Tuning glutamine binding modes in Gd-DOTA-based probes for an improved MRI visualization of tumor cells

Article abstract of DOI:10.1002/chem.200801567

Three new magnetic resonance imaging probes that target glutamine transporters have been synthesized. They consist of a Gd-DOTA-monoamide moiety (DOTA = 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) linked through a six carbon atom chain to a

Full text of DOI:10.1002/chem.200801567

DOI: 10.1002/chem.200801567
Tuning Glutamine Binding Modes in Gd-DOTA-Based Probes for an
Improved MRI Visualization of Tumor Cells
Rachele Stefania,[a] Lorenzo Tei,[b] Alessandro Barge,[c] Simonetta Geninatti Crich,[a]
Ibolya Szabo,[a] Claudia Cabella,[d] Giancarlo Cravotto,[c] and Silvio Aime*[a]
76
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 76 – 85

FULL PAPER
Abstract: Three new magnetic reso-
nance imaging probes that target gluta-
mine transporters have been synthe-
sized. They consist of a Gd-DOTA-
monoamide moiety (DOTA=1,4,7,10-
tetraazacyclododecane-1,4,7,10-tetra-
acetic acid) linked through a six carbon
atom chain to a vector represented by
a glutamine residue bound through a-
carboxylic, g-carboxamidic, or a-amino
functionalities. Their uptake by HTC
(rat hepatocarcinoma) and healthy rat
hepatocytes has shown that the system
containing the glutamine vector bound
through the a-carboxylic group dis-
plays a markedly higher affinity for
tumor cells. The observed behavior is
rationalized in terms of the exploita-
tion of an additional glutamine trans-
porter active in hepatic tumor cells.
Keywords: gadolinium · glutamine ·
hepatoma cells · imaging agents ·
magnetic resonance imaging
Introduction
Recently, we reported that it is possible to accumulate a
sufficient number of small gadolinium-based imaging probes
in tumor cells by exploiting their amino acid transporting
systems, in particular the glutamine (Gln) transporters.[13,14]
The increased expression or up-regulation of the amino acid
transporters, particularly Gln transporters, is strictly related
to cell growth and it can be considered as a good marker for
tumor cells.[15,16] In fact, rapidly growing cells use Gln as a
nitrogen, carbon, and energy source. In a previous study, we
synthesized a Gd-DOTA-monoamide derivative bearing a
Gln residue separated from the chelating moiety by a C6 ali-
phatic chain. In vitro and in vivo experiments demonstrated
that this probe is taken up by tumor cells. In this imaging
probe the Gln vector is linked through an amidic bond be-
tween its a-amino group and the carboxylic residue present
in the aliphatic chain spacer (Gd-DOTAMAC6Gln in
Scheme 1; DOTAMA=DOTA-monoamide). This binding
scheme alters the Gln charge distribution in respect of the
free amino acid as it transforms the primary a-amino group
into a secondary amide. Moreover, the latter group displays
enhanced rigidity and hindrance, which decreases its overall
recognition capabilities towards the Gln transporter. A de-
finitive optimization of the design of the Gln-bearing probe
may come from the knowledge of the three-dimensional
transporter structure, which is, to date, unknown. To over-
come this lack of information we decided to synthesize new
probes in which Gln is bonded to the contrast enhancing
unit through other functional groups (carboxylic, amino, and
carboxamide groups) and to test their targeting capabilities
towards tumor cells to select the most effective system.
Bonding through the a-carboxylic, g-carboxamidic, or a-
amino functionalities led to three new systems, namely, Gd-
DOTAMAC6Glna (Gd-L1), Gd-DOTAMAC6Glng (Gd-
L2), and Gd-DOTAMAC6GlnA (Gd-L3), in addition to the
previously reported Gd-DOTAMAC6Gln (Scheme 1).
Molecular imaging is a rapidly growing field that provides
noninvasive visual representation, characterization, and
quantification of fundamental biological processes in intact
living organisms. Its aims are to widen the detection horizon
of early medical diagnoses and to monitor the effects of
therapeutic treatments by reporting on the earliest physio-
logical and biochemical events that characterize a given dis-
ease.[1–3]
Magnetic resonance (MR) offers the unique opportunity
to obtain images characterized by a superb anatomical reso-
lution and to map simultaneously structure and function in
soft tissues in vivo. MR is therefore considered a particular-
ly important and advantageous modality for molecular imag-
ing applications. To enhance the sensitivity of MR molecular
imaging protocols proper amplification procedures have to
be designed in order to deliver a high number of imaging re-
porting units at the target site.[4] This task of accumulating a
large number of contrast agent (CA) molecules at the target
site has been pursued, for instance, by means of gadolinium-
loaded targeted liposomes,[5–8] targeted iron oxide nanoparti-
cles,[9,10] and perfluorocarbon nanoparticles.[11,12] A potential
problem of these large constructs, however, is that their size
may prevent them from escaping the blood circuit to reach
the sites of interest.
[a] Dr. R. Stefania, Dr. S. Geninatti Crich, Dr. I. Szabo, Prof. S. Aime
Dipartmento di Chimica IFM and Molecular Imaging Center
Universitꢁ di Torino, Via Nizza 52, 10125 Torino (Italy)
Fax : (+39)011-670-6487
E-mail: silvio.aime@unito.it
[b] Dr. L. Tei
Dipartimento di Scienze dellꢂAmbiente e della Vita (DiSAV)
Universitꢁ del Piemonte Orientale “A. Avogadro”
Via Bellini 25/G, 15100 Alessandria (Italy)
[c] Dr. A. Barge, Prof. G. Cravotto
Dipartimento di Scienza e Tecnologia del Farmaco
Universitꢁ di Torino, Via P. Giuria 9, 10125 Torino (Italy)
Results and Discussion
[d] Dr. C. Cabella
Bracco Imaging SpA
via Ribes 5, 10010 Colleretto Giacosa (TO) (Italy)
Probe design: In the previously reported Gd-DOTA-
MAC6Gln probe,[13] the bonding to the imaging reporter im-
plied the transformation of an amino into an amide func-
tionality, which would compromise the contribution of one
of the possible Gln recognition sites on its transporting pro-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801567. It contains a descrip-
tion of HPLC Methods 1–7 used for the purification and characteriza-
tion of final products.
Chem. Eur. J. 2009, 15, 76 – 85
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
77

S. Aime et al.
Scheme 1. Structures of the glutamine-containing probes.
teins. In fact, conversely to what occurred for the neutral
zwitterionic amino acid Gln, the presence of a free carboxyl-
ate group results in an overall negative charge at physiologi-
cal pH. Because the structural details are not yet known, we
have designed three imaging probes in which the Gln
moiety is attached to the DOTAMA unit by one of the
three other functional groups to gain a better understanding
of the recognition target determinants of the membrane
transporter. The new probes, Gd-L1, Gd-L2, and Gd-L3,
maintain the same DOTA-monoamide chelating unit and
the C6 aliphatic spacer of Gd-DOTAMAC6Gln, but the use
of different binding sites endows the Gln residue with quite
different structural properties and overall charge at physio-
logical pH.
In Gd-L1 the Gln is attached to the imaging reporter
through its carboxylic group, thereby forming an amidic
bond, as in Gd-DOTAMAC6Gln. At physiological pH, the
targeting moiety of the probe differs from the parent Gln
amino acid because it displays an overall positive charge,
the a-amino group being protonated and the carboxylic
group being replaced by a secondary amide.
