at SciVerse ScienceDirect

Biomaterials xxx (2012) 1e11Contents lists availableBiomaterials

journal homepage: www.elsevier .com/locate/biomater ia lsMetal-free and MRI visible theranostic lyotropic liquid crystal nitroxide-basednanoparticles

Benjamin W. Muir a,**, Durga P. Acharya a, Danielle F. Kennedy a, Xavier Mulet a,b, Richard A. Evans a,Suzanne M. Pereira a, Kim L. Wark a, Ben J. Boyd b, Tri-Hung Nguyen b, Tracey M. Hinton c,Lynne J. Waddington a, Nigel Kirby d, David K. Wright e, Hong X. Wang e, Gary F. Egan e,Bradford A. Moffat f,*aCSIRO Materials Science and Engineering, Bayview Avenue, Clayton 3168, AustraliabDrug Delivery, Disposition and Dynamics, Monash Institute of Pharmaceutical Sciences, Monash University (Parkville Campus), Parkville 3052, AustraliacCSIRO Livestock Industries, Australian Animal Health Laboratory, East Geelong 3219, AustraliadAustralian Synchrotron, Clayton 3168, AustraliaeHoward Florey Institute, The University of Melbourne, 3010, Australiaf The University of Melbourne, Department of Radiology, Parkville 3050, Australiaa r t i c l e i n f o

Article history:Received 18 November 2011Accepted 6 December 2011Available online xxx

Keywords:CubosomeContrast agentLyotropic liquid crystalCytotoxicityMaximum tolerated doseHigh-throughput* Corresponding author. Fax: 61 393428369.** Corresponding author. Fax: 61 395452515.

E-mail addresses: ben.muir@csiro.au (B.W. Muir)A. Moffat).

0142-9612/$ e see front matter 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.12.018

Please cite this article in press as: Muir BW,Biomaterials (2012), doi:10.1016/j.biomateria b s t r a c t

The development of improved, low toxicity, clinically viable nanomaterials that provide MRI contrasthave tremendous potential to form the basis of translatable theranostic agents. Herein we describea class of MRI visible materials based on lyotropic liquid crystal nanoparticles loaded with a para-magnetic nitroxide lipid. These readily synthesized nanoparticles achieved enhanced proton-relaxivitieson the order of clinically used gadolinium complexes such as Omniscanwithout the use of heavy metalcoordination complexes. Their low toxicity, high water solubility and colloidal stability in buffer resultedin them being well tolerated in vitro and in vivo. The nanoparticles were initially screened in vitro forcytotoxicity and subsequently a defined concentration range was tested in rats to determine themaximum tolerated dose. Pharmacokinetic profiles of the candidate nanoparticles were establishedin vivo on IV administration to rats. The lyotropic liquid crystal nanoparticles were proven to be effectiveliver MRI contrast agents. We have demonstrated the effective in vivo performance of a T1 enhancing,biocompatible, colloidally stable, amphiphilic MRI contrast agent that does not contain a metal.

2011 Elsevier Ltd. All rights reserved.1. Introduction

The development of various nanoparticle formulations aspotential MRI contrast agents is an area of ever growing research.The benefits of developing nanoparticle formulations include thepotential to target these agents to diseased tissues such as tumors.Previously this has been achieved passively via the enhancedpermeability and retention effect (EPR) [1e3] or actively viaattachment of targeting ligands such as folate [4], peptides [5] orantibodies [6]. In addition, the use of nanoparticles may enablemultimodal theranostic agents to be developed for both imagingdiseased tissue, and treating the disease by controlled release ofdrugs [7,8]. Few nanoparticle formulations have been clinically, brad.moffat@rmh.org.au (B.

All rights reserved.

et al., Metal-free and MRI visials.2011.12.018approved due to the immense physico-chemical, biological andregulatory hurdles that confront the nanotechnologist whendevising successful nanoparticle and coating formulations [8].

The main aim of scientists developing new and improved MRItheranostic materials is to produce nanoparticles which providestrong signal contrast while maintaining low toxicity and ease ofintra-venous delivery (small volume bolus, stability in saline,increased blood half life and lowviscosity). Contrast agents are usedclinically when poor contrast of the diseased tissue is observedduring MRI scanning (as seen commonly when imaging brain andliver lesions). Signal enhancement is achieved by contrast agentsthat decrease the water spin-lattice relaxation time (T1) during theacquisition of T1 weighted MR images, while signal suppression isachievedbyagents thatdecrease thewater spin-spin relaxation time(T2) during the acquisition of T2 weighted MR images. The contrastagent relaxivity (r1 or r2) is ameasure of the relative effectiveness ofa given contrast agent and has units of s1mM1. To date mostnanoparticle MRI contrast agent formulations (whether T1 or T2ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

mailto:ben.muir@csiro.aumailto:brad.moffat@rmh.org.auwww.sciencedirect.com/science/journal/01429612http://www.elsevier.com/locate/biomaterialshttp://dx.doi.org/10.1016/j.biomaterials.2011.12.018http://dx.doi.org/10.1016/j.biomaterials.2011.12.018http://dx.doi.org/10.1016/j.biomaterials.2011.12.018

B.W. Muir et al. / Biomaterials xxx (2012) 1e112enhancing) have been used as liver contrast agents due primarily touptake by the reticuloendothelial system (RES) after bolus delivery[8,9]. Most clinically used T1 enhancing agents contain gadolinium(Gd) [10,11]which is highly toxic as a free trivalent ion [12]. Althoughthe Gd metal in clinically used contrast agents are chelated [10] themetal may still potentially leach. Certain agents may cause adversereactions in patients with renal failure resulting in a debilitatingdisease called nephrogenic systemic fibrosis [13]. As a result,phasingout theuseof heavymetal based contrast agents andfindingsuitable alternatives is of great interest to the field. To date, non-heavy metal based materials investigated as T1 enhancing agentshave essentially been limited to 19F [14,15] and nitroxide enrichedcompounds [16e18].

One of the greatest challenges to the materials scientist in thepreparation of suitable MRI contrast agent nanoparticles is theproduction of highly stable colloidal dispersions that have negli-gible cytotoxic properties [8]. To this end we have developed theuse of lyotropic liquid crystal nanoparticles that contain a para-magnetic nitroxide lipid to provide T1 contrast rather than theconventionally used gadolinium based compounds. This may allowthe field to move away from the toxicity issues associated withgadolinium based contrast agents. Previous reports have shown theapplicability of using gadolinium functionalized lipids incorporatedinto lyotropic mesophase liquid crystal nanoparticles as potentialMRI agents in vitro [19,20]. The use of amphiphilic lipids such asphytantriol and glyceryl monooleate (GMO) (structure in Fig. 1)result in the formation of distinct lyotropic mesophases of varyingcomplexity and dimensionality. Glycerol monooleate contains anester group that is susceptible to hydrolysis and therefore biodeg-radation in vivo. The common phases seen when using lyotropicliquid crystal materials include lamellar phase (L) (1-D) comprisedof stacked bilayer sheets, the hexagonal phase (H2) (2-D) which canbe conceptualized as infinitely long hexagonally packed rods withan aqueous interior and finally the cubic phase (Q2) (3-D) con-sisting of a bi-continuous network of hydrophilic and hydrophobicdomains containing two continuous water channels. The Q2 phaserepresents a family of closely related structures, where the under-lying crystal lattice can be described by the gyroid (G), diamond (D)and primitive (P) minimal surfaces, which correspond to the Ia3d(G), Pn3m (D) and Im3m (P) crystallographic space groups,Fig. 1. Chemical structure of phytantriol amphiphile (top structure), glyceryl mono-oleate amphiphile (middle structure, the main component of Myverol) and thenitroxide lipid (bottom structure) used to make lyotropic liquid crystal, MRI contrastagent nanoparticles.

Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018respectively. The 3-dimensional structure affords a self-assembledscaffold with a remarkably high surface area and extensiveporosity. These properties, coupled with the liquid crystallinenature of the phase, result in a structure that was found to not besusceptible to osmotic or mechanical rupture in contrast to theproperties of liposomes [21] or micelles. The thermodynamicstability of the Q2 phase affords a structure that co-exists in equi-librium with excess water over a broad temperature range. Thedispersion of the bulk gel-like cubic phases can be achieved bymechanical or ultrasonic treatment and results in the formation ofnanometer-sized particles that retain the internal cubic structure ofthe parent bulk cubic phase. The incorporation of additives to suchmaterials may result in the formation of other phases such as theinverse hexagonal phase (H2) due to an alteration in the sponta-neous curvature of the lyotropic liquid crystal assemblies. Disper-sions of hexagonal phase nanoparticles are commonly calledhexosomes and cubic phase nanoparticles are referred to ascubosomes [22e26].

The aim of this study was to investigate the benefit of incor-porating a myristic nitroxide lipid (structure in Fig. 1) into lyotropicliquid crystal nanoparticles. Nitroxides are stable, organic freeradicals with an unpaired (paramagnetic) electron and are there-fore capable of shortening the MRI relaxation times [18]. Onceinside the body, an equilibrium exists between the paramagnetic(contrast enhancing) nitroxide and the reduced non-paramagnetic(non-contrast enhancing) hydroxylamine [27]. Previous studieshave shown that these compounds can be useful for imagingintracellular redox metabolism by MRI [28,29], because the ratio ofthe two states was dictated by the local oxygen and redox envi-ronment. In addition these compounds have been shown to havepotential for controlling hypertension and weight, preventingdamage from reperfusion injury, and treating neurodegenerativediseases and ocular damage [29,30]. It was hypothesized thatencapsulating the nitroxide lipid inside the lyotropic mesophaseliquid crystal nanoparticles would extend the nitroxide radicalshalf-life in vivo making it an effective MRI contrast agent withan acceptable cytotoxicity profile. Furthermore, the presence ofconfined water channels in cubic and hexagonal phase nano-particles, and their greater surface area compared to liposomesmayresult in enhanced relaxivities of the nitroxide lipid due to rota-tional correlation constant and proton exchange processes.

To investigate these hypotheses, nanoparticles were synthe-sized using two different bulk cubic phase forming lipids. In thisstudy the effect of nanoparticle structure on relaxivity, cytotoxicity,maximum tolerated dose in rats and efficacy of the contrast agentsin vivo for nitroxide loaded cubosomes and hexosomes, wasinvestigated. The cubic and hexagonal phase nanoparticles havea viscosity approximately equal to water, which makes it desirablefor bolus delivery of MRI contrast agents. Previous formulations ofthese types of nanoparticles have shown them to have highcolloidal stability and low cytotoxicity through the appropriateselection of the amphiphile used to form the lyotropic liquid crystalphase [31].

2. Materials and methods

2.1. Self-assembly of the lyotropic liquid crystal nitroxide containing nanoparticles

Two bulk cubic phase forming lipids were used in this work, phytantriol (DSMNutritional Products, GmbH) and Myverol (Bronson & Jacobs, Sydney) which wereused as received. Myverol was used as a commercially available source of GMOwhich was the main lipid component of this product. Samples for screening the T1and T2 relaxivities were prepared in a high-throughput manner using a ChemspeedAccelerator SLTII robotic synthesis platform equipped with a 4-needle head andprobe sonicator tools. As previously reported [32], stock solutions of the phytantriol,Myverol and nitroxide lipid were prepared in a poly(ethylene) 96 well plate (2 mLcapacity per well) containing chloroform, 0.5 mg of bulk lipid and an appropriateble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

B.W. Muir et al. / Biomaterials xxx (2012) 1e11 3volume of nitroxide lipid. For each bulk lipid solution transfer into the wells an extra100 mL and a 10 mL air-gap were used. To achieve higher accuracy in dispensingnitroxide lipid solutions, no air gap was used for these low volume transfers. Aconcentration gradient of the nitroxide lipid in each bulk lipid multi-well plate zonewas prepared between 0 and 14 wt% nitroxide lipid. The plates were removed fromthe platform in order to remove the solvent using a Genevac centrifugal evapo-rator. The pluronic was then added at a concentration of 10 wt% to each well in 1 mLof water and sonicated on the Chemspeed platform to produce milky dispersions ofnanoparticles.

Bulk samples for relaxivity testing, cytotoxicity testing and injection in animalswere prepared individually. For each sample, 1 g of bulk lipid (phytantriol orMyverol) and the required amount of nitroxide lipid were heated to 60 C andmixeduntil a homogeneous gel was observed. The required amount of pluronic (F127) wasthen added to a final concentration of 15 wt%. The mixture was then heated to 70 Cfor 5 min followed by vortex mixing. This process was repeated 4 or 5 times. This hotmixture was added dropwise to water (at w65 C) under shear generated bya Polytron shear homogenizer using a 20 mm probe rotating at 20,000 rpm. Thedispersion was further homogenized at 65 C with three passes in an Avastinpressure homogenizer to obtain the final dispersed nanoparticle solutions. Thefraction of nonaqueous components in the solution of lipid nanoparticles wasdetermined by thermogravimetric analysis (Mettler). About 35 mg of nanoparticlesolution was placed in a 70 mL silica crucible and analysed in the temperature rangeof 25e800 C. The average of three results was taken to determine the fraction ofnonaqueous components in the solution and the concentration of phytantriol orMyverol in the bulk nanoparticle solution.

2.2. Nitroxide lipid synthesis

The nitroxide lipid was made by dissolving 4-hydroxy-TEMPO (2 g) indichloromethane (DCM, 40 mL) with 2.4 mL of triethylamine. Myristol chloride(2.6 g) in DCM (5 mL) was added dropwise at room temperature. The reaction wasstirred at room temperature under argon for 4 h. The reaction was worked up bywashing with water, dilute hydrochloric acid and brine before drying with sodiumsulphate and evaporation to give 3.6 g of crude oil that partially crystallized. Thematerial was then chromatographed on silica (hexane:ether 70:30). The productfraction was collected and evaporated down to give 2.5 g (45% yield) of red oil. Dueto the paramagnetic nature of the nitroxide the NMR spectra was poor in terms ofresolution, sensitivity and integration. 1H NMR (CDCl3, 400 MHz), d 1.59 (br. s), 1.97(br. s., CH2 CH3), 2.34 (br. s. CH2) 2.99 (2H, CH2CO) ppm. 13C NMR (CDCl3) d 15.4,23.70, 26.00, 30.09, 30.25, 30.31, 30.41, 30.54, 30.58, 30.64, 32.90, 38.57, 172.97(C]O) ppm. Mass Spectroscopy experiments were carried out using a ThermoQuestMAT95XP, employing Electron Impact at 70 eV and perfluorokerosene was used asa reference standard. Mass spectral analysis confirmed the nitroxide lipid had beensuccessfully synthesized with a measured m/z of 382.3318.

