ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11126

October 07, 2014

C 2014 American Chemical Society

Multifunctional Rare-Earth VanadateNanoparticles: Luminescent Labels,Oxidant Sensors, and MRI ContrastAgentsMouna Abdesselem,†,^ Markus Schoeffel,†,^ Isabelle Maurin,‡ Rivo Ramodiharilafy,† Gwennhael Autret,§

Olivier Clement,§ Pierre-Louis Tharaux,§ Jean-Pierre Boilot,‡ Thierry Gacoin,‡ Cedric Bouzigues,† and

Antigoni Alexandrou*,†

†Laboratoire d'Optique et Biosciences, Ecole Polytechnique, CNRS UMR 7645-INSERM U696, 91128 Palaiseau Cedex, France, ‡Laboratoire de Physique de la MatièreCondensée, Ecole Polytechnique, CNRS UMR 7643, 91128 Palaiseau Cedex, France, and §Paris Centre de Recherche Cardiovasculaire (PARCC), INSERM U970, 75015Paris, France. ^These authors contributed equally.

The design and tailoring of nanosizedmaterials opened numerous perspec-tives in the biomedical field. In par-

ticular, their use as drug carriers enjoysincreasing popularity because of their bet-ter pharmacokinetic properties comparedto small-molecule-based drugs.1 Moreover,nanomaterials allow for (i) multiple functio-nalizations and (ii) combined implementa-tions of optical and magnetic propertieswhich led to the development of novelcombined imaging and therapy approachesincluding targeted drug delivery and med-ical imaging features. Such multimodal sys-tems showed promising results for cancertreatment, and some of them are under-going clinical studies (phase 1/2).2

Collecting information on multiple sour-ces is essential in understanding specificpathophysiological processes in a livingorgan-ism in their full complexity and consequentlyin enabling more accurate diagnosis and de-sign of novel therapeutic approaches. Giventhe interindividual variability of pathophysio-logical parameters, detecting multiple signalsin a single experiment is crucial. Additionalbenefits of multimodal probes arise from lesstoxicity as a lower amount of xenobiotics hasto be injected and from cost reduction.2 Thesefeatures have led to widespread developmentof bi- or multifunctional biomedical imagingprobes, especially those suitable for the mostcommon modalities, such as fluorescence3,4

and magnetic resonance imaging (MRI).5�9

* Address correspondence toantigoni.alexandrou@polytechique.edu.

Received for review January 8, 2014and accepted October 7, 2014.

Published online10.1021/nn504170x

ABSTRACT Collecting information on multiple pathophysiological parameters is

essential for understanding complex pathologies, especially given the large inter-

individual variability. We report here multifunctional nanoparticles which are

luminescent probes, oxidant sensors, and contrast agents in magnetic resonance

imaging (MRI). Eu3þ ions in an yttrium vanadate matrix have been demonstrated to

emit strong, nonblinking, and stable luminescence. Time- and space-resolved optical

oxidant detection is feasible after reversible photoreduction of Eu3þ to Eu2þ and

reoxidation by oxidants, such as H2O2, leading to a modulation of the luminescence

emission. The incorporation of paramagnetic Gd3þ confers in addition proton relaxation enhancing properties to the system. We synthesized and characterized

nanoparticles of either 5 or 30 nm diameter with compositions of GdVO4 and Gd0.6Eu0.4VO4. These particles retain the luminescence and oxidant detection

properties of YVO4:Eu. Moreover, the proton relaxivity of GdVO4 and Gd0.6Eu0.4VO4 nanoparticles of 5 nm diameter is higher than that of the commercial Gd3þ

chelate compound Dotarem at 20 MHz. Nuclear magnetic resonance dispersion spectroscopy showed a relaxivity increase above 10 MHz. Complexometric

titration indicated that rare-earth leaching is negligible. The 5 nm nanoparticles injected in mice were observed with MRI to concentrate in the liver and the

bladder after 30 min. Thus, these multifunctional rare-earth vanadate nanoparticles pave the way for simultaneous optical and magnetic resonance detection,

in particular, for in vivo localization evolution and reactive oxygen species detection in a broad range of physiological and pathophysiological conditions.

KEYWORDS: lanthanide nanoparticles . multifunctional probe . oxidant sensor . contrast agent . MRI

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11127

Aiming at the combination of the latter two mod-alities, a first approach consisted of covalently graft-ing a fluorescent moiety to a gadolinium chelate,10,11

Gd3þ ions being the most commonly used ions forT1-based MRI contrast enhancement. Organic near-in-frared fluorophores were also linked to superparamag-netic particles such as iron oxide nanoparticles (NPs),12

which are routinely used as T2-based MRI contrastagents.6 Quantum dots coated with paramagneticlipids (PEGylated Gd-DTPA) and polyethylene glycol(PEG) were successfully used as dual-modalityagents.13 Furthermore, Gd-based nanosized MRI con-trast enhancers were functionalized with coatingscontaining organic fluorophores.14 Recent develop-ments in this field led to the synthesis of all-in-onefluorescent and paramagnetic gadolinium-basednanoparticles.15�17 Stable luminescence propertiesoriginate from doping with rare-earth ions which pro-vide a photostable, nonblinking, narrow emission.18

They may also exhibit up-conversion properties forexcitation in the near-infrared range.17,19 Besides sim-ply detecting the nanoparticles through their lightemission, we have shown that Eu-doped nanoparticlescan be used as oxidant sensors.20 Reactive oxygenspecies (ROS) andmore specifically hydrogen peroxide(H2O2) is produced as a secondary cellular messengerinvolved in a variety of signaling pathways21,22 includ-ing those resulting in cell contraction, migration, andcell injury and healing.23�25 Thus, H2O2 which was for along time considered to be a lethal cellular waste isnow attracting increasing interest. The quest for appro-priate sensors, however, has been restricted by therequirement to provide simultaneously quantitative,time- and space-resolved, intracellular detection.26�29

In our previous work, we demonstrated that Eu-dopedYVO4 nanoparticles uniquely combine all these proper-ties and can quantitatively probe local intracellularconcentrations of H2O2 and their evolution in time.20,30

We have furthermore shown that these particles areideal for single-molecule labeling and tracking andused them for the determination of confinementpotentials in membrane lipid rafts.31,32

Moreover, detecting reactive oxygen species con-centration levels in living organisms is particularlyimportant because ROS signaling and oxidative stressplay a complex role in multiple pathologies. For in-stance, oxidative stress is an essential element ininflammatory processes25,33 and in neurodegenerativediseases, for example, in the degeneration of dopami-nergic neurons in Parkinson's disease.34 ROS furtherregulate the formation and the evolution of tumors:35

cancer cells are indeed known to exhibit high levels ofoxidative stress,36 and ROS are involved in the trigger-ing of tumoral transitions37 but also control the activityof tumor suppressors.38,39 ROS concentration evolu-tion might thus constitute a chemical signature oftumoral tissues and be an efficient diagnosis indicator.

More generally, precise determination of spatiotem-poral patterns of ROS production is likely to providesignificant improvement in the understanding of howROS-related pathologies are triggered. Therefore, amultifunctional imaging probe that combines ROSdetection with complementary modalities is highlydesirable. Despite this, only few probes have beenimplemented for in vivo detection,29,40�42 and to ourknowledge, no multifunctional probes including ROSdetection capability exist.Here, we propose a novel nanosized material which

combines magnetic resonance contrast enhancingproperties with luminescence properties and hydro-gen peroxide sensing features. We designed a versa-tile luminescent MRI contrast enhancer based onGd0.6Eu0.4VO4 nanoparticles of 5 or 30 nm size. Wedemonstrated that these Eu-doped GdVO4 particlesretain the luminescence and hydrogen sensing proper-ties of Eu-doped YVO4

20 in the 1�100 μM rangerelevant for biological signaling. This novel para-magnetic system shows high relaxivities (r1 =8.18 mM�1 s�1, r2 = 9.38 mM�1 s�1) and extremelylow Gd3þ leaching. The relaxation enhancement in-duced by these nanoparticles is easily detected withmagnetic resonance imaging in mice.