Instead, in Gd-L2 the anchor point is the primary g-amide
of Gln, which yields a zwitterionic neutral form at physio-
logical pH. However, although the charge distribution of the
Gln vector is unaltered, in this probe the amino acid loses
its characteristic terminal amide functional group, which in
principle could be important in the recognition by the Gln
transporter.
Finally, Gd-L3 was designed in such a way that the Gln
moiety would remain unaltered as much as possible. For this
reason, the a-amino group of a protected Gln derivative
was subjected to reductive amination with the bifunctional
chelating agent DOTAMAC5CHO to obtain a secondary
amine in place of the primary a-amino group of the parent
amino acid. At physiological pH, the Gd-L3 probe is present
as a zwitterion and has the same charge distribution as in
Gln, although the isoelectric point is slightly higher as a con-
sequence of the occurrence of the secondary, more basic,
amine. Furthermore, the secondary amine is more hindered
than in Gln and this hindrance could interfere in the binding
to the transporter.
Synthesis: Two bifunctional chelating agents containing a
free amino or an aldehydic group separated from the coordi-
nating cage by a hexamethylene aliphatic chain were used
for the synthesis of ligands L1, L2, and L3. Namely,
DOTAMA
ACHTUNGTRENNUNG
and L2 and DOTAMAAHCTUNGTRENNUNG
L3 (Schemes 2 and 3). The synthesis of these DOTAMA-
functionalized building blocks has been recently reported.[17]
L1 and L2 were synthesized by the reaction of
DOTAMAACHTNUGRTNEUNG(OtBu)3C6NH2 with the pentafluorophenyl-acti-
vated esters of N-Boc-l-glutamine and N-Boc-l-glutamic a-
tert-butyl ester, respectively, followed by acid hydrolysis of
the protecting groups (Scheme 2).
L3 was synthesized by reductive amination between
DOTAMAACHTNUTGRNEUNG(tBu)3C5CHO and a Gln derivative protected as
the tert-butyl ester on the carboxylic acid and as the N-trityl
secondary amide (H2N-l-Gln(Tr)-OtBu) (Scheme 3). This
protected Gln was synthesized in three steps starting from
the benzyloxycarbonyl (Cbz)-protected Gln.
Following a literature procedure,[18] the terminal amide
was protected as the trityl amide by reaction with triphenyl-
methanol, acetic anhydride, and sulfuric acid. Then the car-
boxylic acid was protected as the tert-butyl ester by acid-cat-
alyzed transesterification with tert-butyl acetate and the
final H2N-l-Gln(Tr)-OtBu was obtained by hydrogenolysis
of the Cbz group under pressure (50 bar, 508C). The reduc-
tive
amination
between
H2N-l-Gln(Tr)-OtBu
and
DOTAMAAHCUTGNTR(EUNNNG OtBu)3C5CHO was carried out under ultrasound
(US) irradiation for 30 min (20 W, 20 kHz) in MeOH at con-
trolled pH (pH 5); sonication was continued for another
30 min after the addition of triacetoxyborohydride. The re-
action conditions had to be carefully optimized to avoid the
hydrolysis of the amide bond between the spacer and the
DOTA unit, which would result in a series of byproducts
and a reduction in the yield. The yields were invariantly
very low when this reaction was carried out under conven-
78
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 76 – 85

Improved MRI Visualization of Tumor Cells
FULL PAPER
Scheme 2. Synthetic pathways to GdDOTAMAC6NH2, Gd-DOTAMAC6Glna (Gd-L1), and Gd-DOTAMAC6Glng (Gd-L2). i) TFA, CH2Cl2; ii) GdCl3,
H2O, pH 7; iii) Boc-l-Gln-OH, DCC/Pfp, CH2Cl2 to give 1; iv) TFA, CH2Cl2; v) Boc-l-GluACHTNUGTRNEN(UG OtBu)-OH, DCC/Pfp, CH2Cl2 to give 2; vi) TFA, CH2Cl2;
vii) GdCl3, H2O, pH 7.
Scheme 3. Synthetic pathway to Gd-DOTAMAC6GlnA (Gd-L3). i) NaBHACHTUNGTRNEGN(U OAc)3, AcOH, ultrasound (20 W), CH3OH to give 3; ii) TFA, CH2Cl2;
iii) GdCl3, H2O, pH 7.
tional conditions. However, sonication allows a dramatic im-
provement of the yields, as observed in other reported re-
ductive amination procedures.[19,20]
with Gln under hydrolytic conditions, especially in the pres-
ence of bivalent cations and Brønsted or Lewis acids and
bases.[21] In fact, glutamic and pyroglutamic acids (5-oxopyr-
rolidine-2-carboxylate) have been reported to be glutamine
metabolites, which shows that the intramolecular reaction
can also occur in vivo.[22,23] Moreover, Gln as the N-terminal
amino acid in many proteins, hormones, and neurotransmit-
ter peptides has been found to cyclize intramolecularly
either spontaneously or under the action of enzymes such as
glutaminyl cyclase.[24,25] With the intramolecular cyclization
process being such an easy process, the tert-butyl ester hy-
drolysis reaction was monitored by ESI-MS every 30 min in
order to stop it before too much of the cyclic byproduct was
Whereas in the case of L2, cleavage of the Boc and tert-
butyl protecting groups with trifluoroacetic acid gave the
product in good yields, in the cases of L1 and L3, TFA cata-
lyzed an intramolecular transamination reaction between
the a-amino group and the amide of the Gln with the for-
mation of a 5-oxopyrrolidine ring (L4 and L5) and ammonia
(Scheme 4). This clearly causes a dramatic change in the
structural characteristics of the vector moiety.
This cyclization is particularly favored by the formation of
a five-membered pyrrolidinic ring and is known to occur
Chem. Eur. J. 2009, 15, 76 – 85
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
79

S. Aime et al.
L1, Gd-L2, and Gd-L3 were 5.6, 5.6, and 5.5 mmꢀ1 sꢀ1, re-
spectively. These values are in agreement with those previ-
ously reported for a number of analogous gadolinium com-
plexes with variously substituted DOTA-monoamide li-
gands.[27] These values are slightly higher than that reported
for the parent Gd-DOTA complex (r1p =4.7 mmꢀ1 sꢀ1) as a
consequence of the increased molecular weight.[28]
Cell uptake experiments: The affinity of the different Gln-
containing gadolinium complexes for the rat hepatoma cell
line (HTC) was determined and compared with that ob-
served for the previously reported Gd-DOTAMAC6Gln
complex.[13] The procedure used in this study consisted of in-
cubating around 1.5 million cells for 6 h at 378C in a mini-
mum medium (Earlꢂs balanced salt solution: EBSS) contain-
ing the gadolinium complexes at a concentration of 1 mm.
