Dual In Vivo Quantification of Integrin-targeted and Protease-activated Agents in Cancer Using Fluorescence Molecular Tomography (FMT)

B Academy of Molecular Imaging, 2009Published Online: 4 December 2009 DOI: 10.1007/s11307-009-0279-z

Mol ImagingBiol (2010) 12:488Y499

RESEARCH ARTICLE

Dual In Vivo Quantification of Integrin-targetedand Protease-activated Agents in Cancer UsingFluorescence Molecular Tomography (FMT)Sylvie Kossodo,1 Maureen Pickarski,2 Shu-An Lin,2 Alexa Gleason,2 Renee Gaspar,2

Chiara Buono,1 Guojie Ho,1 Agnieszka Blusztajn,1 Garry Cuneo,1 Jun Zhang,1

Jayme Jensen,1 Richard Hargreaves,2 Paul Coleman,2 George Hartman,2

Milind Rajopadhye,1 Le Thi Duong,2 Cyrille Sur,2 Wael Yared,1 Jeffrey Peterson,1

Bohumil Bednar2

1VisEn Medical, Bedford, MA( USA2Merck Research Laboratories, West Point, PA( USA

AbstractPurpose: Integrins, especially αvβ3 and αvβ5, are upregulated in tumor cells and activatedendothelial cells and as such, serve as cancer biomarkers. We developed a novel near-infrared-labeled optical agent for the in vivo detection and quantification of αvβ3/αvβ5.Procedures: A small peptidomimetic αvβ3 antagonist was synthesized, coupled to a near-infrared fluorescent (NIRF) dye, and tested for binding specificity using integrin-overexpressingcells, inhibition of vitronectin-mediated cell attachment, binding to tumor and endothelial cells invitro, and competition studies. Pharmacokinetics, biodistribution, specificity of tumor targeting,and the effect of an antiangiogenic treatment were assessed in vivo.Results: The integrin NIRF agent showed strong selectivity towards αvβ3/αvβ5 in vitro andpredominant tumor distribution in vivo, allowing noninvasive and real-time quantification ofintegrin signal in tumors. Antiangiogenic treatment significantly inhibited integrin signal in vivobut had no effect on a cathepsin-cleavable NIR agent. Simultaneous imaging revealed differentpatterns of distribution reflecting the underlying differences in integrin and cathepsin biologyduring tumor progression.Conclusions: NIRF-labeled integrin antagonists allow noninvasive molecular fluorescent imagingand quantification of tumors in vivo, improving and providing more refined approaches for cancerdetection and treatment monitoring.

Key words: Near-infrared fluorescence imaging, Cancer, Quantification, Optical tomography,Integrin αvβ3

Introduction

I ntegrins are transmembrane glycoproteins consisting oftwo noncovalently bound α and β subunits that mediate

cell-to-cell and cell-to-extracellular matrix interactions

This study conclusively establishes for the first time the ability tosimultaneously and noninvasively quantify integrin αvβ3 and cathepsinactivity using near-infrared fluorescence tomography in vivo.

Correspondence to: Sylvie Kossodo; e-mail: [email protected]

[1, 2]. These adhesive processes are intricately linked to cellmigration, invasion, and extravasation, as well as signaltransduction and apoptosis [2, 3]. The integrins αvβ3 andαvβ5, in particular, appear to play an important role in tumorgrowth, metastasis, and angiogenesis [4]. They are minimallyexpressed in dormant blood vessels but are significantlyupregulated in newly sprouting vasculature [5, 6]. Likewise,αvβ3 and/or αvβ5 are upregulated on commonly used tumor celllines, with expression of these integrins correlating with tumoraggressiveness [7, 8]. On the basis of their involvement in suchcritical aspects of tumor angiogenesis, development, andmetastasis, a significant effort has been devoted to generatemolecules to block their activity in vivo leading to clinical trialsthat have shown the potential of a number of these molecules asvalid antiangiogenic therapeutic options [9–11].