2.3. Cryo-transmission electron microscopy (Cryo-TEM)

A humidity-controlled vitrification system was used to prepare the nano-particles for imaging in a thin layer of vitrified ice using cryo-TEM. Humidity waskept close to 80% for all experiments, and ambient temperature was 22 C. 200-mesh copper grids coated with perforated carbon film (Lacey carbon film: ProSci-Tech, Qld, Australia) were used for all experiments. Aliquots of the sample (4 mL)were pipetted onto each grid prior to plunging. After 30 s of adsorption time the gridwas blotted manually using Whatman 541 filter paper for approximately 2 s. Blot-ting time was optimized for each sample. The grid was then plunged into liquidethane cooled by liquid nitrogen. Frozen grids were stored in liquid nitrogen untilrequired. The samples were examined using a Gatan 626 cryoholder (Gatan,Pleasanton, CA, USA) and Tecnai 12 Transmission Electron Microscope (FEI, Eind-hoven, The Netherlands) at an operating voltage of 120 kV. At all times low doseprocedures were followed, using an electron dose of 8e10 electrons/2 forall imaging. Images were recorded using a Megaview III CCD camera and Ana-lySIS camera control software (Olympus) using magnifications in the range60,000e110,000.

2.4. Dynamic light scattering analysis

Particle size measurements were performed using a Zetasizer-Nano instrument(Malvern, UK). Particle sizes were measured in Milli-Q water using samplesappropriately diluted. The analysis was performed at 25 C and for each sample, themean diameter and polydispersity index (PDI) of six measurements was calculated.

2.5. Synchrotron SAXS measurements

Small-angle X-ray scattering (SAXS) experiments were performed on the SAXS/WAXS beamline at the Australian Synchrotron. Special glass capillaries (1.5 mm)containing sample solutions were placed in a temperature controlled sample holdermaintained at 37 C by a recirculating water bath. Samples were exposed to the12 keV X-ray beam of dimensions 2500 mm 130 mm and a typical flux of51012 photons/s and diffraction patterns were recorded using a Pilatus 1MPlease cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018detector (Dectris, Switzerland). A silver behenate standard was used to calibrate thereciprocal space vector for analysis. Data reduction (calibration and integration) wasperformed using AXcess, a custom-written SAXS analysis program written by Dr.Andrew Heron from Imperial College, London [33]. The bicontinuous cubic phaseswith Pn3m symmetry (diamond) and hexagonal phase observed in this study wereidentified from the positions of diffracted peaks at O2, O3, O4, O6, O8, O9 and at 1, O3,O4 respectively.

2.6. MRI relaxivity measurements

A high throughput MRI screening technique was used to evaluate the NMRrelaxation properties of the nanoparticle solutions at 3 Tesla using a method similarto that previously reported [34]. For initial MRI relaxivity measurements, the platesproduced in the Chemspeed robotic platform were imaged at 23 C in a Siemens(Germany) 3 Tesla TRIO MRI scanner using a body transmit radio frequency coil anda 12 channel radio frequency receiver coil. For MRI relaxivity measurements on bulkproduced samples, 1 mL of each lyotropic liquid crystal nanoparticle formulationwas placed in a 96 well plate and serially diluted to measure T1, T2 and calculaterelaxivities.

For quantifying the spin-spin relaxation rates (R21/T2) a multiple spin echosequence was used to acquire 32 images at echo times (TE) ranging from 11.5 to310.5 ms with a repetition time (TR) of 3 s. To quantify the spin-lattice relaxationrates (R11/T1) a spin echo inversion recovery imaging sequence was used toacquire images at nine different inversion times (TI) ranging from 20 ms to 5 s (TR/TE 10,000/10 ms). All images were acquired with a 3 mm slice thickness, 100 mmFOV, 192154 matrix size and 2 averages. Zero filling was used to reconstruct allimages to a matrix size of 256 256. To calculate R2 the signal from regions ofinterest (>40 voxels) centered within each well was averaged and plotted asa function of TE. The R2 values were then calculated (as the decay constants) bynumerically fitting (using a non-linear least squares algorithm, Matlab, MA, USA)the data to a mono-exponential equation:

S S0expTE R2 (1)where S0 is the signal when TE 1/R2.

A similar approach was used to calculate R1 except the data was plotted asa function of TI and fitted to the following equation:

S S01 2expTI R1 (2)The respective relaxation rates, R1 and R2, were plotted as a function of nitroxidelipid concentration and linear least squares analysis (Matlab, MA, USA) was used toquantify the relaxivities: r1 and r2 respectively.

2.7. Cytotoxicity testing of the lyotropic liquid crystal nitroxide containingnanoparticles

CHO-GFP and HEK293 (ATCC No. CRL1573) cells were seeded at 1104 cells in96-well tissue culture plates in triplicate and grown overnight at 37 C with 5% CO2.Toxicity was measured using the Alamar Blue reagent (Invitrogen USA) according tomanufacturers instructions. In short, 20 mL of Alamar Blue was added to each wellcontaining 200 mL of media and incubated for 4 h at 37 C with 5% CO2. The assaywas read on an EL808 Absorbance microplate reader (BIOTEK, USA) at 540 nm and620 nm. Cell viability was determined by subtracting the optical density measure-ment at 620 nm from that at 540 nm. Results are presented as a percentage ofuntreated cells.

2.8. Pharmacokinetics of phytantriol and Myverol-based lyotropic liquid crystallinenanoparticles

These experiments were approved by the Monash Institute of PharmaceuticalScience Animal Ethics Committee. To assess the circulatory behavior of the lyotropicliquid crystalline nanoparticles, a radiolabelled tracer molecule (3H-dioleyl phos-phatidylcholine) was incorporated into phytantriol and Myverol based liquid crys-talline nanoparticles as was commonly employed to study circulation of liposomes.Two levels of colloidal stabilizer were assessed to determine whether the totalpluronic concentration influenced particle removal from systemic circulation due toinhibition of protein binding. The dispersions consisted of 5% w/v lipid, containing3H-PL equivalent to 0.5 mCi per dose (

Fig. 2. Cryo-TEM images of phytantriol cubosomes containing 2 wt% nitroxide lipid(A) with inset showing periodic internal structure inside a single cubosome andhexosomes containing 14.5 wt% nitroxide lipid (B) with inset showing characteristiconion ring structures (scale bar is 200 nm).

B.W. Muir et al. / Biomaterials xxx (2012) 1e114indicated in results and mixed with scintillation cocktail (2 mL Starscint) andcounted on a Packard Tri-Carb 2000CA liquid scintillation analyzer (Meriden, CT).

2.9. Maximum tolerated dose testing of lyotropic liquid crystal nitroxide containingnanoparticles

Maximum tolerated dose (MTD) testing was performed by MIR (Michigan, USA)and approved by their UCUCA. According to body weight, 10 mL/kg of nanoparticlesolutions were administered to female fisher 344 rats (130e145 g Charles RiverLaboratories, Indiana, USA). The animals were fed irradiated Rodent Diet 5053(LabDiet) and water ad libitum. Animals were housed in static cages with Bed-OCobs bedding inside Biobubble clean rooms that provide HEPA filtered air intothe bubble environment at 100 complete air changes per hour. All treatments andbody weight determinations were carried out in the bubble environment ata temperature of 21 C and a humidity range of 30e70%. Cubosome and hexosomesolutions were administered via tail vein injection into animals at a maximumconcentration of 365 mg/kg and monitored for abnormal effects. If a death occurreddue to nanoparticle toxicity, the nanoparticle solutions were diluted by 25% (v/v)and then to a 10% (v/v) dilution of stock and dosing was repeated in nave animalsuntil tolerated doses were reached for each compound.