After incubation the cells were washed three times with ice-
cold phosphate-buffered saline (PBS) and detached from
the culture flasks mechanically with a cell scraper. Figure 1
Scheme 4. Glutamine cyclization reaction on probe surfaces.
formed. L1 and L4 were recovered after purification by
semipreparative HPLC–MS in about 40 and 25% yields, re-
spectively. A similar result was found for L3 and L5. Be-
cause L1 and L3 are prone to cyclization at both acidic and
basic pHs, we had to find the optimal conditions for HPLC
separation. A reversed-phase column and a gradient with
7 mm ammonium acetate buffer at pH 4 and acetonitrile as
eluents provided the best conditions to minimize the un-
wanted transformation in the separation medium. The pH
had to be kept at slightly acidic values because separation of
the linear and cyclic products was not possible at neutral
pH. However, immediately after separation, the fractions
containing pure L1 or L3 had to be neutralized with a solu-
tion of NH4OH to avoid further cyclization. After complex-
ation with GdCl3, a second HPLC purification procedure
(water/acetonitrile gradient elution) was carried out to
remove salts from the products. Taking into account the ten-
dency of L1 and L3 to cyclize in the presence of metal ions,
it was found useful to skip the purification step for the li-
gands and to just purify the final complexes. These purifica-
tions were carried out by semipreparative HPLC–MS with a
water/acetonitrile gradient elution. Yields of about 40%
were obtained for the linear complexes and of about 25%
for the cyclic ones.
Figure 1. Internalization of gadolinium complexes functionalized with
Gln by HTC cells. For the uptake experiment, about 1.5 million cells
were incubated at 378C for 6 h in EBSS with Gd-L1, Gd-L2, Gd-L3, and
Gd-DOTAMaC6Gln at concentrations of 1 mm.
shows that the compound bearing a free amino group (Gd-
L1) displays an improved uptake by tumor cells, of about
200%, compared with that previously reported for Gd-DO-
TAMAC6Gln. In contrast, the bonding of Gln through the
primary g-amide group in Gd-L2 led to a dramatic decrease
in affinity for tumor cells. Also the uptake of Gd-L3, al-
though present as a zwitterion and with the same functional
groups and charge distribution as Gln, was significantly
lower.
T1-weighted spin echo MR images of cells (Figure 2)
showed that only after the incubation of Gd-L1 and Gd-
DOTAMAC6Gln were the amounts of internalized gadolini-
um sufficient to generate hyperintensity in the correspond-
ing MR images.
Synthesis and relaxometric characterization of gadolinium
complexes: The GdIII complexes were prepared by mixing
stoichiometric amounts of the DOTAMA-containing probes
and GdCl3 solution. In the case of Gd-L2, unchelated Gd3+
ions were eliminated by precipitation of the hydroxide at
basic pH by adding aliquots of a concentrated NaOH solu-
tion. The amount of free GdIII was determined by UV spec-
trophotometry.[26] Finally, excess gadolinium was eliminated
by adding a stoichiometric amount of the ligand. In the
cases of Gd-L1 and Gd-L3, to avoid the cyclization of Gln
catalyzed by basic pH and/or trivalent cations, unchelated
Gd3+ was eliminated by the addition of 0.5 equiv of iminodi-
acetic acid followed by LC–MS purification using H2O and
CH3CN as eluents (see the Supporting Information). The re-
laxivities (the proton relaxation enhancement of water pro-
tons in the presence of the paramagnetic complex at a con-
centration of 1 mm) measured at 20 MHz and 298 K for Gd-
To assess whether the high affinity of Gd-L1 for tumor
cells relies on the presence of a positive charge, the nonspe-
cific precursor Gd-DOTAMAC6NH2 was incubated with
HTC cells under the same conditions (6 h, 378C, EBSS).
The number of moles of gadolinium per mg of protein taken
80
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 76 – 85

Improved MRI Visualization of Tumor Cells
FULL PAPER
with that observed with hepatoma cells. To this end, rat hep-
atocytes and HTC cells were incubated for 6 h in the pres-
ence of 1 mm Gd-L1. Figure 3A shows that HTC cells la-
Figure 2. T1-weighted spin-echo MR image (measured at 7 T) of an agar
phantom containing HTC cells incubated for 6 h at 378C with 1) Gd-L1,
2) Gd-L3, 3) Gd-DOTAMA-C6-Gln, and 4) Gd-L2.
up was 50 times lower (about 7ꢃ10ꢀ10) than that measured
after incubation with Gd-L1, which indicates that the pres-
ence of Gln on the complex surface is essential for its specif-
ic interaction to tumor cells. Further evidence that Gln was
the vehicle for the internalization through the amino acid
transporting system was obtained by incubating HTC cells
for 3 h in the presence of 0.25 mm Gd-L1 with or without an
excess of Gln. In the presence of 10 mm Gln the number of
moles of gadolinium taken up per mg of protein was 75%
lower than that measured in its absence, which demonstrates
the specific internalization pathway.
The unexpected behavior of the two probes that have the
same charge distribution and functional groups in Gln (Gd-
L2 and Gd-L3) could be ascribed to the significant modifica-
tion introduced into the vector structure by the binding to
the MRI reporter. We suggest that such a modification of
the Gln structure compromises its interaction with appropri-
ate transporters. However, the presence of a positive or neg-
ative localized charge on the other two probes (Gd-L1 and
Gd-DOTAMAC6Gln) may reduce this unfavorable effect,
probably because of the formation of electrostatic interac-
tions between the vector and transporter.
Hepatoma cells transport Gln at a rate 10–20 times faster
than normal hepatocytes[29] as a consequence of the expres-
sion of an additional transport system that does not occur in
normal hepatocytes. In fact, in human hepatoma cell lines,
Gln is transported through the cytoplasmatic membrane by
both the system commonly expressed on hepatocytes
(system N) and by a different system that has only subse-
quently been elucidated.[30] This system differs from ASC
(the transport system that mediates Gln transport in many
tumor cell lines) due to its interaction with basic amino
acids. This observation may account for the increased inter-
nalization of the basic Gd-L1 complex compared with other
probes investigated in this work. In fact, the presence of the
free amino group appears to facilitate the interaction with
the negatively charged domains of the protein transporter
on the cell membrane.
Figure 3. a) T1-weighted spin-echo MR image (measured at 7 T) of an
agar phantom containing A) unlabeled hepatocytes, B) hepatocytes la-
beled with Gd-L1, C) unlabelled HTC, and D) HTC labeled with Gd-L1.
b) Comparison of the internalization of Gd-L1 and Gd-DOTAC6Gln into
rat hepatocytes and hepatoma (HTC). Cells were incubated for 6 h in the
presence of 1 mm gadolinium complex solutions in EBSS.
beled with Gd-L1 display a hyperintense signal in the T1-
weighted spin-echo image compared with the unlabeled
cells. In contrast, the signal intensity of the healthy rat hepa-
tocytes is only slightly greater than the signal intensity of
the unlabeled cells. In fact, in this case, the amount of gado-
linium internalized (1 nmol of gadolinium per mg of pro-
tein) is about 35 times lower than that found for hepatoma
cells (Figure 3B). As a consequence of the increased uptake
of Gd-L1 by tumor cells, there is around a 350% enhance-
ment in the intensity of the hepatocytes of the rat hepatoma
when compared with the pellet of cells derived from healthy
rat liver.
Comparison of Gd-L1 with the previously reported nega-
tively charged Gd-DOTAMAC6Gln showed a significant in-
crease in the tumor/healthy cell ratio of internalized gadoli-
nium. In fact, the amount of Gd-L1 taken up by hepatoma
cells is about 35 times higher than the amount of Gd-L1
taken up by healthy hepatocytes, whereas the amount of
Gd-DOTAMAC6Gln taken up was about 10 times lower.
