nucl. Chem. 2000, 246, 689-692.
(34) Powell, D. H.; Dhubhghaill, O. M. N.; Pubanz, D.; Helm, L.; Lebedev, Y.
S.; Schlaepfer, W.; Merbach, A. E. J. Am. Chem. Soc. 1996, 118, 9333-
9346.
9
16482 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

A Potential Agent for Bone Targeting
A R T I C L E S
Figure 3. Determination of the time required to establish the thermody-
namic equilibrium of sorption of Tb-BPAMD on HA.
Figure 4. Adsorption isotherm of labeled Tb-BPAMD on the surface of
HA. The solid line is obtained by linear regression.
Table 2. Affinity Constants of Bis(phosphonates) with HA (298 K)
of the zero-field splitting modulation (298τv) for the title ligand
are close to those reported for similar ligands (Table 1). The
values for the low-field limit of the electronic relaxation time,
τs0, can be calculated to be 132 and 473 ps for Gd-BPAMD
and Gd-DOTA, respectively. This difference can be ascribed
to the difference in symmetry between these two complexes.1,28
The rotational correlation time (298τr) was found to be 88 ps, as
should be expected for a Gd(III) complex of this size. The
relaxivity (5.34 s-1 mM-1 at 20 MHz, 298 K, pH ) 7.5) was
found to be somewhat higher than that for the Gd-DOTA
K/103 [mol-1 L]
HEDPa,b
3.3
0.7
MDPa,b
Tb-BPAMD
210
a HEDP ) 1-hydroxyethane-1,1-bis(phosphonic acid); MDP ) methyl-
enediphosphonic acid. b Ref 42.
presence of fresh apatite. Having confirmed the full revers-
ibility of the sorption/desorption process, we treated the corre-
sponding data on the basis of a linearized Langmuir adsorption
isotherm:40
complex despite the rather long water residence time (298τM
)
1.2 µs) and the short electronic relaxation time. The negative
effect of these two parameters on the relaxivity is compensated
by an increase in relaxivity due to the somewhat larger τr.
Complexes of related ligands often have a substantial
contribution to the relaxivity by water molecules in the second
sphere as a result of an extensive hydrogen-bond network around
the phosphonic/phosphinic acid moieties.20,35-37 Introduction of
second-sphere parameters into the model did not result in better
fits of the experimental data, which suggests that the second-
sphere contribution is not important in the present case.
In addition, we measured the pH dependence of the relaxivity
to obtain the typical shape for relaxivity influenced by proto-
tropic exchange38,39 of water protons in both acidic and basic
regions of the profile (for more details, see Supporting Informa-
tion and Figure S5).
Sorption Experiments with Tb-BPAMD. Since the title
ligand was designed for bone targeting, an investigation of its
sorption abilities was of primary importance. As a model system
of bone tissue, we choose an aqueous suspension of hydroxy-
apatite (HA). The sorption of the Tb-BPAMD complex labeled
with the 160Tb radioisotope proved to be remarkably fast: more
than 95% of the complex was adsorbed during the first hour of
the experiment (Figure 3).
c/X ) 1/KXm + c/Xm
where c represents the equilibrium concentration and X is the
equilibrium-specific adsorbed amount (Figure 4). Linear regres-
sion afforded the affinity constant K and the maximal adsorption
capacity Xm (application of this model is reasonable due to
negligible dissociation of the Ln(III) DOTA-like complexes in
the presence of HA41). Comparison of the data obtained with
those reported for related systems (Table 2) indicates that the
sorption of labeled Tb-BPAMD is extremely favorable: K was
found to be 2.1 × 105 L mol-1, which is about 2 orders of
magnitude higher than the values reported for other simple bis-
(phosphonates).42 To estimate the thermodynamic stability of
the (Tb-BPAMD)-HA adduct, desorption experiments with
methylenediphosphonic acid (MDP) were carried out. Impor-
tantly, the presence of a 100-fold molar excess of MDP resulted
in only 6% desorption of labeled Tb-BPAMD from the HA
surface.
(35) Botta, M. Eur. J. Inorg. Chem. 2000, 399-407.
(36) Rudovsky´, J.; C´ıgler, P.; Kotek, J.; Hermann, P.; Vojt´ısˇek, P.; Lukesˇ, I.;
Peters, J. A.; Vander Elst, L.; Muller, R. N. Chem. Eur., J. 2005, 11, 2373-
2384.
Definitely, the reversibility of the sorption process is one of
the crucial points, since it determines the in vivo applicability
of Gd-BPAMD as MRI CA. To verify the reversibility, we
performed a competition experiment between already adsorbed
labeled complex and free nonlabeled complex. After 3 days,
the equilibrium activity was the same as that of a mixture of
labeled/nonlabeled complex of the same concentrations in the
(37) Rudovsky´, J.; Kotek, J.; Hermann, P.; Lukesˇ, I.; Mainero, V.; Aime, S.