RESULTS AND DISCUSSION

Particle Size, Shape, and Microstructure. GdVO4 and Eu-doped GdVO4 nanoparticles of about 5 and 30 nmwere synthesized using protocols already optimizedfor Y1�xEuxVO4 particles of similar size. The 30 nmparticles were obtained by a standard coprecipitationroute in water,43 whereas the addition of citrate ionsduring the synthesis leads to the formation of smaller5 nm particles.44 The size obtained from dynamic lightscattering measurements and the zeta-potential forboth types of nanoparticles are reported in the Sup-porting Information Table S1. We analyzed the trans-mission electron microscopy (TEM) micrographs ofover 400 GdVO4 nanoparticles prepared by the stan-dard route and found that their two-dimensionalprojection is an ellipse with the two major axes havinglengths of 13.1 ( 1.1 and 26.6 ( 4.8 nm (mean (standard deviation), respectively (Figure 1A�C). High-resolution TEM images (Figure 1B) show local fluctua-tions of the contrast within a particle that may be dueto particle porosity, to a nonsmooth surface texture,and/or to the presence of an internal substructure.However, even if several crystallites are present in eachnanoparticle, they apparently retain a preferentialcrystallographic orientation as lattice fringes werefound to extend over the whole particle.

This latter conclusion was confirmed by the analysisof X-ray diffraction (XRD) data, which give access to thestructural coherence length or crystallite size(Supporting Information Figure S1). Complementedby Rietveld analysis, themicrostructural study suggests

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11128

that the nanoparticles can be described as compress-ed prolate spheroids with 16 nm thickness and 25 nmlength (Supporting Information Figure S1), in verygood agreement with the dimensions found by TEM.This speaks either for an absence of internal substruc-ture or for the presence of a 3D ordered polygranularsubstructure.

The high-resolution TEM images indicate a highspecific surface, similarly to what was reported forY0.6Eu0.4VO4 samples synthesized with the same pro-tocol and which show a similar microstructure.19 Inaddition, gas adsorption measurements on these Eu-doped YVO4 nanoparticles analyzed with the Bru-nauer�Emmett�Teller theory showed that their spe-cific surface was almost 20 times higher than the oneexpected from the geometric size of the particles.19

The high specific surface suggested by the high-reso-lution TEM images is probably related to the low(ambient) temperature of the synthesis and may playa key role in the sensing and relaxation properties ofGd0.6Eu0.4VO4 by providing solvent access to a greaternumber of Eu3þ and Gd3þ ions.

The structural and microstructural characterizationof the citrate-route nanoparticles is discussed in theSupporting Information. Briefly, TEM measurementsyield a particle size of 5 ( 0.2 nm, and the X-raydiffraction analysis yields coherence length valuesranging between 3 and 6 nm depending on the Braggreflections.

After silica or dextran coating, the nanoparticleswere characterized with Fourier transform infrared(FT-IR) spectroscopy measurements (Supporting Infor-mation Figure S3), which confirmed the efficiency ofthe coating procedures by detecting specific absorp-tion peaks of Si�O�Si or C�H, C�O, and C�C bonds.

Luminescence and Hydrogen Peroxide Detection. The lu-minescence excitation and emission spectra of a sus-pension of Gd0.6Eu0.4VO4 nanoparticles are shown inFigure 2A,B. Gd0.6Eu0.4VO4 absorbs strongly in the UVrange, and an efficient energy transfer is possible fromVO4

3� ions to Eu3þ ions.45 The emission spectrum issimilar to that measured for Y0.6Eu0.4VO4 nano-particles43 and corresponds to that found for Eu-doped

GdVO4 in the literature45,46 with a peak at 593 nm dueto the 5D0�7F1 transition, the strong principal doubleemission peak at 616 nm (5D0�7F2), a weak peak at650 nm (5D0�7F3), and another double peak at 699 nm(5D0�7F4). A quantum yield of 3.8% was determinedusing rhodamine as a reference. Similar emission spec-tra are observed for direct excitation of the Eu3þ ions at396 and 466 nm (Supporting Information Figure S4).We measured similar excitation and emission spectrafor the citrate-route nanoparticles (Supporting Infor-mation Figure S5) and a quantum yield of 10%.

Solutions of 30 nm Gd0.6Eu0.4VO4 nanoparticleswere spin-coated on a glass coverslip after dilution toensure single-particle imaging (Figure 2B) and illu-minated at 466 nm at relatively high laser intensity(1.6 kW/cm2). We observed a decrease of the lumines-cence after 250 s due to photoreduction of Eu3þ toEu2þ similar to the one detected in Eu-doped YVO4.

20

The luminescence decrease is due to the fact that wedetect the emission of Eu3þ ions with a filter centeredat 617 nm, and the emission of Eu2þ is shifted to lowerwavelengths.20 Moreover, different single nanoparti-cles show a similar photoreduction and reoxidationbehavior (Supporting Information Figure S6). The lumi-nescence evolution averaged for 20 nanoparticlesshows a biexponential decay during photoreduc-tion (T1 = 3.6 ( 0.1 s and T2 = 49.7 ( 2.5 s) and aluminescence signal decrease of 30% (Figure 2C). Thecharacteristic decay times and the signal decrease arecomparable to those observed for Y0.6Eu0.4VO4

nanoparticles20 under a similar illumination. Uponaddition of H2O2, we observe a luminescence signalrecovery due to chemical reoxidation of Eu2þ ions backto their initial Eu3þ state. Additions of H2O2 concentra-tions as small as 1 μM can be detected (Figure 2D). Thefinal luminescence value after a photoreduction/chemical reoxidation cycle is a monotonous functionof the H2O2 concentration (Figure 2D). Gd0.6Eu0.4VO4

nanoparticles are thus efficient, quantitative oxidantlocal sensors in the physiological signaling range of1�100 μM H2O2. Note also that the photoreduction/reoxidation process is reversible rendering time-resolved detection possible (Supporting Information

Figure 1. (A) TEM image of normal-route GdVO4 nanoparticles, scale bar: 50 nm. (B) High-magnification TEM image, scale bar:10 nm. (C) Width and length distributions obtained from 420 GdVO4 particles. The solid lines are log-normal fits of the sizedistributions.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11129

Figure S7). Moreover, the response to hydrogen per-oxide in physiological medium reveals that it is af-fected neither by the presence of proteins nor by thatof buffering ions (Figure S8). These properties aresimilar to those of Y0.6Eu0.4VO4 particles

20,21 and openthe possibility of using Gd3þ-based nanoparticles forspace- and time-resolved oxidant detection.

The oxidant sensing properties of citrate-routeGd0.6Eu0.4VO4 nanoparticles are discussed in the Sup-porting Information (Figure S9). We found similar res-ponses to hydrogen peroxide addition with a typicalluminescence recovery of 8% after photoreductionand subsequent addition of 50 μM H2O2. However,we observed a higher interparticle variability of theH2O2-induced luminescence recovery (8% dispersioncompared to 5% for 30 nm nanoparticles) and a slowerrecovery dynamics. These effects may be explained bythe smaller size rendering the role of the surface moreimportant or by differences of the surface propertiesdue to the citrate complexes or of the crystallinity. Thisindicates that normal-route nanoparticles are a betterchoice for local hydrogen peroxide sensing at the celllevel than citrate-route ones.

Relaxation Times at 20 MHz. Magnetization measure-ments carried out for both GdVO4 and Eu-dopedGdVO4

nanoparticles indicate a paramagnetic behavior at roomtemperature. The main difference between bulk and

nanoparticle samples is a slight decrease of the ex-change energy at low temperatures and a decrease ofthe paramagnetic to antiferromagnetic transition tem-perature (Supporting Information Figure S10). We de-termined the proton relaxation times of solutions ofthese rare-earth vanadate nanoparticles and found alinear dependence of the inverse relaxation times withtheGd3þ ion concentration contained in the respectiveparticles for all examined samples. Data are shown forcitrate-route Gd0.6Eu0.4VO4 nanoparticles (Figure 3A),and the measured relaxivities for all studied samplesare listed in Table 1. We paid particular attention to aprecise measurement of the Gd3þ ion concentration(see Supporting Information) because it determinesthe precision of the relaxivity values. All samples showrelaxivity ratios r2/r1 similar to those found for Dotaremand free Gd3þ ions of about 1.2 and are thus positivecontrast agents.

Silica-coated nanoparticles compared to pristineones show similar r1

ion and r2ion relaxivity values

(Table 1). We can thus conclude that the silica coatingdoes not impede contact of water protons with thegadolinium ions, which is consistent with the fact thatthe silicate layer has been shown to be partially con-densed and permeable to small molecules.20,47 Thisopens theway for further functionalization of the silica-coated particles which is required for active targeting.