The uptake of the more efficient Gd-L1 was determined
with the primary culture of rat hepatocytes and compared
Chem. Eur. J. 2009, 15, 76 – 85
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
81

S. Aime et al.
This is a direct consequence of the contrasting behavior ob-
served with the positively charged amino acid of transport
systems expressed on tumor and normal cells (system N), re-
spectively.
the target cells due to the exploitation of high capacity
transporters. The new insights gained in this work add fur-
ther support to the overall armory of tools available for MR
visualization of tumor cells.[31–33]
Afterwards, the internalization by hepatoma cells of Gd-
L1 was compared with that observed with its cyclic deriva-
tive (Scheme 4) as a function of complex concentration in
the incubation media. Figure 4 shows that intramolecular
Gln cyclization reduces significantly the amount of gadolini-
um internalized in the range of concentrations investigated
as a consequence of the change in the vector structure.
Experimental Section
All chemicals were purchased from Sigma-Aldrich Co. and were used
without purification unless otherwise stated. NMR spectra were recorded
on JEOL Eclipse Plus 400 and Bruker Avance 600 spectrometers (oper-
ating at 9.4 and 14 T, respectively). The sonochemical apparatus used in
this work was developed by the authors in collaboration with Danaca-
merini s.a.s. (Torino).[19,34] ESI mass spectra were recorded on a Waters
Micromass ZQ spectrometer. Analytical and preparative HPLC–MS
were carried out on
equipped with Waters 2996 diode array and Micromass ZQ (ESCI ioniza-
tion mode) detectors. DOTAMA(OtBu)3C6NH2 and DOTAMA-
(OtBu)3C5CHO were prepared following a reported procedure.[17] The
a Waters FractionLynx autopurification system
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
longitudinal water proton relaxation rate was measured on a Stelar Spin-
master spectrometer (Stelar, Mede, Italy) operating at 20 MHz using the
standard inversion–recovery technique (16 experiments, 2 scans). A typi-
cal 908 pulse width was 3.5 ms and the reproducibility of the T1 data was
ꢁ0.5%.
Tri-tert-butyl ester of (S)-1-[2,11-dioxo-12-(tert-butoxycarbonylamino)-
14-carbamoyl-3,10-diazatetradecanyl]-1,4,7,10-tetraazacyclododecane-
4,7,10-triacetic acid (1): A solution of dicyclohexylcarbodiimide (DCC)
(1.3 g, 6.56 mmol) in CH2Cl2 (20 mL) was added to a stirred solution of
N-Boc-l-Gln-OH (1.47 g, 5.96 mmol) in CH2Cl2 (80 mL) that had been
cooled to 08C. Pentafluorophenol (Pfp) (1.2 g, 6.56 mmol) was added and
the mixture was stirred at 08C for 3 h and then for another 16 h at room
temperature. The solvent was partly removed and precipitated dicyclo-
hexylurea (DCU) was filtered off. The filtrate was added dropwise to a
&
Figure 4. Internalization by HTC cells of Gd-L1 ( ) and its cyclic deriva-
tive Gd-L4 (&). Cells were incubated for 6 h at 378C with increasing con-
centrations (0.07–1 mm) of the two gadolinium complexes.
solution of DOTAMAACHTUNGTRNENUG(OtBu)3C6NH2 (4 g, 5.96 mmol) in CH2Cl2
(50 mL). The mixture was allowed to react overnight at room tempera-
ture, the solvent was removed under reduced pressure, and the residue
purified by column chromatography (silica gel, elution gradient: CH2Cl2/
MeOH, 95:5!9:1!8:2!7:3; TLC: CH2Cl2/MeOH, 9:1 (v/v), Rf =0.22)
to yield a slightly yellow oil (2.4 g, 44.8%). 1H NMR (400 MHz, CDCl3):
d=8.63 (br, CONH, 1H), 7.93 (br, CONH, 1H), 7.56, (br, CONH, 1H),
4.29 (m, CHa, 1H), 3.50–3.20, 3.10–2.07 (br, CH2 macrocycle, NCH2CO,
This observation is particularly important in the develop-
ment of systems that exploit Gln as a carrier for tumor cells
because undetected cyclization can reduce markedly the
probe uptake in target cells. For this reason it is useful to
obtain an estimate of the ratio of the cyclic and linear form
of Gd-L1 by an appropriate analytical HPLC method (see
Method 7 in the Supplementary Information). Furthermore,
the asymptotic shape of the curve observed for Gd-L1 indi-
cates the occurrence of an active internalization mechanism.
CONHCH2, CH2b, CH2g, 36H), 1.53–1.28 ppm (br, C
CH2, 40H); 13C NMR (400 MHz, CDCl3): d=176.1, 172.4, 172.2, 172.1,
171.4, 155.8 (CONH, COO, 6C), 81.8, 79.1 (C(CH3)3, 4C), 56.1, 55.7,
ACHTUGNTRENN(UNG CH3)3, aliphatic
AHCTUNGTRENNUNG
55.6, 54.5, 53.5 (NCH2CO, NCH2 macrocycle, 12C), 50.0 (CHa,1C), 38.8,
38.7 (CONHCH2, 2C), 32.4, 29.6, 29.0, 28.9 (aliphatic CH2, CH2b, CHg,
4C), 28.4, 27.9 (CACTHUNTRGNE(UNG CH3)3, 12C), 26.0, 25.9 ppm (aliphatic CH2, 2C); MS
(ESI+): m/z: calcd for C44H83N8O11 [M+H]+: 899.62; found: 899.91.
Conclusion
(S)-1-(2,11-Dioxo-12-amino-14-carbamoyl-3,10-diazatetradecanyl)-
1,4,7,10-tetraazacyclododecane-4,7,10-triacetic acid (L1): Compound
1
(1.00 g, 1.11 mmol) was dissolved in CH2Cl2 (10 mL), trifluoroacetic acid
(35 mL) and triisopropylsilane (0.6 mL) were added, and the mixture was
stirred at room temperature for 2 h. The reaction was monitored by ESI-
MS every 30 min to avoid any delay that would favor cyclization. The
mixture was then evaporated in vacuo and the product precipitated with
excess diethyl ether. The product was isolated by centrifugation, washed
thoroughly with diethyl ether, and dried in vacuo. It was then dissolved
in H2O (5 mL) and the pH neutralized by addition of dilute NH4OH at
08C. The crude product was purified by preparative HPLC–MS by using
a Waters Atlantis RPdC18 19/100 column by Method 1 using 7 mm
CH3COONH4 (pH 4) (A) and CH3CN (B) as eluents (see the Supporting
Information). A second fast chromatographic separation was needed to
eliminate salts (Method 2, see the Supporting Information). The pure
Modulation of the binding scheme through which a Gln resi-
due is bound to the surface of a MR imaging reporter has
allowed a system that is avidly taken up by hepatoma tumor
cells to be identified. The differences observed in the inter-
nalization of this probe into HTC and healthy hepatocytes
illustrates the important role of a transporter devoted to the
uptake of basic amino acids that is peculiar to hepatic
tumors. The amount of gadolinium-based probe Gd-L1 that
is taken up by HTC cells is twice as large as that previously
reported for Gd-DOTAC6Gln and markedly higher than
that for the other Gln-based probes developed in this work.