2.10. In vivo imaging

Approval for this was obtained from the Florey Neuroscience Institutes animalethics committee. An adult male SpragueeDawley rat weighing 423 g was anes-thetized with 5% isoflurane in a 1:1 mixture of medical grade air and oxygen andprepared for surgery. Anesthesia was maintained throughout the procedure andsubsequent imaging with 1e2.5% isoflurane delivered through a nosecone placedover the animals snout. The femoral vein was exposed, cannulated and the incisionclosed with silk sutures while the animal was immediately prepared for MRI.

Anesthetized animals were laid supinely in an animal holder with respirationcontinuously monitored throughout the experiment using a pressure sensitiveprobe positioned over the rats diaphragm. The cradle was inserted into a transmit/receive coil fixed inside a BGA12S gradient set for imaging with a 4.7 Tesla BrukerBiospec 47/30 scanner. The scanning protocol consisted of a three-plane localizersequence followed bymulti-slice axial, coronal and sagittal scout images to ascertainthe orientation and position of the kidneys and liver. Two oblique slices wereselected; the first orientated to image the midline of both kidneys while the seconddissected the liver. T1-weighted imageswere acquired prior to, and after, injection ofcontrast enhancing hexosomes using a rapid acquisition, relaxation enhanced(RARE) sequence with the following imaging parameters: repetition time (TR)500 ms, RARE factor 4, effective echo time (TEeff) 15.63 ms, in-plane reso-lution 137137 mm, number of slices 2, slice thickness 2 mm, averages(NEX) 4 and a total scan time of 4 min and 16 s.

2.11. Intravenous administration of contrast agent

After acquiring a baseline scan, 600 mL of a 40 mg/mL solution of Myverolhexosomes containing 14.5 wt% nitroxide lipid in 0.1 M NaCl was intravenously(10 mL/s) injected followed by 0.3 mL of saline to flush the cannula line. This wasfollowed by a post injection scan, acquired immediately after administration of thecontrast agent.

3. Results and discussion

3.1. Self-assembly of the lyotropic liquid crystal nitroxidecontaining nanoparticles

The paramagnetic nitroxide lipid (Fig.1) capable of providingMRIcontrast used in this workwas synthesized using a fatty reactive acylchloride with 4-hydroxy TEMPO as described in the experimentalsection. The compoundwas easily chromatographically purified andthen incorporated into two lyotropic liquid crystal forming bulk lipidmaterials, namely phytantriol and Myverol. Cubosome and hex-osomedispersionswere easily generated from these bulk gels via theaddition of F127 in water acting as a steric stabilizer followed bysonication of the solution. Phytantriol-pluronic-nitroxide lipid andmyverol-pluronic-nitroxide lipid dispersions were prepared atvarious concentrations of nitroxide lipid from 0 to 14 wt % in a high-throughputmanner using a ChemspeedAccelerator SLTII robot. Anautomated formulation method allowed for the rapid productionand screening of nanoparticles to determine the effects of smallchanges in additive composition upon the lyotropic liquid crystalnanoparticle properties. The average particle size measured byPlease cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018dynamic light scattering (DLS, intensity vs. size) of all nanoparticleformulations used in this workwere found to be consistently around200 30 nm in dimension and have polydispersity indices ofbetween 0.13 and 0.18 (data not shown). The nanoparticle sizes andpolydispersity indices did not vary significantly between the twobulk lipids (phytantriol and Myverol) or incorporated nitroxideconcentration and no trend in size or polydispersity index wasobserved.

Cryo-TEM was further used to investigate the dispersed lyo-tropic liquid crystalline nanoparticles. The cryo-TEM images (Fig. 2)confirm the presence of discrete colloidal nanoparticles spanninga size range of 60 nm and 300 nm which was typical for thesesystems and was consistent with the DLS traces observed. Thephytantriol sample containing 2% nitroxide lipid shows nano-structured particles displaying a periodic internal structure typicalof cubic phase materials. As the percentage loading of nitroxidelipid was increased, distinctly different structures known as onionrings and multi-faceted nanoparticles were visible, typical ofhexagonal phase materials. This indicated that a cubic to hexagonalphase change occurred upon increased loading of nitroxide lipid.

To investigate this phase change in more detail and gain furtherinsight into the nanostructural changes occurring in these nano-particles from increasing the nitroxide lipid incorporation, SAXSwas performed on the different formulations. This techniqueble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

B.W. Muir et al. / Biomaterials xxx (2012) 1e11 5allowed a detailed investigation of the possible internal lamellar,cubic and hexagonal phase symmetries of the dispersed nano-particles produced from phytantriol and Myverol (Fig. 3). Thelattice parameters of cubic (Q2) and hexagonal (H2) phasesdetected (Fig. 3) were calculated from the scattering curves atdifferent loadings of the nitroxide lipid (data not shown). The SAXSpatterns of nanoparticles made from phytantriol (lattice parameter67.3 ) and Myverol (lattice parameter 86.3 ) without any addi-tion of nitroxide lipid indicated a cubic lattice symmetry witha Pn3m space group (double diamond) typical of these bulk lipids.As the concentration of nitroxide lipid increased, a Q2 to H2 tran-sition as indicated by a two phase region commencing at a nitro-xide lipid concentration of 4 wt% in phytantriol and 2 wt% inMyverol was observed (Fig. 3). The lattice parameter of both the Q2and H2 phases decreased with increasing concentration of nitro-xide lipid. The cubic and hexagonal lattice parameter of the78

80

82

84

86

88

0 2 4 6

Lattice p

aram

eter o

f Q

2 p

hase (A

)

Wt % of nitroxide li

two phase region

A

B

60

62

64

66

68

70

0 2 4 6

Lattice p

aram

eter o

f Q

2 p

hase (A

)

Wt % of nitroxide lip

two phase region two phase region

two phase region

Fig. 3. Synchrotron source SAXS lattice parameter () analysis and liquid crystal phase datconcentration. SAXS lattice parameter data of selected bulk gel compositions are shown as

Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018phytantriol nanoparticles was around 1e2 nm smaller than that ofthe Myverol nanoparticles. The nitroxide lipid appears to drive anincrease in curvature in both bulk lipids which results in the inversecubic (Q2) to inverse hexagonal (H2) phase change observed. Thisdecrease in lattice parameter with increasing additive loading(increasing curvature) has been observed for other systems andwas typical of aliphatic additives [32,35]. These results indicate thatthe nitroxide lipid was more readily incorporated into phytantriolthan Myverol. The robustness of phytantriol to retain its originalcubic phase in the presence of higher levels of additives whencompared to Myverol has been observed by us previously [32]. Insummary, the DLS, Cryo-TEM and SAXS data demonstrate thataddition of small amounts of the nitroxide lipid result in significantnanostructural changes (specifically a Q2 to H2 phase transitionand decrease in lattice parameter) within the lyotropic liquidcrystal nanoparticle population.56

58

60

62

64

66

8 1 0 1 2 1 4

Lattice p

aram

eter o

f H

2 p

hase (A

)

pid in Myverol

Q2 Phase

H2 phase

Q2 Phase

H2 phase

45

47

49

51

53

55

8 1 0 1 2 1 4

Lattice p

aram

eter o

f H

2 p

hase (A

)

id in Phytantriol

a of phytantriol (A) and Myverol (B) nanoparticles with increasing wt% nitroxide lipida comparison.

ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

B.W. Muir et al. / Biomaterials xxx (2012) 1e1163.2. Magnetic resonance relaxion rates and relaxivities

The nitroxide lipid in this work was paramagnetic because itcontains a stable unpaired electron. Previous work has shown thatliposomes with a nitroxide lipid incorporated into their structuremay induce a slight enhancement in MRI relaxation rates [17].However, to date there are limited reports on the observed relax-ivities of nitroxide nanoparticles. The use of these liposomal-basedmaterials has been limited in the past [36,37] due to the fact thenitroxyl group was rapidly reduced in vivo producing diamagneticN-hydroxy compounds that will not provide extended and effectivecontrast in vivo [17]. Therefore it was hypothesized that lyotropicmesophase nanoparticles containing nitroxides should be capableof also enhancing theMRI relaxation rateswhilst also protecting thenitroxide groups from oxidation in vivo for an extended period[38e40]. The use of a fatty nitroxide lipid inside a cubic or hexagonalphase nanoparticles, as reported here, was expected to showa significant increase in relaxivity compared to the free nitroxylgroup. It is postulated that this may be due to either or both rota-tional correlation constant effects from the reduced tumbling of the0

0.5

1

1.5

2

2.5

3

0 2 4 6

Wt% ni

Lo

ng

itu

din

al re

la

xa

tio

n ra

te

, R

1 (s

ec

-1

) Phytantriol

Myverol

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 2 4 6

Wt% nitr

Tran

sv

erse

re

lax

atio

n rate

, R

2 (se

c-1

) Phytantriol

Myverol

A

B

Fig. 4. Longitudinal (A) and Transverse (B) relaxation rates of phytantriol and

Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018nitroxide group inside the nanoparticle and water confinement/exchange effects in the bi-continuous water channels throughoutthe cubosomes. Previously [17], maximal in vivo contrast enhance-ment of the nanoparticles was achieved when these tumbling rateswere similar to the MRI Larmor precessional frequency.

The effect of the nitroxide lipid concentration on the R1 and R2relaxation rates in nanoparticle dispersions made with phytantrioland Myverol is shown in Fig. 4A and B respectively. In the phy-tantriol system, a rapid rise of R1 was observed from 0 to 1 wt%nitroxide lipid followed by a linear increase of R1 with increasingnitroxide lipid (Fig. 4A). The phytantriol system also demonstrateda possible plateau of R1 between 4 and 8 wt% nitroxide lipidloading which corresponds with the two phase region (Q2 and H2)as detected via SAXS. Above 8 wt% there was a further increase inrelaxivity with increasing nitroxide lipid content. In the Myverolsystem no significant increase in R1 was observed until a concen-tration of 4 wt% was reached. The rate of increase in R1 withnitroxide concentration in the Myverol systemwas less than that ofthe phytantriol system. The R2 relaxation rates observed in bothnanoparticle systems followed a similar trend as for the R1s as8 10 12 14

troxide lipid

8 10 12 14

oxide lipid

Myverol nanoparticles with increasing wt% nitroxide lipid concentration.

ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

Table 1Relaxivity values of selected phytantriol and Myverol self assembled lyotropic liquid crystal nanoparticles containing varying levels of a nitroxide lipid. The nanoparticleformulation liquid crystal phases as determined via SAXS are also reported.

Bulk lipid Nitroxide lipid concentration (wt%) r1 (mM 1s1) r2 (mM1 s1) r1/r2 Liquid crystal phase

Phytantriol 4.0 0.59 1.35 0.43 Cubic (Q2)Phytantriol 14.5 0.27 0.43 0.43 Hexagonal (H2)Myverol 2 0.14 0.32 0.44 Cubic (Q2)Myverol 14.5 0.08 0.17 0.49 Hexagonal (H2)

B.W. Muir et al. / Biomaterials xxx (2012) 1e11 7a function of the nitroxide lipid concentration (Fig. 4B). To furtherinvestigate the two bulk lipid systems, the r1 and r2 relaxivities(mM1 s1) of two select nitroxide containing samples (in the cubicand hexagonal phase region respectively) of both bulk lipid systemswere calculated (Table 1).

The relaxivity was calculated by measuring the gradient of thelinear best fit of the R1 and R2 versus nitroxide lipid concentration(mM) plots. The relaxivity of the cubic phase nanoparticles forboth bulk lipids was approximately double that of the hexagonalphase nanoparticles in each system. The highest r1 relaxivity(0.59 mM1 s1) was measured in the phytantriol system at a nitro-xide lipid concentration of 4 wt%. As a comparison the relaxivities ofsmall molecule nitroxyl contrast agents used in functional MRIimaging were on the order of 0.1 mM1 s1 (4.7 T) with half lives ofa few minutes [41]. In a study by Gussoni et al. poly(amidoamine)basedpolymers conjugatedwith amino-TEMPOrecorded relaxivitiesCHO-GFP cells

0

10

20

30

40

50

60

70

80

90

100

050

75

100

125

150

175

200

225

250

275

300

325

350

400

Myverol Lipid Concentration (ug/ml)

Ce

ll V

ia

bilit

y (%

) no Nitroxide lipid (Q2)2.1% Nitroxide lipid (Q2)14.6% Nitroxide lipid (H2)

CHO-GFP cells

0

10

20

30

40

50

60

70

80

90

100

0 510

20

40

60

80

Phytantriol Lipid Concentration (ug/ml)

Cell V

iab

ility (%

) no Nitroxide lipid (Q2)

3% Nitroxide lipid (Q2)

6% Nitroxide lipid (Q2 & H2)

9% Nitroxide lipid (H2)

14% Nitroxide lipid (H2)

A

B

Fig. 5. Results of Alamar blue cytotoxicity assay testing (expressed as a percentage of viabnanoparticle formulations on Green Fluorescent Protein expressing Chinese Hamster Ovardisplay the wt% of nitroxide lipid in the nanoparticle formulations and the nanoparticleH2Hexagonal). (For interpretation of the references to colour in this figure legend, the re

Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018similar to those reported here on the order of 0.4 mM1 s1 in vitrobut nodata onefficacy in vivohasbeen reportedof such systems [42].

It was quite striking that the cubic phase nanoparticles con-taining less nitroxide lipid (4 wt%) were significantly more effectiveat enhancing the relaxation rates compared to the hexagonal phasenanoparticles containing greater concentrations of nitroxide lipid(14.5 wt%). This indicates a strong effect of lyotropic mesophaseupon nanoparticle relaxivities. As the nanoparticles have verysimilar size distributions and polydispersities, rotational correla-tion constant effects should not be a significant factor in theenhanced relaxivity observed in the cubic phase nanoparticles.Therefore the nanostructure of the cubic phase nanoparticles mustbe playing an important role in enhancing the MRI relaxation rates.This observation was consistent with the plateau observed in thephytantriol nanoparticles R1 at nitroxide loadings between 4 and8 wt% where there was a cubic to hexagonal phase nanoparticle120

Phytantriol Lipid Concentration (ug/ml)

Ce

ll V

ia

bilit

y (%

)

HEK cells

0

20

40

60

80

100

0 510

20

40

60

80

no Nitroxide lipid (Q2)

3% Nitroxide lipid (Q2)

6% Nitroxide lipid (Q2 & H2)

9% Nitroxide lipid (H2)

14% Nitroxide lipid (H2)

HEK cells

0

10

20

30

40

50

60

70

80

90

100

050

75

100

125

150

175

200

225

250

275

300

325

350

400

Myverol Lipid Concentration (ug/ml)

Ce

ll V

iab

ility

(%

) no Nitroxide lipid (Q2)

2.1% Nitroxide lipid (Q2)14.6% Nitroxide lipid (H2)

le cells relative to a positive control) of nitroxide lipid (A: phytantriol & B: Myverol)y cells (CHO-GFP) and Human Embryonic Kidney cells (HEK293). The figure legendsliquid crystal phase determined via SAXS analysis is shown in brackets (Q2 Cubic,ader is referred to the web version of this article.)

ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

B.W. Muir et al. / Biomaterials xxx (2012) 1e118transition with significantly lower relaxivities. In the two phaseregion with increasing wt% nitroxide lipid more of the nano-particles being generated consist of the lower relaxivity hexagonalphase material. Therefore, as the proportion of higher relaxivitycubosomes decreases despite the increasing nitroxide loading, theaverage R1 in the system plateaus. Finally, as the total concentrationof nitroxides in the system increases significantly past 8 wt%nitroxide lipid, the relaxivity of the nanoparticles begins toincrease. The higher relaxivity of the cubic phase nanoparticles waslikely due in part to the greater surface area of the bi-continuouswater channels in these structures compared to the hexagonalphase nanoparticles.

The lattice parameter of the phytantriol cubosomes used in thisstudy was around 1e1.5 nm smaller than the Myverol cubosomesand the relaxivity of phytantriol cubosomes was significantlygreater. This may be due to a higher rate of water exchange in theinner and/or second coordination sphere of the nitroxide lipidheadgroup in the smaller water channels of the phytantriol cubicphase compared to the Myverol cubic phase. The diameter of thewater channels in the bulk cubic phase nanoparticles in phytantriolcan be theoretically calculated to be approximately 2.3 nmcompared to 4.3 nm in theMyverol system, based on themethod ofBriggs et al. [43]. In more recent work Ritman et al. confirmed thiscalculation experimentally using atomic force microscopy tomeasure the internal water channels in excess water of phytantrioland monoolein [44]. To summarise, the increased confinement ofwater inside the bi-continuous water channels of the bulk cubicphase of phytantriol may result in an increased relaxivity due tofaster exchange of the inner sphere water molecules with bulkwater, compared to the hexagonal phase water that was located incolumns. In addition, the smaller bi-continuous water channels inthe cubic phase of phytantriol compared to Myverol may create anenvironment that enhances lifetimes (decreased mobility) of innersphere water protons. Interestingly, this effect of increasing relax-ivity in confined systems has recently been reported for gadoliniumbased MRI contrast agents [45].Tim

0 100 200

DPM

/mL

10000

100000

Fig. 6. Plasma concentration (disintegrations per minute (DPM)) versus time pharmacoparticles after intravenous administration in rats. Circles denote dispersions prepared3H-dioleylphosphatidylcholine as a radioactive tracer.

Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.0183.3. Cytotoxicity and maximum tolerated dose study

The toxicity of lyotropic mesophase nanoparticles is uncleardespite the large number of in vitro and in vivo studies published,since only a limited number of studies have reported exper-imental data on the toxicity of such nanoparticles [46e49]. Prior totesting of lead compounds in animals, cytotoxicity testing wasperformed in vitro using Chinese hamster ovary green fluorescentprotein expressing (CHO-GFP) cells and human embryonic kidney(HEK293) cells. The results of an Alamar blue cytotoxicity assay onphytantriol and Myverol nanoparticles with increasing concentra-tions of nitroxide lipid are shown in Fig. 5A. Both cell linesdemonstrated that the phytantriol nanoparticles were toxic ata concentration of 20 mg/mL and were highly toxic above 40 mg/mL.The cytotoxicity of phytantriol based nanoparticles was observed toincreasewith nitroxide lipid concentration. Shen et al. [46] reporteda similar cytotoxicity of phytantriol containing cubosomes in L929fibroblast cells at concentrations greater than 40 mg/mL using anMTT assay. Lyotropic mesophase lipid nanoparticles may interactwith the lipidbilayerof cells and result inmembrane fusion and lipidexchange that can disrupt membrane integrity and cause cell lysis.

Whilst in vitro testing is considered the gold standard in theinitial phase of toxicity testing, it is also limited in its ability toaddress complex, living systems. Hence it was deemed necessary toperform in vivo testing based on the preliminary data to determinethe MTD. The static nature of in vitro tests and the addedcomplexity of the system once these nanoparticles were deliveredin vivo, not to mention their mode of delivery (oral, buccal, dermal,bolus), make it very difficult to ascertain from in vitro cell basedassays the maximum nanoparticle concentration that can safely bedelivered in vivo. A significant difficulty in assessing nanoparticletoxicity is the absence of a standard assay or cell line. As such weperformed MTD testing in rats on phytantriol based nanoparticlesand found that the MTD in rats of phytantriol cubosomes was350 mg/kg while for phytantriol cubosomes containing 4 wt%nitroxide lipid the MTD decreased to 80 mg/kg. Phytantriole (min)

300 400 500

Phytantriol + 8% PF127Myverol + 8% PF127Phytantriol + 15% PF127Myverol + 15% PF127

kinetic profiles for phytantriol (filled symbols) and Myverol (open symbols) nano-using 8 wt% F127 and triangles denote 15 wt% F127. Particles contained 0.5 mCi

ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

Fig. 7. False color T1-weighted MR images of a rat liver prior to, and after, intravenousinjection of Myverol hexosomes containing 14.5 wt % nitroxide lipid (left and rightimages, respectively). Signal intensity is normalized to the mean muscle tissue signal.Color bar shows signal intensity in arbitrary units. No contrast enhancement wasobserved in the kidneys (data not shown).

B.W. Muir et al. / Biomaterials xxx (2012) 1e11 9hexosomes containing more nitroxide lipid were not tested. Allsubsequent tests on animals were performed below this concen-tration (80 mg/kg). The increased level of toxicity in the cubosomesafter addition of the nitroxide lipid correlated qualitatively with theAlamar blue cytotoxicity assay data. The cytotoxicity of the Myverolnanoparticles in the Alamar blue assay was significantly less thanthat of the phytantriol nanoparticles. Cytotoxicity effects were notobserved until concentrations above 75 mg/mL were reached(Fig. 5B). Nanoparticles made from Myverol were also found to bebetter tolerated by the CHO-GFP cells compared to the HEK293cells. In addition, no toxic effects were observed in MTD testing ofnitroxide loaded Myverol nanoparticles up to the maximumconcentration tested (365 mg/kg).

3.4. Pharmacokinetic behavior of lyotropic liquid crystallinenanoparticles

To investigate the behavior after IV injection of nanoparticlesmade from Myverol and phytantriol a study of their pharmacoki-netic behavior (blood half life) was performed in rats. The bloodplasma profiles in Fig. 6 demonstrate the relative difference inpharmacokinetic behavior of Myverol and phytantriol nano-particles with two different levels of pluronic F127 stabiliser (8 and15 wt%). The phytantriol nanoparticles were rapidly removed fromcirculation on administration, whereas the Myverol nanoparticlesdemonstrated a significantly longer circulation profile with a halflife of approximately 90 min. At 2 h post injection there wasdouble the concentration of Myverol nanoparticles in circulationrelative to phytantriol nanoparticles. At longer times the concen-tration of nanoparticles in circulation converge, indicating thatsome residual nanoparticles were still in circulation, albeit at a lowlevel (

B.W. Muir et al. / Biomaterials xxx (2012) 1e1110heavymetal containing compounds such as gadolinium chelates. Inaddition, the in vitro and in vivo showed they had low cytotoxicityand were readily biodegradable.