Figure 2. Luminescence and hydrogen peroxide detection properties of normal-route Gd0.6Eu0.4VO4 nanoparticles. (A)Luminescence excitation spectrum (detection at 615 nm). (B) Emission spectrum (excitation at 280 nm). The peak positions aswell as the corresponding transitions are indicated. (C) Gd0.6Eu0.4VO4NPsdetected individually on aglass coverslip. Excitationat 466 nm, detection at 617 nm. Laser intensity: 1.6 kW/cm2; acquisition time: 1 s. (D) Luminescence intensity evolution of asingle nanoparticle. Left panel: Photoreduction under strong illumination (1.6 kW/cm2, excitation at 466 nm, detection at617 nm) during t = 250 s. Acquisition time: 1 s. Right panel: Excitation intensity was lowered to 0.3 kW/cm2, and 100 μMhydrogen peroxide was added to the medium (pure water) causing luminescence recovery. Acquisition time: 3 s. Theluminescence decrease and recovery are fitted with a biexponential and a monoexponential, respectively (solid black lines).(E) Luminescence recovery averaged for∼20 individually detected nanoparticles for different hydrogen peroxide concentra-tions. Particles were photoreduced prior to H2O2 addition as shown in C. Laser intensity: 0.3 kW/cm2; excitation at 466 nm;acquisition time: 3 s.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11130

In the case of GdVO4cit and Gd0.6Eu0.4VO4

cit nano-particles, we find relaxivities higher than the relaxivitiesof the commercial compound Dotarem (Table 1). Acomparison of the samples GdVO4

cit to GdVO4 andGd0.6Eu0.4VO4

cit to Gd0.6Eu0.4VO4, shows that relaxivityvalues per Gd3þ ion are higher for the smaller particlesprepared by the citrate route. A priori, according tothe Solomon�Bloembergen equations9,48 for the pro-ton relaxation in the presence of a contrast agent, thisdependence with the particle size results from theinterplay between two quantities: (i) the number ofaccessible Gd3þ ions and (ii) the rotational correlationtime of the contrast agent.

Indeed, the proton relaxation times according tothe Solomon�Bloembergen equations9,48 strongly de-pend on τci (i = 1, 2 for the longitudinal and transversecomponent), the dipole�dipole interaction time, givenby the following equation:

1τci

¼ 1τM

þ 1Tei

þ 1τR

(i ¼ 1, 2) (1)

where τM is the mean residence time of the boundwater molecule, Tei is, respectively, the longitudinaland transverse electron-spin relaxation time, and τR isthe rotational correlation time. It is well-established forGd chelates, at 20 MHz, that relaxivity increases in thesame fashion as the rotational correlation time, that is,as the chelate size.49 This trend, however, is no longer

valid for species with τR values above 10 ns where thethird term in eq 1 becomes negligible and the electron-spin relaxation and/or residence lifetime terms be-come dominant.48,49 For spherical nanoparticles witha diameter d = 12 nm at 25 �C in water with a viscosityof η ≈ 0.9 mPa 3 s,

50 we obtain τR = 200 ns (in contrast,τR for Dotarem is only 80 ps51). Assuming that thesurface Gd3þ ions in GdVO4 particles retain the sameelectronic and water coordination properties as inGd3þ chelates, the previous discussion suggests thatτR only weakly influences the relaxivity of our samples.Thus, the higher relaxivity measured in the smaller NPsmay result from a greater number of accessible Gd3þ

sites in citrate-route particles. This number scales as thesurface to volume ratio (i.e., as 1/d for spherical NPs).This justifies quite well the difference observed be-tween Gd0.6Eu0.4VO4 and Gd0.6Eu0.4VO4

cit, where thediameter ratio of 3 (42 nm/14 nm) is consistentwith therelaxivity ratios of 8.18/2.97 = 2.75 and 9.38/3.47 = 2.70,respectively, for r1 and r2 (see Table 1). Moreover,changes in relaxivity between the two different syn-thesis routes may also result from a different degree ofporosity and/or from different surface states (includingsurface charge and solvation shell structure) and thusdifferent accessibility to Gd3þ ions. Given that small-er nanoparticles exhibit higher relaxivity values perGd3þ ion and assuming that the toxicity of the nano-particles is only determined by the total Gd3þ quantity

Figure 3. (A) Relaxation times for citrate-route Gd0.6Eu0.4VO4 nanoparticles. The solid lines are linear fits to the data. (B)Comparison of the NMRD profiles for several Gd-containing nanoparticles. The superscript “cit” stands for particlessynthesized by the citrate route.

TABLE 1. Relaxivities for Rare-Earth Vanadate Nanoparticles Measured at 20 MHz and 37 �C in Watera

composition d Ænæ (nm) sample r1ion (mM�1 s�1) r2

ion (mM�1 s�1) r1NP r2

NP r2/r1

GdVO4 41 this work 0.95 1.31 420000 570000 1.08Gd0.6Eu0.4VO4 42 this work 2.97 3.47 840000 980000 1.17Gd0.6Eu0.4VO4/SiO2 57 this work 2.52 3.03 710000 860000 1.20GdVO4

cit 11 this work 4.60 5.52 39000 47000 1.20Gd0.6Eu0.4VO4

cit 14 this work 8.18 9.38 86000 98000 1.15Dotarem CC Guerbet 3.59 4.44 1.23GdCl3 ion Prolabo 10.4 12.1 1.16

a Particle sizes are given as the number-average diameter from dynamic light scattering measurements. riion is the relaxivity relative to the molar quantity of Gd3þ ions. Given

the fact that the Eu3þ ions also contribute to the relaxivity (see below), the relaxivity values of Eu-doped nanoparticles relative to the molar quantity of both rare-earth ions(Gd3þ and Eu3þ) can be obtained from the values of the table by multiplying with 0.6. The relaxivity for an average nanoparticle is denoted by ri

NP (i = 1,2). The superscript“cit” denotes nanoparticles formed via the citrate route. CC stands for coordination complex.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11131

administered to the patient, it is thus preferable tointroducemore nanoparticles of smaller size for a givenquantity of material.

Interestingly, in Gd0.6Eu0.4VO4 and Gd0.6Eu0.4VO4cit

nanoparticles, replacing the paramagnetic Gd3þ ionsby the nonmagnetic Eu3þ ions does not diminishproportionally the relaxivity of these compounds com-pared to Gd-based nanoparticles (Table 1). We there-fore examined the contribution of Eu3þ to the relaxivitywith frequency-dependent relaxation measurements.

Frequency-Dependent Relaxation Times. Longitudinal re-laxation times were measured as a function of theLarmor frequency in the range from ω/2π = 10 kHz to15 MHz (2.3 mT to 0.35 T). The corresponding Gd3þ

concentration normalized relaxivities are plotted inFigure 3B for GdVO4 nanoparticles, doped or not withEu3þ, from the normal and citrate routes. The relaxiv-ities are nearly constant for all samples up to a fre-quency of about 5 MHz. A similar behavior was alsofound for Gd3þ chelates where the plateau rangeextends up to 1�10 MHz depending on the actualcontrast agent.9 For high frequencies beyond theplateau range, the relaxivity decreases dramaticallyfor Gd chelates, such as Dotarem, from 1 to 10 MHz,9

whereas the relaxivities for our Eu-doped nanoparticlesare constant from 0.01 to 10 MHz with an increasesetting in for frequencies above 10 MHz for all nano-particles except normal-route GdVO4 (Figure 3). Thisdifference in frequency dependence between Gd che-lates and nanoparticles can be explained by the differ-ence in rotational correlation time and corresponds tothe predictions of Lauffer et al.9,48 Indeed, Lauffer pre-dicts a relaxivity maximum between 10 and 100 MHzfor long rotational correlation times.48 These calcula-tions were done for rotational correlation times of τR =20 ns; however, as indicated above, the effect of therotational correlation time on the relaxivity behaviorbecomes negligible for τR values above 10 ns. Ascommercial MRI scanners operate at a magnetic fieldof 1.5 T or now also 3 T, corresponding respectively to aproton resonance frequency of 64 or 128 MHz, acontrast agent that shows a marked relaxation maxi-mum in this range is of great interest.

