PEG-covered ultra-small Gd₂O₃ nanoparticles for positive contrast at 1.5T MR clinical scanning

Publisher’s version / Version de l'éditeur:

Nanotechnology, 18, 39, pp. 395501-1-395501-9, 2007-09-04

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.

https://nrc-publications.canada.ca/eng/copyright

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la

première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

NRC Publications Archive

Archives des publications du CNRC

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

https://doi.org/10.1088/0957-4484/18/39/395501

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

PEG-covered ultra-small Gd

_{₂O₃ nanoparticles for positive contrast at}

1.5T MR clinical scanning

Fortin, Marc-André; Petoral Jr., Rodrigo M.; Söderlind, Fredrik; Klasson, A.;

Engstöm, Maria; Veres, Teodor; Käll, Per-Olod; Uvdal, Kajsa

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

NRC Publications Record / Notice d'Archives des publications de CNRC:

https://nrc-publications.canada.ca/eng/view/object/?id=d204616c-f9e6-4780-bb1a-d767a762af42

https://publications-cnrc.canada.ca/fra/voir/objet/?id=d204616c-f9e6-4780-bb1a-d767a762af42

Nanotechnology 18 (2007) 395501 (9pp) doi:10.1088/0957-4484/18/39/395501

Polyethylene glycol-covered ultra-small

Gd

₂

O

₃

nanoparticles for positive contrast

at 1.5 T magnetic resonance clinical

scanning

Marc-Andr´e Fortin

_{, Rodrigo M Petoral Jr}

_{, Fredrik S¨oderlind}

_,

A Klasson

_{, Maria Engstr¨om}

_{, Teodor Veres}

_{, Per-Olof K¨all}

_and

Kajsa Uvdal

1_{Unit´e de Bioing´enerie et de Biotechnologie, Centre Hospitalier Universitaire de Qu´ebec}

(CHUQ) and Laval University, QC, G1K 7P4, Canada

2_{Department of Physics, Chemistry and Biology (IFM), Link¨oping University, SE-58183}

Link¨oping, Sweden

3_{Center for Medical Image Science and Visualization (CMIV), Link¨oping University/US,}

SE-58185 Link¨oping, Sweden

4_{National Research Council of Canada (CNRC-IMI) 75, boulevard de Mortagne,}

Boucherville, QC, J4B 6Y4, Canada E-mail:marc-andre.fortin@gmn.ulaval.ca

Received 29 May 2007, in final form 25 July 2007

Published 4 September 2007

Online at

stacks.iop.org/Nano/18/395501

Abstract

The size distribution and magnetic properties of ultra-small gadolinium oxide crystals

(US-Gd

O

) were studied, and the impact of polyethylene glycol capping on the

relaxivity constants (r

, r

) and signal intensity with this contrast agent was investigated.

Size distribution and magnetic properties of US-Gd

O

nanocrystals were measured

with a TEM and PPMS magnetometer. For relaxation studies, diethylene glycol

(DEG)-capped US-Gd

O

nanocrystals were reacted with PEG-silane (MW 5000).

Suspensions were adequately dialyzed in water to eliminate traces of Gd

and

surfactants. The particle hydrodynamic radius was measured with dynamic light

scattering (DLS) and the proton relaxation times were measured with a 1.5 T MRI

scanner. Parallel studies were performed with DEG–Gd

O

and PEG-silane–SPGO

(Gd

O

, < 40 nm diameter). The small and narrow size distribution of US-Gd

O

was

confirmed with TEM (∼3 nm) and DLS. PEG-silane–US-Gd

O

relaxation parameters

were twice as high as for Gd–DTPA and the r

/r

ratio was 1.4. PEG-silane–SPGO gave

low r

relaxivities and high r

/r

ratios, less compatible with positive contrast agent

requirements. Higher r

were obtained with PEG-silane in comparison to DEG–Gd

O

.

Treatment of DEG–US-Gd

O

with PEG-silane provides enhanced relaxivity while

preventing aggregation of the oxide cores. This study confirms that PEG-covered Gd

O

nanoparticles can be used for positively contrasted MR applications requiring stability,

biocompatible coatings and nanocrystal functionalization.

S

Supplementary data are available from

stacks.iop.org/Nano/18/395501

(Some figures in this article are in colour only in the electronic version)

5 _{Address for correspondence: Unit´e de Bioing´enerie et de Biotechnologie,}

Nanotechnology 18 (2007) 395501 M-A Fortin et al

1. Introduction

Gadolinium chelates coupled to T1-weighted imaging proce-dures are widely used in magnetic resonance imaging (MRI) for providing positive contrast of the vascular system. Chelates such as Gd–diethylenetriaminepentaacetic acid (Gd–DTPA) are fast diffusing small molecules containing paramagnetic metal ions. The fast diffusion and low specific signal of Gd chelates can, however, represent drawbacks in several contrast agent applications. In fact, the MR signal provided by the re-laxation of protons in the vicinity of a singly chelated Gd ion is too low to compete with the signal and detection capacity of more sensitive imaging modalities such as positron emission tomography (PET). Poor sensitivity of MRI and insufficient signal per unit of contrast agent have long been considered a barrier for the development of MR-based molecular imaging diagnostics, despite better observation depths and intrinsically higher spatial resolutions than nuclear methods.