4. Conclusions

In summary, T1 enhancing nitroxide containing lyotropic liquidcrystal nanoparticles have been produced that provide effectivein vivo MRI contrast enhancement without containing potentiallyhazardous heavy metals such as gadolinium. Phytantriol basednanoparticles were found to have poor pharmacokinetic behavior(low blood half life) when pluronic F127 was used as stabilizerwhen compared to Myverol nanoparticles. In addition, phytantriolbased nanoparticles were found to be significantly more toxic thanMyverol nanoparticles in vitro and in vivo. Unlike nitroxide loadedliposomes that have been previously reported, lyotropic mesophasenanoparticles loaded with nitroxide lipids were extremely effectiveat providing MRI contrast in vivo. Interestingly, incorporation ofa nitroxide lipid into the liquid crystal nanoparticle structureallowed for a greater enhancement of relaxivity when a Pn3m spacegroup cubic phase was present compared to the hexagonal phasethat appeared at higher nitroxide loadings. The Myverol basednanoparticles were well tolerated in cells and live animals, and thephysical properties of the nanoparticles can be tuned by carefulcontrol over additive composition.

Acknowledgements

The authors would like to acknowledge the work of CharlesRiver Discovery and Imaging Services for performing MTD testingand in particular Vinod Kaimal, Patrick McConville and Lisa Repke.The authors would also like to thank Charlotte Conn for usefuldiscussions and calculations on the size of the continuous waterchannels in the cubic phase nanoparticles and Adrian Hawley forhelp with SAXS analysis. This research was undertaken on theSAXS/WAXS beamline at the Australian Synchrotron, Victoria,Australia. This work was supported by Australian NHMRC Fellow-ships to GFE (#1003993) and BAM (#454790).

References

[1] Chertok B, David AE, Moffat BA, Yang VC. Substantiating in vivo magneticbrain tumor targeting of cationic iron oxide nanocarriers via adsorptivesurface masking. Biomaterials 2009;30(35):6780e7.

[2] Chertok B, Moffat BA, David AE, Yu F, Bergemann C, Ross BD, et al. Iron oxidenanoparticles as a drug delivery vehicle for MRI monitored magnetic targetingof brain tumors. Biomaterials 2008;29(4):487e96.

[3] Moffat BA, Reddy GR, McConville P, Hall DE, Chenevert TL, Kopelman RR, et al.A novel polyacrylamide magnetic nanoparticle contrast agent for molecularimaging using MRI. Mol Imaging 2003;2(4):324e32.

[4] Lin JJ, Chen JS, Huang SJ, Ko JH, Wang YM, Chen TL, et al. Folic acid-pluronicF127 magnetic nanoparticle clusters for combined targeting, diagnosis, andtherapy applications. Biomaterials 2009;30(28):5114e24.

[5] Reddy GR, Bhojani MS, McConville P, Moody J, Moffat BA, Hall DE, et al.Vascular targeted nanoparticles for imaging and treatment of brain tumors.Clin Cancer Res 2006;12(22):6677e86.

[6] Farokhzad OC, Cheng JJ, Teply BA, Sherifi I, Jon S, Kantoff PW, et al. Targetednanoparticle-aptamer bioconjugates for cancer chemotherapy in vivo. ProcNatl Acad Sci USA 2006;103(16):6315e20.

[7] Kono K, Nakashima S, Kokuryo D, Aoki I, Shimomoto H, Aoshima S, et al.Multi-functional liposomes having temperature-triggered release andmagnetic resonance imaging for tumor-specific chemotherapy. Biomaterials2011;32(5):1387e95.

[8] Shubayev VI, Pisanic TR, Jin SH. Magnetic nanoparticles for theragnostics. AdvDrug Del Rev 2009;61(6):467e77.

[9] Hahn PF, Saini S. Liver-specific MR imaging contrast agents. Radiol Clin NorthAm. 1998;36(2):287.

[10] Caravan P, Ellison JJ, McMurry TJ, Lauffer RB. Gadolinium(III) chelates as MRIcontrast agents: structure, dynamics, and applications. Chem Rev 1999;99(9):2293e352.

[11] Caravan P. Strategies for increasing the sensitivity of gadolinium based MRIcontrast agents. Chem Soc Rev 2006;35(6):512e23.Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018[12] Laurent S, Elst LV, Muller RN. Comparative study of the physicochemicalproperties of six clinical low molecular weight gadolinium contrast agents.Contrast Media Mol Imaging 2006;1(3):128e37.

[13] Sadowski EA, Bennett LK, Chan MR, Wentland AL, Garrett AL, Garrett RW,et al. Nephrogenic systemic fibrosis: risk factors and incidence estimation.Radiology 2007;243(1):148e57.

[14] Srinivas M, Cruz LJ, Bonetto F, Heerschap A, Figdor CG, de Vries IJM. Cus-tomizable, multi-functional fluorocarbon nanoparticles for quantitativein vivo imaging using F-19 MRI and optical imaging. Biomaterials 2010;31(27):7070e7.

[15] Ahrens ET, Flores R, Xu HY, Morel PA. In vivo imaging platform for trackingimmunotherapeutic cells. Nat Biotechnol 2005;23(8):983e7.

[16] Keana JFW, Pou S, Rosen GM. Nitroxides as potential contrast enhancingagents for MRI application: influence of structure on the rate of reduction byrat hepatocytes, whole liver homogenate, subcellular fractions, and ascorbate.Magn Reson Med 1987;5(6):525e36.

[17] Keana JF, Pou S. Nitroxide-doped liposomes containing entrapped oxidant: anapproach to the reduction problem of nitroxides as MRI contrast agents.Physiol Chem Phys Med NMR 1985;17(2):235e40.

[18] Hyodo F, Chuang KH, Goloshevsky AG, Sulima A, Griffiths GL, Mitchell JB, et al.Brain redox imaging using blood-brain barrier-permeable nitroxide MRIcontrast agent. J Cereb Blood Flow Metab 2008;28(6):1165e74.

[19] Conn CE, Panchagnula V, Weerawardena A, Waddington LJ, Kennedy DF,Drummond CJ. Lanthanide phytanates: liquid-crystalline phase behavior,colloidal particle dispersions, and potential as medical imaging agents.Langmuir 2010;26(9):6240e9.

[20] Moghaddam MJ, de Campo L, Waddington LJ, Drummond CJ. Chelatingphytanyl-EDTA amphiphiles: self-assembly and promise as contrast agents formedical imaging. Soft Matter 2010;6(23):5915e29.

[21] Mulder WJM, Strijkers GJ, van Tilborg GAF, Griffioen AW, Nicolay K. Lipid-based nanoparticles for contrast-enhanced MRI and molecular imaging. NMRBiomed 2006;19(1):142e64.

[22] Spicer PT. Progress in liquid crystalline dispersions: cubosomes. Curr OpinColloid Interface Sci 2005;10(5-6):274e9.

[23] Larsson K, Fontell K, Krog N. Structural relationships between lamellar, cubicand hexagonal phases in monoglyceride-water systems: possibility of cubicphases in biological systems. Chem Phys Lipids 1980;27:321e8.

[24] Hyde ST, Andersson S, Ericsson B, Larsson K. A cubic structure consisting ofa lipid bilayer forming an infinite periodic minimum surface of the gyroidtype in the glycerolmonooleat-water system. Z Kristallogr 1984;168:213e9.

[25] Gustafsson J, Ljusberg-Wahren H, Almgren M, Larsson K. Cubic lipid-waterphase dispersed into submicron particles. Langmuir 1996;12:4611e3.

[26] Andersson S, Jacob M, Lidin S, Larsson K. Structure of the cubosome-a closedlipid bilayer aggregate. Z Kristallogr 1995;210:315e8.

[27] Wilcox CS, Pearlman A. Chemistry and antihypertensive effects of tempol andother nitroxides. Pharmacol Rev 2008;60(4):418e69.