These data confirm the trends observed at a singlefrequency (20MHz, Table 1). Over the whole frequencyrange, the suspensions containing nanoparticles fromthe citrate route show higher relaxivities than thoseobtained from the normal synthesis path. Furthermore,Eu-containing nanoparticles possess a higher r1 relax-ivity than those without Eu for all measured frequen-cies. Trivalent europium ions at the ground state arenot expected to influence the magnetic properties ofthe nanoparticles as J= 0 (Eu3þ, 4f6, L= S= 3). However,the energy separation between the ground state andthe first excited state is only 255 cm�1,52 which iscomparable to the energy kT corresponding to themeasurement temperature, 37 �C. This peculiar

configuration results in the so-called Van Vleck para-magnetism52 that predicts a nonzero magnetic mo-ment of trivalent europium ions at room temperature.The effectivemagneticmoment at 300 K of Eu3þ-basedmaterials shows values of approximately 3.4 μB (Bohrmagneton) per Eu3þ ion depending on the oxide.50,51

The moment of Gd3þ-based nanoparticles is 7.9 μBper Gd3þ ion (J = S = 7/2). We can thus estimate thesquared magnetic moment of Gd0.6Eu0.4VO4 particlesby adding the weighted contributions of Eu and Gdtrivalent ions which yields an effective squared mag-netic moment of 42 μB

2 per rare-earth ion.Therefore, reporting the relaxivity per Gd3þ ion and

neglecting the contribution of Eu3þ ions to the mag-netic moment, and thus to the relaxivity, leads to anoverestimation of the r1

ion and r2ion relaxivity values for

nanoparticles containing 40% of Eu3þ ions. It is well-known that the inner-sphere contribution to the re-laxation times and thus the relaxivities depends on theeffective magnetic moment as follows: 1/Ti

DD � g2μB2

S(Sþ 1) = μeff2, where Ti

DD refers to the longitudinal andtransverse dipole�dipole relaxation times, g is theelectronic Landé factor, μB the Bohr magneton, S thespin number, and μeff the effectivemagneticmoment.9

For citrate-route nanoparticles, the relaxivity valuesbelow 1 MHz are 6.4 mM�1 s�1 for the Eu-doped NPsrelative to the molar quantity of Gd3þ ions (i.e.,3.8 mM�1 s�1) relative to the molar quantity of bothrare-earth ions (Gd3þ and Eu3þ) and 4.6 mM�1 s�1 forthe nondoped NPs. The relaxivity ratio is thus 4.6/3.8 =1.2, which is compatible with the ratio of the squaredeffectivemagneticmoments: (μGdVO4

2)/(μGd0.6Eu0.4VO4

2) =(62/42)2 = 1.5. This clearly demonstrates the contribu-tion of Eu3þ ions in the relaxation properties ofGd0.6Eu0.4VO4.

Assessment of Rare-Earth Leaching from Nanoparticles.Free rare-earth ions are cytotoxic, and the medianlethal dose (DL50) in mice is 100�200 and 550 mg/kgfor GdCl3

53,54 and EuCl3,55 respectively. We therefore

measured the kinetics of rare-earth leaching from adispersion of normal-route Gd0.6Eu0.4VO4 nanoparti-cles in ultrapure water using xylenol orange to detectthe free rare-earth concentration for an initial nano-particle concentration corresponding to 67mM in van-adate ions (Figure 4). To assess the free rare-earthconcentration, we centrifuged to separate the nano-particles and then titrated the free rare-earth ions inthe supernatant (Methods in the Supporting Infor-mation). We performed, after synthesis, five purifica-tion cycles to remove most of the unreacted rare-earthions. The decrease in rare-earth concentration aftereach washing step shows their clearance from thesolution (Figure 4). Moreover, the relatively slow de-crease suggests that not only unreacted rare-earths insolution but also adsorbed rare-earth ions at the nano-particle surface are removed. These adsorbed specieshave been observed at the surface of lanthanide

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11132

vanadates56 and could contribute to particle stability insolution. Further centrifugation steps do not lead to afurther decrease of the residual free rare-earth con-centrationwhich remains stable at approximately 2 μMfor normal-route particles (Figure 4). The detectedvalue of 2 μM represents ∼30 ppm of the total rare-earth ion content in the nanoparticles. We then per-formed another series of experiments at room tem-perature, after the initial purification cycles, wherecentrifugations were performed after a variable num-ber of days up to 10 days. After each centrifugation, therare-earth concentration in the supernatant was as-sessed (inset Figure 4); no increase of free rare-earthions in solution was observed. Similar measurementsover a period of 10 days at room temperature forcitrate-route nanoparticles (initial concentration of67 mM in vanadate ions) led to rare-earth concentra-tion values below our detection limit (�1.6 ( 2.6 μM).Note, however, that rare-earth leaching may be higherin biological environments, in particular, due to thepresence of Ca2þ ions and to transmetalation reac-tions. The potential role of this effect will be addressedin future work.

Moreover, free vanadate ions have also been shownto be carcinogenic due to interference with redoxregulation and signaling pathways through interactionwith phosphate binding sites of protein phos-phatases.57 A significant number of deaths were foundin mice for sodium orthovanadate doses higher than30 mg/kg/day.58 Given the fact that vanadate ions arereleased from the solid matrix where they are presentin the same concentration as rare-earth ions, we mayassume that they are released in equally smallconcentrations.

We thus demonstrated negligible rare-earth leach-ing from the Eu-doped GdVO4 particles for both syn-thesis routes, significantly smaller than the valuesreported in the literature. For comparison, Hifumiet al.59 reported for dextran-coated GdPO4 nanoparti-cles, with an initial concentration of 2.2 mM in Gd3þ

ions, a leached concentration of 6.5 μMafter 2weeks at4 �C (which corresponds to 3000 ppm), and Huanget al.60 found 33 and 43 μM of free gadolinium after8 days at 37 �C for hollow and porous Gd2O3 nano-particles with an initial concentration of 2.2 mM inGd3þ ions, respectively (corresponding to 15 000 ppmand 19 500 ppm, respectively). Given their good relax-ivity properties and their very low leaching, theseGd0.6Eu0.4VO4 nanoparticles are suitable for in vivo

injections and are expected to produce contrast en-hancement in MRI.

Toxicity Measurements. We evaluated the potentialcell mortality induced by incubation with dextran-coated nanoparticles at a similar concentration (200 μLat 10 mM in vanadate ions added to 2 mL of culturemedium) as the one resulting from nanoparticle injec-tion (200 μL at 10 mM in vanadate ions) in the mouseblood circulation, assuming a blood volume of 2 mL.We used endothelial progenitor cells (EPCs), which areendothelial cells likely to encounter intravenously in-jected xenobiotics. Cell mortality MTT tests (see Meth-ods in the Supporting Information) on EPCs showed nocytotoxicity of citrate-route Gd0.6Eu0.4VO4 particlesafter 3 h and overnight incubation (102 ( 5% and99 ( 3% versus 100 ( 3% for control cells). Normal-route particles induced limited cell death (13% after 3 hand 25% after an overnight incubation, SupportingInformation Figure S11). These results indicate a rea-sonable tolerance of GdVO4:Eu nanoparticles, notablyof citrate-route nanoparticles, which are more suitablefor in vivo experiments.

MRI. We injected dextran-coated citrate-routeGd0.6Eu0.4VO4 nanoparticles antero-orbitally into threeBL6 mice and monitored the MRI contrast enhance-ment for 2 h. During all experiments, all physiologicalconstants (breathing and heart rate) of injected miceremained normal, indicating the absence of harmfuleffects of Gd0.6Eu0.4VO4 nanoparticles. T1-weightedMRI images demonstrate that the Gd0.6Eu0.4VO4 colloidis an efficient contrast enhancer as significant signalchanges up to 45% were measured after injection of0.06 mmol/kg Gd0.6Eu0.4VO4 (the value indicated cor-responding to that for Gd3þ ions), an amount typical ofthat used for experiments with Dotarem (usual recom-mended dose of Dotarem is 0.1 mmol/kg). Monitoringof the bladder areas showed a signal increase duringthe first 30min after injection, which we attribute to anaccumulation of the nanoparticles in the bladder(Figure 5B,C). This means that the circulation time ofthe nanoparticles in mouse blood is at least 30min. Forcomparison, Dotarem typical circulation time is 10 minapproximately.61 Moreover, during our observationtime, the volume of the bladders increased and theirshapes tended to change from round to ellipsoidal.Considering that the third dimension of the bladder isequal to the average of the two axes of the ellipseobserved in the 2D projection, we find an average

Figure 4. Concentration in free rare-earth ions after succes-sive centrifugation steps to remove unreacted precursorsafter a normal-route synthesis for an initial rare-earth con-centration of 67 mM in vanadate ions. The solid line is aguide for the eye. Inset: Rare-earth concentration as afunction of the time separating two purification steps afterthe removal of residual unreacted species.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11133

volume increase of 80 ( 54% for the three mice in-jected with citrate-route Gd0.6Eu0.4VO4 nanoparticlesprobably due to the liquid injected.