More specific contrast agents have been developed to offer proton relaxation effects per unit of contrast agent that could be several orders of magnitude higher than with conventional Gd chelates. In particular, the emergence of ultra-small iron oxide nanoparticles (USPIO, core ∅ _{< 10 nm) [1, 2]}

has led to the clinical establishment of a variety of T2 (negative) contrast strategies. Large-molecular agents such as USPIO could enable imaging at lower temporal resolutions because of the prolonged circulation times of large-molecular compared with small-molecular agents [3]. Enhanced T2 relaxivity with USPIO is due to susceptibility effects: the magnetization differences, which occur as a result of non-homogeneous distribution of magnetized material in vivo, give rise to local field gradients which tend to accelerate the loss of phase coherence of the spins contributing to the MR signal. Adequately covered with biocompatible polymers preventing the oxide cores’ agglomeration, USPIO can also remain in the blood for longer durations than Gd chelates. This allows longer times for imaging the vascular system, and the possibility to more effectively deliver contrast agents to targeted areas in the body.

Depending on the concentration and the fine optimization of MR acquisition sequences, fine USPIOs can also provide positive contrast in certain diagnostics. The ratio between the transverse and longitudinal relaxivities (expressed as r2/r1) is the parameter defining the potential of a contrast agent in providing positive or negative contrast. When reachingr2/r1 close to unity, brightening is observed and can be clearly shown inT1-weighted images [4,5]. Adequately capped with biocompatible molecules, fine USPIOs can therefore provide positive contrast (T1) and high signal intensities, which can be valuable assets for vascular imaging [5–7], cell tracking [8] and also molecular targeting studies [9]. The signal intensity onT1 -weighted images of contrast agent solutions obtained by spin– echo MR sequences can be estimated by using bothr1(1/ T1) andr2(1/ T2), according to the following equation [10]:

S∝ρ

1−e−TR/T1

e−TE/T2 ₍₁₎

where ρ is the spin density, TR is the repetition time and TE is the echo time. It is clear from equation (1) that higher longitudinal relaxation rates(1/ T1)enhance the signal

intensity, while an increase in transversal relaxation rates

(1/ T2)has an opposite effect.

A disadvantage of the use of superparamagnetic contrast agents (e.g. USPIO) is that T2 relaxation may negatively affect T1-weighted MR imaging [11, 12]. Although the T1 relaxivity of iron oxide nanoparticles is typically three or four times higher than for small gadolinium chelates [4], the signal intensity created by the relaxing protons in the USPIOs vicinity is limited by the large magnetization and relatedT2effects [13]. At local iron concentrations higher than 500 µmol l−1, a marked contrast decrease is often noted [14]; at concentrations of 1.5 mmol Fe, susceptibility artefacts can even interfere with the diagnostic of endogenous accumulations of contrast agent (e.g. in small haemorrhages) [15]. Those limitations can severely affect T1-based perfusion imaging [16, 17] and contrast-enhanced MRA [18], requiring the development of new contrast agents coupled with dose optimization methods. In cell staining studies, which is one of the more promising applications of magnetic nanoparticles, the intracellular compartmentalization of USPIOs substantially decreases the T1 relaxivity, leading to predominant negative contrast [8]. Hence the high r2/r1 ratio and related signal intensity effects have a restrictive impact on the useful concentration range of USPIOs. Fine optimization of the acquisition parameters are also required in order to produce reliable images from which quantitative data could be extracted based on the local contrast effects’ interpretation.

In the context of cell labelling and molecular targeting studies, there is a strong interest for developing MR positive contrast agents providing high signal per unit. Nanoparticulate contrast agents containing a large number of magnetic ions, and showingr2/r1 ratios close to unity, could offer higher signal-to-noise ratios and better anatomic resolutions in T1 -weighted images [19]. Those characteristics would be particularly well suited for targeted molecular imaging and cell labelling studies. The magnetic tags should ideally be small enough (<10 nm) to roughly match the size of common targeting molecules such as antibodies. Ultra-small gadolinium oxide crystals (US-Gd2O3,∼3 nm) could provide both a high number of paramagnetic atoms per unit of contrast agent, fine diameters and the lowr2/r1ratios necessary for positive contrast and high specific signal. Due to the exceptional fraction of surface atoms per crystal, nanoscale magnetism effects in very small metallic clusters can lead to much larger average magnetic moment per atom [20], which in turn could translate into enhanced proton relaxation in the contrast agent’s surroundings.

Recent improvements in chemical reduction nanoparticle synthesis have led to the establishment of reliable methods for synthesizing Gd2O3nanoparticles of very fine and narrow size distributions [21, 22]. We recently reported on the characterization of a new contrast agent made of ultra-small gadolinia cores (US-Gd2O3) capped with diethylene glycol (DEG) [23] and providing enhanced proton relaxivities compared with Gd–DTPA. In the present study, we provide a comprehensive characterization of the size and magnetic properties of US-Gd2O3; we measure the proton relaxivities of the solutions and we suggest a surface treatment that prevents the oxide cores’ agglomeration. Capping the nanoparticles with nonimmunogenic, nonantigenic and protein-resistant 2

layers of a biocompatible material such as polyethylene glycol (PEG) could help in preventing aggregation, enhancing the blood retention, facilitating nanoparticle cell uptake and providing binding sites for targeting agents [24, 25]. Also, our preliminary studies with DEG–Gd2O3showed indications that diethylene glycol-covered US-Gd2O3 could not perform adequately when submitted to electrolyte conditions, to long dialysis times and to repeated freeze–thawing cycles (gel-like sedimentation). PEG-silane grafting at the surface of US-Gd2O3 could be a promising treatment to stabilize Gd2O3 surfaces. We therefore quantified the impact of residual DEG on proton relaxivities and MR contrast enhancement, and we have found paths towards the establishment of an optimal and functional gadolinium oxide surface treatment for systemically injectable nanoparticles.