The results obtained represent a step forward in the devel-
opment of MR molecular imaging applications based on the
accumulation of a high number of gadolinium chelates at
product was obtained as
a
white powder (270 mg, 39%). 1H NMR
(600 MHz, D2O): d=3.92 (m, CHa,1H), 3.76–3.70 (br, CH2 macrocycle,
4H), 3.49, 3.32 (br, CH2 macrocycle, 12H), 3.18–2.97 (br, NCH2CO,
CONHCH2, CH2NH, 12H), 2.36 (m, CH2b, 2H), 2.08 (m, CH2g, 2H),
82
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 76 – 85

Improved MRI Visualization of Tumor Cells
FULL PAPER
1.46, 1.26 ppm (m, aliphatic CH2, 8H); 13C NMR (600 MHz, D2O): d=
176.4, 175.6, 170.3, 169.2, 167.8 (CONH and COO, 5C), 55.6, 54.6, 54.3,
51.6 (NCH2CO, CHa, 5C), 50.6, 49.9, 47.2 (NCH2 macrocycle, 8C), 38.4,
38.1 (CONHCH2, 2C), 29.1, 27.1, 25.5, 24.7, 22.8, 22.4 ppm (aliphatic
CH2, CH2b, CHg, 6C); MS (ESI+): m/z: calcd for C27H51N8O9 [M+H]+
631.38; found: 631.62.
hexane (200 mL). The precipitated product was separated by centrifuga-
tion to obtain
CDCl3): 7.26 (m, 20H), 5.09 (s, 2H), 4.16 (m, 1H), 2.51 (m, 1H), 2.38
(m, 1H), 2.02 ppm (m, 2H); MS (ESI+): m/z: calcd for C32H31N2O5
[M+H]+: 523.22; found: 523.31.
a
white powder (0.94 g, 50%). 1H NMR (400 MHz,
N-Cbz-l-Gln(Tr)-OtBu: N-Cbz-l-Gln(Tr)-OH (5 g, 9.55 mmol) was sus-
pended in tBuOAc (64 mL, 0.478 mol) in a round-bottomed flask. HClO4
(0.164 mL, 1.91 mol) was slowly added dropwise into the reaction mix-
ture, which was stirred for 20 h at RT. A light-yellow suspension resulted
and the reaction was monitored by TLC (silica, CH2Cl2/MeOH, 9:1,
(S)-1-[6-(5-Oxopyrrolidin-2-ylcarbonylamino)hexylcarbamoylmethyl]-
1,4,7,10-tetraazacyclododecane-4,7,10-triacetic acid (L4)—cyclic product:
The cyclized product (153 mg, 25% yield) was recovered by HPLC sepa-
ration. 1H NMR (600 MHz, D2O): d=4.23 (m, CHa, 1H), 3.81–3.68 (br,
CH2 macrocycle, 4H), 3.43, 3.38 (br, CH2 macrocycle, 12H), 3.18–3.01
(br, NCH2CO, CONHCH2, CH2NH, 12H), 2.97 (m, CH2b, 2H), 2.37 (m,
CH2g, 2H), 1.46, 1.26 ppm (m, aliphatic CH2, 8H); MS (ESI+): m/z:
calcd for C27H48N7O9 [M+H]+: 614.35; found: 614.58.
R
f,product =0.95). Aqueous Na2CO3 (10%, 40 mL) was slowly added to the
reaction mixture. The solution was transferred to a separating funnel and
EtOAc (100 mL) was added. The water phase was separated, its pH was
adjusted to 10 by adding aqueous NaOH (1m, ca. 22 mL), and then it
was returned to the organic phase in the separating funnel and thorough-
ly shaken. The organic phase was washed with H2O (3ꢃ100 mL), dried
with Na2SO4, and the solvent was removed under reduced pressure. The
product was purified by column chromatography (stationary phase:
silica; mobile phase: gradient starting from CH2Cl2 (100 mL), then
CH2Cl2/MeOH 95:5). Yield: 74%. 1H NMR(400 MHz, CDCl3): d=7.26
(m, 20H), 5.09 (s, 2H), 4.22 (m, 1H), 2.41 (m, 1H), 2.33 (m, 1H), 2.16
(m, 1H), 1.86 (m, 1H), 1.42 ppm (s, 9H); MS (ESI+): m/z: calcd for
C36H39N2O5 [M+H]+: 579.28; found: 579.3.
Tetra-tert-butyl ester of (S)-1-[2,11-dioxo-14-(tert-butoxycarbonylamino)-
14-carboxy-3,10-diazatetradecanyl]-1,4,7,10-tetraazacyclododecane-4,7,10-
triacetic acid (2): A solution of DCC (1.4 g, 7.26 mmol) in CH2Cl2
(80 mL) was added to a stirred solution of Boc-l-GluACHTUNTRGNEU(GN OtBu)-OH (2.0 g,
6.6 mmol) in CH2Cl2 (80 mL) under cooling at 08C. Then pentafluoro-
phenol (1.3 g, 7.26 mmol) was added and the mixture stirred at 08C for
3 h and for another 16 h at room temperature. After filtration of the pre-
cipitated DCU, the solvent was partly removed and the solution was fil-
tered again to remove some more DCU. The filtrate was added dropwise
to a solution of DOTAMA
(OtBu)3C6NH2 (4.4 g, 6.6 mmol) in CH2Cl2
H2N-l-Gln(Tr)-OtBu: N-Cbz-l-Gln(Tr)-OtBu (2 g, 3.47 mmol) was
placed in a PARR reactor, dissolved in dioxane (60 mL), and water-mois-
tened 10% Pd/C (0.3 g) was added. The mixture was stirred under hydro-
gen pressure (50 bar) at 508C for about 24 h. The reaction was checked
by TLC (silica, CH2Cl2/MeOH, 9:1, Rf,product =0.32, Rf,reactant =0.95). The
resulting solution was first filtered, then run through a short column of
silica gel. This was washed with methanol to recover all the product re-
tained on the silica gel. The solvent was removed under reduced pres-
sure. Yield: 84%. 1H NMR (400 MHz, CDCl3): d=7.26 (m, 15H), 3.30
(m, 1H), 2.41 (m, 2H), 2.08 (m, 1H), 1.75 (m, 1H), 1.41 ppm (s, 9H);
MS (ESI+): m/z: calcd for C28H33N2O3 [M+H]+: 445.25; found: 445.32.
(50 mL). After the mixture had been allowed to react overnight at room
temperature, the solvent was removed under reduced pressure and the
residue was purified by column chromatography (silica gel, elution gradi-
ent: CH2Cl2/MeOH, 95:5!9:1!8:2!7:3; TLC: CH2Cl2/MeOH, 9:1
(v/v), Rf =0.20) to yield a slightly yellow oil (3.2 g, 53.9%). 1H NMR
(400 MHz, CDCl3): d=8.73 (br, CONH, 1H), 8.35 (br, CONH, 1H),
7.82, (br, CONH, 1H), 4.05 (m, CHa,1H), 3.45–3.10, 2.95–1.80 (br, CH2
macrocycle, NCH2CO, CONHCH2, CH2b, CH2g, aliphatic CH2, 36H),
1.51–1.28 ppm (br, C
CDCl3): d=174.0, 173.9, 172.6, 172.5, 172.1, 172.0, 156.9 (CONH, COO,
7C), 81.5, 82.4, 80.7 (C(CH3)3, 5C), 58.2, 58.1, 56.1, 56.2, (NCH2CO,
ACHTUNGTRENNUNG
(CH3)3, aliphatic CH2, 49H); 13C NMR (400 MHz,
AHCTUNGTRENNUNG
Tri-tert-butyl ester of (S)-1-[2-oxo-11-(tert-butoxycarbonyl)-13-(N-trityl-
carbamoyl)-3,10-diazaterdecanyl]-1,4,7,10-tetraazacyclododecane-4,7,10-
CHa, 5C), 55.5, 55.0, 54.0, 53.5 (NCH2 macrocycle, 8C), 40.4, 40.3
(CONHCH2, 2C), 34.2, 30.6, 30.4, (aliphatic CH2, CH2b, CH2g, 4C), 30.1,
triacetic acid (3):
A solution of DOTAMAACTHUNTGNRUEGN(OtBu)3C5CHO (1.8 g,
28.9, 28.8 (CACHTUNGTRENNUNG(CH3)3, 15C), 28.0, 27.9 ppm (aliphatic CH2, 2C); MS
2.73 mmol) in CH3OH (10 mL) was adjusted to pH 5 with glacial acetic,
then NH2-l-Gln(Tr)-OtBu (1.2 g, 2.73 mmol) dissolved in CH3OH
(10 mL) was added. The mixture was transferred to a glass cylinder and
(ESI+): m/z: calcd for C48H89N7NaO12 [M+Na]+: 978.65; found: 979.11.