In the liver, we observe a signal decrease (instead ofan increase) of 45%which ismost probably due to rapiduptake of large amounts of NPs (Figure 5C). For large NPconcentrations, the contrast agent accumulation causeslocal field heterogeneities which result in the so-calledT2* artifact, that is, a decrease of the transverse (T2)relaxation time of protons in this organ leading to adecrease of the signal in T1-weightedMRI images.17,62,63

This hypothesis is confirmed by luminescencemeasure-ments that show the characteristic red luminescence ofGd0.6Eu0.4VO4 in extracted liver tissue homogenates(see Supporting Information Methods and Figure S12).The signal in the kidneys remains stable over themonitoring time, which suggests that nanoparticles donot accumulate in this organ.

Note that the nanoparticle uptake in the liver isfaster than the arrival in the bladder (average charac-teristic times of 21 ( 9 min versus 37 ( 9 min). This isdue to the different uptake mechanisms involved.Nanoparticle uptake in the liver is based on vasculardiffusion followed by endocytosis and phagocytosis byKupfer cells. Filtration by the kidney, however, requiresseveral circulation cycles before the whole concentra-tion is eliminated toward the bladder. Provided theirsize is small enough, typically only 20% of the nano-particles will be filtered by the first passage throughthe kidneys.64

We also performed an injection of 0.06 mmol/kg ofnormal-route Gd0.6Eu0.4VO4 particles in the same condi-tions as above (Supporting Information Figure S13).Thirty minutes after injection, no contrast changes weremeasured in the bladder and kidneys, whereas we ob-served an important signal decrease (�55%) in the liver.

Luminescence detection in tissue homogenatesconfirms the presence of NP accumulation in the liverand the absence of nanoparticles in urine and kidneys(Supporting Information Figure S12). These experi-ments show that citrate-route nanoparticles are de-tected both in the urine and in the liver, whereasnormal-route particles are only detected in the liver.This means that the smaller particles are at least partlycleared by the kidney and the bladder, whereas biggerparticles are eliminated from the blood circulation intothe liver because of the glomeruli filtering threshold.

These results are in agreement with the data re-ported in the case of quantum dots, where efficientand rapid uptake in the mouse bladder followed byurinary excretion and elimination was found to takeplace for nanoparticles with hydrodynamic diameterssmaller than 5.5 nm.65 For larger diameters, accumula-tion in the liver and other organs was observed.

CONCLUSIONS

We have synthesized nanoparticles made of pureGdVO4 and Gd0.6Eu0.4VO4 compositions. On the basisof well-documented synthesis routes, we produced forthe two compositions nanoparticles of 5 and 30 nm

Figure 5. (A) T1-weighted MRI monitoring at different times before and after 0.06 mmol/kg injection of dextran-coatedcitrate-route Gd0.6Eu0.4VO4 nanoparticles. The three rows correspond to a selection of coronal planes of the mouse. Whitearrows show organs of interest. (B) Contrast evolution in bladder, liver, and kidney after injection of the nanoparticles for themouse shown in (A). The signal was averagedover the area of the organ of interest and normalized to the reference signal of aphantom tube containing water. (C) Signal evolution 30 min after injection averaged for three mice. Error bars are standarderror of the mean.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11134

diameter. A combination of X-ray diffraction and elec-tron microscopy experiments showed that the lattersynthesis route yields GdVO4 nanoparticles with ashape best described as compressed prolate spheroidswith about 25 nm length and 16 nm height. TEMobservations suggest that these particles have aninternal structure which allows access of the solventto Eu3þ and Gd3þ surface ions and thus favors bothoxidant sensing and MRI contrast enhancement.We have demonstrated that Gd0.6Eu0.4VO4 nanopar-

ticles retain the property of luminescence modulationby the oxidant concentration and thus allow quantita-tive, time- and space-resolved detection of intracellularH2O2. Quantitative ROS detection using these probesinside living organisms should prove to be a valuabletool for assessing critical parameters of various ROS-related pathologies.The advantage of gadolinium ions being incorpo-

rated as an integral component of the nanoparticlecrystal structure lies in a very low leaching of free ionsover a period of 10 days at room temperature. Theseparticles are more stable than other MRI contrastenhancement particles reported in the literature.We found that the relaxivity of pure or Eu-doped

5 nm GdVO4 nanoparticles is higher than that of thecommercial chelate compound Dotarem, makingthem promising candidates for biomedical applica-tions. In addition, relaxivities for nanoparticle disper-sions increase in the frequency range of interest formedical imaging while those of commercial gadoli-nium chelates decrease steeply. MRI imaging in mice

showed accumulation of the particles in the bladderand in the liver. A full biodistribution study of thenanoparticles will be performed in future work.In conclusion, rare-earth vanadate nanoparticles

containing simultaneously europium and gadoliniumions are multifunctional probes allowing detection oftheir localization using luminescence and MRI as wellas detection of ROS concentrations produced in phy-siological or pathophysiological conditions. We de-monstrated these multiple modalities for two dif-ferent synthesis methods: the so-called normal andcitrate synthesis path. Resulting nanoparticles retainedthese multicontrast abilities for biological imaging,even though they do not present the same perfor-mances for each modality. Citrate-route particles seemindeed more suitable for MRI, while ROS detection incell culture is more efficient with normal-route nano-particles. We thus have the possibility to choose be-tween these different types of particles for furtherapplications, in order to optimize multimodal imagingdepending on the studied biological systems.The Eu3þ excitation wavelength at 466 nm limits the

optical imaging depth in vivo because of tissue absor-bance. Nevertheless, ROS detection in superficial le-sions or optical-fiber-based endoscopy at lesion sitesremains possible and is highly relevant for numerouspathologies. Future work will exploit these multiple-modality features to detect recruitment dynamics ofmacrophages loaded with these particles to inflamma-tion sites combined with the evaluation of the role ofoxidative stress in this pathology.

EXPERIMENTAL METHODS

Nanoparticle Synthesis. Rare-earth vanadate nanoparticlesweresynthesized using two well-documented procedures, namely,standard coprecipitation43 and citrate-assisted44 routes, yieldingparticles of approximately 30 and 5 nm in size, respectively.Details are reported in the Supporting Information.

Silica Coating. Nanoparticles prepared through the standardor citrate routes were diluted in ultrapure water to a vanadateconcentration of about 20 mM, and the dispersions were soni-cated for several minutes. Then, the same volume of a 3 wt %sodium silicate solution (Na2SiO3, Riedel-de Haën) was imme-diately added under vigorous stirring. The mixture was stirredovernight at ambient temperature and then dialyzed in SpectraPor regenerated cellulose dialysis membranes (MWCO 12�14 kDa) against ultrapure water for several days. Dialysis wasstopped after reaching a conductivity of the nanoparticledispersion between 50 and 100 μS/cm. The presence of a silicacoating was validated by IR spectroscopy.

Dextran Coating. Nanoparticles were alternatively coatedwith carboxymethyl-dextran (CM-dextran) following the proce-dure described by Dutz et al.:66 colloidal solutions of 20 mMvanadate concentration were heated at 45 �C under vigorousstirring. CM-dextran powder (40 000 kDa, Sigma-Aldrich) wasadded in a dextran/nanoparticle mass proportion of 2:1. After60 min condensation, the solutions were purified by successivecentrifugations. The hydrodynamic radius of these nanoparti-cles before and after CM-dextran coating was measured bydynamic light scattering to be 23 and 35 nm (number-averagevalues), respectively.

Standard Characterizations. Dynamic light scattering measure-ments were performed using a Malvern Nano ZS Zetasizer ondiluted dispersions in water (25 �C, 633 nm). For all rare-earthvanadate materials, the refractive index was assumed to be thesame as that of YVO4, that is, 1.99 at 25 �C and 633 nm.67

For observations by transmission electron microscopy, weused dilute solutions of nanoparticles deposited on a carbongrid. Imaging was performed with a Philips CM30 microscopeoperating at 300 kV with a point resolution of 0.235 nm. Imageswere analyzed using the ImageJ software by manual selection.