2. Materials and methods

2.1. Synthesis of Gd2O3nanocrystals

Gd2O3 nanocrystals were synthesized via the polyol route [21,22] using Gd(NO3)3·6H2O dissolved in diethylene glycol (DEG) and heated until a clear solution was obtained (at 90–100◦_{C). Amounts of the hydrolysis agent (NaOH) were} added, and the temperature was raised to 180–210◦

C for 4 h. After cooling, the nanocrystal-containing solution was centrifuge-filtered for 30 min at 40◦

C and 2000 rpm (filters: polyethersulfone, 0.2 µm, Vivascience Sartorius, Hannover, Germany), until complete collection of the fluid. Agglomerations or large-size particles that could interfere with dynamic light scattering measurements were therefore eliminated.

2.2. Dialysis procedure and treatment with PEG-silane

During preliminary studies we found that 85–95% of the Gd3+ precursors used to prepare the synthesis bath remained in the DEG–Gd2O3colloidal suspension after reduction. Therefore, the as-synthesized nanoparticulate contrast agent must be cleared from excess Gd3+ ions, and the DEG solvent must be replaced by water. After filtration, fractions of 4 ml (A), 2 ml (B) and 2 ml (C) were sampled from the DEG suspension. Each sample was put in cylindrical dialysis membranes (1000 MW, regenerated cellulose, Spectra/Por, SpectrumLabs, Rancho Dominguez, CA) and dialyzed against deionized water in separate tanks, following a contrast agent:water ratio of 1:1000. For economical and environmental reasons, a preliminary dialysis bath of smaller volume (1:10, 20 min) could also be used, enabling the recuperation and recycling of a large fraction of unreacted Gd3+ions prior to the 1:1000 ratio dialysis. Water was refreshed three times on day 1, and then two times per day. Sample B was collected after 4 days and sample C after 6 days. Both samples were stored at 4◦

C until the measurement of relaxation parameters was performed (within 48 h). Sample A was removed from the bath after 2 days only and mPEG-silane (MW 5901, Nektar Therapeutics, Huntsville, AL) was added (15 mg ml−1) to the Gd2O3 suspension. The mixture was vigorously vortexed during 15 min (Vortex Genie, G-560E, Sci. Ind. Inc., Bohemia, NY), and then sonicated for 2 h at 40◦

C. The suspension was placed back into a fresh 1000 MW dialysis

membrane and dialyzed against deionized water for 24 h. Then, the membrane was changed for a 12–14 000 MW membrane, allowing the removal of excess PEG. The dialysis was prolonged for 24 h and then stopped. Procedure A was performed in duplicate, and the compound was identified as PEG–US-Gd2O3. Samples A were stored in the freezer until measurement of relaxation parameters. Binding of the silane group to the gadolinium oxide surface was assessed by x-ray photoelectron spectroscopy.

2.3. Nanoparticle size study

Dynamic light scattering measurements (AV/DLS-5000 sys-tem, Lange) were performed on as-synthesized DEG–Gd2O3 (50 times diluted in water), on dialyzed DEG–Gd2O3samples A (2 days), B (4 days) and C (6 days) and also on PEG– US-Gd2O3. The optimal counting rate was about 250 mHz, and normalized intensity correlation function curves were care-fully fitted with an exponential algorithm of the second order (200 grid points). A comparative size distribution study of the oxide cores was performed with high-resolution transmis-sion electron microscopy (HRTEM). Samples for TEM anal-ysis were prepared by dispersing as-synthesized, non-dialyzed DEG–Gd2O3in methanol (50×vol.). The dispersion was dried on amorphous carbon-covered copper grids and analyzed with a FEI Tecnai G2 electron microscope (200 kV).

2.4. PEG-stabilized SPGO

In order to compare with a coarser Gd2O3system, suspensions of <40 nm Gd2O3 crystals (SPGO) [26] were prepared by using commercially available gadolinia nanoparticles (99.999% pure, Sigma-Aldrich). One (1) g of Gd2O3(<40 nm) was mixed with 10 ml of deionized water and sonicated for 2 h with mPEG-silane (15 mg ml−1_{) at 40}◦_{C. Agglomerations} of nanoparticles were eliminated by centrifugation (2000 rpm, 30 min) and the suspension was dialyzed against water, in the same two-step procedure as described above (1000 MW membrane, 24 h; 12–14 000 MW membrane, 24 h).

2.5. Relaxivity measurements

The following contrast agent preparations were placed in a series of 1.5 ml tubes and diluted to concentrations ranging from 0.1 to 1.6 mM Gd:DEG–US-Gd2O3 (4 and 6 days’ dialysis), PEG–US-Gd2O3 and PEG–SPGO. Final preparations had pHs ranging from 6.0 to 7.2. Contrast agents were measured for relaxivity with a 1.5 T Philips Achieva whole-body system, according to an established method [23]. The tubes were immersed in a basin containing saline solution kept at 19.5 ±1.5◦_{C and inserted in a head coil. A 2D} mixed spin–echo (SE) sequence interleaved with an inversion recovery (IR) sequence was used [23,27]. Images from four echoes were acquired in each sequence. Measurements were performed on 7 mm cross-section slices of the vials. In order to cover the entire concentration range of the contrast agent, we used two sets of parameters (TE, TR(SE), TR(IR), IR delay, set 1: 30, 500, 1150, 150 ms; set 2: 50, 760, 2290, 370): longitudinal and transverse relaxation rates (1/ T1and 1/ T2) results showing the lowest standard deviation were selected to calculate transversal and longitudinal relaxivity parameters

Nanotechnology 18 (2007) 395501 M-A Fortin et al

(a) (b)

Figure 1.US-Gd2O3particle diameter distribution (high resolution

TEM study) in frequency units (the total fraction of the nanoparticle sample contained in a given diameter range: sum of the columns is equal to 1). Chemically synthesized nanocrystals have a narrow size distribution.