Clearly, the ability to spatially and temporally visualize aswell as quantify αvβ3/αvβ5 in vivo would greatly improve ourability to follow the expression of these integrins during tumordevelopment and metastasis, develop novel anti-integrintherapies, monitor treatment efficacy, and aid in patientselection [12]. Both of these integrins bind extracellular matrixproteins through the exposed arginine–glycine–aspartic acid(RGD) tripeptide in their ligands. For this reason, an assortmentof radionuclide-labeled molecules, many comprising RGDpeptides, have been developed for positron emission andsingle-photon emission computed tomography imaging[13–17]. While these long-established and noninvasiveimaging modalities are sensitive and provide functionalinformation [13, 18], they are limited by their use of ionizingradiation (with its associated costs, complexity, short half-lives,and radioactive material handling/disposal) as well as thereadout of a single biological process at a time. Alternativesinclude using contrast-enhanced ultrasound to image micro-bubbles targeted with RGD-containing peptides or anti-αvβ3antibodies. This targeting strategy has been used to detect earlytumor angiogenesis, changes in blood volume, and therapyresponse [19, 20]. However, ultrasound is not strictlyquantitative, cannot distinguish between multiple targets,and can potentially cause soft tissue heat and cavitation.Conventional magnetic resonance imaging (MRI) offersthe highest morphological resolution but is hindered byrelatively limited sensitivity, long scanning time, and highinstrumentation cost. Nonetheless, the use of paramagneticnanoparticles targeted to endothelial αvβ3 via the LM609monoclonal antibody provided enhanced and detailedimaging of rabbit carcinomas and imaging of angiogenic“hot spots” that are not seen by standard MRI [21].

Optical molecular imaging using NIRF contrast agents hasbecome a valuable new platform for noninvasive imaging dueto advances in the development of targeted and activatableagents and improved imaging systems [13, 18]. Opticalmolecular imaging allows us to visualize and quantifybiological processes in vivo at a relatively low equipment cost,follow the spatial and temporal biodistribution of one or moretargets simultaneously, detect subtle changes otherwiseunidentifiable by methods relying on gross morphological

detection, and perform repetitive imaging in vivo of the samesubjects, thus, reducing the need for sacrificing animals andalso the number of animals needed per study. With this inmind, near-infrared fluorescence (NIRF)-labeled RGD agentshave been used successfully to image, albeit in a semi-quantitative manner, integrin expression in a variety of tumormodels [13, 22–27]. Optical imaging is also compatible withthe complementary use of multiple modality imaging. A studyby Cai et al. used RGD-modified quantum dots for dualpositron emission tomography (PET) and NIRF integrinimaging. The authors found an excellent correlation betweenthe results measured by in vivo PET imaging and thosemeasured by ex vivo NIRF imaging [28].

Quantification of NIRF agents in vivo can now beperformed with fluorescence molecular tomography (FMT).FMT quantifies fluorescence by tomographic reconstruction ofa series of fluorescence images excited at different sourcelocations, using a model of photon propagation in a scatteringmedium. By normalizing each fluorescence image to itscorresponding excitation image, FMT overcomes the opticalheterogeneity of biological tissue resolving the ambiguityabout depth, size, and concentration of fluorescence thatafflicts planar imagers [29, 30]. Quantification of knownconcentrations of fluorochrome using the FMT have shownstrong linearity between actual and reconstructed FMT dyeconcentrations and a detection threshold of approximately 100femtomoles [29, 30]. FMT has been used for concurrentimaging and quantification of different targets and/or biologicalprocesses [31, 32]. In particular, Nahrendorf et al. were able tosimultaneously quantify a fluorescent activatable cathepsinagent, ProSense, and a fluorescently tagged CLIO particleuptaken preferentially by macrophages in a model of myocar-dial infarction [33].

The present work sought to overcome the specificity andaffinity limitations of RGD-based integrin agents by adopting asmall peptidomimetic antagonist-based approach. We tookadvantage of a noninvasive and quantitative imaging paradigmto image and quantify αvβ3/αvβ5 and cathepsin activity in tumorxenografts. Cathepsins and in particular, cysteine proteasessuch as cathepsin B are proteases involved in protein turnoverand degradation of the extracellular matrix and as such, play akey role in tumor cell infiltration, growth, metastasis, andangiogenesis [34]. We used dual channel imaging to quantifyand appraise the localization of two distinct biologicalprocesses in vivo, namely, tumor/host phagocyte cathepsinactivity and tumor/vasculature integrin expression, in responseto a standard antiangiogenic therapeutic treatment.