(S)-1-(2,11-Dioxo-14-amino-14-carboxy-3,10-diazatetradecanyl)-1,4,7,10-
tetraazacyclododecane-4,7,10-triacetic acid (L2): Compound 2 (1.50 g,
1.66 mmol) was dissolved in CH2Cl2 (15 mL), trifluoroacetic acid (15 mL)
and triisopropylsilane (0.6 mL) were added, and the mixture was stirred
at room temperature overnight. The liquid was then evaporated in vacuo
and the product precipitated with excess diethyl ether. The product was
isolated by centrifugation, washed thoroughly with diethyl ether, and
dried in vacuo to yield the desired product as an amorphous white solid.
This was purified by preparative HPLC–MS by using a Waters Atlantis
RPdC18 19/100 column by Method 3 and H2O/0.1% TFA (A) and
CH3CN/0.1% TFA (B) as eluents (see the Supporting Information). The
pure product was obtained as a white powder (0.9 g, 89.9%). 1H NMR
(600 MHz, D2O): d=3.68 (m, CHa, 1H), 3.40–2.57 (br, CH2 macrocycle,
NCH2CO, CONHCH2, CH2NH, 16H), 2.91–2.55 (br, CH2 macrocycle,
12H), 2.10 (m, CH2b, 2H), 1.85(m, CH2g, 2H), 1.15, 0.95 ppm (m, aliphat-
ic CH2, 8H); 13C NMR (600 MHz, D2O): d=173.7, 171.6, 170.3, 169.5,
164.8 (CONH, COO, 6C), 65.97, 54.7 (NCH2CO, CHa, 5C), 52.5, 52.0,
51.3, 47.5 (NCH2 macrocycle, 8C), 39.2, 39.0 (CONHCH2, 2C), 30.9,
27.9, 27.8, 26.9, 25.7, 25.4, 25.3 ppm (aliphatic CH2, CH2b, CH2g, 6C); MS
(ESI+): m/z: calcd for C27H50N7O10 [M+H]+: 632.36; found: 632.57.
sonicated in a cup horn at 20 W for 20 min. NaBHACTHNURGTNEUNG(OAc)3 was then
added portionwise and the mixture sonicated for another 30 min. The re-
sulting solution was evaporated in vacuo and the residue dissolved in
H2O/CHCl3 (30 mL, 1:1 (v/v)). Dilute aqueous NH4OH was added under
stirring at 08C to adjust the pH to 8, then the organic phase was separat-
ed and the aqueous phase extracted with CHCl3 (3ꢃ30 mL). The organic
phases were collected, dried with anhydrous Na2SO4, filtered, and evapo-
rated in vacuo. The residue was purified by column chromatography
(silica gel, eluent: CHCl3/MeOH, 9:1; TLC: Rf =0.36) to yield a light-
yellow oil (1.4 g, 49%). 1H NMR (400 MHz, CDCl3): d=8.51 (br,
CONH, 1H), 8.08 (br, CONH, 1H), 7.26–7.18 (m, ArH, 15H), 3.46 (m,
CHa, 1H), 3.41–3.26 (br, CH2 macrocycle, NCH2CO, CONHCH2,
CH2NH, 28H), 2.34 (m, CH2b, 2H), 2.29 (m, CH2g, 2H), 1.41 (m, C-
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
4C), 70.6 (C(Ph)3, 1C), 61.5 (CHa, 1C), 56.4, 56.0, 55.9, 53.9 (NCH2CO,
NCH2 macrocycle, 12C), 48.0 (NHCH2, 1C), 39.5 (CONHCH2, 1C), 34.3,
32.8, 29.7 (aliphatic CH2, CH2b, 3C), 28.7, 28.4, 28.3, 28.2 (CACHTUNTRGNEUNG(CH3)3, 12C),
27.6, 27.2, 26.3 ppm (aliphatic CH2, CH2g, 3C); MS (ESI+): m/z: calcd
for C62H96N7O10 [M+H]+: 1098.72; found: 1099.03; calcd for
C62H95N7NaO10 [M+Na]+: 1120.70; found: 1121.02.
N-Cbz-l-Gln(Tr)-OH: Cbz-L-Gln-OH (1 g, 3.57 mmol) and Tr-OH
(1.86 g, 7.14 mmol) were suspended in glacial acetic acid (11 mL). Ac2O
(0.7 mL, 7.14 mmol) and H2SO4 (0.02 mL, 0.36 mmol) were added and
the reaction mixture was stirred for 1 h under N2 at 508C (bath tempera-
ture) to form a clear yellow solution. The reaction mixture was dropped
into cold water (110 mL) and the precipitate was filtered, dissolved in
EtOAc (30 mL), and washed with H2O (3ꢃ10 mL). The organic phase
was dried with Na2SO4 and the solvent was evaporated under reduced
pressure. The product was dissolved in EtOAc (50 mL) and dropped into
(S)-1-(2-Oxo-11-carboxy-13-carbamoyl-3,10-diazaterdecanyl]-1,4,7,10-tet-
raazacyclododecane-4,7,10-triacetic acid (L3): Trifluoroacetic acid
(20 mL) was added dropwise to a solution of 3 (337 mg, 0.30 mmol) and
triisopropylsilane (0.6 mL) in CH2Cl2 (5 mL) cooled to 0–58C. The solu-
tion was stirred at room temperature for 1 h, then evaporated, and the
residue dissolved in the same volumes of CH2Cl2 and trifluoroacetic acid.
Chem. Eur. J. 2009, 15, 76 – 85
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
83

S. Aime et al.
After 2 h, the solvent volume was reduced and the product precipitated
by adding diethyl ether at 0–58C. The precipitate was dissolved in H2O
(5 mL), the solution neutralized at 0–58C by adding dilute aqueous
NH4OH, and finally lyophilized. The product was then purified by prepa-
rative HPLC–MS by using a Waters Atlantis RPdC18 19/100 column by
Method 4 and 7 mm CH3COONH4 (pH 4) (A) and CH3CN (B) as eluents
(see the Supporting Information). A second fast chromatographic separa-
tion was needed to eliminate salts (Method 2, Supporting Information).
The pure product was obtained as a white powder (270 mg, 39%).
