Review Article

Molecular Imaging and Therapy of AtherosclerosisWith Targeted Nanoparticles

Samuel A. Wickline, MD,1,2* Anne M. Neubauer,2 Patrick M. Winter, PhD,1,2

Shelton D. Caruthers, PhD,2,3 and Gregory M. Lanza, MD, PhD1,2

Advances in bionanotechnology are poised to impact thefield of cardiovascular diagnosis and therapy for decades tocome. This review seeks to illustrate selected examples ofnewly developed diagnostic and therapeutic nanosystemsthat have been evaluated in experimental atherosclerosis,thrombosis, and vascular biology. We review a variety ofnanotechnologies that are capable of detecting early car-diovascular pathology, as well as associated imaging ap-proaches and conjunctive strategies for site-targeted treat-ment with nanoparticle delivery systems.

Key Words: nanotechnology; contrast agents; imaging;drug therapy; atherosclerosisJ. Magn. Reson. Imaging 2007;25:667–680.© 2007 Wiley-Liss, Inc.

UNLIKE MANY OTHER DISEASES, atherosclerosisand/or vulnerable/unstable plaques are often diag-nosed only after an acute, sometimes fatal event. Of the�700,000 cardiac deaths reported per year in America,approximately 60% are “sudden deaths” that occurwithout any advance warning of pathology (1). Predict-ing if and when a plaque might rupture and cause acuteinfarction or sudden death is an uncertain business.Atherosclerotic plaques grow in discrete stages that in-volve repeated episodes of rupture, hemorrhage, throm-bosis, and healing, which lead inevitably to a final rup-ture event and complete vascular obstruction (2).Exposure of the lipid core, even through a small, local-ized rupture, can induce a clotting cascade through theinteraction of serum clotting factors with locally ex-

pressed tissue factor (TF) (3). The fibrin matrix andhemorrhagic components are incorporated into theplaque mass and extend the dimensions of the lesion,which means that vulnerable plaques will likely rupture(i.e., be unstable or disrupted) during their life cycle (4).The accumulation of macrophages and other inflamma-tory cells that secrete high levels of metalloproteinases(MMPs) also undermines the fibrous cap, potentiallyexposing the thrombotic lipid core (5,6). Up-regulationof angiogenesis can lead to erosion of the extracellularmatrix and replacement with physically fragile neovas-cular beds, weakening the fibrous cap and promotingplaque rupture (7,8). These and other cellular pro-cesses are suitable targets for molecular imaging andtargeted drug therapy.

Recent developments in the fields of cellular and mo-lecular imaging promise to allow noninvasive detectionof the molecular components of pathologic processes,such as image-based identification of specific mole-cules associated with inflammation or angiogenesis.Molecular and cellular imaging techniques are avail-able for most imaging modalities, including nuclear(9,10), optical (10,11), ultrasound (US) (12,13), andmagnetic resonance imaging (MRI) (14,15). These meth-ods are nondestructive in vivo analogs of traditionalimmunocytochemical techniques (16). We review a se-lection of the advanced imaging methods and new tar-geted nanoparticle contrast agents for early character-ization of atherosclerosis and cardiovascular pathologyat the cellular and molecular level that may representthe next frontier for combining imaging and rationaldrug delivery (14).

Investigators in the field of nanotechnology seek todevelop and combine new materials by precisely engi-neering atoms and molecules to yield new molecularassemblies on the scale of individual cells, organelles,or even smaller components. These components havebeen practically classified as ranging from 1 to 100 nm,although somewhat larger agents are often includedunder the rubric of “nanoparticle” because of their po-tential utility and in accordance with the traditionalscientific definition of the nanoscale regime (14,15,17).The specific organization of such nanoscale materials isanticipated to confer unique chemical and biologicalproperties based on interactions that occur at theirsurfaces. These materials may mimic or substitute for

1Department of Medicine, Washington University, St. Louis, Missouri,USA.2Department of Biomedical Engineering, Washington University, St.Louis, Missouri, USA.3Philips Medical Systems, Best, The Netherlands.Contract grant sponsor: NIH; Contract grant numbers: HL-42950; HL-59865; EB-01704; HL-073646; NO1-CO-07121; U54-CA-119342; Con-tract grant sponsors: Philips Medical Systems; Olin Foundation; Edithand Alan Wolf Charitable Trust.*Address reprint requests to: S.A.W., Washington University School ofMedicine, Campus Box 8086, 660 South Euclid Ave., St. Louis, MO63110. E-mail: saw@wuphys.wustl.eduReceived December 23, 2005; Accepted October 5, 2006.DOI 10.1002/jmri.20866Published online 8 March 2007 in Wiley InterScience (www.interscience.wiley.com).

JOURNAL OF MAGNETIC RESONANCE IMAGING 25:667–680 (2007)

© 2007 Wiley-Liss, Inc. 667

many existing features of cell behavior that alreadyoperate at the nanoscale level. The synthesis of suchmaterials may occur from a “top-down” approach en-tailing the miniaturization of existing microscopic ma-terials, or more likely from a “bottom-up” approachinvolving “self assembly” of molecules into reproducibleand well-defined nanoscale constructs (18). To achieveclinically effective cellular and molecular imaging, tar-geted nanoscale contrast agents must be designed toaccomplish a long circulating half-life, sensitive andselective binding to the epitope of interest, prominentcontrast-to-noise ratio (CNR) enhancement, acceptabletoxicity, ease of clinical use, and applicability with stan-dard commercially available imaging systems. The useof nanoparticles or targeted nanoemulsions as carriersis advantageous because large payloads of imaging ortherapeutic agents can be selectively delivered to thetissue site.

CONTRAST AGENTS AND IMAGINGMETHODOLOGIES

The array of nanomaterials that can be used as contrastagents for molecular imaging and drug delivery isbroad, and the following description is intended to beillustrative rather than comprehensive. Size consider-ations dictate both the mechanism and rate of clear-ance, as well as access to molecular targets. For exam-ple, for agents larger than 50 nm, intravascular targetsappear to be the most appropriate, whereas smallerparticles may penetrate the intact or leaky vascularendothelium by endothelial permeability and retention(EPR) mechanisms, and thus directly target tissues.“Functionalization” (or preparation of the particle sur-face for binding targeting ligands or drug delivery li-gands) can be achieved by various chemical means(e.g., provision of reactive moieties on the surface) andthrough avidin–biotin or electrostatic interactions (e.g.,DNA binding to cationic lipids in particle membranes).To improve particle stability and permit adequate cir-culation times, some types of particles (such as lipo-somes and or polymers) may require surface-compo-nent cross-linking to enhance structural integrity,and/or incorporation of polyethylene glycol to avoidimmediate sequestration by the reticuloendothelial sys-tem.

One can achieve multifunctional activity by incorpo-rating combinations of one or more targeting ligands,imaging agents, and/or drugs into the formulation si-multaneously. Materials can be covalently or nonco-valently linked to the particle surface, dissolved in thecoating (e.g., lipophilic drugs deposited in lipid mem-brane layers), or carried in the particle interiors forcellular deposition and activation. For most types ofparticles, the reticuloendothelial system is the predom-inant clearance pathway. Toxicity is influenced not onlyby chemical composition and dose of the agent, but alsoby additional factors such as its size and shape, and theroute of administration.

Targeting Ligands and Specificity

A variety of different types of targeting ligands can beutilized, including antibodies or antibody fragments,

small peptides, peptidomimetics, polysaccharides, andaptamers. To complex these elements with particles,one can functionalize the particle surfaces by includingstandard linking groups. Common methods for associ-ating targeting ligands with particles include 1) nonco-valent avidin–biotin interactions, 2) covalent complex-ation via reactive groups, or 3) nonspecific surfaceadsorption. Avidin–biotin interactions are extremelyuseful, high-affinity, noncovalent targeting systemsthat have been incorporated into many biological andanalytical systems and selected in vivo applications.Additionally, avidin, which has four independent bi-otin-binding sites, provides signal amplification andimproves detection sensitivity. One can employ avidin–biotin interactions to create a “one-step” system by per-forming the avidin–biotin conjugations in vitro prior toinjection. Investigators have used this approach to suc-cessfully target vascular epitopes in vivo (19,20).

For in vivo use, targeting ligands are preferably at-tached chemically to the contrast agent by a variety ofmethods depending on the nature of the particle sur-face (21). Conjugations may be performed before orafter the particle is created, depending on the ligandemployed and its tolerance to the chemical processingconditions required. Direct chemical conjugation of li-gands to proteinaceous agents often takes advantage ofnumerous amino groups (e.g., lysine) that are inher-ently present within the surface. Alternatively, func-tionally active chemical groups, such as pyridyldithio-propionate, maleimide, amino, and aldehyde, may beincorporated into the surface as chemical “hooks” forligand conjugation after the particles are formed. An-other common postprocessing approach is to activatesurface carboxylates with carbodiimide prior to addi-tion of the ligand. The selected covalent linking strategyis primarily determined by the chemical nature of theligand.

Multiple copies of ligands can be incorporated de-pending on the size of the particle, which serves toenhance avidity and target detectability by reducing theparticle dissociation rate and securing the agent at theintended site. Specificity is conferred by the targetingligand itself and generally should be in the nanomolarrange, although high-avidity agents may in part over-come this limitation by multivalent interactions. To en-sure high ligand binding integrity and maximize tar-geted particle avidity, flexible polymer spacer arms(e.g., polyethylene glycol or simple caproate bridges)can be inserted between an activated surface functionalgroup and the targeting ligand. These extensions can be10 nm or longer, and minimize interference of ligandbinding by particle surface interactions. Regardless ofthe targeting, it is clear that most such agents willexhibit a modicum of nonspecific targeting related ei-ther to nonspecific ligand attachment, the EPR effect, orsequestration in immature vasculature (angiogenesis).