Materials and MethodsSynthesis of IntegriSense

The parent non-peptide αvβ3 antagonist [35] was converted to the3-cyano derivative on the tetrahydro-1,8-naphthyridine moietyusing procedures reported in the literature [36] and then reducedto the 3-aminomethyl analog. This 3-aminomethyl αvβ3 antagonist

S. Kossodo, et al.: Dual In Vivo Quantification Using FMT 489

derivative was conjugated with VivoTag-S680, an amine-reactivenear-infrared fluorochrome (VisEn Medical, Bedford, MA, USA),to provide the imaging agent that was purified by RP-HPLC(Phenomenex Luna, 5 µ phenyl-hexyl, 250×10 mm, flow rate4.7 ml/min, using an aqueous triethylamine bicarbonate (25 mM)/acetonitrile gradient: 20% to 30% acetonitrile over 15 min).Evaporation of the appropriate fractions gave IntegriSense680(IntegriSense) as a dark blue solid. Molecular weight wasconfirmed by MS (Waters Micromass ZQ, Milford, MA, USA).Excitation and emission spectra as well as the ε were generated inaqueous solution (PBS) at neutral pH using a Varian Cary 50 UV-VIS spectrophotometer and a Varian Cary Eclipse Fluorescencespectrophotometer (Varian, Palo Alto, CA) respectively.

Cell Lines and Culture Conditions

Human umbilical vein endothelial cells (HUVEC), human colorectalHCT-116, human colorectal HT29, human breast carcinoma MDA-MB231, and human rhabdomyosarcoma A673 cell lines wereobtained from ATCC (Manassas, VA). HUVECs were cultured inF-12K supplemented with 0.1 mg/ml heparin; 0.03–0.05 mg/mlendothelial cell growth supplement (ECGS), 10% fetal bovine serum,HCT-116, HT29 and A673 cells were grown in McCoy’s supple-mented with 10% fetal bovine serum, MDA-MB231 cells were grownin Dulbecco’s Modified Eagle’s Medium (DMEM) and 10% fetalbovine serum, and human embryonic kidney cells (HEK293)transfected with integrin constructs [37] were grown in MinimalEssential Medium (MEM), 1 mg/ml kanamycin and 2 mML-glutamine, supplemented with 10% fetal bovine serum, at 37ºC ina humidified atmosphere containing 5% CO2.

Cell Binding Assays

HEK293 cells stably transfected with αvβ3 (HEK293-αvβ3 cells)were plated onto poly-L-lysine coated Lab-Tek II chamberedcoverglass (Nunc, Rochester, NY, USA) the day before the assay.Cells were first incubated with DAPI (10 μg/ml) for 3 h, followedby CellMask (Invitrogen, Carlsbad, CA, USA) staining for 5 min.Cells were then washed 2× with PBS and incubated withIntegriSense (200 nM in PBS containing 5% serum) at 4°C or37°C for 30 min. For the competition study, cells were incubatedwith parent compound (200 μM) for 1 h before IntegriSenseincubation. Confocal microscopy was performed on a Leica SP5confocal system (Leica, Wetzlar, Germany). In separate experimentsaimed at validating the tumor model used for the antiangiogenicexperiment, HUVEC or A673 (50,000 cells) were seeded unto 8-wellLab-Tek chamberslides (Nalge-Nunc International, Naperville, IL,USA) and left to adhere overnight at 37ºC in a humidifiedatmosphere containing 5% CO2. The following day, the mediumwas changed to fresh complete medium containing IntegriSense orfree dye (20 nM), incubated for 30 min at 37ºC, washed 2× with PBS,and fixed with 2% paraformaldehyde for 30 min at 4ºC. Competitionstudies were performed by preincubating the cells for 5 min withparent compound (200 nM) before addition of IntegriSense. Thecells on microscopy slides were then allowed to air dry protectedfrom light. The slides were mounted with ProLong® Gold antifadereagent with DAPI (Molecular Probes, Eugene, OR, USA) and curedovernight at RT protected from light. Merged images of the nuclear

staining (DAPI) and fluorescent signals were collected usingfluorescence microscopy (Zeiss, Thornwood, NY, USA).

Kd and koff Determination

HEK293-αvβ3 cells (1×105) were incubated with 0, 3.12, 6.25,12.5, 20, 25, 40, 50, 80, 100, and 200 nM of IntegriSense at 4°C for30 min in “flow cytometric buffer” (145 nM NaCl, 5 mM KCl,1 mM MgCl2, and 10 mM HEPES ph 7.4). The amount of probebound to integrins on HEK293-αvβ3 cells was determined by flowcytometry (FACSCalibur, BD Biosciences, CA, USA). Data wasanalyzed using FlowJo software, and Kd values were calculatedusing SigmaPLot 10 software. For koff determination, HEK293-αvβ3 cells (1×10

6) were incubated with 100 nM of IntegriSense at4°C for 30 min, and labeled cells were then transferred into PBScontaining 10 mM of parent compound. The amount of probebound to the integrins on HEK293-αvβ3 cells was determined byflow cytometry (FACSCalibur, BD Biosciences, CA, USA) beforemixing with unlabeled compound and 1, 1.5, 2.5, 3.5, 4.5, 5.5, and24 h after mixing. Data was analyzed using FlowJo software andkoff and t1/2 values calculated using SigmaPLot 10 software.