Org. Biomol. Chem. 2005, 3, 112-117.
(38) Aime, S.; Botta, M.; Fasano, M.; Terreno, E. Acc. Chem. Res. 1999, 32,
941-949.
(39) Aime, S.; Gianolio, E.; Barge, A.; Kostakis, D.; Plakatouras, I. C.;
Hadjiljadis, N. Eur. J. Inorg. Chem. 2003, 2045-2048.
(40) Langmuir, I. J. Am. Chem. Soc. 1918, 40, 1361-1403.
(41) Li, W. P.; Ma, D. S.; Higginbotham, C.; Hoffman, T.; Ketring, A. R.;
Cutler, C. S.; Jurisson, S. S. Nucl. Med. Biol. 2001, 28, 145-154.
(42) Claessens, R. A. M. J.; Kolar, I. Z. Langmuir 2000, 16, 1360-1367.
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16483

A R T I C L E S
Kub´ıcˇek et al.
Figure 6. Longitudinal 1H relaxation rates of the solution and the HA layer
obtained after sedimentation of an equilibrated suspension of HA containing
Gd-BPAMD and Gd-DTPA in water (298 K, 300 MHz).
Figure 5. Competitive sorption experiment of the ligand BPAMD and
labeled Tb-BPAMD complex: (f) initial concentrations: labeled Tb-
BPAMD 0.55 mmol L-1, BPAMD 0-1.67 mmol mL-1; (9) initial
concentrations: labeled Tb-BPAMD 1.1 mmol L-1, BPAMD 0-3.33 mmol
L-1; (O) initial concentration: labeled Tb-BPAMD 0.17-3.33 mol L-1
.
The area on the HA surface occupied by one molecule of
the complex was estimated by combining the fitted value of
the maximal adsorption capacity (Xm ) 4.0 × 10-5 mol g-1
)
with the known active surface size of the HA sample (53 m2
g-1). The calculated value 221 Å2 is in excellent agreement with
an estimation obtained with molecular modeling:43 by assuming
free rotation of the complex around the anchored bis(phospho-
nic) group, the surface occupied by one molecule was calculated
to be 227 Å2. Hence, it can be concluded that a monomolecular
layer was formed on HA by sorption of the Tb-BPAMD
complex.
Figure 7. 1H NMRD profiles of pure HA (+), (Gd-BPAMD)-HA (4),
the paramagnetic contribution of the (Gd-BPAMD)-HA slurry (b), and
Gd-BPAMD complex in aqueous solution (O). The Gd(III) concentration
in the two latter samples is 0.8 mM.
To further elucidate the origin of the favorable sorption ability
of the ligand BPAMD, we designed a competitive co-sorption
experiment between free BPAMD ligand and labeled Tb-
BPAMD complex at variable concentrations of the components
in the mixture. Plots of the activity obtained versus the sum of
initial ligand and complex concentrations were almost identical
for all ratios of free ligand and complex applied (Figure 5).
This reveals that the affinity constants and maximum adsorption
capacity of both species are similar. It may be concluded that
the bis(phosphonate) moiety attached to the macrocycle via the
amide nitrogen atom is almost exclusively responsible for the
unusually strong adsorption of BPAMD complexes to HA.
Relaxometric Experiments with Gd-BPAMD Adsorbed
on the HA Surface. As a complementary method for monitoring
the sorption processes of Gd-BPAMD, we applied a slightly
modified relaxometric procedure used for qualitative charac-
terization of the interaction of Gd-DOTP interaction with HA
by Sherry et al.14 A solution of the Gd-BPAMD complex in
TRIS buffer was shaken overnight with HA. The suspension
was then transferred into an NMR tube. After sedimentation
(completed in 4 h), the supernatant was decanted off and filtered
into another NMR tube. The relaxivities of both the supernatant
and the remaining slurry were measured. The experiment was
arranged to obtain approximately the same volume of the slurry
and the supernatant, allowing a direct comparison of the
relaxation rates of both samples. The results show that the Gd-
BPAMD complex was almost quantitatively adsorbed on the
HA surface (Figure 6). The concentration of Gd-BPAMD in
the supernatant was very low (0.06 mmol L-1, i.e., 6% of the
initial concentration), which was confirmed by a measurement
of the Gd(III)-induced bulk magnetic susceptibility shift.17
Additionally, control samples of Gd-DTPA under the same
conditions were measured (negligible sorption abilities of the
Gd-DTPA complex were previously reported41). The slightly
different relaxation rates of Gd-DTPA in the slurry and the
supernatant (Figure 6) should be ascribed only to the different
physicochemical environments of the Gd-DTPA in both phases.