Nanoparticles

Liposomes (50–700 nm, uni-or multilammelar vesiclesthat consist of lipid bilayer membranes surrounding anaqueous interior) have been approved to enhance theefficacy and safety of various drugs, such as doxorubi-

668 Wickline et al.

cin (e.g., Doxil™; ALZA Corp., Tibotec Therapeutics, NJ,USA). Applications of liposomal technology as molecu-lar imaging agents for both US and MRI have beenreported (20,22).

Emulsions, which are chemically distinct from lipo-somes, are oil-in-water-type mixtures that are stabi-lized with surfactants to maintain their size and shape.Perfluorocarbon core emulsions (200–400 nm) havebeen used for molecular imaging with MRI, US, fluores-cence, nuclear, and computed tomography imaging(CTI) (13–15,23). For example, when vast numbers ofparamagnetic gadolinium (Gd) complexes (�50,000)are incorporated onto emulsion particles, the possiblesignal enhancement for each binding site is magnifieddramatically, by a factor of �106 over conventionalparamagnetic extracellular contrast agents (24,25).Modified micellar particles, such as high-density li-poprotein (HDL) or low-density lipoprotein (LDL) parti-cles, have been utilized as molecular imaging agents forMRI (26,27).

Polymers (40–200 nm) are used in a wide variety offlexible “designer approaches” to construct molecularimaging agents and therapeutic delivery devices (18).They may be linear, branched, or globular, and com-prise single or multiple molecular components (copoly-mers). Their size and shape can be tightly controlled,and functionalization of their surface permits bindingof myriad targeting and therapeutic moieties for imag-ing, as well as drug and gene delivery. Polymers madefrom polyhydroxy acids, such as the copolymer of poly-(lactic acid) (PLA) and poly(D,L-lactide-co-glycolide)(PLGA), have been investigated for localized drug andgene delivery. Dendrimers, or cascade polymers, arehighly branched polymeric structures that are globularin configuration. Their cores are varied, and thebranches are sequentially assembled in covalent inter-actions that produce layers referred to as “generations.”Paramagnetic polyamidoamine (PAMAM) and diami-nobutane (DAB) dendrimers have been employed forMRI applications (28,29). The multivalent surface com-prises a number of functional sites that can undergoreactions to add drugs, imaging agents, and targetingligands.

Metallic particles, such as iron oxide nanoparticles(15–60 nm), generally comprise a class of superpara-magnetic agents that can be coated with dextran, phos-pholipids, or other compounds to inhibit aggregationand enhance stability for use as passive or active tar-geting agents. The iron in monocrystalline iron oxidenanoparticles (MIONs), small particles of iron oxide(SPIO, 50–500 nm), and ultrasmall particles of ironoxide (USPIO, 10–50 nm) produces strong local disrup-tions in the magnetic field of MRI scanners, which leadsto increased T2* relaxation and hence a decrease inimage intensity in areas with iron particle accumula-tion (“susceptibility” effects). These particles exhibit avery long circulating half-life (24� hours) and may besequestered by tissue macrophages. These propertieshave allowed dextran-coated USPIO nanoparticles to beemployed for passive targeted imaging of pathologicalinflammatory processes, such as unstable atheroscle-rotic plaques, by MRI (30). Alternatively, other similartypes of particles (e.g., cross-linked iron oxide (CLIO)

particles complexed with retroviral “tat” protein li-gands) have been used for localization and transcellulardeposition (31).

Other metal-based agents, such as gold shell nano-particles (�120 nm), have been used for both imagingand therapy (32–34). Carbon nanotubes and fullerenes(4 nm) have been utilized as particulate systems whosesurfaces also can be functionalized for tissue binding(35,36). Native fluorescent properties have been re-ported (37,38). Quantum dots (2–8 nm) are constructedfrom semiconductor materials (e.g., cadmium selenide)that manifest stable (nonquenching) fluorescent prop-erties at various wavelengths depending on the exactcomposition of the materials (39–42). These propertiesallow registration of multiple simultaneous signalsfrom distinct particles that bear unique spectra. Foruse in vivo, they must be coated with materials (poly-mers) that both allow solubilization and prevent leach-ing of the toxic heavy metals.

Synthetic viral capsid structures (e.g., cowpea virus)that self-assemble as protein cage particles were origi-nally developed by Allen et al (43). These types of nano-particles can be manipulated under certain chemicalconditions to create pores that permit encapsulation ofimaging agents or drugs. Approximately 180 bindingsites are available for Gd, as compared to 50,000 ormore for other active particle systems in which pay-loads are critical for performance (44). Calcium and Gdcompete for the same sites that exhibit approximatelymicromolar dissociation constants, which raises theissue of undesirable transmetallation potential and freeGd release.

“Smart” Imaging Agents

So-called “smart” contrast agents represent an effort todevelop selectively excitable agents that only show sig-nal upon some specific form of activation, such thatbackground signal does not interfere with detection ofthe molecular target. The problem is exemplified by theuse of iron oxide particulates with a serum half-life ofmany hours to days. This creates negative contrast inthe blood pool that can interfere with detection of vas-cular wall targets, such as macrophages that ingest theparticles. In this case, imaging generally must be per-formed more than 24 hours after contrast injection toallow blood-pool clearance in order to avoid confound-ing susceptibility artifacts produced by the circulatingagent.

Examples of the various “smart” agents available in-clude agents that 1) are selectively activated by targetedenzymic processes, 2) become apparent only throughselective spectral activation, 3) exhibit little appreciablesignal until they accumulate in high concentrations atspecific targets of interest, and 4) exhibit multifunction-ality to enhance the specificity of diagnosis. The firsttype was initially proposed by Louie et al (45) and in-volves uncaging a sequestered Gd ion by the action of�-galactosidase action on a galactopyranose cappingmolecule, which then permits free water access to thelanthanide, a tripling of proton relaxation, and detec-tion of the presence of a transfected lacZ gene thatproduces the enzyme. The second type includes chem-

Molecular Imaging With Nanoparticles 669

ical exchange saturation transfer (CEST) agents. Withthese agents, saturation of exchangeable protons usingalternative lanthanide chelates, such as europium, al-low bulk water signal suppression after selective exci-tation at proton resonances that are removed from thebulk water frequency (46). Another example is fluorineimaging, which allows selective spectral excitation ofvarious fluorine moieties incorporated into nanopar-ticles (47). The third type is represented by agents thathave weak signals in circulating blood but enhanceupon accumulation (e.g., multivalent targeting of para-magnetic ions, and polymerization of iron oxide com-pounds) at a selected site (15,48,49). The fourth type istypified by agents that can simultaneously produce sig-nals that are detectable by more than one imaging mo-dality (e.g., fluorescence and superparamagnetic effects(50) or selected combinations of nuclear, US, and para-magnetic effects (44)).

IMAGING MODALITIES AND METHODS

Nuclear/PET Imaging

The role of nanoparticles in imaging cardiovascular pa-thology is diverse and varies among the available imag-ing modalities (Table 1). In the case of nuclear (gamma/SPECT) imaging or positron emission tomography(PET), for example, the general approach has been toutilize very small tracer quantities of contrast agents(e.g., radionuclide-labeled antibodies, peptides, orsmall molecules) rather than large payload particles.However, the high sensitivity, unique spectral signa-tures of the tracer elements, and potential for localquantification of these tracers based on the emissioncount rate enhance their role in early detection andserial evaluation of pathology. For example, folate re-ceptor-targeted polymeric shell cross-linked nanopar-ticles containing 64Cu were recently used for PET im-aging of tumors (51). More typical approaches for

characterizing atherosclerosis include imaging of apo-ptosis by annexin-phosphatidyl serine targeting (52),unstable carotid plaque imaging with metabolic (FDG)readouts (53), and macrophage chemotaxis imaging (9).

Optical Imaging

Optical approaches are promising, especially for morelocalized detection of pathologies. Quantum dots thathome to vascular endothelial targets have been re-ported to be useful for identifying selected tissue zip-codes (39,40). Gold nanoshells directly injected intotumors have been used to evaluate and treat tumorswith application of exogenous thermal energy (32–34).Carbon nanotubes also emit detectable fluorescencethat varies depending on the composition of the localenvironment (35–38). In some cases, multifunctionalityhas been designed into nanoparticles for combined im-aging (54). Tissue autofluorescence can obscure diag-nosis at some wavelengths; however, near-infraredwavelengths appear to be useful for enhanced sensitiv-ity and specificity in this regard (55,56). The lack oflarger-scale noninvasive imaging systems for patientsis a drawback.