Determination of IC50

HEK293-αvβ3 cells were lifted with trypsin-EDTA and washed fourtimes with serum-free MEM. Cells (25,000 cells/well) were addedto microtiter wells coated with vitronectin or fibronectin andallowed to attach for 2 h at 37ºC in a humidified incubator in theabsence or presence of increasing concentrations of IntegriSense orparent compound. Nonattached cells were gently washed away.Similar experiments were carried out with HEK293 cells stablytransfected with αvβ5, αIIbβ3, or α5β1. The attached cells werequantified by colorimetric detection of hexosaminidase enzymaticactivity [37, 38] in a Vmax plate reader (Molecular Devices, MenloPark, CA, USA). The number of attached cells was quantified usinga standard curve for each cell line assayed and expressed as a meanvalue of triplicate samples.

Pharmacokinetics

A bolus intravenous (i.v.) injection of IntegriSense (2 nmoles inPBS) was administered to 24 female retired breeder BALB/c mice(Charles River Laboratories, Wilmington, MA, USA). Terminalblood samples were collected by cardiac puncture from each mouse(following carbon dioxide asphyxiation). Three mice were sampledfor each time point. Plasma was collected by centrifugation, diluted1:2 in DMSO, and fluorescence read using a fluorescence platereader (Molecular Devices, Sunnyvale, CA, USA). Data wasnormalized to a standard curve prepared with known concentrationsof IntegriSense.

Biodistribution

Female NU/NU mice 6–8 weeks old (Charles River Laboratory,Wilmington, MA, USA) were injected subcutaneously (s.c.) withA673 cells (5×106) in the first mammary fat pads. Once tumorsreached the desired volume (measured with calipers using the

490 S. Kossodo, et al.: Dual In Vivo Quantification Using FMT

formula volume mm3 ¼ length� width2� �

=2), mice were injected i.v.with IntegriSense (2 nmoles). Mice were sacrificed by carbondioxide asphyxiation 24 h later, and tissues were excised, rinsedwith saline, blotted dry, and the imaged on VisEn’s FMT2500 usingthe reflectance mode. Regions of interest (ROIs) were drawn aroundeach organ using the FMT Software and the mean fluorescence(reported as counts/energy) determined for each organ and normal-ized to the mean fluorescence in tumors taken to be 100%. All animalstudies were performed according to protocols approved by VisEn’sand Merck’s Institutional Animal Care and Use Committee.

Pharmacodynamics

Female NU/NU mice 6–8 weeks old (Charles River Laboratory,Wilmington, MA, USA) were injected s.c. with MDA-MB231,HT29, or A673 cells in the first mammary fat pads. Once tumorsreached the desired volume, mice were injected i.v. with Integ-riSense in the presence or absence of parent compound (competi-tion) and imaged by FMT2500 at various time points. ROIs weredrawn over the tumors to access the picomole quantity of the probein the tumors, and then the values were normalized to the highestamount of probe (2 h time point).

In Vivo Fluorescence Molecular Tomography(FMT) Measurements

For in vivo imaging, mice were anesthetized by gas anesthesia(isoflurane/oxygen mixture), placed in VisEn’s FMT2500 systemimaging chamber one at a time and imaged. A single fiber-coupledlaser diode scans the surface of the animal at a user-adjustable pitchand span in X and Y. The laser power and exposure time at each scanpoint are automatically adjusted by the system to provide high signal

to noise while avoiding saturation. Light transilluminates the animaland is detected on the opposite side by a low-noise thermoelectricallycooled CCD detector via an array of passband filters and focusingoptics. Total scan times are on the order of 2–4 min. Quantificationaccuracy of the FMT systemwas assessed as previously described [30]with VivoTag-S680. Briefly, data was acquired from phantomscontaining known amounts of fluorochrome. Average reconstructedFMT concentrations were extracted from ROIs encompassing theentire phantom and plotted against the actual dye concentration. Underthese conditions, a highly linear relationship was obtained betweenknown concentrations and the reconstructed values with over 99.5%linearity across a broad range of concentrations spanning 50–1,000 nM (data not shown). Three-dimensional ROIs were drawnaround tumor regions, and a threshold was applied equal to 50% of themaximum value of fluorescence in the adjoining nontumor area. Thetotal amount (in picomoles) of fluorochrome was automaticallycalculated relative to internal standards generated with knownconcentrations of the appropriate dye.