X-ray powder diffraction measurements were recorded inBragg�Brentano parafocusing geometry using a X'Pert MPDpowder diffractometer (λ = 1.5418 Å). Diffracted intensity wasmonitored using an X'Celerator area detector (PANalytical)which covers a scattering angle segment of 2.546� through anickel filter. Sample eccentricity and peak positions werecalibrated using bulk GdVO4. Details related to the microstruc-tural analysis of these XRD data are reported as SupportingInformation.

Magnetization measurements were performed using acryogenic SX600 magnetometer (see Supporting InformationFigure S4). Powder samples were mounted in polycarbonatecapsules with the powder blocked by paraffin wax. Magneticdata were systematically corrected for the diamagnetic con-tributions of the wax and polycarbonate capsule.

We performed FT-IR spectroscopy to characterize nanopar-ticle coatings with silica or dextran. A drop of colloid wasdeposed on a piece of silicon wafer and dried at 90 �C forseveral hours. The spectrum was measured with a BrukerEquinox infrared spectrometer from 500 to 4000 cm�1. Data

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11135

were normalized to a reference silicon substrate (see Support-ing Information Figure S2).

Luminescence Spectroscopy. Luminescence excitation andemission spectra were recorded using a Hitachi F-4500 fluores-cence spectrophotometer with 2.5 nm spectral resolution. Forexcitation spectra, luminescence was excited in the wavelengthrange of 250�600 nm and emission was collected at 617 nm.For emission spectra, luminescence was excited at 280 nm. Theluminescence quantum yield was determined by comparison ofthe spectrally integrated emission intensity to that of rhoda-mine 6G (Q = 94% in ethanol) in the wavelength range of500�750 nm after excitation at 280 nm for solutions havingsimilar optical density (OD < 0.3). We obtained the colloidquantum yield as follows: QNP = (nNP

2 /nR2) � (INP/IR) � (AR/ANP),

where n is the refractive index, I is the emission intensity, and Ais the absorbance of the solutions.

Luminescence Microscopy Imaging. Single nanoparticle observa-tion relies on the direct excitation of Eu3þions.20 Imaging iscarried out on an inverted microscope (Axiovert S100TV CarlZeiss) under wide-field illumination by an Innova 300 argonion laser (Coherent) at 466 nm. Nanoparticle luminescence iscollected by an Apochromat 63�, 1.40 numerical aperture oil-immersion objective (Zeiss) and selected by a 500DCLP dichroicmirror (Chroma Technology) and a D617/8m band-pass filter(fwhm 8 nm, Chroma). Image acquisition is performed by aQuantEM 512SC CCD camera (Photometrics). Diluted solutionsof nanoparticles are spin-coated on glass coverslips previouslycleaned in a plasma cleaner. Their response to hydrogenperoxide is assessed as follows: the nanoparticles are photo-reduced at a laser intensity of 1.6 kW/cm2 for 5 min (acquisitionrate: 1 image/s). The stability of the luminescence level afterphotoreduction is checked at a laser intensity of 0.3 kW/cm2 for5 min (acquisition rate: 1 image every 3 s). The luminescencerecovery is observed after addition of hydrogen peroxide(Sigma-Aldrich) in the solution for several minutes at a laserintensity of 0.3 kW/cm2 (acquisition rate: 1 image every 3 s).

Measurement of Relaxation Times. Relaxation times were deter-mined using a Bruker NMS 120 minispec relaxometer operatingat a proton resonance frequency of ω/2π = 20 MHz (0.47 T) andat 37 �C. T1 relaxation timeswere determined from the inversionrecovery pulse sequence with a repetition time TR = 5 s. Thepulse separation time TI was adjusted until the condition TI ≈0.6 T1 was fulfilled. To measure the T2 relaxation time, theCarr�Purcell�Meiboom�Gill pulse sequence was employedusing a repetition time TR = 8 s. Generally, 100 echoes with anecho time TE between 0.5 and 2 ms depending on the con-centration of the sample were recorded. TE was adjustedmanually in order to record the complete magnetization decayduring the 100 echoes.

Nuclear Magnetic Resonance Dispersion Spectroscopy. 1H nuclearmagnetic resonance dispersion (NMRD) spectroscopy experi-ments were carried out with a Stelar Spinmaster FFC2000 fastfield cycling relaxometer at 25 �C. Relaxation times weremeasured in the 10 kHz to 15 MHz frequency range for avanadate concentration of 10 mM. A polarization field corre-sponding to 15 MHz was applied for 2 s. The transition to therelaxation field was performed within 2 ms, and the magnetiza-tion relaxed for a time span Δ varied within the interval 0�5 T1.The remaining magnetization after Δ was measured after aquick transition within 1 ms to the acquisition field correspond-ing to 9 MHz and by applying a π/2 detection pulse.

In VivoMRI. In vivo experiments were performed on BL6wild-type mice anesthetized by isoflurane inhalation. Respirationwas monitored during experiments, and isoflurane delivery wasadjusted as a consequence to preserve stable vital signs inanimals. Prior to contrast agent (Gd0.6Eu0.4VO4) injection, weperformed a first image series that represents the innate tissuecontrast (measurement at time zero). T1-weighted imaging wasperformed with a BioSpec Bruker 47/40 USR apparatus at a fieldof 4.7 T and a corresponding frequency of 200 MHz. Theparameters of the gradient echo sequence and the imagingconfiguration are as follows: TR/TE = 150/5.4 ms, R = 40�, FOV =5� 4 cm, detectionmatrix = 256� 256, NEX = 4, slice thickness =1 mm. The signal was measured in the liver, kidney, andbladder and normalized to a reference region (tube containing

water). Animals received humane care, and the study protocolscomplied with PARCC guidelines for the care and use oflaboratory animals.

Conflict of Interest: The authors declare no competingfinancial interest.

Acknowledgment. The research described here has beenpartly supported by Triangle de la Physique (Contract 2008-051T). We are indebted to Pierre Levitz (LPMC) for the fre-quency-dependent measurements. We acknowledge financialsupport from CNRS Défi G3N 2012 and CNRS Défi Nano 2014.

Supporting Information Available: Dynamic light scatteringand zeta-potential data for all samples mentioned, microstruc-tural analysis for normal-route nanoparticles, structural charac-terization by TEM imaging and X-ray diffraction for citrate-routenanoparticles, infrared spectroscopy characterization of silicaand dextran surface coatings, magnetic characterization ofGdVO4 and Gd0.6Eu0.4VO4 NPs. Information is also given onthe luminescence properties, the photoreduction and oxidantsensing capabilities of individual nanoparticles and of nanopar-ticle assemblies for citrate- and normal-route particles, thereversibility of the photoreduction/reoxidation reaction and theabsence of modification of the oxidant sensing response inphysiologicalmedium.We also provide cell viability assay results,MRI measurements after injection of 30 nm Gd0.6Eu0.4VO4 NPs,and luminescence measurements on urine, kidney, and livertissuehomogenates after NP injection. Finally, further descriptionis given on experimental procedures for nanoparticle synthesisand purification, the determination of vanadate concentration,rare-earth leaching in solution, cell viability experiments, andluminescence imaging of tissue homogenates. This material isavailable free of charge via the Internet at http://pubs.acs.org.

REFERENCES AND NOTES1. Malam, Y.; Loizidou, M.; Seifalian, A. M. Liposomes and

Nanoparticles: Nanosized Vehicles for Drug Delivery inCancer. Trends Pharmacol. Sci. 2009, 30, 592–599.

2. Cheng, Z.; Al Zaki, A.; Hui, J. Z.; Muzykantov, V. R.; Tsourkas,A. Multifunctional Nanoparticles: Cost versus Benefit ofAdding Targeting and Imaging Capabilities. Science 2012,338, 903–910.

3. Goldys, E. Fluorescence Applications in Biotechnology andLife Sciences; John Wiley & Sons, Inc.: Hoboken, NJ, 2009.

4. Michalet, X.; Pinaud, F. F.; Bentolila, L. A.; Tsay, J. M.; Doose,S.; Li, J. J.; Sundaresan, G.; Wu, A. M.; Gambhir, S. S.; Weiss, S.Quantum Dots for Live Cells, In Vivo Imaging, and Diag-nostics. Science 2005, 307, 538–544.

5. Westbrook, C.; Roth, C.; Talbot, J. MRI in Practice, 4th ed.;Wiley-Blackwell: Oxford, 2011.

6. Laurent, S.; Forge, D.; Port, M.; Roch, A.; Robic, C.; VanderElst, L.; Muller, R. N. Magnetic Iron Oxide Nanoparticles:Synthesis, Stabilization, Vectorization, PhysicochemicalCharacterizations, and Biological Applications. Chem.Rev. 2008, 108, 2064–2110.