CTI

Given the large installed base of imaging equipmentworldwide and the ease of use of CTI and US, the incor-poration of nanoparticle-based imaging agents in thesetechniques hypothetically shows great potential. A ma-jor challenge for CT researchers is to produce robustand potent contrast agents. Because X-ray absorptiondepends directly on the potency of the material used ascontrast agent (the “Z” value) and exponentially on thethickness of the layer of material deposited, sensitivityis expected to be only modest for nanoparticles thataccumulate in small concentrations when used as mo-lecular imaging tools. Nevertheless, lipid emulsion

Table 1Examples of Nanosystems Reported for Cardiovascular Imaging and/or Therapy

Nanoparticle class andsize

Imaging modalityCardiovascular target Localization Therapeutics

US MRI Optical Nuc/PET CT

SPIO (50-500 nm) SP Stem cell labeling P DoxorubicinUSPIO (5-50 nm) SP Macrophages, ischemic

stroke,P

CLIO (40 nm) SP NIR Macrophages, VCAM,selectins, stem celllabeling, apoptosis

A, P

PFC emulsions (200-300 nm)

x PM, fluorine F Gamma x Plaque angiogenesis,tissue factor,Integrins, fibrinthrombi, matrixcollagen, stem celllabeling

A Doxorubocin,paclitaxel,fumagillan,ultrasoundactivation

Liposomes (50-300 nm) x PM F Angiogenesis, adhesionmolecules

A, P Doxorubicin, paclitaxelsteroids,methotrexate

Micelles (50-150 nm) PM Plaque targeting A, P

SPIO � superparamagnetic iron oxide particles, SP � superparamagnetic, USPIO � ultrasmall superparamagnetic iron oxides, PM �paramagnetic, CLIO � cross-linked iron oxides, NIR � near infrared fluorescent, PFC � perfluorocarbon, F � fluorescent, HDL � highdensity lipoproteins, Nuc � nuclear imaging, A � active targeting, LDL � low density lipoproteins, P � passive targeting.

670 Wickline et al.

nanoparticles that contain radio-opaque iodinated tri-glycerides have been described for passive hepatic tar-geting (57), and work in this field is ongoing (58). Anunavoidable drawback for serial use remains the con-siderable radiation dose required.

US Imaging

US offers many advantages, such as benign imagingenergy (i.e., compressional waves), flexibility, highthroughput, low cost, and excellent patient tolerance.However, it is more highly operator-dependent thanother tomographic methods and cannot image all areasof the body. Available US contrast agents include siz-able shell-stabilized gas-filled microbubbles, and manyhave been developed that can be targeted to vascularepitopes (57). In the regime of smaller particles, acous-tically active emulsion nanoparticles for both imagingand therapy, and reflective liposomes for imaging(12,59,60) have received the most attention (Fig. 1).Although US is exquisitely sensitive for detecting micro-bubbles, it is less so for nanoparticles because of thesize dependency (r6) and relative incompressibility ofliquid particles, which preclude the use of availableharmonic resonance-based imaging techniques thatare typically applied to microbubble detection. A bal-ancing consideration is the unique potential for pre-cisely depositing large amounts of highly focused en-ergy in a convenient manner that can facilitatenanoparticle-based imaging and therapeutics with ex-ogenous US activation, as was recently described byseveral groups (61–63). The advent of mathematicalmodels to characterize the fundamental scattering be-havior from newer classes of nanoparticles (e.g., emul-

sions) raises the potential for extracting more quantita-tive information from the reflected signals (64,65).

MRI and MRS

MRI offers several advantages over other modalities,including high resolution, high anatomical contrast,high signal-to-noise ratios (SNRs), widespread clinicalavailability, and lack of ionizing radiation (15,66). How-ever, the comparatively modest MR contrast enhance-ment achievable with targeted contrast agents for mo-lecular imaging necessitates the delivery of higherpayloads of contrast materials, which can be providedby novel nanotechnologies. Because molecular epitopesof interest may reside on or inside cells in very sparsequantities at low nanomolar or picomolar concentra-tions, one can considerably amplify the local contrasteffect by incorporating large amounts of paramagneticor superparamagnetic materials as the payload. Forparamagnetic agents (e.g., in T1-weighted imaging), thesimple attachment of a few Gd atoms to an antibody foruse as a targeting ligand may not provide enough signalif micromolar concentrations of the lanthanide are re-quired to elicit contrast enhancement based on conven-tional T1 relaxation mechanisms (see below) (67). In thecase of T1-weighted imaging, the surface of the particlecan be decorated with numerous copies of Gd chelates(up to 100,000) to achieve the micromolar concentra-tions required per voxel (Fig. 2). Imaging with super-paramagnetic agents takes advantage of the fact thatenough material can be packed into the core of thenanoparticle to exert a prominent T2* effect and pro-duce a localized signal reduction that can be detectedwith potentially greater sensitivity than is possible with

Figure 1. US molecular imaging with nanoparticles. a: Fibrin-targeted nanoparticles enhance contrast in thrombi formed in thecarotid arteries of pigs with the use of clinical 7.5-MHz linear-array transducers. Top: Carotid artery lumen with an echogenicanode (arrowhead) to induce fibrin-platelet thrombus, which remains invisible at 7.5 MHz. Bottom: After fibrin-targetednanoparticles bind to the thrombus, backscatter is augmented (brighter) throughout and along the extent of the clot (multiplearrows) (reprinted from Lanza GM, Wallace KD, Scott MJ, et al. A novel site-targeted ultrasonic contrast agent with broadbiomedical application. Circulation 1996;95:3334–3340, with permission). b: TF imaging after balloon injury to a porcine carotidartery. Top: Scanning electron micrograph of TF expression in vitro on smooth muscle cells targeted with nanoparticlescontaining mAb ligands to TF. Bottom: 30-MHz intravascular ultrasound imaging of TF induced in medial smooth muscle cellsby balloon stretch injury. Note the contrast enhancement (brighter) heterogeneously distributed throughout the media of thevessel representing binding of TF-targeted nanoparticles (see arrows in targeted site) (reprinted from Lanza GM, AbendscheinDR, Hall CS, et al. In vivo molecular imaging of stretch-induced tissue factor in carotid arteries with ligand-targeted nanopar-ticles. J Am Soc Echocardiography 2000;13:608–614, with permission). c: Fibrin-targeted echogenic liposomes binding to andenhancing contrast (brighter) from LV thrombi in four-chamber (top) and parasternal (bottom) views (left: preinjection; right:postinjection) (reprinted from Hamilton A, Huang SL, Warnick D, et al. Left ventricular thrombus enhancement after intravenousinjection of echogenic immunoliposomes: studies in a new experimental model. Circulation 2002;105:2772–2778, with permis-sion).

Molecular Imaging With Nanoparticles 671

paramagnetic agents. Recently, quantitative ap-proaches have been described for molecularly targetedparamagnetic emulsions that allow the concentrationof bound nanoparticles to be computed under certaincircumstances (68).

Recent advances in imaging techniques have enabledhot-spot detection of iron oxide-based particles. Ex-ploiting the same inherent dipole of magnetic particlesthat causes signal dropout on typical MRI, Cunning-ham et al (69) and Stuber et al (70) developed tech-niques for off-resonance imaging that can insteadproduce bright signals in regions surrounding the ac-cumulation of particles. These techniques require spe-cialized excitation pulses that image the water mole-cules in close proximity to an accumulation of particles.Once optimized, these techniques offer potential for lo-calizing sources of extraneous magnetic dipoles (i.e.,superparamagnetic particles) and, via their signal in-

tensity, tracking their size and distribution and provid-ing a method for relative quantification.

Alternatively, the signal generated by the fluorine at-oms in the perfluorocarbon core of perfluorocarbon-based nanoparticles has been introduced as a uniquesignature for molecular MRI (16,71). Because biologicaltissues contain little endogenous fluorine, measure-ment of the fluorine component of targeted particlesmay allow definitive confirmation of nanoparticle dep-osition at the site. This recent approach has been dem-onstrated for imaging and spectroscopy of fibrin at4.7T, and for quantifying the concentration of nanopar-ticle binding to a selected site based on localized fluo-rine spectroscopy (16) and subsequently for stem cellimaging (see below) in vivo (72).

Another use for fluorine imaging and spectroscopy ofperfluorocarbon particles is to identify different tar-geted moieties on the same sample. Due to differences

Figure 3. VCAM-1 imaging in apoE –/– mice with peptide-targeted CLIO nanoparticles (reprinted from Kelly KA, Allport JR,Tsourkas A, Shinde-Patil VR, Josephson L, Weissleder R. Detection of vascular adhesion molecule-1 expression using a novelmultimodal nanoparticle. Circ Res 2005;96:327–336, with permission). a: Twenty-four hours after injection, MRI signal lossoccurs where the nanoparticles localize to aortic plaque in vivo (arrows). Note the loss of signal resulting from the susceptibilityeffect of accumulated iron oxide particles. (b) Before and (c) 24 hours after injection of targeted nanoparticles showing aorticcross sections with signal loss (darker) at the plaque cap. d: Ex vivo MR image of signal loss (darker) in the aorta afternanoparticle binding. e: Matched epifluorescent image of dual-function particles containing fluorescent probe.

Figure 2. MR image of thrombi with paramagnetic nanoparticles targeted to fibrin. a: Thrombus formed in vivo in a caninejugular vein imaged at 1.5T. Note the “hot spot” at the site of nanoparticle binding. b: “Disrupted” carotid endarterectomyspecimens incubated with fibrin-targeted nanoparticles binding to small amounts of fibrin at the shoulder regions (note the “hotspots” indicated by yellow arrows) of ruptured plaque cap imaged ex vivo at 1.5T (reprinted from Flacke S, Fischer S, Scott MJ,et al. Novel MRI contrast agent for molecular imaging of fibrin: implications for detecting vulnerable plaques. Circulation2001;104:1280–1285, with permission).

672 Wickline et al.

in their local nuclear environments (e.g., electronshielding, J-coupling, etc.), different fluorine atoms res-onate at slightly different frequencies and thus are of-ten readily distinguishable on an NMR spectrum (73).This means that by using perfluorocarbon nanopar-ticles formulated with different perfluorocarbon spe-cies, one can target them to the same sample and quan-tify their presence separately with one spectroscopicscan (16). While spectroscopy is ultimately the mostuseful for quantification, other techniques allow imag-ing of the different perfluorocarbon particles as well.Such techniques include frequency-selective excitationso that only the perfluorocarbon species of interest pro-duces a signal, and other forms of chemical shift imag-ing (CSI) that have been developed to differentiate fatfrom water in clinical imaging (74,75).