Immunohistochemical Analysis and ConfocalMicroscopy

Tumor-bearing animals were euthanized at selected times (1, 3, and48 h) following IntegriSense bolus injection. Tumors were excised,embedded in OCT, snap frozen, and prepared into 10-µm-thick frozensections. FITC-labeled pan cadherin antibody (Abcam) was used as aplasma membrane stain. Confocal microscopy was performed on aLeica SP5 confocal system (Leica, Wetzlar, Germany).

Western Blot Analysis

Cells or tumors were homogenized using a hand-held homogenizer(Biospec, Bartlesville, OK, USA) in RIPA lysis buffer (Sigma, StLouis, MO, USA) supplemented with 0.5 mM phenylmethylsufonylfluoride, 10 μg/ml each of leupeptin, aprotinin, and pepstatin (all fromSigma, St Louis, MO, USA). Samples were centrifuged at 13,000 rpmfor 30 min at 4ºC. Protein concentration in the supernatants wasquantified using detergent-compatible protein assay kit (Bio-Rad DCProtein Assay, Bio-Rad, Hercules, CA, USA). Samples (50 μg ofprotein) were boiled for 10 min in Laemmli Sample Buffer (Bio-Rad,Hercules, CA, USA) and loaded onto 10% precast polyacrylamidegels (Bio-Rad, Hercules, CA, USA). After migration, proteins wereelectrophoretically transferred onto a nitrocellulose membrane (Bio-Rad, Hercules, CA, USA). The membrane was washed 3×5 min inwash buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% Tween-20), blocked 2×15 min in wash buffer/3% nonfat dry milk (Bio-Rad,Hercules, CA, USA), incubated overnight at 4ºC on a rotator in washbuffer containing a polyclonal rabbit antihuman Integrin αV antibody(number AB1930, Millipore, Billerica, MA, USA) at a 1:5,000 finaldilution, washed 3×5 min, incubated for 1 h at room temperature withan HRP-conjugated goat anti-rabbit IgG, (number 12–348, Millipore,Billerica, MA, USA) at a 1:5,000 final dilution in wash buffer andwashed 5×5 min. The blot was then incubated with the chemilumi-nescent Super Signal West Femto Max Sensitivity Substrate (ThermoScientific–Pierce Protein Research Products, Rockford, IL, USA) for1 min at room temperature and the signal detected using anAlphaImager 3400 (Alpha Innotech, San Leandro, CA, USA).

Parent non-peptide αvβ3 antagonist

HN N COOH

NN

CH3

IntegriSense

HN N COOH

NN

CH3

HN

O

NIRF

Fig. 1 Chemical structures of the parent nonpeptide αvβ3antagonist compound and IntegriSense.


Competition and Anti-Angiogenic TreatmentStudies in a Human Tumor Xenograft Model

A673 tumor-bearing mice were randomized once their tumors reachedthe desired volume and injected i.v. with either IntegriSense (2 nmoles)or free dye (2 nmoles). For competition studies, mice were injectedwith the parent compound (100 nmoles) followed 5 min later byIntegriSense (2 nmoles). For therapy studies, mice were treated with2 mg/kg Avastin (Bevacizumab, Genentech, CA, USA) intraperito-neally (i.p.) 2× a week, and injected 7 days after beginning of thetreatment with 2 nmoles each IntegriSense and ProSense (ProS-ense750, VisEn Medical, Bedford, MA, USA) in a final volume of200 μl PBS. ProSense is a protease-activatable fluorescence agent

comprised of fluorochromes linked to a lys–lys peptide scaffold andpartially methoxypegylated graft copolymers. Proteolytic cleavage ofthe scaffold by cathepsin B activity (and to a lesser extent cathepsinsK, L, and S) results in extensive fluorescence [via dequenching; 33].Imaging was performed 24 h after agent administration using theFMT2500. The wavelength of the laser is selectable at 680 and750 nm, corresponding to the absorption spectra of the fluorescentagents administered to the animal, enabling simultaneous imaging, andquantification of two spectrally distinct agents. Three-dimensionalROIs were drawn around tumor and adjacent background areas and thetotal amount of fluorescence (in picomoles) determined after applyinga uniform background thresholding to all animals (using 1× the meancontrol background mean fluorescence as the background thresholdfor both experimental and control mice).

a 4 0C

IntegriSense IntegriSense IntegriSense Cell Mask Cell Mask Parent compound

DAPI DAPI DAPI DAPI

37 0C IntegriSense, DAPI

b c

Concentration (nM)