(r1 and r2). Dilutions of GdCl3 (Gd+ ions), and chelated gadolinium (Gd–DTPA, diethylenetriaminepentaacetic acid, Magnevist®) in deionized water were also measured. Gd contents in all suspensions and solutions were assessed by ICP-SFMS (ELEMENT, Finnigan MAT, Bremen) and the results were plotted in figures2and3.

2.6. Estimation of MRI signal enhancement

In order to compare the signal intensity given by the different contrast agents, we simulated intensity curves by using equation (1) and results from the proton relaxation rates as well as the following values:ρ =1, TR=350 ms, TE=10 ms.

2.7. Magnetization study

The magnetization of US-Gd2O3 was measured with a Quantum Design physical property measurement system (PPMS). Ultra-small nanoparticles were collected by adding amounts of citric acid to the cooled synthesis bath: the sediment was thoroughly rinsed with methanol and dried. About 30 mg of powder was used per scan. ICP-SFMS analysis was performed to evaluate the mass fraction of Gd contained in citric acid-capped US-Gd2O3which gave a value of 35.8% (2.27 mmol Gd per gram of CA–Gd2O3). In PPMS measurements, the applied magnetic field ranged from

−60 000 to 60 000 Oe, with a sensitivity of 2×10−5 _emu for DC measurements. The temperature dependence of magnetization was measured in an applied magnetic field of 100 Oe between 5 and 300 K, using zero-field-cooling (ZFC) and field-cooling (FC) procedures.

3. Results

3.1. Particle-size studies

3.1.1. TEM. High-resolution TEM images of the

nanocrystals exhibited lattice fringes and indicated a high crystallinity (figure 1). The particles form a spherical shape of uniform size, and the mean particle diameter was about 3 nm, as calculated from measurements performed over more than 200 nanoparticles. In our previous study, we performed XRPD studies on US-Gd2O3 nanoparticles produced by the polyol method and identified the 222 peak (at 2θ ≈ 28.6◦

) corresponding to cubic Gd2O3[22]. Louis et al [28] measured

(a)

(b)

(c)

Figure 2.(a) R1(1/ T1) and (b) R2(1/ T2) relaxation rates with

PEG–US-Gd2O3nanoparticles (∼3 nm crystal cores) compared with

results from PEG–SPGO (<40 nm Gd2O3crystal cores). Relaxation

rates with Gd3+_{ions (GdCl}

3dissolved in water) and Gd–DTPA

(Magnevist) are also given. (c) Simulated signal intensities with PEG–US-Gd2O3, PEG–SPGO, Gd–DTPA and Gd3+ions. High T1

relaxivity with US-Gd2O3generates a steep signal increase at low

concentration (0.1–0.4 mM Gd). The simulated signal intensity peaks at ∼0.8 mM Gd without sharp signal loss at concentrations higher than 1.0 mM. Error bars are the sum of the absolute pipetting error and MR measurement standard deviation.

the inter-reticular planes from HRTEM experiments obtained with terbium-doped Gd2O3nanocrystals also produced by the polyol method and found values in perfect agreement with those expected in a cubic phase with a cell parameter of around 10.81 ˚A [29].

3.1.2. DLS. The DLS intensity decay curves obtained with PEG–US-Gd2O3(sample A) gave sharp and reproducible peaks at 2.8± 1.1 nm. The scattered photon signal was constant and no sign of large aggregates was found in the fluids. However, with as-synthesized, diluted suspensions 4

(a)

(b)

(c)

Figure 3.(a) R1(1/ T1), (b) R2(1/ T2) relaxation rates and (c) signal

intensities obtained with PEG–US-Gd2O3nanoparticles (∼3 nm

crystal cores) compared with results from DEG–US-Gd2O3dialyzed

for 4 and 6 days. Optimal relaxivities were obtained with

PEG–US-Gd2O3. Elimination of diethylene glycol from the surface

of Gd2O3crystal cores has a significant impact both on the relaxivity

and on the signal intensity. Error bars are the sum of the absolute pipetting error and MR measurement standard deviation.

containing large amounts of DEG, the signal was too irregular to allow the generation of reproducible fits. One may speculate that grafting PEG at the surface of US-Gd2O3 restrains the hydrodynamic particle radius while favouring the elimination of DEG from the nanoparticulate system (as suggested in supporting information, figure A (available at stacks.iop.org/Nano/18/395501)). However, establishing a correlation between reproducibility of DLS decay curves and the presence of diethylene glycol at the surface of rare-earth oxide nanoparticles is a task that exceeds the scope of the present work and we will address this question in future studies. With PEG–SPGO, the particle size distribution was clearly broader. Signs of coarse aggregations

were evident on the decay curves, despite treatment with PEG-silane and repeated centrifuge–filtering procedures prior to DLS measurement. Although we did not observe sedimentation at the macroscopic level, those results illustrate the difficulty to obtain stable suspensions of monodisperse Gd2O3nanoparticles from agglomerated precursor powders.

3.2. Relaxation studies

Relaxivity parameters (r1, r2) of DEG–Gd2O3, PEG–US-Gd2O3, PEG–SPGO, Gd3+ and Gd–DTPA were calculated from the slope of 1/ T1and 1/ T2graphs (figures2(a) and (b) and 3(a) and (b)) and are listed in table 1. Among all gadolinium oxide suspensions, PEG–US-Gd2O3revealed the highest r1 relaxivity (9.4 mM−1 s−1) and a r2/r1 ratio of 1.4, close enough to unity to represent a significant advantage for positively contrasted applications [4, 5, 19]. PEG–US-Gd2O3suspensions can express longitudinal proton relaxivities at least 2.2 times higher than common Gd chelates such as Gd–DTPA. Figure 3 reveals the significant impact of DEG overlayers on the proton relaxivities as well as on the predicted MR signal intensity. DEG–Gd2O3relaxivities increased by a factor of∼4 (figures3(a) and (b)) by prolonging the dialysis procedure to 6 days (instead of 4), and ther2/r1ratio remained above two which is sufficiently high compared with PEG–US-Gd2O3to cause a significant decrease in signal intensity. The comparative results obtained with PEG-silane-treated coarser Gd2O3nanopowders (PEG–SPGO<40 nm) gave longitudinal proton relaxivities on a very different range (∼0.1 mM−1s−1) and with a r2/r1 ratio as high as ∼81 (figure 2). Those results are similar to previously reported data for SPGO [26]. In our studies, SPGO gives a predominantr2 (1/ T2) effect in water, while keeping the spin–lattice relaxation time (T1) almost undisturbed.