The sensitivity of MRI for detecting paramagnetic orsuperparamagnetic nanoparticulate imaging agentsdepends on the specific field strength, pulse sequence,coil sensitivities, epitope prevalence, and contrastagent concentration used. In general, the MR signalstrength is modest compared to nuclear imaging appli-cations. For example, at clinical imaging fieldstrengths, micromolar concentrations of paramagneticagents (e.g., Gd and CEST agents) are required. Forsuperparamagnetic agents that elicit susceptibility ar-tifacts, nanomolar concentrations may suffice. CNRs of5 or better generally produce readily identifiable (diag-nostic) qualitative signal enhancement. Examples ofoptimizing pulse sequences and paramagnetic and flu-orinated nanoparticle concentrations to enhance sen-sitivity have been presented by Morawski et al (16,68)with MR signal modeling approaches.

IMAGING APPLICATIONS

Atherosclerotic Plaques

Rupturing atherosclerotic plaques are frequently man-ifested at various stages in arteries with only modest(40–60%) stenosis (76), and they remain diagnosticallyelusive with routine clinical imaging techniques. Serumbiomarkers may offer information about the generalstate of the vasculature, but provide no useful informa-tion about the propensity for any given lesion to rup-ture. Accordingly, one major motivation for molecularimaging is the recognition and localization of telltalemolecular elements of unstable or disrupted plaquesthat might provide a window of opportunity extendingfrom days to weeks or months to intervene before moreserious clinical sequelae ensue (2). A sine qua non ofdisrupted plaque is fibrin deposition. Not only is fibrindeposition one of the earliest signs of plaque rupture orerosion, it also accounts for (along with intraplaquehemorrhage) a considerable part of the core of growinglesions (77). The ability to diagnose disrupted plaque bydetecting small deposits of fibrin in erosions or micro-fractures could allow characterization of a potential“culprit” lesion before a high-grade stenosis has beenformed that is detectable by cardiac catheterization.

The possibility of performing nanoparticle targetedfibrin imaging with US or paramagnetic MR contrastagents was first demonstrated by Lanza and Wickline

(12) and Lanza et al (78) as early as 1996. In this case,the ligand comprises an antibody fragment that ishighly specific for certain cross-linked fibrin peptidedomains, which can be complexed to the particle eitherthrough avidin–biotin linkages, or covalently to thefunctionalized nanoparticle, as has been shown for TFtargeting (71,79). For US imaging, thrombi formed insitu in canine carotid arteries were detectable within 30minutes with commercially available 7.5 MHz linear-array imaging transducers (Fig. 1) (78).

TF is a prothrombotic transmembrane glycoproteinthat is expressed within plaques, is upregulated follow-ing vascular injury or stent placement, and contributesas a mitogen to restenosis (80). TF in the core of plaquesis exposed during plaque rupture and is the proximatecause of local thrombosis that leads to vessel occlusionor distal embolization. TF imaging has been demon-strated in vivo for molecular imaging with US (Fig. 1)and in vitro with MRI (12,68). These observations illus-trate the potential use of the first reported molecularimaging agent for US to delineate molecules that areinvolved in plaque instability and restenosis. In fact,the ability to image TF-targeted paramagnetic nanopar-ticles bound to smooth muscle cell monolayers in cellculture at 1.5T attests to the potency of nanoparticlesagents that carry 50,000 or more Gd chelates.

Echogenic liposomes (ELIPS), in contrast to nanopar-ticles or emulsions, are composed of alternating layersof aqueous fluid and lipid bilayers that are formulatedto produce a US signal. Hamilton et al (59,81) usedthese liposomes to target thrombi (Fig. 1) and variousvascular signatures associated with atheroma develop-ment in injured vessels of miniswine for intravascularUS imaging (22). By targeting intercellular adhesionmolecule-1 (ICAM-1), vascular cell adhesion molecule-1(VCAM-1), fibrin, fibrinogen, and TF, they were able toproduce targeted enhancement in the vessel walls fiveminutes after intravenous administration of the lipo-somes.

In MRI studies, perfluorocarbon particles loaded with50–90,000 Gd atoms per particle yielded a substantialamplification of signal from fibrin clots at 1.5T both invitro and in vivo (15,64). Furthermore, the detection ofdisrupted plaque was illustrated in actual human ca-rotid endarterectomy specimens obtained from patientssymptomatic with transient ischemic attacks, stroke,or bruits (Fig. 2) (24). These data provided early evi-dence that disrupted dangerous atheroscleroticplaques can be detected noninvasively with MRI. MRI ofVCAM-1 was recently performed with the use of pep-tide-targeted superparamagnetic nanoparticles in theaortas of cholesterol-fed ApoE null mice by Kelly et al(54) (Fig. 3), indicating that early molecular mecha-nisms that are important in the evolution of atheroscle-rosis can be defined noninvasively with MRI.

EPIX Pharmaceuticals (Cambridge, MA, USA) utilizedphage display methods to produce a peptide ligand spe-cific for fibrin (EP-2104R), which may be useful forimaging thrombi in various body locations such as theleft atrium, pulmonary arteries, and coronary arteriesin experimental preparations (82–84). It contains fourGd-DTPA chelates per peptide moiety and thus pro-vides signal enhancement on an MR image. Despite the

Molecular Imaging With Nanoparticles 673

low Gd load per binding site, the excess of fibrinepitopes in fresh or chronic clots allows the accumula-tion of contrast agent concentrations that are sufficientto achieve micromolar levels of the lanthanide, whichenables ready detection of the clot after the signal in theblood pool is sufficiently decreased (one to two hours).

Alternatively, the fluorine component of perfluorocar-bon-based nanoparticles can be utilized to advantagefor molecular imaging. Our group originally demon-strated the concept of targeted nanoparticle fluorineimaging at 1.5T or 4.7T for detection of experimentalthrombi or small fibrin deposits in disrupted humancarotid arteries using particles made with perfluoroc-tylbromide (PFOB), crown ether (CE), or other PFC corematerials (Fig. 4) (16,71). We recently demonstrated theuse of rapid steady-state free precession (SSFP) 19Fimaging of combined PFOB and CE fibrin-targetednanoparticles on fibrin clots and endarterectomy spec-imens in vitro at 1.5T (47). This approach uses spectral-selective excitation to delineate various types of nano-particles that contain different perfluorocarbon cores,

and produces no appreciable confounding backgroundsignal.

Angiogenesis

While fibrin and TF can be utilized to delineate unstablecardiovascular diseases, the ���3-integrin is a generalmarker of angiogenesis and plays an important role in awide variety of disease states, including atherosclerosis(85) and cancer. The ���3-integrin is a well-character-ized heterodimeric adhesion molecule that is widely ex-pressed by endothelial cells, monocytes, fibroblasts,and vascular smooth muscle cells. In particular, ���3-integrin plays a critical part in smooth muscle cell mi-gration and cellular adhesion (86,87), both of which arerequired for the formation of new blood vessels. The���3-integrin is expressed on the luminal surface ofactivated endothelial cells, but not on mature quiescentcells (88). We have demonstrated the utility of ���3-integrin targeted nanoparticles for the detection andcharacterization of angiogenesis associated with

Figure 4. Fluorine imaging with fibrin-targeted perfluorocarbon-based nanoparticles (reprinted from Morawski AM, Winter PM,Yu X, et al. Quantitative “magnetic resonance immunohistochemistry” with ligand-targeted 19F nanoparticles. Magn Reson Med2004;52:1255–1262, with permission). a: Optical image of an excised human carotid endarterectomy sample shows asymmet-rical plaque distribution, with areas of intimal fat deposition (yellow). b: Matched fluorine projection image at 4.7T showsheterogeneous nanoparticle binding where fibrin is present (bright signal with no background) on the plaque. c: With NMRspectroscopy the fluorine image can be converted into a false color map of the nanoparticle binding, and coregistered andoverlaid on the proton image (shown in grayscale), corresponding to the intravoxel concentration of nanoparticles bound to fibrinepitopes. d: Quantitative false color mapping scale illustrating quantification of signal enhancement as a nanoparticle concen-tration expressed in nanomolar per voxel.

Figure 5. Detection of plaque neovascularization in cholesterol-fed rabbits (reprinted from Winter PM, Morawski AM, CaruthersSD, et al. Molecular imaging of angiogenesis in early-stage atherosclerosis with alpha(v)beta(3)-Integrin-targeted nanoparticles.Circulation 2003;108:2270–2274, with permission). a: Aortic cross sections imaged at 1.5T with �v�3-integrin targeted nano-particles. Note the heterogeneous distribution in aortic cross sections (false colored contrast enhancement), but little enhance-ment in nontargeted rabbits (�v�3-) or rabbits on a standard diet (Chol-). b: MRI signal modeling illustrates that picomolar (�100pM) intravoxel concentrations of perfluorocarbon-based paramagnetic nanoparticles are required to achieve a diagnostic CNRratio of �5 (blue dashed line). c: Immunochemical staining for �v�3-integrin at the media-adventitia border of aortic segments.Note the abundant red-brown vascular segments indicating the presence of integrin.

674 Wickline et al.

growth factor expression (19), tumor growth (89,90),and atherosclerosis (91).