7. Wang, Y. X.; Hussain, S. M.; Krestin, G. P. Superparamag-netic Iron Oxide Contrast Agents: Physicochemical Char-acteristics and Applications in MR Imaging. Eur. Radiol.2001, 11, 2319–2331.

8. Geraldes, C. F. G. C.; Laurent, S. Classification and BasicProperties of Contrast Agents for Magnetic ResonanceImaging. Contrast Media Mol. Imaging 2009, 4, 1–23.

9. Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B.Gadolinium(III) Chelates asMRI Contrast Agents: Structure,Dynamics, and Applications. Chem. Rev. 1999, 99, 2293–2352.

10. Hüber, M. M.; Staubli, A. B.; Kustedjo, K.; Gray, M. H.; Shih, J.;Fraser, S. E.; Jacobs, R. E.; Meade, T. J. FluorescentlyDetectable Magnetic Resonance Imaging Agents. Biocon-jugate Chem. 1998, 9, 242–249.

11. Zhang, X.; Jing, X.; Liu, T.; Han, G.; Li, H.; Duan, C. Dual-Functional Gadolinium-Based Copper(II) Probe for Selec-tive Magnetic Resonance Imaging and Fluorescence Sens-ing. Inorg. Chem. 2012, 51, 2325–2331.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11136

12. Kelly, K. A.; Allport, J. R.; Tsourkas, A.; Shinde-Patil, V. R.;Josephson, L.; Weissleder, R. Detection of Vascular Adhe-sion Molecule-1 Expression Using a Novel MultimodalNanoparticle. Circ. Res. 2005, 96, 327–336.

13. Mulder, W. J. M.; Koole, R.; Brandwijk, R. J.; Storm, G.; Chin,P. T. K.; Strijkers, G. J.; de Mello Donegá, C.; Nicolay, K.;Griffioen, A. W. Quantum Dots with a Paramagnetic Coat-ing as a Bimodal Molecular Imaging Probe. Nano Lett.2006, 6, 1–6.

14. Bridot, J.-L.; Dayde, D.; Rivière, C.; Mandon, C.; Billotey, C.;Lerondel, S.; Sabattier, R.; Cartron, G.; Le Pape, A.; Blondiaux,G.; et al. Hybrid Gadolinium Oxide Nanoparticles Combin-ing Imaging and Therapy. J. Mater. Chem. 2009, 19, 2328.

15. Shi, Z.; Neoh, K. G.; Kang, E. T.; Shuter, B.; Wang, S.-C.Bifunctional Eu3þ-Doped Gd2O3 Nanoparticles as a Lumi-nescent and T1 Contrast Agent for Stem Cell Labeling.Contrast Media Mol. Imaging 2010, 5, 105–111.

16. Jung, J.; Kim, M. A.; Cho, J.-H.; Lee, S. J.; Yang, I.; Cho, J.; Kim,S. K.; Lee, C.; Park, J. K. Europium-Doped GadoliniumSulfide Nanoparticles as a Dual-Mode Imaging Agent forT1-Weighted MR and Photoluminescence Imaging. Bio-materials 2012, 33, 5865–5874.

17. Zhou, J.; Sun, Y.; Du, X.; Xiong, L.; Hu, H.; Li, F. Dual-ModalityIn Vivo Imaging Using Rare-Earth Nanocrystals with Near-Infrared to Near-Infrared (NIR-to-NIR) Upconversion Lumi-nescence and Magnetic Resonance Properties. Biomater-ials 2010, 31, 3287–3295.

18. Bouzigues, C.; Gacoin, T.; Alexandrou, A. Biological Appli-cations of Rare-Earth Based Nanoparticles. ACS Nano2011, 5, 8488–8505.

19. Mialon, G.; Türkcan, S.; Alexandrou, A.; Gacoin, T.; Boilot,J.-P. New Insights into Size Effects in Luminescent OxideNanocrystals. J. Phys. Chem. C 2009, 113, 18699–18706.

20. Casanova, D.; Bouzigues, C.; Nguyen, T.-L.; Ramodiharilafy,R. O.; Bouzhir-Sima, L.; Gacoin, T.; Boilot, J.-P.; Tharaux, P.-L.;Alexandrou, A. Single Europium-Doped NanoparticlesMeasure Temporal Pattern of Reactive Oxygen SpeciesProduction inside Cells. Nat. Nanotechnol. 2009, 4, 581–585.

21. D'Autréaux, B.; Toledano, M. B. ROS as Signalling Mol-ecules: Mechanisms That Generate Specificity in ROSHomeostasis. Nat. Rev. Mol. Cell Biol. 2007, 8, 813–824.

22. Rhee, S. G. Cell Signaling. H2O2, a Necessary Evil for CellSignaling. Science 2006, 312, 1882–1883.

23. Clempus, R. E.; Griendling, K. K. Reactive Oxygen SpeciesSignaling in Vascular SmoothMuscle Cells. Cardiovasc. Res.2006, 71, 216–225.

24. Sundaresan, M.; Yu, Z. X.; Ferrans, V. J.; Irani, K.; Finkel, T.Requirement for Generation of H2O2 for Platelet-DerivedGrowth Factor Signal Transduction. Science 1995, 270,296–299.

25. Hensley, K.; Robinson, K. A.; Gabbita, S. P.; Salsman, S.;Floyd, R. A. Forum on Therapeutic Applications of ReactiveOxygen and Nitrogen Species in Human Disease. FreeRadical Biol. Med. 2000, 28, 1449–1450.

26. Wolfbeis, O.; Dürkop, A. A Europium Ion Based Lumines-cent Sensing Probe for Hydrogen Peroxide. Angew. Chem.,Int. Ed. 2002, 1, 4495–4498.

27. Belousov, V. V.; Fradkov, A. F.; Lukyanov, K. A.; Staroverov,D. B.; Shakhbazov, K. S.; Terskikh, A. V.; Lukyanov, S.Genetically Encoded Fluorescent Indicator for IntracellularHydrogen Peroxide. Nat. Methods 2006, 3, 281–286.

28. Miller, E. W.; Tulyathan, O.; Tulyanthan, O.; Isacoff, E. Y.;Chang, C. J. Molecular Imaging of Hydrogen PeroxideProduced for Cell Signaling. Nat. Chem. Biol. 2007, 3,263–267.

29. Lee, D.; Khaja, S.; Velasquez-Castano, J. C.; Dasari, M.; Sun,C.; Petros, J.; Taylor, W. R.; Murthy, N. In Vivo Imaging ofHydrogen Peroxide with Chemiluminescent Nanoparti-cles. Nat. Mater. 2007, 6, 765–769.

30. Bouzigues, C. I.; Nguyen, T.-L.; Ramodiharilafy, R.; Claeson,A.; Tharaux, P.-L.; Alexandrou, A. Regulation of the ROSResponse Dynamics and Organization to PDGF MotileStimuli Revealed by Single Nanoparticle Imaging. Chem.Biol. 2014, 21, 647–656.

31. Türkcan, S.; Masson, J.-B.; Casanova, D.; Mialon, G.; Gacoin,T.; Boilot, J.-P.; Popoff, M. R.; Alexandrou, A. Observing theConfinement Potential of Bacterial Pore-Forming ToxinReceptors inside Rafts with Nonblinking Eu3þ-DopedOxide Nanoparticles. Biophys. J. 2012, 102, 2299–2308.

32. Türkcan, S.; Alexandrou, A.; Masson, J.-B. A BayesianInference Scheme To Extract Diffusivity and PotentialFields from Confined Single-Molecule Trajectories. Bio-phys. J. 2012, 102, 2288–2298.

33. Cuzzocrea, S.; Riley, D. P.; Caputi, A. P.; Salvemini, D.Antioxidant Therapy: A New Pharmacological Approachin Shock, Inflammation, and Ischemia/Reperfusion Injury.Pharmacol. Rev. 2001, 53, 135–159.

34. Chinta, S. J.; Andersen, J. K. Redox Imbalance in Parkinson'sDisease. Biochim. Biophys. Acta 2008, 1780, 1362–1367.

35. Gupta, S. C.; Hevia, D.; Patchva, S.; Park, B.; Koh, W.;Aggarwal, B. B. Upsides and Downsides of Reactive Oxy-gen Species for Cancer: The Roles of Reactive OxygenSpecies in Tumorigenesis, Prevention, and Therapy. Anti-oxid. Redox Signal. 2012, 16, 1295–1322.