0 10 20 30 40 50 60

Flu

ores

cenc

e

0

70

140

210

280

350

420

Time (Hours)0 1 52 3 4 6 25

400

450

500

550

600

650

Flu

ores

cenc

e

Fig. 2 Binding of IntegriSense to HEK293-αvβ3 cells. a Confocal microscopy of HEK293-αvβ3 cells incubated with IntegriSense(red) CellMask (green) as a membrane marker and DAPI (blue) for nuclear staining. Preincubation with parent compound inhibitsIntegriSense binding. Colocalization of IntegriSense with the membrane marker indicates accumulation on the surfacemembrane. Incubation of IntegriSense at 37°C shows internalization of the probe (arrows). b Similarity of equilibrium bindingisotherm values (Kd=4.2±0.6 nM) in the presence (filled square), or absence (filled circle) of Mn2+. Binding to non-αvβ3expressing HEK cells is significantly lower (filled triangle). c Dissociation (koff=1.08×10

−4 s−1, t1/2=107 min) of IntegriSense fromthe αvβ3 integrin expressed on HEK293 cells.


Statistical Evaluation

All values were expressed as mean ± SEM. Statistical significancewas determined using the two-tailed unpaired Student’s t-test.Differences were considered significant at pG0.05.

ResultsSynthesis and Characterization of IntegriSense IntegriSensewas synthesized by converting the parent nonpeptide αvβ3antagonist to the 3-cyano derivative, reducing the derivative tothe 3-aminomethyl analog, and reacting the resultingcompound with VivoTag-S680, an amine-reactive NIRfluorochrome which allows for deeper tissue penetration,reduced autofluorescence, and minimal scattering (Fig. 1).The absorption and emission spectra in aqueous solution werefound to be 674 nm/692 nm and the ε=2.2×105 M−1cm−1. Themolecular weight as calculated by LC/Mass was 1,430.4 forC67H82N8O17S5; found 1,431.5 M.

+1

In Vitro Binding Assays To determine the affinity of Integri-Sense for αvβ3, we carried out equilibrium binding anddissociation measurements using HEK293 cells stablyexpressing human αvβ3. As depicted in Fig. 2a, IntegriSensebinds to the receptors on the surface of cells, and binding can beprevented by the parent compound. Receptor recycling(endocytosis) results in IntegriSense internalization at 37°C,but not at 4°C. Equilibrium binding experiments provided,using a single binding site model, the value of apparentequilibrium binding constant, Kd=4.2±0.6 nM (Fig. 2b).IntegriSense dissociates from the αvβ3 receptors with the ratedescribed by the dissociation rate constant koff=1.08×10

−4 s−1

(t1/2=107 min, Fig. 2c).To evaluate IntegriSense in a functional cell assay and

determine the selectivity for αvβ3 when compared to αvβ5,α5β1, and αIIbβ3, we employed an in vitro cell attachmentassay. Table 1 depicts the IC50 values for the attachment ofHEK293 cells stably expressing tested integrins to vitronec-tin- or fibronectin-coated surfaces as measured for Integ-riSense and the parent compound. As compared to the parentcompound, IntegriSense shows only a fivefold loss ofaffinity to αvβ3 and a 25-fold loss of affinity to αvβ5,retaining higher potency in inhibiting binding to αIIbβ3 andlittle affinity for α5β1. IntegriSense exhibits stronger affinity

for αvβ3 integrin, as compared to αIIbβ3 (ninefold higher) andαvβ5 (eightfold higher), while no specific binding to α5β1 isdetected.

Pharmacokinetics and Biodistribution of IntegriSense Theconcentration dependence of IntegriSense in plasma wasfitted into a two compartmental model with a calculatedt(1)1/2=6 min and t(2)1/2=210 min (Fig. 3). Thus, it islikely that t(1)1/2 represents free IntegriSense clearingquickly from circulation while the second much longert(2)1/2 likely corresponds to clearance of bound IntegriSense.To better pinpoint the biodistribution of the agent in vivo,tumor-bearing mice were injected i.v. with IntegriSense andsacrificed 24 h later. At this timepoint, fluorescence is widelydistributed into the tissues with the highest concentration foundin tumors, followed by the liver, small intestine, and skin(Fig. 3b).