1H NMR (600 MHz, D2O): d=3.82–3.69 (br, CH2 macrocycle, 4H), 3.57
(m, CHa, 1H), 3.42–3.34 (br, CH2 macrocycle, 12H), 3.14–2.96 (br,
NCH2CO, CONHCH2, CH2NH, 12H), 2.38 (m, CH2b, 2H), 2.15–2.02 (m,
CH2g, 2H), 1.66, 1.47, 1.32 ppm (m, aliphatic CH2, 8H); 13C NMR
(600 MHz, D2O): d=178.8, 176.9, 175.9, 175.8, 170.9 (CONH, COO, 6C),
60.2 (CHa, 1C), 59.9, 55.0 (NCH2CO, 4C), 53.5, 51.1, 48.1, 46.6 (NCH2
macrocycle, 8C), 41.8 (CH2NH, 1C), 39.2 (CONHCH2, 1C), 30.4, 29.6,
28.0, 27.8, 25.8, 25.7 ppm (aliphatic CH2, CH2b, CH2g, 6C); MS (ESI+):
m/z: calcd for C27H50N7O10 [M+H]+: 632.36; found: 632.51; calcd for
C27H49N7NaO10 [M+Na]+: 654.34; found: 654.5.
sin, glutamine, and all other chemicals were purchased from Sigma
Chemical Co., St. Louis, MO, USA. Rat hepatocytes were isolated by fol-
lowing the previously reported method[35,36] and were used for the uptake
experiments after 24 h of culture. HTC (rat hepatoma tissue culture)
were grown in DMEM F-12 medium supplemented with 5% FBS, 2 mm
glutamine, 100 UmLꢀ1 penicillin, and 100 mgmLꢀ1 streptomycin. HTC
were cultured in 75 cm2 flasks in a humidified incubator in CO2/air (5:95,
v/v) at 378C. For the uptake experiments HTC cells were seeded at a
density of 1.5ꢃ104 cellscmꢀ2. Cells were ready for the uptake experi-
ments 24 h after seeding.
Uptake experiments: To carry out the uptake experiments with the gado-
linium complexes, HTC and rat hepatocytes were seeded in 6 and 10 cm
diameter culture dishes, respectively. After removal of the culture
medium, cells were washed three times with phosphate saline buffer
(PBS) (5 mL) and then Earlꢂs Balanced Salt Solution (2 or 5 mL; EBSS:
CaCl2 0.266 gLꢀ1, KCl 0.4 gLꢀ1, NaCl 6.8 gLꢀ1, glucose 1 gLꢀ1, MgSO4
0.204 gLꢀ1, NaH2PO4 0.144 gLꢀ1, NaHCO3 1.1 gLꢀ1, pH 7.4) were added
to the 6 or 10 cm dishes, respectively. Cells were incubated (5% CO2 at
378C) for 6 h with the different compounds dissolved in EBSS at a 1 mm
concentration. At the end of the incubation, the medium was removed
and cells were washed three times with ice-cold PBS. Then cells were re-
covered from the Petri dishes in 250 mL PBS with a cell scraper. The ga-
dolinium content of the HTC was determined by using inductively cou-
pled plasma mass spectrometry (ICP-MS, Element-2, Thermo-Finnigan,
Rodano (MI), Italy). Sample digestion was performed with concentrated
HNO3 (70%, 2 mL) under microwawe heating (Milestone MicroSYNTH
Microwave labstation equipped with an optical fiber temperature control
and a HPR-1000/6M six-position high-pressure reactor, Bergamo, Italy).
After digestion the volume of each sample was brought to 2 mL with ul-
trapure water, and the samples were analyzed by ICP-MS. Three repli-
cates of each sample solution were analyzed. The protein concentration
of each sample was determined from cell lysates by the Bradford method
using bovine serum albumin as the standard.
(S)-1-[6-(2-Carboxy-5-oxopyrrolidin-1-yl)hexylcarbamoylmethyl]-1,4,7,10-
tetraazacyclododecane-4,7,10-triacetic acid (L5)—cyclic product: The cy-
clized product (123 mg, 20%) was recovered by HPLC separation.
1H NMR (600 MHz, D2O): d=4.40 (m, CHa, 1H), 3.92–3.64 (br, CH2
macrocycle, 4H), 3.42–3.10 (br, CH2 macrocycle, NCH2CO, CONHCH2,
CH2NH, 24H), 2.45–2.34 (m, CH2b, 2H), 2.12–2.02 (m, CHg, 2H), 1.46,
1.26 ppm (m, aliphatic CH2, 8H); MS (ESI+): m/z: calcd for
C27H47N6O10 [M+H]+: 615.33; found: 615.41.
GdIII complexes (Gd-L1, Gd-L2, Gd-L3, Gd-L4, Gd-L5, and Gd-DOTA-
MAC6NH2: An equimolar amount of GdCl3 solution (about 120 mm in
water) was slowly added to a 40 mm ligand solution; the solutions were
maintained at pH 7 with NaOH.
In the cases of Gd-L1 and Gd-L3, iminodiacetic acid (0.5 equiv) was
added after 2 h to complex free Gd3+ ions. The mixtures were stirred
overnight at RT and lyophilized to give a white solid. Then Gd-L1 and
Gd-L3 were purified by preparative HPLC–MS by Methods 5 and 6, re-
spectively, using a Waters Atlantis RPdC18 19/100 column and H2O (A)
and CH3CN (B) as eluents (see the Supporting Information). Starting
from crude L1 and L3 (i.e., also containing L4 and L5, respectively), the
same protocol was used, but this time Gd-L4 and Gd-L5 were also ob-
tained after HPLC purification.
MRI: MR images of cell pellets were acquired on a Bruker Avance 300
(7 T) instrument equipped with a microimaging probe. The system was
endowed with two birdcage resonators of 30 and 10 mm inner diameter,
respectively. 1.5ꢃ106 HTC cells were incubated for 6 h with equal
amounts (1 mm) of the different gadolinium complexes. 5ꢃ106 rat hepa-
tocytes were incubated for 6 h with Gd-L1. At the end of the incubation,
cells were washed three times with ice-cold PBS, detached with EDTA
(1 mm), and transferred into glass capillaries for the analysis. MR imaging
was performed by using a standard T1-weighted multislice multiecho se-
quence (TR/TE/NEX=200:3.3:16, field of view=1.2 cm, one slice=
1 mm, in-plane resolution=94ꢃ94 mm). T1 measurements of cells were
performed by using a standard saturation recovery sequence.
In the cases of Gd-L2 and Gd-DOTAMAC6NH2, the mixtures were
stirred overnight at RT. Then the pH was raised to 8.5 and the mixtures
stirred for 2 h. Centrifugation at 7000 rpm for 5 min at 108C allowed the
separation of Gd(OH)3 from solution.
The amount of residual free Gd3+ ions was assessed by the Orange Xyle-
nol UV method.[26] All complexes used in this work were found to con-
tain less than 0.3% (mol/mol) of residual free Gd3+ ions. The overall ga-
dolinium contents were determined by 1H NMR T1 measurement of the
mineralized complex solution (in 6m HCl at 1208C for 16 h).
Gd-L1: MS (ESI+): m/z: calcd for C27H48GdN8O9 [M+H]+: 786.28;
found: 786.74; HPLC: Method 7 (see the Supporting Information), reten-
tion time 2.83 min, purity 96%.
Gd-L4: MS (ESI+): m/z: calcd for C27H45GdN7O9 [M+H]+: 769.25;
found: 769.65; HPLC: Method 7, retention time 4.70 min, purity 80%.