36. Toyokuni, S.; Okamoto, K.; Yodoi, J.; Hiai, H. PersistentOxidative Stress in Cancer. FEBS Lett. 1995, 358, 1–3.

37. Ishikawa, K.; Takenaga, K.; Akimoto, M.; Koshikawa, N.;Yamaguchi, A.; Imanishi, H.; Nakada, K.; Honma, Y.; Hayashi,J.-I. ROS-Generating Mitochondrial DNA MutationsCan Regulate Tumor Cell Metastasis. Science 2008, 320,661–664.

38. Ramsey, M. R.; Sharpless, N. E. ROS as a Tumour Suppres-sor? Nat. Cell Biol. 2006, 8, 1213–1215.

39. Liu, B.; Chen, Y.; Clair, D. S. ROS and p53: A VersatilePartnership. Free Radical Biol. Med. 2008, 44, 1529–1535.

40. Van de Bittner, G. C.; Bertozzi, C. R.; Chang, C. J. Strategy forDual-Analyte Luciferin Imaging: In Vivo BioluminescenceDetection of Hydrogen Peroxide and Caspase Activity in aMurine Model of Acute Inflammation. J. Am. Chem. Soc.2013, 135, 1783–1795.

41. Van de Bittner, G. C.; Dubikovskaya, E. A.; Bertozzi, C. R.;Chang, C. J. In Vivo Imaging of Hydrogen Peroxide Produc-tion in a Murine Tumor Model with a ChemoselectiveBioluminescent Reporter. Proc. Natl. Acad. Sci. U.S.A. 2010,107, 21316–21321.

42. Lippert, A. R.; Keshari, K. R.; Kurhanewicz, J.; Chang, C. J. AHydrogen Peroxide-Responsive Hyperpolarized 13C MRIContrast Agent. J. Am. Chem. Soc. 2011, 133, 3776–3779.

43. Huignard, A.; Gacoin, T.; Boilot, J.-P. Synthesis and Lumi-nescence Properties of Colloidal YVO4:Eu Phosphors.Chem. Mater. 2000, 12, 1090–1094.

44. Huignard, A.; Buissette, V.; Laurent, G.; Gacoin, T.; Boilot,J.-P. Synthesis and Characterizations of YVO4:Eu Colloids.Chem. Mater. 2002, 14, 2264–2269.

45. Jin, D.; Yang, H.; Ding, G.; Yu, X.; Wang, L.; Zheng, Y.Hydrothermal Synthesis and Photoluminescence Behaviorof Eu-Doped GdVO4. Inorg. Mater. 2008, 44, 1121–1124.

46. Zheng, F.; Wang, W.; Yang, P. GdVO4Ln (Ln = Eu, Dy, andSm) Microstructures: Solvothermal and LuminescentProperties. J. Optoelectron. Adv. Mater. 2011, 5, 596–599.

47. Aslan, K.; Wu, M.; Lakowicz, J. R.; Geddes, C. D. FluorescentCore�Shell Ag@SiO2 Nanocomposites for Metal-Enhanced Fluorescence and Single Nanoparticle SensingPlatforms. J. Am. Chem. Soc. 2007, 129, 1524–1525.

48. Lauffer, R. B. Paramagnetic Metal Complexes as WaterProton Relaxation Agents for NMR Imaging: Theory andDesign. Chem. Rev. 1987, 87, 901–927.

49. Aime, S.; Botta, M.; T, E. Gd(III)-Based Contrast Agents forMRI. Adv. Inorg. Chem. 2005, 57, 173–237.

50. Sengers, J. V.; Watson, J. T. R. Improved InternationalFormulations for the Viscosity and Thermal ConductivityofWater Substance. J. Phys. Chem. Ref. Data 1986, 15, 1291.

51. Powell, D. H.; Dhubhghaill, O. M. N.; Pubanz, D.; Helm, L.;Lebedev, Y. S.; Schlaepfer, W.; Merbach, A. E. Structural andDynamic Parameters Obtained from 17O NMR, EPR, andNMRD Studies of Monomeric and Dimeric Gd3þ Com-plexes of Interest in Magnetic Resonance Imaging: AnIntegrated and Theoretically Self-Consistent Approach 1.J. Am. Chem. Soc. 1996, 118, 9333–9346.

ARTIC

LE

ABDESSELEM ET AL. VOL. 8 ’ NO. 11 ’ 11126–11137 ’ 2014

www.acsnano.org

11137

52. Van Vleck, J. The Theory of Electric and Magnetic Suscept-ibilities; Oxford University Press: Oxford, 1932.

53. Caillé, J. M.; Lemanceau, B.; Bonnemain, B. Gadolinium as aContrast Agent for NMR. Am. J. Neuroradiol. 1983, 4, 1041–1042.

54. Penfield, J. G.; Reilly, R. F. What Nephrologists Need ToKnow about Gadolinium. Nat. Clin. Pract. Nephrol. 2007, 3,654–668.

55. Haley, T. J.; Komesu, N.; Colvin, G.; Koste, L.; Upham, H. C.Pharmacology and Toxicology of Europium Chloride.J. Pharm. Sci. 1965, 54, 643–645.

56. Buissette, V.; Giaume, D.; Gacoin, T.; Boilot, J.-P. AqueousRoutes to Lanthanide-Doped Oxide Nanophosphors.J. Mater. Chem. 2006, 16, 529.

57. Beyersmann, D.; Hartwig, A. Carcinogenic Metal Com-pounds: Recent Insight intoMolecular and Cellular Mecha-nisms. Arch. Toxicol. 2008, 82, 493–512.

58. Sanchez, D.; Ortega, A.; Domingo, J. L.; Corbella, J. Devel-opmental Toxicity Evaluation of Orthovanadate in theMouse. Biol. Trace Elem. Res. 1991, 30, 219–226.

59. Hifumi, H.; Yamaoka, S.; Tanimoto, A.; Citterio, D.; Suzuki, K.Gadolinium-Based Hybrid Nanoparticles as a Positive MRContrast Agent. J. Am. Chem. Soc. 2006, 128, 15090–15091.

60. Huang, C.-C.; Liu, T.-Y.; Su, C.-H.; Lo, Y.-W.; Chen, J.-H.; Yeh,C.-S. Superparamagnetic Hollow and Paramagnetic Por-ous Gd2O3 Particles. Chem. Mater. 2008, 20, 3840–3848.

61. Port, M.; Meyer, D.; Bonnemain, B.; Corot, C.; Schaefer, M.;Rousseaux, O.; Simonot, C.; Bourrinet, P.; Benderbous, S.;Dencausse, A.; et al. P760 and P775: MRI Contrast AgentsCharacterized by New Pharmacokinetic Properties. MAG-MA 1999, 8, 172–176.

62. Na, H. B.; Song, I. C.; Hyeon, T. Inorganic Nanoparticles forMRI Contrast Agents. Adv. Mater. 2009, 21, 2133–2148.

63. Corot, C.; Robert, P.; Idée, J.-M.; Port, M. Recent Advances inIron Oxide Nanocrystal Technology for Medical Imaging.Adv. Drug Delivery Rev. 2006, 58, 1471–1504.

64. Moeller, M. J.; Tenten, V. Renal Albumin Filtration: Alter-native Models to the Standard Physical Barriers. Nat. Rev.Nephrol. 2013, 9, 266–277.

65. Choi, H. S.; Liu,W.; Misra, P.; Tanaka, E.; Zimmer, J. P.; Itty Ipe,B.; Bawendi, M. G.; Frangioni, J. V. Renal Clearance ofQuantum Dots. Nat. Biotechnol. 2007, 25, 1165–1170.

66. Dutz, S.; Andrä, W.; Hergt, R.; Müller, R.; Oestreich, C.;Schmidt, C.; Töpfer, J.; Zeisberger, M.; Bellemann, M. E.Influence of Dextran Coating on the Magnetic Behaviourof Iron Oxide Nanoparticles. J. Magn. Magn. Mater. 2007,311, 51–54.

67. Bass, M.; DeCusatis, C. M.; Enoch, J. M.; Lakshminarayanan,V.; Li, G.; MacDonald, C.; Mahajan, V. N.; Van Stryland, E.Handbook of Optics. Optical Properties of Materials, Non-linear Optics, Quantum Optics; McGraw-Hill: Columbus,OH, 2010; Vol. IV.

ARTIC

LE