[28] Woldman YY, Semenov SV, Bobko AA, Kirilyuk IA, Polienko JF, Voinov MA,et al. Design of liposome-based pH sensitive nanospin probes: nano-sizedparticles with incorporated nitroxides. Analyst 2009;134(5):904e10.

[29] Swartz HM. Use of nitroxides to measure redox metabolism in cells andtissues. J Chem Soc, Faraday Trans 1987;83(1):191e202.

[30] Soule BP, Hyodo F, Matsumoto KI, Simone NL, Cook JA, Krishna MC, et al.Therapeutic and clinical applications of nitroxide compounds. Antioxid RedoxSignal 2007;9(10):1731e43.

[31] Drummond CJ, Fong C. Surfactant self-assembly objects as novel drug deliveryvehicles. Curr Opin Colloid Interface Sci 1999;4(6):449e56.

[32] Mulet X, Kennedy DF, Conn CE, Hawley A, Drummond CJ. High throughputpreparation and characterisation of amphiphilic nanostructured nano-particulate drug delivery vehicles. Int J Pharm 2010;395(1, 2):290e7.

[33] Seddon JM, Squires AM, Conn CE, Ces O, Heron AJ, Mulet X, et al. Pressure-jump X-ray studies of liquid crystal transitions in lipids. Philos Trans R SocLondon, Ser A 2006;364(1847):2635e55.

[34] Muir BW, Moffat BA, Harbour P, Coia G, Zhen GL, Waddington L, et al.Combinatorial discovery of novel amphiphilic polymers for the phase transferof magnetic nanoparticles. J Phys Chem C 2009;113(38):16615e24.

[35] Conn CE, Mulet X, Moghaddam MJ, Darmanin C, Waddington LJ, Sagnella SM,et al. Enhanced uptake of an integral membrane protein, the dopamine D2Lreceptor, by cubic nanostructured lipid nanoparticles doped with Ni(II)chelated EDTA amphiphiles. Soft Matter 2011;7(2):567e78.

[36] Glockner JF, Chan HC, Swartz HM. In vivo oximetry using a nitroxide-liposomesystem. Magn Reson Med 1991;20(1):123e33.

[37] Chan HC, Magin RL, Swartz HM. Delivery of nitroxide spin label to culturedcells by liposomes. Magn Reson Med 1988;8(2):160e70.

[38] Kinoshita Y, Yamada K, Yamasaki T, Mito F, Yamato M, Kosem N, et al. In vivoevaluation of novel nitroxyl radicals with reduction stability. Free Radic BiolMed 2010;49(11):1703e9.

[39] Khan N, Blinco JP, Bottle SE, Hosokawa K, Swartz HM, Micallef AS. The eval-uation of new and isotopically labeled isoindoline nitroxides and an aza-phenalene nitroxide for EPR oximetry. J Magn Reson 2011;211(2):170e7.

[40] Davis RM, Mitchell JB, Krishna MC. Nitroxides as cancer imaging agents. Anti-Cancer Agents Med Chem 2011;11(4):347e58.

[41] Matsumoto K. Utility decay rates of T1-weighted magnetic resonance imagingcontrast based on redox-sensitive paramagnetic nitroxyl contrast agents. BiolPharm Bull 2009;32(4):711e6.ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

B.W. Muir et al. / Biomaterials xxx (2012) 1e11 11[42] Gussoni M, Greco F, Ferruti P, Ranucci E, Ponti A, Zetta L. Poly(amidoamine)scarrying TEMPO residues for NMR imaging applications. New J Chem 2008;32(2):323e32.

[43] Briggs J, Chung H, Caffrey M. The temperature-composition phase diagramand mesophase structure characterization of the monoolein/water system.Journal De Physique Ii 1996;6(5):723e51.

[44] Rittman M, Frischherz M, Burgmann F, Hartley PG, Squires A. Direct visual-isation of lipid bilayer cubic phases using atomic force microscopy. Soft Matter2010;6(17):4058e61.

[45] Ananta JS, Godin B, Sethi R, Moriggi L, Liu XW, Serda RE, et al. Geometricalconfinement of gadolinium-based contrast agents in nanoporous particlesenhances T1 contrast. Nature Nanotech 2010;5(11):815e21.

[46] Shen HH, Crowston JG, Huber F, Saubern S, McLean KM, Hartley PG. Theinfluence of dipalmitoyl phosphatidylserine on phase behaviour of andcellular response to lyotropic liquid crystalline dispersions. Biomaterials2010;31(36):9473e81.

[47] Murgia S, Falchi AM, Mano M, Lampis S, Angius R, Carnerup AM, et al.Nanoparticles from lipid-based liquid crystals: emulsifier influence onmorphology and cytotoxicity. J Phys Chem B 2010;114(10):3518e25.Please cite this article in press as: Muir BW, et al., Metal-free and MRI visiBiomaterials (2012), doi:10.1016/j.biomaterials.2011.12.018[48] Gong XJ, Moghaddam MJ, Sagnella SM, Conn CE, Mulet X, Danon SJ, et al.Nanostructured self-assembly materials from neat and aqueous solutions ofC18 lipid pro-drug analogues of Capecitabine-a chemotherapy agent: focus onnanoparticulate cubosomes of the oleyl analogue. Soft Matter 2011;7(12):5764e76.

[49] Fong C, Weerawardena A, Sagnella SM, Mulet X, Krodkiewska I, Chong J, et al.Monodisperse nonionic isoprenoid-type hexahydrofarnesyl ethylene oxidesurfactants: high throughput lyotropic liquid crystalline phase determination.Langmuir 2011;27(6):2317e26.

[50] Dong Y-D, Larson I, Bames TJ, Prestidge CA, Boyd BJ. Adsorption of nonlamellarnanostructured liquid-crystalline particles to biorelevant surfaces forimproved delivery of bioactive compounds. ACS Appl Mater Interfaces 2011;3(5):1771e80.

[51] Kaminskas LM, McLeod VMH, Porter CJ, Boyd BJ. Differences in colloidalstructure of pegylated nanomaterials dictate the likelihood of acceleratedblood clearance. J Pharm Sci 2011;100(11):5069e77.

[52] Allen TM, Hansen C, Rutledge J. Liposomes with prolonged circulation times:factors affecting uptake by reticuloendothelial and other tissues. BiochimBiophys Acta 1989;981(1):27e35.ble theranostic lyotropic liquid crystal nitroxide-based nanoparticles,

Metal-free and MRI visible theranostic lyotropic liquid crystal nitroxide-based nanoparticles1. Introduction2. Materials and methods2.1. Self-assembly of the lyotropic liquid crystal nitroxide containing nanoparticles2.2. Nitroxide lipid synthesis2.3. Cryo-transmission electron microscopy (Cryo-TEM)2.4. Dynamic light scattering analysis2.5. Synchrotron SAXS measurements2.6. MRI relaxivity measurements2.7. Cytotoxicity testing of the lyotropic liquid crystal nitroxide containing nanoparticles2.8. Pharmacokinetics of phytantriol and Myverol-based lyotropic liquid crystalline nanoparticles2.9. Maximum tolerated dose testing of lyotropic liquid crystal nitroxide containing nanoparticles2.10. In vivo imaging2.11. Intravenous administration of contrast agent

3. Results and discussion3.1. Self-assembly of the lyotropic liquid crystal nitroxide containing nanoparticles3.2. Magnetic resonance relaxion rates and relaxivities3.3. Cytotoxicity and maximum tolerated dose study3.4. Pharmacokinetic behavior of lyotropic liquid crystalline nanoparticles3.5. In vivo MRI study

4. ConclusionsAcknowledgementsReferences