3.3. Estimation of MRI signal enhancement

The calculated signal intensity profile with PEG–US-Gd2O3 (figures 2(c) and 3(c)) is similar to that of paramagnetic relaxation agents [30]. According to equation (1), short T1 increases the signal whereas shortT2leads to a signal decrease, and a signal peak occurs at intermediate concentrations. In figure2(c), the signal intensity curve profiles obtained with Gd3+and PEG–US-Gd2O3(free from Gd3+) almost overlap. Compared with Gd–DTPA, PEG–US-Gd2O3 signal intensity curves express a steep signal increase at low Gd concentrations, which suggests that lower molar quantities of Gd would be necessary to provide the same contrast effect in terms of signal intensity. Such signal increase was not observed with PEG–SPGO (figure2(c)). Due to poor 1/ T1relaxation rates and to much higherr2/r1 ratios, signal intensity projections with PEG–SPGO gave signal intensities more than four times lower than with PEG–US-Gd2O3. Finally, prolonged dialysis times significantly enhance the signal intensity of DEG-capped Gd2O3(figure 3(c)); however, diethylene glycol capping has an overall restrictive effect on relaxivity parameters compared with PEG–US-Gd2O3.

3.4. Magnetization study

The ZFC–FC loop obtained for US-Gd2O3 powders is displayed in figure4(a). The shape of the curve is similar to

Nanotechnology 18 (2007) 395501 M-A Fortin et al

(a) (b)

Figure 4.(a) ZFC–FC curves for the citric acid capped US-Gd2O3(the two curves are superposed ZFC and FC curves). Even at temperatures

as low as 2 K, we found no evidence of the blocking temperature. The inverse susceptibility, H /M, is also plotted and indicates the

Curie–Weiss-like behaviour of the compound. (b) Magnetization (emu g−1_{) versus H / T for datasets collected at 5, 152 and 300 K exhibiting}

a characteristic paramagnetic behaviour. The inset presents a magnification of the graph near the origin.

Table 1. Relaxivity parameters (r1, r2) in s−1mM−1for PEG–US-Gd2O3, DEG–US-Gd2O3, PEG–SPGO, Gd–DTPA and Gd ions (GdCl3)

measured in deionized water at 1.5 T and 19.5 ± 1.5◦

C. Crystal core size (nm) Dialysis time (days) r1 (mM−1_s−1₎ r2 (mM−1_s−1₎ r2/r1 (mM−1_s−1₎ PEG-Gd2O3 <3 4 9.4 13.4 1.4 DEG-Gd2O3 <3 4 1.6 3.2 2.1 DEG-Gd2O3 <3 6 6.4 15.2 2.4 PEG-SPGO <40 2 0.1 7.6 81.6 Gd–DTPA — — 4.1 4.7 1.1 Gd ions (GdCl3) — — 10.5 12.4 1.2

previously reportedM/H graphs obtained with paramagnetic Gd2O3 powders [31] and the inverse susceptibility (H/M) graph shows the Curie–Weiss-like behaviour of the compound (with a small deviation below∼10 K). At 5 K the saturation magnetization was of 70–75 emu g−1 _citric acid-capped-US-Gd2O3 while at 300 K and 15 kOe the compound reaches a magnetization of 0.75 emu g−1_{. No evident} split of the ZFC and FC curves was found, that could have indicated superparamagnetism. However, magnetization curves (figures 4(b) and (c)) show a lack of hysteresis, and data of different temperatures superpose onto a universal curve of M versus H/ T [32]; no such superposition was obtained with SPGO nanopowders of size ∼20– 40 nm (supporting information, figure B (available at

stacks.iop.org/Nano/18/395501)).

4. Discussion

Lowr2/r1(super)paramagnetic nanoparticles may offer higher signal-to-noise ratios and better anatomic resolutions in T1 -weighted images than conventional nanoparticulate agents. This could be achieved by developing small-sized particles containing a large number of paramagnetic atoms such as the lanthanide element gadolinium. The first attempts to use Gd2O3 crystals for MRI purposes resulted in the synthesis of micrometric-sized particles and for which concentrations

needed for effective image enhancement were too high to prevent toxicity issues related to liver, spleen and bone accumulation [33]. Later, McDonald et al suggested the use of small particulate Gd2O3 (SPGO, <40 nm diameter) as a multimodal imaging tool [26]. The first results with SPGO indicated rather low longitudinal relaxivities, but improvements in the synthesis method (by additions of dextran) lead to more stable particle suspensions. However

r2/r1 ratios were still in the 3–4 range [34], limiting applications of the contrast agent inT1-weighted images. The first studies with US-Gd2O3obtained from colloidal synthesis in DEG revealed signal enhancement properties at low Gd concentrations which could allow advanced applications in positively contrasted MRI procedures [23]. Synthesis of ultra-fine Gd2O3nanoparticles inside single-wall carbon nanotubes could also be potentially useful for producing positive MR contrast agents [35].