Pharmacodynamics and Tumor Localization Pharmaco-dynamics and biodistribution of IntegriSense were assessed inseveral human tumor xenograft models. Tumors were chosen

Time after administration (min)0 100 200 300 400 1450

Inte

griS

ense

(M

)

0.0

0.5

1.0

1.5

2.0

2.5a

b

0 20 40 60 80 100 120

Lung

Skin

Kidney

Gut

Muscle

Bone

Liver

Fat

Tumor

% of tumor fluorescence

Time after administration (min)0 100 200 300 400 1450

Inte

griS

ense

(μM

)

0.0

0.5

1.0

1.5

2.0

2.5a

b

0 20 40 60 80 100 120

Lung

Skin

Kidney

Gut

Muscle

Bone

Liver

Fat

Tumor

0 20 40 60 80 100 120

Lung

Skin

Kidney

Gut

Muscle

Bone

Liver

Fat

Tumor

% of tumor fluorescence

Fig. 3 Pharmacokinetics and biodistribution of IntegriSense.a Pharmacokinetic dependence of the IntegriSense concen-tration plasma and the calculated fit using a two compart-mental model shows rapid distribution (t(1)1/2=6 min) and along terminal half-life (t(2)1/2=200 min). b Biodistribution ofIntegriSense in the organs of A673 tumor-bearing athymicmice 24 h after agent injection shows predominant tumorlocalization.

Table 1. IC50 values of integrisense and parent compound on integrin-transfected HEK cells

IC50 (nM)

IntegriSense Parent compound

αvβ3 4.1±0.3 0.82±0.2αvβ5 33 1.3±0.2αIIbβ3 39 910,000α5β1

a 92,000 910,000

aBinding assays to vitronectin were performed for all integrins except α5β1which was performed on fibronection


based on cell lines positive for αvβ3 (MDA-MB231, HCT116,and A673) or negative for αvβ3 but positive for αvβ5 (HT29).As shown in Fig. 4a with MDA-MB231 and HT29,fluorescence signal in tumors drops within the first 6 h afterIntegriSense injection to a level which remains stable, withinthe experimental error, at least up to 24 h, with specificfluorescence still detectable up to 5–10 days. MDA-MB231tumors have significantly higher levels of IntegriSense tumorsignal as compared to HT29, from 4–24 h postinjection, a

finding we attribute to the expression of αvβ5. Injection ofIntegriSense in the presence of 100-fold excess of the parentcompound lowers the fluorescence in the MDA-MB231xenograft tumors by 60% at 24 h after the injection. A similareffect of the parent compound is observed in tumors derivedfrom other cancer cell lines positive for αvβ3 (not shown).Based on the dissociation rate (koff) and the clearance of theprobe from plasma, IntegriSense should be cleared completelywithin ~30 h Thus, the probe must be distributed later inanother compartment, likely internalized in cells. To test thishypothesis, we analyzed tumor sections dissected fromsacrificed animals 1, 3, and 48 h after probe injection usingconfocal microcopy. Images of the sections for tumors fromone of the studied cancer cell lines (A673) indicate that at1 h, most of the probe is located on the surface of the cells(bound to receptors on the surface), while at 3 h, some ofthe probe is already detected in the intracellular pools, andinternalization is almost complete by 48 h (Fig. 4b).

Cell Integrin Binding Assays To validate the specificity ofIntegriSense to target αvβ3/αvβ5-expressing cells used in thein vivo studies, we assessed the ability of endothelial andtumor cells to bind IntegriSense. Human umbilical veinendothelial cells (HUVEC, used as a surrogate for primaryendothelial cells) and human rhabdomyosarcoma A673tumors cells were seeded unto chamber slides and culturedin complete medium containing IntegriSense or free dye.Competition studies were performed by preincubating thecells for 5 min with a 10× molar excess of parentcompound. Significant fluorescent signal is observed forboth HUVEC and A673 cells labeled with IntegriSense(Fig. 5a), and the observed fluorescence intensity is reducedin the presence of a blocking dose of parent compound.Reduced or no fluorescence is detected in cells incubatedwith free dye (same as used to synthesize IntegriSense)only. Western blot analysis confirms the expression of theαv subunit both in the A673 cell line and in A673 tumorxenografts ex vivo (Fig. 5b), supporting the premise thatIntegriSense specifically binds to integrin-expressing tumorand endothelial cells in vivo.