Angiogenesis plays a critical role in plaque growthand rupture (92,93). In regions of atherosclerotic le-sions, angiogenic vessels proliferate from the vasa va-sorum to meet the high metabolic demands of plaquegrowth (94,95). Inflammatory cells within the lesionstimulate angiogenesis through local molecular signal-ing, which in turn promotes neovascular growth andprovides an avenue for more inflammatory cells to enterthe plaque (92).

Molecular imaging of expanded vasa vasorum in ath-erosclerotic lesions in cholesterol-fed rabbits was firstdemonstrated for MRI by Winter et al (91) with the useof paramagnetic nanoparticles targeted to ���3-integrinexpressing endothelial cells (Fig. 5). Animals on a con-trol diet exhibited no increased signal, and backgroundwas minimal. Expression of ���3-integrins in the adven-

titial layer and beyond was confirmed by co-localizedhistological staining of ���3-integrin and PECAM, ageneral endothelial marker. This work was the first todemonstrate the potential of MRI for the noninvasivedetection and quantification of angiogenesis in athero-sclerotic plaque.

Cyrus et al (96) recently employed ���3-integrin tar-geted and collagen-III targeted paramagnetic nanopar-ticles to image arteries subjected to balloon stretch in-jury (angioplasty). They observed a high degree ofbinding associated with the exposure of native smoothmuscle cell integrins after injury, and the upregulationof integrins as a component of the inflammatory re-sponse. The molecular targeting extended far beyondthe length of the actual balloon itself, indicating thepotential for extensive yet occult injury beyond the con-fines of angioplasty and potential stent deployment,which may have consequences for restenosis. Both col-

Figure 6. Macrophage imaging in cholesterol-fed rabbits with untargeted superparamagnetic nanoparticles (reprinted fromRuehm SG, Corot C, Vogt P, Kolb S, Debatin JF. Magnetic resonance imaging of atherosclerotic plaque with ultrasmall superpara-magnetic particles of iron oxide in hyperlipidemic rabbits. Circulation 2001;103:415–422). a: SPIO nanoparticles taken up by plaquemacrophages depict atherosclerosis in cholesterol-fed rabbits according to magnetic susceptibility effects (dark regions) that areseen distributed along the aorta in T2-weighted images acquired more than 24 hours after injection. b: Iron stain of plaquedemonstrating uptake of particles by intimal macrophages (blue stain).

-

Figure 7. Stem-cell labeling and imaging with iron-oxide nanoparticles (reprinted from Kraitchman DL, Heldman AW, Atalar E,et al. In vivo magnetic resonance imaging of mesenchymal stem cells in myocardial infarction. Circulation 2003;107:2290–2293,with permission). Left: Stem cells incubated with nontargeted iron oxide nanoparticles that undergo endocytosis are injectednear the apex in a porcine heart and imaged at 1.5T. Note the dark spot due to the susceptibility effect (arrow). Right: Iron stain(blue) of heart tissue illustrating stem cells containing abundant nanoparticles in the cytoplasmic compartment.

Molecular Imaging With Nanoparticles 675

lagen-III and ���3-integrins were detectable specifically,although the MRI integrin signal exceeded that of thecollagen signal.

Other Plaque Components

Macrophage imaging was first performed with nontar-geted USPIO by Schmitz et al (30) in Watanabe rabbits,and by Rheum et al (97) in cholesterol-fed atheroscle-rotic rabbits (Fig. 6). These are among the first illustra-tions of passive targeting of important early cell typesinvolved in atherosclerosis. Because macrophages areabundant in plaques throughout the vascular tree, andthey are known to ingest particulate matter, the use ofsuperparamagnetic agents to delineate macrophagesand foam cells has been pursued in both animal modelsand clinical trials (98). The demonstration of macro-phage targeting in vivo in rabbits required a waitingperiod of one to three days to allow for both passiveuptake of sufficient numbers of particles and blood-stream clearance of the long-circulating particles. Ingeneral, the susceptibility artifacts produced extendedbeyond the confines of the plaque macrophages andappeared as heterogeneously distributed signal voidsup and down the aorta.

In similar clinical trials conducted by Kooi et al (99)and Trivedi et al (100) with patients undergoing carotidendarterectomy, USPIO particles accumulated in themacrophages in plaques and were optimally imaged assignal reductions 24 hours after injection. Kooi et al(99) also noted that more contrast change was observedfor ruptured than for stable plaques. USPIO-labeledmacrophages have been imaged and localized to unsta-ble and ruptured plaques (75% demonstrating uptake),but not in stable lesions (only 7% showing USPIO up-take).

Frias et al (26) recently reported the development ofrecombinant paramagnetic HDL-like particles that canenhance atherosclerotic regions in apoE-deficient mice.These particles are formed through the delipidation ofnormal isolated human HDL particles, followed by re-constitution with phospholipids and the addition of aphospholipid-based conjugate of Gd-DTPA (15–20 mol-ecules of Gd included in each 9-nm particle) for signalenhancement. Nonselective accumulation in athero-sclerosis has been demonstrated.

This group also demonstrated the use of conventionalnontargeted agents, such as gadofluorine, that appearto preferentially label the fatty cores of plaques (101).Gadofluorine is a lipophilic chelate of Gd (Gd-DO3Aderivative) with a fluorinated side-chain that forms5-nm-sized micelles in aqueous solution. The small sizeand lipophilic nature of this contrast agent allows it toaccumulate in lipid-rich areas of plaque in cholesterol-fed rabbits.

The Weissleder group (102) has developed other MRIsusceptibility agents to image plaque components. “Mag-netic switches” (iron oxide particles) that contain numer-ous copies of high-affinity ligands have been used to de-tect myeloperoxidase activity in plaques. This enzyme hasbeen implicated as a product of inflammatory cells thatmay incite plaque instability and lead to myocardial in-farction. Components of thrombi important for stabilizing

clots (Factor 13) have been detected with targeted ironoxide probes consisting of a dextran-coated caged ironoxide particle (CLIO) conjugated to a2AP peptide frag-ments that are recognized by activated Factor 13 (103). Invitro experiments with human plasma thrombi incubatedwith F13-CLIO revealed thrombus contrast enhancementas compared with control probes.

Stem-Cell Imaging Methods Relevant toAtherosclerosis

Stem-cell imaging with MRI is another emerging areathat might fit under the rubric of molecular imagingwith targeted nanoparticle contrast agents. Stem cellsloaded with superparamagnetic nanoparticles in vitrocan be engrafted into the selected location by local in-jection. The stem cells are induced in vitro to ingestnanoparticles through endocytosis by various strate-gies, such as coating the particles with dendrimers,transfection agents, or antibodies/peptides (104–106).This results in the intracellular accumulation of signif-icant amounts of intact nanoparticles, which can thenexert a local susceptibility effect for detection in vivo.

Original work by Frank et al (105), Bulte and Kraitch-man (107), and others has demonstrated the use of MRImethods for stem-cell tracking based on detection ofsusceptibility artifacts created by the superparamag-netic nanoparticles. Using this approach, Kraitchmanet al (108) were able to detect and track mesenchymalstem cells injected into necrotic regions of a pig heart at1.5T (Fig. 7). This group more recently used a combi-nation of SPECT/CT to track stem-cell migration in theheart after local injections (109). Additional applica-tions likely include detecting and tracking stem-cellactivity in vascular inflammation. These approachesoffer sensitive and robust detection of important cellu-lar vehicles for cardiovascular tissue regeneration.

Ahrens et al (72) recently followed the fluorine molec-ular imaging approaches demonstrated originally by Yuet al (25) and Morawski et al (16) by employing a crownether preparation of nanoparticles to load dendritic im-mune cells with no loss of viability, and demonstratedan extension of the use of fluorine imaging at researchfield strengths (11.7T) to track cells after local and sys-temic injections (Fig. 8). Again, the advantage to thisapproach is that no background signal exists becausethere is no appreciable amount of fluorine in the body toconfound the signal from the targeted cells. Recentmethods for labeling of proangiogenic endothelial pre-cursor cells with multiple types of perfluorocarbonnanoparticles and rapid imaging at clinical (1.5T) andresearch (11.7T) field strengths have been reported(110,111).

CONJUNCTIVE NANOPARTICLE THERAPEUTICS

The potential dual use of nanoparticles for both imag-ing and targeted delivery of therapeutic agents to sitesof cardiovascular disease offers great promise for indi-vidualizing therapeutics. Image-based therapeuticswith site-selective agents should enable conclusive as-surance that the drug is reaching the intended targetand a molecular effect is occurring. Traditional phar-

676 Wickline et al.

macokinetic and pharmacodynamic analyses used topredict drug efficacy and toxicity are rooted in monitor-ing serum concentrations as inputs to linked differen-tial equations to describe the transport of drug from onecompartment (e.g., serum) to another (i.e., the extracel-lular space where the drug binds to its target). In thecase of particulate agents, however, the mechanisms of

drug delivery become more complicated and hence se-rum concentrations are not necessarily indicative of theamount of drug that is accessible to the desired site(112). Furthermore, when the carrier is targeted to thetissue of interest, the drug release is also localized tothat area, resulting in a much higher effective drugconcentration (or area under the curve (AUC)) at the site

Figure 8. In vivo MR image of perfluorocarbon-labeled dendritic cells in a mouse (reprinted from Ahrens ET, Flores R, Xu H,Morel PA. In vivo imaging platform for tracking immunotherapeutic cells. Nat Biotechnol 2005;23:983–987, with permission). a:Electron micrograph of dendritic cells labeled with PFC nanoparticles. Note the clear spherical nanoparticles scattered through-out the cytoplasm. b: The 19F intensity is displayed on a “hot-iron” intensity scale, and the 1H images are shown in grayscale.c: Mouse quadriceps after intramuscular injection of perfluorocarbon-labeled dendritic cells (asterisk indicates injection site).Shown (from left to right) are 19F and 1H images and a “composite” 19F/1H image.