In the present study, we demonstrated that ultra-small Gd2O3 nanoparticles synthesized by a colloidal method can provide nanoparticle suspensions of very fine and narrow particle size distributions. Our results indicate that US-Gd2O3 consists of a single-crystal core having a mean diameter of 3 nm, corresponding to a total number of Gd atoms of the order of∼200. This represents one of the highest densities achieved for a Gd-containing contrast agent (paramagnetic ions nm−3_{). A recent quantum chemical computational} study on Gd12O18 clusters [36] suggested that US-Gd2O3 6

could exhibit superparamagnetic behaviour similar to that observed with monodisperse iron oxide nanoparticles [37]. We performed experimental PPMS studies with US-Gd2O3 in order to find significant deviations from the behaviour of typical paramagnetic materials, which could possibly be correlated with the enhanced proton relaxivities observed in

T1andT2measurements. As briefly mentioned in the results section, US-Gd2O3 nanoparticles do not entirely fulfil the superparamagnetism criteria [20, 38], as the ZFC–FC graph (figure 4(a)) of a superparamagnetic nanoparticulate system should also provide the signature of the blocking temperature,

TB, as a clear split between both ZFC and FC curves. The existence of a blocking temperature could not be shown on a temperature scale down to 2 K. The Curie–Weiss fit to inverse susceptibility versus temperature gives a Curie temperature close to 0, confirming in this way the paramagnetic behaviour of our particles [29]. Moreover, the magnetization versus

H/ T dependence in figure4(b) shows a Brillouin-like shape consistent with the presence of Gd3+ _{ionic ground states} as reported in other studies performed with Gd-containing nanocrystals [39]. Surface interactions are very important with nanoparticles at that scale, since the coordination number of a large fraction of atoms is smaller than within the crystal bulk: in the case of ultra-small nanoparticles made of ferromagnetic materials such as iron, nickel and cobalt, the surface magnetic moments are enhanced by 10–30% over their bulk values [20]. Therefore, nanomagnetism surface effects with US-Gd2O3 should also be studied and correlated to the enhanced proton relaxivities obtained with derivative compounds in hydrogen proton-rich solutions.

We treated DEG–Gd2O3nanoparticles with a stabilizing and biocompatible agent, PEG-silane. PEG can allow relatively strong binding of water molecules along the polymer chains. Pegylated surfaces are biocompatible, not immunogenic nor antigenic [40]. We assessed the binding of silane groups to the oxide surface by using XPS. In order to optimize signal detection, nanoparticles must be uniformly dispersed on clean substrates; they should present a high total substrate surface coverage while not being contaminated with excess polymer. In our studies, we found evidence of silane groups on PEG-silane–US-Gd2O3 nanoparticles (supporting information, figure C (available at

stacks.iop.org/Nano/18/395501)). The binding energy value (about 103 eV) of the Si 2p peak can be compared to that of metal oxide surfaces treated with organosilanes [41] and corresponds to a silane functionality bound to an oxide surface. Our study revealed that PEG-coated US-Gd2O3nanoparticles provided higher proton relaxivities compared with DEG– Gd2O3. Although polyethylene glycol chains are hydrophilic and prevent aggregation of the nanoparticles, one could argue that capping at the Gd2O3interface with PEG through silane coupling could severely impede the water–Gd binding in the coordination sphere, which is a key factor of T1-reducing paramagnetic contrast agents [42]. Not onlyr1 parameters obtained with PEG–US-Gd2O3 were clearly higher than with DEG–Gd2O3(with lowerr2/r1ratios), but macroscopic results from simulated curves suggested that PEG–US-Gd2O3acts on the signal intensity in a very similar manner in vitro as Gd3+ paramagnetic ions. Higher signal intensities than Gd–DTPA were obtained. The nanoparticulate contrast agent was diluted

and imaged on a large Gd concentration range (0.1–1.6 mmol). Even at high concentrations (1.0–1.5 mmol), PEG–US-Gd2O3 did not exhibit a signal intensity decrease characteristic of iron oxideT1 agents, and no evidence was found of susceptibility artefacts in the MR images. A PEG–US-Gd2O3unit containing about 200 Gd atoms would therefore provide a much higher contrast effect than 200 units of Gd–DTPA, with the advantage of being highly localized on the molecule, or the cell to be tracked. US-Gd2O3 cell-internalization studies, related contrast enhancement properties and cell toxicity assessments will be published in a dedicated paper.

Relaxivity measurements performed with PEG-silane– SPGO (<40 nm) revealed lowr1and rather highr2/r1ratios. As previously reported by McDonald and Watkin, it can be difficult to produce contrast agents of sufficiently fine and narrow particle size distributions by using 20–40 nm Gd2O3 nanopowders [26, 34]. In the present study, despite careful elimination of larger size aggregates from PEG– SPGO suspensions prior to DLS and MRI measurement, the broader DLS particle size distributions and possible particle aggregation present in this compound seem to strongly affect theT1-reducing capacity of the contrast agent.

Dynamic light scattering (DLS) is a faster and comple-mentary technique to TEM to evaluate the mean size (hydrody-namic radius) of macromolecules and nanosized particles in a fluid. However, particle size assessment by laser methods can be influenced by the presence of miscible liquids in the fluid suspension. In fact, diethylene glycol has a high affinity for the gadolinium oxide surfaces which tends to create a diffuse liquid overlayer at the water/Gd2O3interface (schematic repre-sentation provided as supporting information, figure A (avail-able atstacks.iop.org/Nano/18/395501)). In non-dialyzed con-ditions, we faced difficulties to measure DEG–Gd2O3. Also, the presence at the interface of two liquid phases of different freezing points could be the reason for the strong aggregation as gel-like sediments in DEG-capped Gd2O3systems after stor-age below the freezing point. The hydrodynamic radius and proton relaxivities of PEG–US-Gd2O3suspensions remained unaffected after treatments below 0◦_{C. Therefore, our study} points to the importance of an adequate dialysis procedure to eliminate both excess diethylene glycol molecules and residual Gd3+_{ions in as-synthesized DEG–Gd}

2O3nanoparticle suspen-sions. As suggested by the DLS results, the PEG-silane treat-ment may enhance the elimination of DEG (as indicated by clear DLS fits) while preventing particle aggregation.