To further investigate the impact of sorption on the relaxivity
of Gd-BPAMD, the 1H NMRD profile of the HA slurry upon
sorption was measured. In fact, the profile obtained is a
superposition of a paramagnetic contribution coming from the
Gd(III) complex and a diamagnetic contribution coming from
HA (Figure 7). The paramagnetic contribution to the 1H NMRD
1
profile was obtained by subtraction of the H NMRD profile
measured on a slurry of pure HA.44 The resulting profile showed
a local maximum at 10-20 MHz, which is typical for Gd(III)
complexes with a relatively large τr. Notably, upon adsorption
of the Gd-BPAMD on the HA surface, the relaxivity at 20
MHz is 24.0 s-1 mM-1, which is 4.5 times higher than that for
(43) HyperChem Release 7.5 for Windows, Molecular Modeling System;
Hypercube, Inc., 2002.
9
16484 J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005

A Potential Agent for Bone Targeting
A R T I C L E S
“free” complex. This observation is in contrast with results
reported for Gd-DOTP (tetraphosphonated analogue of DOTA
with no water molecule in the first sphere), where an expulsion
of the second-sphere water molecules damped its relaxivity
totally.14
swiftly and strongly. Additionally, the sorption of the complexes
was demonstrated to be reversible. The H NMRD of Gd-
1
BPAMD immobilized on HA afforded a profile with a high-
field dispersion typical for a slow-tumbling Gd(III) complex.
This significant increase of rotational correlation time upon
binding of the Gd-BPAMD complex on a hydroxyapatite
surface results in a significant increase in relaxivity. Upon
adsorption, the relaxivity is 24.0 s-1 mM-1 (20 MHz, 25 °C).
These in vitro results demonstrate that BPAMD is a promising
ligand for bone targeting. By virtue of favorable properties, its
lanthanide(III) complexes have great potential for applications
in MRI, radiodiagnosis, and radiotherapy.
Conclusion
We prepared the new DOTA monoamide analogue with a
distant bis(phosphonic acid) moiety and a series of its Ln(III)
complexes. The ligation process proceeds stepwise. Initially,
the Ln(III) ion is coordinated to the ligand through the
phosphonic acid groups. In a second step, insertion of the metal
ion into the macrocyclic cavity takes place. This step was found
to be strongly dependent on the pH and is reasonably fast only
in alkaline solutions. The solution structure of the Ln(III)
complexes resembles that of complexes of DOTA monoam-
ides: an octadentate binding fashion of the organic ligand, one
water molecule in the first coordination sphere, and similar m/M
ratios.
The relaxometry study of the Gd(III) complex revealed a
relatively slow water exchange (298τM ) 1.2 µs) typical for
DOTA monoamides, but nevertheless a relaxivity (r1 ) 5.3 s-1
mM-1, 20 MHz, 25 °C) that is slightly higher than that for
DOTA.
Acknowledgment. Thanks are due to the EU for financial
support via a Marie Curie training site host fellowship (QLK5-
CT-2000-60062). Supports from the Grant Agency of the Czech
Republic (No. 203/03/0168) and from FNRS of the French
Community of Belgium are acknowledged. This work was done
in the frame of COST D18 Action and the EU Network of
Excellence (NoE) “European Molecular Imaging Laboratory”
(EMIL).
Supporting Information Available: 31P NMR spectra of Eu-
BPAMD measured shortly after mixing of the components, 1H
EXSY spectrum of the Eu-BPAMD complex, 31P spectra of
the Dy and Yb complexes, variable-temperature 31P NMR
spectra of the Eu complex, equations used for the fitting of the
NMRD and 17O NMR data, and full list of best-fit parameters
for Gd-BPAMD and a discussion on the pH dependence of
the relaxivity of Gd-BPAMD. This material is available free
of charge via the Internet at http://pubs.acs.org.
Sorption experiments with hydroxyapatite demonstrated that
both the free BPAMD and its Ln(III) complexes are adsorbed
(44) The high relaxivity of HA at low field can be explained by the long
correlation time modulation of dipole-dipole relaxation of protons. This
very long correlation time is due to the adsorption equilibrium of the water
molecules on the surface of HA and to their subsequent slow motion. The
same trend is observed with colloidal suspensions of silica beads. Roose,
P.; van Craen, J.; Finsy, R.; Eisendraht, H. J. Magn. Reson. A 1995, 115,
20.
JA054905U
9
J. AM. CHEM. SOC. VOL. 127, NO. 47, 2005 16485