Figure 9. Augmenting therapeutics with US applied to bound targeted nanoparticles (reprinted from Crowder KC, Hughes MS,Marsh JN, et al. Sonic activation of molecularly-targeted nanoparticles accelerates transmembrane lipid delivery to cancer cellsthrough contact-mediated mechanisms: implications for enhanced local drug delivery. Ultrasound Med Biol 2005;31:1693–1700, with permission). a: Perfluorocarbon-based nanoparticles can be loaded with lipophilic drug in the outer lipid monolayerand delivered to the cell by lipid mixing and/or lipid vesicle-cell membrane fusion. As an example, rhodamine-labeled lipids (redstream from the nanoparticle) rapidly mix into the cell membrane (see arrow in inset) upon interaction with C32 melanoma cells(cells are transfected with GFP to label endosomes), and then rapidly distribute into the cytoplasm without requiring endocytosisof the intact particles. b: US potentiation of lipid uptake by cells in culture. FITC-labeled nanoparticles targeted to C32 cellsdemonstrate lipid mixing and fusion of particles and cells, and cytoplasmic delivery (note the green lipid-conjugated fluorophoresdistributed in the cytoplasm). c: After a five-minute insonification of cells in culture with a 2.5-MHz clinical phased-arraytransducer at medium power, a marked increase in cytoplasmic lipid delivery is achieved with no untoward effects on theviability of targeted cells.

Molecular Imaging With Nanoparticles 677

than is indicated by serum levels alone. The uniqueability to image these particles could be of great benefitfor estimating local drug concentrations and developingnew pharmacokinetic and dynamic paradigms to de-scribe this new class of agents.

As an example of this new paradigm for drug delivery,Lanza et al (23,71) treated smooth muscle cells in cul-ture with TF-targeted nanoparticles that were loadedwith paclitaxel. The smooth muscle cells were har-vested from pig aorta and constitutively expressed TFepitopes in vitro. Binding of the drug-free nanoparticlesto the cells yielded no alterations in growth character-istics of the cultured cells. However, when paclitaxel-loaded nanoparticles were applied to the cells, specificbinding elicited a substantial reduction in smooth mus-cle cell proliferation. Nontargeted paclitaxel-loaded par-ticles applied to the cells (i.e., with no binding of nano-particles to cells) resulted in normal cell proliferation,indicating that selective targeting may be a requirementfor effective drug delivery for these emulsions. Similarbehavior has been demonstrated for doxorubicin-con-taining particles (71). Recent reports indicate that in-travenous delivery of fumagillin-loaded nanoparticles(an antiangiogenic agent) targeted to �v�3-integrinepitopes on the vasa vasorum in growing plaques re-sults in marked inhibition of plaque angiogenesis incholesterol-fed rabbits (113). Kolodgie et al (114) alsoutilized Taxol-containing albumin nanoparticles tolimit the restenotic response after angioplasty and stentplacement in experimental animals.

The unique mechanism of drug delivery for highlylipophilic agents (e.g., paclitaxel) contained withinemulsions depends on close apposition between thenanoparticle carrier and the targeted cell membrane,and has been described as “contact facilitated drugdelivery” (71). In contrast to liposomal drug delivery,which generally requires endocytosis, the mechanismof drug transport in this case involves lipid exchange orlipid mixing between the emulsion vesicle and the tar-geted cell membrane (Fig. 9) (61,62), which depends onthe extent and frequency of contact between two lipidicsurfaces (61,62,71). The rate of lipid exchange and drugdelivery can be greatly increased by the application ofclinically safe levels of US energy. This increases thepropensity for fusion or enhanced contact between thenanoparticles and the targeted cell membrane by stim-ulating these interactions between nanoparticles andcell membranes (Fig. 9) (62). These methods offer addi-tional mechanisms for facilitating targeted drug deliv-ery with the application of exogenous, safe US energy inconjunction with therapeutic targeted nanoparticles.

In conclusion, the combination of targeted drug de-livery and molecular imaging with MRI has the potentialto revolutionize the detection and treatment of cardio-vascular disease. Drug-delivery agents that are alsoquantifiable at the targeted site based on imaging read-outs may ultimately permit serial characterization ofthe molecular epitope expression and confirmation oftherapeutic efficacy. Rapid developments in genomics,molecular biology, and nanotechnology are contribut-ing to the multidisciplinary field of molecular imaging,and clinical trials are beginning.

REFERENCES

1. Zheng ZJ, Croft JB, Giles WH, Mensah GA. Sudden cardiac deathin the United States, 1989 to 1998. Circulation 2001;104:2158–2163.

2. Ojio S, Takatsu H, Tanaka T, et al. Considerable time from theonset of plaque rupture and/or thrombi until the onset of acutemyocardial infarction in humans: coronary angiographic findingswithin 1 week before the onset of infarction. Circulation 2000;102:2063–2069.

3. Petit L, Lesnik P, Dachet C, Moreau M, Chapman MJ. Tissue factorpathway inhibitor is expressed by human monocyte-derived mac-rophages: relationship to tissue factor induction by cholesteroland oxidized LDL. Arterioscler Thromb Vasc Biol 1999;19:309–315.

4. Kolodgie FD, Gold HK, Burke AP, et al. Intraplaque hemorrhageand progression of coronary atheroma. N Engl J Med 2003;349:2316–2325.

5. Newby AC, Zaltsman AB. Fibrous cap formation or destruction—the critical importance of vascular smooth muscle cell prolifera-tion, migration and matrix formation. Cardiovasc Res 1999;41:345–360.

6. Shah PK. Role of inflammation and metalloproteinases in plaquedisruption and thrombosis. Vasc Med 1998;3:199–206.

7. de Boer OJ, van der Wal AC, Teeling P, Becker AE. Leucocyterecruitment in rupture prone regions of lipid-rich plaques: a prom-inent role for neovascularization? Cardiovasc Res 1999;41:443–449.

8. McCarthy MJ, Loftus IM, Thompson MM, et al. Angiogenesis andthe atherosclerotic carotid plaque: an association between symp-tomatology and plaque morphology. J Vasc Surg 1999;30:261–268.

9. Britz-Cunningham SH, Adelstein SJ. Molecular targeting with ra-dionuclides: state of the science. J Nucl Med 2003;44:1945–1961.

10. Herschman HR. Molecular imaging: looking at problems, seeingsolutions. Science 2003;302:605–608.

11. Tsien RY. Imagining imaging’s future. Nat Rev Mol Cell Biol 2003;Suppl:SS16–SS21.

12. Lanza G, Wickline S. Targeted ultrasonic contrast agents for mo-lecular imaging and therapy. Prog Cardiovasc Dis 2001;44:13–31.

13. Lanza GM, Wickline SA. Targeted ultrasonic contrast agents formolecular imaging and therapy. Curr Prob Cardiol 2003;28:625–653.

14. Wickline SA, Lanza GM. Molecular imaging, targeted therapeu-tics, and nanoscience. J Cell Biochem 2002;39:90–97.

15. Wickline SA, Lanza GM. Nanotechnology for molecular imagingand targeted therapy. Circulation 2003;107:1092–1095.

16. Morawski AM, Winter PM, Yu X, et al. Quantitative “magneticresonance immunohistochemistry” with ligand-targeted 19Fnanoparticles. Magn Reson Med 2004;52:1255–1262.

17. Buxton DB, Lee SC, Wickline SA, et al. Recommendations of theNational Heart, Lung, and Blood Institute Nanotechnology Work-ing Group. Circulation 2003;108:2737–2742.

18. Hawker CJ, Wooley KL. The convergence of synthetic organic andpolymer chemistries. Science 2005;309:1200–1205.

19. Anderson SA, Rader RK, Westlin WF, et al. Magnetic resonancecontrast enhancement of neovasculature with alpha(v)beta(3)-tar-geted nanoparticles. Magn Reson Med 2000;44:433–439.

20. Sipkins DA, Cheresh DA, Kazemi MR, Nevin LM, Bednarski MD, LiKC. Detection of tumor angiogenesis in vivo by alphaVbeta3-tar-geted magnetic resonance imaging. Nat Med 1998;4:623–626.

21. Hermanson GT. Bioconjugate techniques. San Diego: AcademicPress; 1996.

22. Demos SM, Alkan-Onyuksel H, Kane BJ, et al. In vivo targeting ofacoustically reflective liposomes for intravascular and transvas-cular ultrasonic enhancement. J Am Coll Cardiol 1999;33:867–875.

23. Lanza GM, Winter P, Caruthers S, et al. Novel paramagnetic con-trast agents for molecular imaging and targeted drug delivery.Curr Pharm Biotechnol 2004;5:495–507.

24. Flacke S, Fischer S, Scott MJ, et al. Novel MRI contrast agent formolecular imaging of fibrin: implications for detecting vulnerableplaques. Circulation 2001;104:1280–1285.

25. Yu X, Caruthers SD, Love SM, et al. Rapid and sensitive thrombusdetection with a fibrin-targeted nanoparticle MRI contrast agent.Circulation 2001;104:1635.

678 Wickline et al.

26. Frias JC, Williams KJ, Fisher EA, Fayad ZA. Recombinant HDL-like nanoparticles: a specific contrast agent for MRI of atheroscle-rotic plaques. J Am Chem Soc 2004;126:16316–16317.

27. Li H, Gray BD, Corbin I, et al. MR and fluorescent imaging of low-density lipoprotein receptors. Acad Radiol 2004;11:1251–1259.