5. Conclusion

The present results demonstrate that PEG–US-Gd2O3 exerts contrast-enhancement properties that could open new diag-nostic possibilities for positively contrasted targeted molecu-lar imaging and cell tracking MR applications with common 1.5 T clinical systems. PEG-silane-Gd2O3exhibited enhanced proton relaxivities and higher signal intensities compared to DEG–Gd2O3 and Gd chelates, while preventing particle ag-gregation. The contrast agent provided a high signal on a large Gd concentration range, with more than doubled relaxivities compared to chelated Gd. The very small size and optimal Gd

Nanotechnology 18 (2007) 395501 M-A Fortin et al

packing density allows the labelling of matching size target-ing molecules (∼10 nm), which would represent a strong as-set compared to larger-sized Gd-containing paramagnetic con-trast agents. The narrow particle size distribution of ultra-small Gd2O3nanocrystals was confirmed and PPMS measurements revealed the paramagnetic character of the nanocrystal cores. In future in vivo studies, heterobifunctional PEG chains could be used to offer binding sites for targeting molecules.

Acknowledgments

This work was supported by the Wenner–Gren Foundation of Sweden. F Normandin, L Savitchi, M Ahr´en, S Jonsson and D Ivansson are acknowledged for their valuable contribution to this work.

References

[1] Weissleder R, Elizondo G, Wittenberg J, Lee A S, Josephson L and Brady T J 1990 Ultrasmall

superparamagnetic iron oxide: an intravenous contrast agent for assessing lymph nodes with MR imaging Radiology 175 494–8

[2] Weissleder R, Elizondo G, Wittenberg J, Rabito C A, Bengele H H and Josephson L 1990 Ultrasmall superparamagnetic iron oxide: characterization of a new class of contrast agents for MR imaging Radiology 175 489–93

[3] Kroft L J and de Roos A 1999 Blood pool contrast agents for cardiovascular MR imaging J. Magn. Reson. Imaging

10395–403

[4] Koenig S H and Kellar K E 1996 Theory of proton relaxation in solutions of magnetic nanoparticles, including the

superparamagnetic size range Acad. Radiol.3S273–6

[5] Kellar K E, Fujii D K, Gunther W H, Briley-Saebo K, Spiller M and Koenig S H 1999 ‘NC100150’ a preparation of iron oxide nanoparticles ideal for positive-contrast MR angiography MAGMA8207–13

[6] Chambon C, Clement O, Le Blanche A,

Schouman-Claeys E and Frija G 1993 Superparamagnetic iron oxides as positive MR contrast agents: in vitro and in

vivoevidence Magn. Reson. Imaging11509–19

[7] Canet E, Revel D, Forrat R, Baldy-Porcher C, de Lorgeril M, Sebbag L, Vallee J P, Didier D and Amiel M 1993 Superparamagnetic iron oxide particles and positive enhancement for myocardial perfusion studies assessed by subsecond T1-weighted MRI Magn. Reson. Imaging 111139–45

[8] Simon G H, Bauer J, Saborovski O, Fu Y, Corot C, Wendland M F and Daldrup-Link H E 2006 T1and T2

relaxivity of intracellular and extracellular USPIO at 1.5 T and 3 T clinical MR scanning Eur. Radiol.16738–45

[9] Baio G, Fabbi M, de Totero D, Ferrini S, Cilli M,

Derchi L E and Neumaier C E 2006 Magnetic resonance imaging at 1.5 T with immunospecific contrast agent in vitro and in vivo in a xenotransplant model MAGMA19313–20

[10] In den Kleef J J and Cuppen J J 1987 RLSQ: T1, T2, and rho

calculations, combining ratios and least squares Magn.

Reson. Med.5513–24

[11] Turetschek K, Roberts T P, Floyd E, Preda A, Novikov V, Shames D M, Carter W O and Brasch R C 2001 Tumor microvascular characterization using ultrasmall superparamagnetic iron oxide particles (USPIO) in an experimental breast cancer model J. Magn. Reson. Imaging

13882–8

[12] Muller R N, Roch A, Colet J M, Ouakssim A and Gollis P 2001 Particulate magnetic contrast agents The Chemistry of

Contrast Agents in Medical Magnetic Resonance Imaginged A Merbach and E Toth (West Sussex, UK: Wiley) pp 417–35

[13] Bjornerud A, Johansson L O, Briley-Saebo K and

Ahlstrom H K 2002 Assessment of T1and T2∗effects in vivo

and ex vivo using iron oxide nanoparticles in steady state; dependence on blood volume and water exchange Magn.

Reson. Med.47461–71

[14] Lawaczeck R, Menzel M and Pietsch H 2004

Superparamagnetic iron oxide particles: contrast media for magnetic resonance imaging Appl. Organomet. Chem.