Gd-L2: MS (ESI+): m/z: calcd for C27H47GdN7O10 [M+H]+: 787.26;
Acknowledgements
This work was supported by the EC-FP6 (MEDITRANS: NMP4-CT-
2006-026668; DiMI: LSHB-CT-2005-512146, and EMIL: LSHC-CT-2004-
503569) and by the Italian Government (project nos. FIRB
RBNE03PX83_006, FIRB RBIP06293N, and PRIN 2005039914). EC-
COST D38 Action is also acknowledged.
found: 787.68; HPLC: Method 7, retention time 3.16 min, purity 95%.
Gd-L3: MS (ESI+): m/z: calcd for C27H47GdN7O10 [M+H]+: 787.26;
found: 787.51; HPLC: Method 7, retention time 2.08 min, purity 84%.
Gd-L5: MS (ESI+): m/z: calcd for C27H44GdN6O10 [M+H]+: 770.24;
found: 770.61; HPLC: Method 7, retention time 3.27 min, purity 80%.
[1] T. F. Massoud, S. S. Gambhir, Trends Mol. Med. 2007, 13, 183–191.
[2] R. Weissleder, M. J. Pittet, Nature 2008, 452, 580–589.
[3] K. Brindle, Nat. Rev. Cancer 2008, 8, 94–107.
[4] S. Aime, S. Geninatti Crich, E. Gianolio, G. B. Giovenzana, L. Tei,
E. Terreno, Coord. Chem. Rev. 2006, 250, 1562–1579.
[5] W. J. Mulder, G. J. Strijkers, G. A. van Tilborg, A. W. Griffioen, K.
Nicolay, NMR Biomed. 2006, 19, 142–164.
For all complexes the signals observed show isotopic distribution consis-
tent with a gadolinium complex. The products were stored at ꢀ208C to
prevent cyclization.
Cell culture: The culture media DMEM F-12, M199, and fetal bovine
serum (FBS) were purchased from Cambrex, East Rutherford, NJ, USA.
The penicillin–streptomycin mixture, serum bovine albumin (BSA), tryp-
[6] N. Kamaly, T. Kalber, A. Ahmad, M. H. Oliver, P. W. So, A. H. Her-
lihy, J. D. Bell, M. R. Jorgensen, A. D. Miller, Bioconjugate Chem.
2008, 19, 118–129.
84
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 76 – 85

Improved MRI Visualization of Tumor Cells
FULL PAPER
[7] S. Erdogan, Z. O. Medarova, A. Roby, A. Moore, V. P. Torchilin, J.
Magn. Reson. Imaging 2008, 27, 574–580.
[8] K. F. Pirollo, J. Dagata, P. Wang, M. Freedman, A. Vladar, S. Fricke,
L. Ileva, Q. Zhou, E. H Chang, Mol. Imaging 2006, 5, 41–52.
[9] S. Laurent, D. Forge, M. Port, A. Roch, C. Robic, L. Vander Elst,
R. N. Muller, Chem. Rev. 2008, 108, 2064–2110.
[10] C. Corot, P. Robert, J. M. Idee, M. Port, Adv. Drug Delivery Rev.
2006, 58, 1471–1504.
[11] S. D. Caruthers, P. M. Winter, S. A. Wickline, G. M. Lanza, Methods
Mol. Med. 2006, 124, 387–400.
[23] G. Chen, J. B. Russell, J. Bacteriol. 1989, 171, 2981–2985.
[24] R. W. Garden, T. P. Moroz, J. M. Gleeson, P. D. Floyd, L. Li, S. S.
Rubakhin, J. V. Sweedler, J. Neurochem. 1999, 72, 676–681.
[25] W. H. Busby, G. E. Quackenbush, J. Humm, W. W. Youngblood, J. S.
Kizer, J. Biol. Chem. 1987, 262, 8532–8536.
[26] A. Barge, G. Cravotto, E. Gianolio, F. Fedeli, Contrast Media Mol.
Imaging 2006, 1, 184–188.
[27] D. Thonon, V. Jacques, J. F. Desreux, Contrast Media Mol. Imaging
2007, 2, 24–34.
[28] P. Caravan, J. J. Ellison, T. J. McCurry, Chem. Rev. 1999, 99, 2293–
2352.
[12] T. D. Tran, S. D. Caruthers, M. Hughes, J. N. Marsh, T. Cyrus, P. M.
Winter, A. M. Neubauer, S. A. Wickline, G. M. Lanza, Int. J.
Nanomed. 2007, 2, 515–526.
[29] B. P. Bode, D. L. Kaminski, W. W Souba, A. P. Li, Hepatology 1995,
21, 511–520.
[13] S. Geninatti Crich, C. Cabella, A. Barge, S. Belfiore, C. Ghirelli, L.
Lattuada, S. Lanzardo, A. Mortillaro, L. Tei, M. Visigalli, G. Forni,
S. Aime, J. Med. Chem. 2006, 49, 4926–4936.
[30] J. D. McGivan, Biochem. J. 1998, 330, 255–260.
[31] C. Li, M. F. Penet, P. Winnard, Jr., D. Artemov, Z. M. Bhujwalla,
Clin. Cancer Res. 2008, 14, 515–522.
[14] L. Lattuada, S. Demattio, V. Vincenzi, C. Cabella, M. Visigalli, S.
Aime, S. Geninatti Crich, E. Gianolio, Bioorg. Med. Chem. Lett.
2006, 16, 4111–4114.
[15] W. W. Souba, Ann. Surg. 1993, 218, 715–728.
[16] M. A. Medina, J. Nutr. 2001, 131, 2539S-2542S.
[17] A. Barge, L. Tei, D. Upadhyaya, F. Fedeli, L. Beltrami, R. Stefanꢄa,
S. Aime, G. Cravotto, Org. Biomol. Chem. 2008, 6, 1176–1184.
[18] P. Sieber, B. Rineker, Tetrahedron Lett. 1991, 32, 739–742.
[19] G. Cravotto, G. Omiccioli, L. Stevanato, Ultrason. Sonochem. 2005,
12, 213–217.
[32] P. M. Winter, K. Cai, S. D. Caruthers, S. A. Wickline, G. M. Lanza,
Expert Rev. Med. Devices 2007, 4, 137–145.
[33] A. S. Krishnan, A. A. Neves, M. M. de Backer, D. E. Hu, B. Davle-
tov, M. I. Kettunen, K. M. Brindle, Radiology 2008, 246, 854–862.
[34] G. Cravotto, A. Dimetri, G. M. Nano, G. Palmisano, A. Penoni, S.
Tagliapietra, Eur. J. Org. Chem. 2003, 4438–4444.
[35] G. Gardini, C. Cabella, C. Cravanzola, C. Vargiu, S. Belliardo, G.
Testore, S. P. Solinas, A. Toninello, M. A. Grillo, S. Colombatto, J.
Hepatol. 2001, 35, 482–489.
[36] I. Probst, Eur. J. Biochem. 1985, 153, 347–353.
[20] A. Barge, G. Cravotto, B. Robaldo, E. Gianolio, S. Aime, J. Inclu-
sion Phenom. Macrocyclic Chem. 2007, 57, 489–495.
[21] V. R. Young, A. M. Ajami, J. Nutr. 2001, 131, 2449S–2459S.
[22] B. D. Ross, S. Bullock, Biochem. J. 1978, 170, 177–179.
Received: July 31, 2008
Published online: December 4, 2008
Chem. Eur. J. 2009, 15, 76 – 85
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
85