28. Kobayashi H, Kawamoto S, Jo SK, Bryant Jr HL, Brechbiel MW,Star RA. Macromolecular MRI contrast agents with small den-drimers: pharmacokinetic differences between sizes and cores.Bioconjug Chem 2003;14:388–394.

29. Sato N, Kobayashi H, Hiraga A, et al. Pharmacokinetics and en-hancement patterns of macromolecular MR contrast agents withvarious sizes of polyamidoamine dendrimer cores. Magn ResonMed 2001;46:1169–1173.

30. Schmitz SA, Coupland SE, Gust R, et al. Superparamagnetic ironoxide-enhanced MRI of atherosclerotic plaques in Watanabe he-reditable hyperlipidemic rabbits. Invest Radiol 2000;35:460–471.

31. Koch AM, Reynolds F, Merkle HP, Weissleder R, Josephson L.Transport of surface-modified nanoparticles through cell mono-layers. Chembiochem 2005;6:337–345.

32. Hirsch LR, Stafford RJ, Bankson JA, et al. Nanoshell-mediatednear-infrared thermal therapy of tumors under magnetic reso-nance guidance. Proc Natl Acad Sci USA 2003;100:13549–13554.

33. Loo C, Lin A, Hirsch L, et al. Nanoshell-enabled photonics-basedimaging and therapy of cancer. Technol Cancer Res Treat 2004;3:33–40.

34. O’Neal DP, Hirsch LR, Halas NJ, Payne JD, West JL. Photo-ther-mal tumor ablation in mice using near infrared-absorbing nano-particles. Cancer Lett 2004;209:171–176.

35. Cherukuri P, Bachilo SM, Litovsky SH, Weisman RB. Near-infra-red fluorescence microscopy of single-walled carbon nanotubes inphagocytic cells. J Am Chem Soc 2004;126:15638–15639.

36. Tsyboulski DA, Bachilo SM, Weisman RB. Versatile visualizationof individual single-walled carbon nanotubes with near-infraredfluorescence microscopy. Nano Lett 2005;5:975–979.

37. Barone PW, Baik S, Heller DA, Strano MS. Near-infrared opticalsensors based on single-walled carbon nanotubes. Nat Mater2005;4:86–92.

38. Hertel T, Hagen A, Talalaev V, et al. Spectroscopy of single- anddouble-wall carbon nanotubes in different environments. NanoLett 2005;5:511–514.

39. Akerman ME, Chan WC, Laakkonen P, Bhatia SN, Ruoslahti E.Nanocrystal targeting in vivo. Proc Natl Acad Sci USA 2002;99:12617–12621.

40. Chen L, Zurita AJ, Ardelt PU, Giordano RJ, Arap W, Pasqualini R.Design and validation of a bifunctional ligand display system forreceptor targeting. Chem Biol 2004;11:1081–1091.

41. Gao X, Nie S. Quantum dot-encoded beads. Methods Mol Biol2005;303:61–71.

42. Michalet X, Pinaud FF, Bentolila LA, et al. Quantum dots for livecells, in vivo imaging, and diagnostics. Science 2005;307:538–544.

43. Allen M, Bulte JW, Liepold L, et al. Paramagnetic viral nanopar-ticles as potential high-relaxivity magnetic resonance contrastagents. Magn Reson Med 2005;54:807–812.

44. Wickline S, Neubauer A, Winter P, Caruthers S, Lanza G. Applica-tions of nanotechnology to atherosclerosis, thrombosis, and vascularbiology. Arterioscler Thromb Vasc Biol 2006;26:435–441.

45. Louie AY, Huber MM, Ahrens ET, et al. In vivo visualization of geneexpression using magnetic resonance imaging. Nat Biotechnol2000;18:321–325.

46. Winter PM, Cai K, Chen J, et al. Molecular imaging of fibrin with atargeted PARACEST nanoparticle contrast agent. Magn ResonMed 2006;56:1384–1388.

47. Caruthers SD, Neubauer AM, Hockett FD, et al. In vitro demon-stration using 19F magnetic resonance to augment molecularimaging with paramagnetic perfluorocarbon nanoparticles at 1.5Tesla. Invest Radiol 2006;41:305–312.

48. Bogdanov AJ, Matuszewski L, Bremer C, Petrovsky A, WeisslederR. Oligomerization of paramagnetic substrates result in signalamplification and can be used for MR imaging of molecular tar-gets. Mol Imaging 2002;1:16–23.

49. Perez JM, Josephson L, O’Loughlin T, Hogemann D, Weissleder R.Magnetic relaxation switches capable of sensing molecular inter-actions. Nat Biotechnol 2002;20:816–820.

50. Quinti L, Weissleder R, Tung C. A fluorescent nanosensor forapoptotic cells. Nano Lett 2006;6:488–490.

51. Rossin R, Pan D, Qi K, et al. 64Cu-labeled folate-conjugated shellcross-linked nanoparticles for tumor imaging and radiotherapy:synthesis, radiolabeling, and biologic evaluation. J Nucl Med2005;46:1210–1218.

52. Blankenberg FG, Strauss HW. Noninvasive strategies to imagecardiovascular apoptosis. Cardiol Clinics 2001;19:165–172.

53. Rudd JHF, Warburton EA, Fryer TD, et al. Imaging atheroscleroticplaque inflammation with [18F]-fluorodeoxyglucose positronemission tomography. Circulation 2002;105:2708–2711.

54. Kelly KA, Allport JR, Tsourkas A, Shinde-Patil VR, Josephson L,Weissleder R. Detection of vascular adhesion molecule-1 expres-sion using a novel multimodal nanoparticle. Circ Res 2005;96:327–336.

55. Chen J, Tung C, Mahmood U, et al. In vivo imaging of proteolyticactivity in atherosclerosis. Circulation 2002;105:2766–2771.

56. Jaffer FA, Tung CH, Wykrzykowska JJ, et al. Molecular imaging offactor XIIIa activity in thrombosis using a novel, near-infraredfluorescent contrast agent that covalently links to thrombi. Cir-culation 2004;110:170–176.

57. Weber SM, Peterson KA, Durkee B, et al. Imaging of murine livertumor using microCT with a hepatocyte-selective contrast agent:accuracy is dependent on adequate contrast enhancement. J SurgRes 2004;119:41–45.

58. Winter PM, Shukla HP, Caruthers SD, et al. Molecular imaging ofhuman thrombus with computed tomography. Acad Radiol 2005;12:S9–S13.

59. Hamilton AJ, Huang SL, Warnick D, et al. Intravascular ultra-sound molecular imaging of atheroma components in vivo. J AmColl Cardiol 2004;43:453–460.

60. Morawski AM, Lanza GA, Wickline SA. Targeted contrast agentsfor magnetic resonance imaging and ultrasound. Curr Opin Bio-technol 2005;16:89–92.

61. Crowder KC, Hughes MS, Marsh JN, et al. Sonic activation ofmolecularly-targeted nanoparticles accelerates transmembranelipid delivery to cancer cells through contact-mediated mecha-nisms: implications for enhanced local drug delivery. UltrasoundMed Biol 2005;31:1693–1700.

62. Crowder KC, Hughes MS, Marsh JN, et al. Augmented and selec-tive delivery of liquid perfluorocarbon nanoparticles to melanomacells with noncavitational ultrasound. Proc IEEE Symp03CH37476C: 532–535, 2003.

63. Zhao S, Borden M, Bloch SH, Kruse D, Ferrara KW, Dayton PA.Radiation-force assisted targeting facilitates ultrasonic molecularimaging. Mol Imaging 2004;3:135–148.

64. Lanza GM, Trousil RL, Wallace KD, et al. In vitro characterizationof a novel, tissue-targeted ultrasonic contrast system with acous-tic microscopy. J Acoust Soc Am 1998;104:3665–3672.

65. Marsh JN, Hall CS, Scott MJ, et al. Improvements in the ultra-sonic contrast of targeted perfluorocarbon nanoparticles using anacoustic transmission line model. IEEE Trans Ultrason FerroelectFreq Control 2002;49:29–38.

66. Gupta H, Weissleder R. Targeted contrast agents. Magn ResonImaging Clin N Am 1996;4:171–184.

67. Ahrens ET, Rothbacher U, Jacobs RE, Fraser SE. A model for MRIcontrast enhancement using T1 agents. Proc Natl Acad Sci USA1998;5:8443–8448.

68. Morawski AM, Winter PM, Crowder KC, et al. Targeted nanopar-ticles for quantitative imaging of sparse molecular epitopes withMRI. Magn Reson Med 2004;51:480–486.

69. Cunningham CH, Arai T, Yang PC, McConnell MV, Pauly JM,Conolly SM. Positive contrast magnetic resonance imaging of cellslabeled with magnetic nanoparticles. Magn Reson Med 2005;53:999–1005.

70. Stuber M, Gilson WD, Schaer M, Bulte JW, Kraitchman DL. Shed-ding light on the dark spot with IRON: a method that generatespositive contrast in the presence of superparamagnetic nanopar-ticles. In: Proceedings of the 13th Annual Meeting of ISMRM,Miami Beach, FL, USA, 2005 (Abstract 2608).

71. Lanza GM, Yu X, Winter PM, et al. Targeted antiproliferative drugdelivery to vascular smooth muscle cells with an MRI nanoparticlecontrast agent: implications for rational therapy of restenosis.Circulation 2002;106:2842–2847.

72. Ahrens ET, Flores R, Xu H, Morel PA. In vivo imaging platform fortracking immunotherapeutic cells. Nat Biotechnol 2005;23:983–987.