18506–13

[15] Cunningham C H, Arai T, Yang P C, McConnell M V, Pauly J M and Conolly S M 2005 Positive contrast magnetic resonance imaging of cells labeled with magnetic

nanoparticles Magn. Reson. Med.53999–1005

[16] Panting J R, Taylor A M, Gatehouse P D, Keegan J, Yang G Z, McGill S, Francis J M, Burman E D, Firmin D N and Pennell D J 1999 First-pass myocardial perfusion imaging and equilibrium signal changes using the intravascular contrast agent NC100150 injection J. Magn. Reson. Imaging

10404–10

[17] de Lussanet Q G, Backes W H, Griffioen A W, van Engelshoven J M and Beets-Tan R G 2003 Gadopentetate dimeglumine versus ultrasmall superparamagnetic iron oxide for dynamic

contrast-enhanced MR imaging of tumor angiogenesis in human colon carcinoma in mice Radiology229429–38

[18] Wikstrom L J, Johansson L O, Ericsson B A, Borseth A, Akeson P A and Ahlstrom K H 1999 Abdominal vessel enhancement with an ultrasmall, superparamagnetic iron oxide blood pool agent: evaluation of dose and echo time dependence at different field strengths Acad. Radiol.

6292–8

[19] Harika L, Weissleder R, Poss K and Papisov M I 1996 Macromolecular intravenous contrast agent for MR lymphography: characterization and efficacy studies

Radiology 198365–70

[20] Sorensen C M 2001 Magnetism Nanoscale Materials

in Chemistryed K J Klabunde (Heidelberg: Wiley) pp 169–221

[21] Bazzi R, Flores-Gonzalez M A, Louis C, Lebbou K, Dujardin C, Brenier A, Zhang W, Tillement O,

Bernstein E and Perriat P 2003 Synthesis and luminescent properties of sub-5 nm lanthanide oxides nanoparticles

J. Lumin.102445–50

[22] S¨oderlind F, Pedersen H, Petoral R M, K¨all P O and Uvdal K 2005 Synthesis and characterisation of Gd2O3

nanocrystals functionalised by organic acids J. Colloid

Interface Sci.288140–8

[23] Engstr¨om M, Klasson A, Pedersen H, Vahlberg C, K¨all P O and Uvdal K 2006 High proton relaxivity for gadolinium oxide nanoparticles MAGMA19180–6

[24] Golander C G, Herron J N, Lim K, Claesson P and Andrade J D 1992 Polyethylene Glycol Chemistry, Biotechnical and

Biomedical Applicationsed J M Harris (New York: Plenum) pp 221–45

[25] Zhang Y, Kohler N and Zhang M 2002 Surface modification of superparamagnetic magnetite nanoparticles and their intracellular uptake Biomaterials231553–61

[26] McDonald M A and Watkin K L 2003 Small particulate gadolinium oxide and gadolinium oxide albumin

microspheres as multimodal contrast and therapeutic agents

Invest. Radiol.38305–10

[27] Denkleef J J E I and Cuppen J J M 1987 Rlsq-T1, T2, and rho-calculations, combining ratios and least-squares Magn.

Reson. Med.5513–24

[28] Louis U, Bazzi R, Marquette C A, Bridot J L, Roux S, Ledoux G, Mercier B, Blum L, Perriat P and

Tillement O 2005 Nanosized hybrid particles with double luminescence for biological labeling Chem. Mater.

171673–82

[29] Moon R M and Koehler W C 1975 Magnetic properties of Gd2O3Phys. Rev.B111609–22

[30] Hendrick R E and Haacke E M 1993 Basic physics of MR contrast agents and maximization of image contrast J. Magn.

Reson. Imaging3137–48

[31] Bud’ko S L and Canfield P C 2006 Evaluation of a long-time temperature drift in a commercial quantum design MPMS SQUID magnetometer using Gd2O3as a standard J. Magn. Magn. Mater.299281–7

[32] Bean C P and Jacob I S 1956 Magnetic granulometry and super-paramagnetism J. Appl. Phys.271448–52

[33] Burnett K R, Wolf G L, Shumacher H R Jr and Goldstein E J 1985 Gadolinium oxide: a prototype agent for contrast enhanced imaging of the liver and spleen with magnetic resonance Magn. Reson. Imaging365–71

[34] McDonald M A and Watkin K L 2006 Investigations into the physicochemical properties of dextran small particulate gadolinium oxide nanoparticles Acad. Radiol.13421–7

[35] Miyawaki J, Yudasaka M, Imai H, Yorimitsu H, Isobe H, Nakamura E and Iijima S 2006 Synthesis of ultrafine Gd2O3

nanoparticles inside single-wall carbon nanohorns J. Phys.

Chem.B1105179–81

[36] Pedersen H and Ojama¨e L 2006 Towards biocompatibility of RE2O3nanocrystals-water and organic molecules

chemisorbed on Gd2O3and Y2O3nanocrystals studied by

quantum-chemical computations Nano Lett.62004–8

[37] Shen T, Weissleder R, Papisov M, Bogdanov A Jr and Brady T J 1993 Monocrystalline iron oxide nanocompounds (MION): physicochemical properties Magn. Reson. Med.

29599–604

[38] Leslie-Pelecky D L and Rieke R D 1996 Magnetic properties of nanostructured materials Chem. Mater.81770–83

[39] Majetich S A, Artman J O, Mchenry M E, Nuhfer N T and Staley S W 1993 Preparation and properties of carbon-coated magnetic nanocrystallites Phys. Rev. B4816845–8

[40] Moghimi S M, Hunter A C and Murray J C 2001

Long-circulating and target-specific nanoparticles: theory to practice Pharmacol. Rev. 53 283–318

[41] Petoral R M, Yazdi G R, Spetz A L, Yakimova R and Uvdal K 2007 Organosilane-functionalized wide band gap semiconductor surfaces Appl. Phys. Lett.90(22)

[42] Caravan P, Ellison J J, McMurry T J and Lauffer R B 1999 Gadolinium(III) chelates as MRI contrast agents: structure, dynamics, and applications Chem. Rev.992293–352