73. Bovey F, Jelinski L, Mirau P. Nuclear magnetic resonance spec-troscopy. San Diego: Academic Press; 1988.

Molecular Imaging With Nanoparticles 679

74. Doyle M, Mansfield P. Chemical-shift imaging: a hybrid approach.Magn Reson Med 1987;5:255–261.

75. Noth U, Jager LJ, Lutz J, Haase A. Fast 19F-NMR imaging in vivousing FLASH-MRI. Magn Reson Imaging 1994;12:149–153.

76. Goldstein JA. Multifocal coronary plaque instability. Prog Cardio-vasc Dis 2002;44:449–454.

77. Constantinides P. Plaque fissuring in human coronary thrombo-sis. J Atheroscler Res 1966;6:1–17.

78. Lanza GM, Wallace KD, Scott MJ, et al. A novel site-targetedultrasonic contrast agent with broad biomedical application. Cir-culation 1996;95:3334–3340.

79. Lanza GM, Abendschein DR, Hall CS, et al. In vivo molecularimaging of stretch-induced tissue factor in carotid arteries withligand-targeted nanoparticles. J Am Soc Echocardiography 2000;13:608–614.

80. Oltrona L, Speidel CM, Recchia D, Wickline SA, Eisenberg PR,Abendschein DR. Inhibition of tissue factor-mediated coagulationmarkedly attenuates stenosis after balloon-induced arterial injuryin minipigs. Circulation 1997;96:646–652.

81. Hamilton A, Huang SL, Warnick D, et al. Left ventricular throm-bus enhancement after intravenous injection of echogenic immu-noliposomes: studies in a new experimental model. Circulation2002;105:2772–2778.

82. Botnar RM, Buecker A, Wiethoff AJ, et al. In vivo magnetic reso-nance imaging of coronary thrombosis using a fibrin-binding mo-lecular magnetic resonance contrast agent. Circulation 2004;110:1463–1466.

83. Spuentrup E, Fausten B, Kinzel S, et al. Molecular magneticresonance imaging of atrial clots in a swine model. Circulation2005;112:396–399.

84. Spuentrup E, Katoh M, Wiethoff AJ, et al. Molecular magneticresonance imaging of pulmonary emboli with a fibrin-specific con-trast agent. Am J Respir Crit Care Med 2005;172:494–500.

85. Kerr JS, Mousa SA, Slee AM. Alpha(v)beta(3) integrin in angiogen-esis and restenosis. Drug News Perspect 2001;14:143–150.

86. Bishop GG, McPherson JA, Sanders JM, et al. Selective alpha(v)-beta(3)-receptor blockade reduces macrophage infiltration and re-stenosis after balloon angioplasty in the atherosclerotic rabbit.Circulation 2001;103:1906–1911.

87. Corjay MH, Diamond SM, Schlingmann KL, Gibbs SK, StoltenborgJK, Racanelli AL. alphavbeta3, alphavbeta5, and osteopontin arecoordinately upregulated at early time points in a rabbit model ofneointima formation. J Cell Biochem 1999;75:492–504.

88. Brooks PC, Stromblad S, Klemke R, Visscher D, Sarkar FH,Cheresh DA. Antiintegrin alpha v beta 3 blocks human breastcancer growth and angiogenesis in human skin. J Clin Invest1995;96:1815–1822.

89. Winter PM, Caruthers SD, Kassner A, et al. Molecular imaging ofangiogenesis in nascent Vx-2 rabbit tumors using a novel alpha-(nu)beta3-targeted nanoparticle and 1.5 tesla magnetic resonanceimaging. Cancer Res 2003;63:5838–5843.

90. Schmieder AH, Winter P, Caruthers S, et al. Molecular MR imag-ing of melanoma angiogenesis with avb3-targeted paramagneticnanoparticles. Magn Reson Med 2005;53:621–627.

91. Winter PM, Morawski AM, Caruthers SD, et al. Molecular imagingof angiogenesis in early-stage atherosclerosis with alpha(v)beta(3)-Integrin-targeted nanoparticles. Circulation 2003;108:2270–2274.

92. Moulton KS, Heller E, Konerding MA, et al. Angiogenesis inhibi-tors endostatin or TNP-470 reduce intimal neovascularization andplaque growth in apolipoprotien E-defficient mice. Circulation1999;99:1653–1655.

93. Tenaglia AN, Peters KG, Sketch Jr MH, Annex BH. Neovascular-ization in atherectomy specimens from patients with unstableangina: implications for pathogenesis of unstable angina. AmHeart J 1998;135:10–14.

94. Wilson SH, Hermann J, Lermann LO, et al. Simvastatin preservesthe structure of coronary adventitial vasa vasorum in experimen-tal hypercholesterolemia independent of lipid lowering. Circula-tion 2002;105:415–418.

95. Zhang Y, Cliff WJ, Schoefl GI, Higgins G. Immunohistochemicalstudy of intimal microvessels in coronary atherosclerosis. Am JPathol 1993;143:164–172.

96. Cyrus T, Abendschein DR, Caruthers SD, et al. MR three-dimen-sional molecular imaging of intramural biomarkers with targetednanoparticles. J Cardiovasc Magn Reson 2006;8:1–7.

97. Ruehm SG, Corot C, Vogt P, Kolb S, Debatin JF. Magnetic reso-nance imaging of atherosclerotic plaque with ultrasmall super-paramagnetic particles of iron oxide in hyperlipidemic rabbits.Circulation 2001;103:415–422.

98. Corot C, Petry KG, Trivedi R, et al. Macrophage imaging in centralnervous system and in carotid atherosclerotic plaque using ul-trasmall superparamagnetic iron oxide in magnetic resonanceimaging. Invest Radiol 2004;39:619–625.

99. Kooi ME, Cappendijk VC, Cleutjens KB, et al. Accumulation ofultrasmall superparamagnetic particles of iron oxide in humanatherosclerotic plaques can be detected by in vivo magnetic reso-nance imaging. Circulation 2003;107:2453–2458.

100. Trivedi RA, U-King-Im JM, Graves MJ, et al. In vivo detection ofmacrophages in human carotid atheroma: temporal dependenceof ultrasmall superparamagnetic particles of iron oxide-enhancedMRI. Stroke 2004;35:1631–1635.

101. Sirol M, Itskovich VV, Mani V, et al. Lipid-rich atheroscleroticplaques detected by gadofluorine-enhanced in vivo magnetic res-onance imaging. Circulation 2004;109:2890–2896.

102. Perez JM, Josephson L, O’Loughlin T, Hogemann D, Weissleder R.Magnetic relaxation switches capable of sensing molecular inter-actions. Nature Biotechnology 2002;20:816–820.

103. Jaffer FA, Tung CH, Houng AK, O’Loughin T, Reed GL, WeisslederR. MRI of blood coagulation factor XIII activity using a novelpeptide-derivatized caged iron oxide nanoparticle (F13-CLIO) [Ab-stract]. Mol Imaging 2002;1:217–218.

104. Arbab AS, Bashaw LA, Miller BR, Jordan EK, Bulte JW, Frank JA.Intracytoplasmic tagging of cells with ferumoxides and transfec-tion agent for cellular magnetic resonance imaging after celltransplantation: methods and techniques. Transplantation 2003;76:1123–1130.

105. Frank JA, Miller BR, Arbab AS, et al. Clinically applicable labeling ofmammalian and stem cells by combining superparamagnetic ironoxides and transfection agents. Radiology 2003;228:480–487.

106. Funovics MA, Kapeller B, Hoeller C, et al. MR imaging of theher2/neu and 9.2.27 tumor antigens using immunospecific con-trast agents. Magn Reson Imaging 2004;22:843–850.

107. Bulte JW, Kraitchman DL. Monitoring cell therapy using iron oxideMR contrast agents. Curr Pharm Biotechnol 2004;5:567–584.

108. Kraitchman DL, Heldman AW, Atalar E, et al. In vivo magneticresonance imaging of mesenchymal stem cells in myocardial in-farction. Circulation 2003;107:2290–2293.

109. Kraitchman DL, Tatsumi M, Gilson WD, et al. Dynamic imaging ofallogeneic mesenchymal stem cells trafficking to myocardial in-farction. Circulation 2005;112:1451–1461.

110. Chen J, Crowder K, Brant J, et al. Labeling and imaging stem/progenitor cells with multiple unique nanoparticulate fluorinemarkers: the potential for multispectral stem cell detection with19F MRI. In: Proceedings of the 14th Annual Meeting of ISMRM,Seattle, WA, USA, 2006 (Abstract 187).

111. Neubauer AM, Crowder K, Caruthers SD, et al. Unique andrapid 19F imaging of stem cells labeled with perfluorocarbonnanoparticles at 11.7T and 1.5T. In: Proceedings of the 14thAnnual Meeting of ISMRM, Seattle, WA, USA, 2006 (Abstract354).

112. Harashima H, Iida S, Urakami Y, Tsuchihashi M, Kiwada H. Op-timization of antitumor effect of liposomally encapsulated doxo-rubicin based on simulations by pharmacokinetic/pharmacody-namic modeling. J Control Release 1999;61:93–106.

113. Winter PM, Morawski AM, Caruthers SD, et al. Antiangiogenictherapy of early atherosclerosis with paramagnetic alpha(v)-beta(3)-integrin-targeted fumagillin nanoparticles. J Am Coll Car-diol 2004;43:322A–323A.

114. Kolodgie FD, John M, Khurana C, et al. Sustained reduction ofin-stent neointimal growth with the use of a novel systemic nano-particle paclitaxel. Circulation 2002;106:1195–1198.

680 Wickline et al.