In Vivo Imaging of Tumor Integrins To further corroboratethe specificity of integrin targeting in vivo, the agent(2 nmoles) was injected in A673 tumor-bearing mice inthe absence or presence of excess, competing parentcompound (100 nmoles). Control mice received free dye.Fluorescence is readily detected in tumors (Fig. 6a) and theunlabeled parent compound effectively reduces the totalamount of IntegriSense signal in tumors (similar to theobservations in MDA-MB231 xenografts as previouslydescribed). Under the same imaging conditions, nodetectable fluorescence is observed in the tumors of miceinjected with free dye. The fluorescent signal was quantifiedby drawing a three-dimensional ROI encompassing eachtumor. As shown in Fig. 6b, the total amount of IntegriSense

b 1 hr 3 hr 48 hr

Time after admnistration (hours)0 4 8 12 16 20 24 100 200

Nor

mal

ized

Flu

ores

cenc

e

0.0

0.2

0.4

0.6

0.8

1.0

MDA-MB231 -IntegriSenseMDA-MB231 -IntegriSense + Parent CompoundHT29 -IntegriSens

a

Fig. 4. Pharmacodynamics and tumor localization. a Phar-macodynamics of IntegriSense in xenograft tumor models inthe absence (filled circle and filled triangle) or presence(filled square) of a molar excess of parent compound.b Confocal microscopy of tumor sections taken 1, 3, and48 h after agent administration (IntegriSense—red, anti-pancadherin membrane marker—green, and DAPI-nucleusstaining—blue).


fluorescence is significantly decreased in tumors of micehaving received the parent compound (by 67%, p=0.011) ordye only (p=0.00038) as compared to tumors of mice injectedwith IntegriSense. These results demonstrate the targetingspecificity using an integrin-sensing agent in vivo.

Simultaneous In Vivo Imaging of Integrin-targeted andProtease-activatable Agents and their Response to Anti-angiogenic Therapy We next assessed the possibility ofsimultaneously imaging and quantifying separate biologicalprocesses and their response to treatment with a known anti-VEGF monoclonal antibody, Avastin. Mice bearing A673tumors of comparable size were randomized and treatedbiweekly with either 2 mg/kg Avastin i.p. or vehicle (PBS).A week later, mice received IntegriSense and ProSenseconcurrently and imaged 24 h later by FMT2500 (Fig. 7a).Because these two agents are spectrally distinct, it is possible toimage them simultaneously without spectral interference (i.e.,“crosstalk” between the two fluorescence channels). Both

IntegriSense and ProSense signal in tumors are significantlydecreased following Avastin treatment (by 57%, p=0.0008 and51%, p=0.036, respectively, Fig. 7b), reflecting a decrease intumor volume (by 58.3%). Moreover, the simultaneousimaging of these two agents reveals different patterns ofdistribution (Fig. 7).

DiscussionOptical molecular imaging has opened the door to visualiz-ing specific biomarkers, non invasively and in real-time, indisease diagnosis, prognosis, staging, and therapy monitor-ing [18]. One class of biomarkers is integrins, which haveemerged as significant indicators in cancer, cardiovasculardisease, and bone remodeling [39]. Given their role in suchdiverse pathologies, a considerable effort has been investedin developing antagonists for use in patients [9–11, 40] andin synthesizing molecules that are able to trace theirlocalization and regulation in vivo. Most of these effortshave concentrated on radiolabeled RGD-based moleculeswhich, while effective, are burdened by their high cost, long

201kd150kd

126kd90kd

30ng 10ng 5ngA673Cells

A673Tumor

V Positive Control

HUVEC

A673

IntegriSense +Parent Compounda

b

150kd

90kd

30ng 10ng 5ngA673Cells

A673Tumor

V Positive Control

HUVEC

A673

IntegriSenseIntegriSense +

Parent Compound VivoTag-S680

α

Fig. 5. Specific binding of IntegriSense by endothelial andA673 tumor cells in vitro and expression of αv subunit byA673 tumors ex vivo. a Fluorescence microscopy images ofHUVECs and A673 tumor cells incubated with 20 nMIntegriSense for 30 min at 37ºC, showing endocytosis ofthe agent (IntegriSense in red, nuclei counterstained withDAPI in blue) with a 10× excess (200 nM) parent compoundshowing effective blocking of the NIRF signal and incubationwith control free dye only, showing no detectable uptake inHUVECs and decreased uptake in A673 cells. Note thespecificity of IntegriSense binding to both cell types. bWestern blot analysis of the expression of the αv subunit inA673 tumors ex vivo confirming that A673 tumors havesignificant levels of αv subunit, common to αvβ3 and αvβ5.



a

bnM

0

5

10

15

20

25

30


Parent Compound

Flu

ores

cenc

e (p

mol

es)

VivoTag-S680



a

bnM

200

150

100

50

0

200

150

100

50

0

0

5

10

15

20

25

30


Parent Compound

Flu

ores

cenc

e (p

mol

es)

VivoTag-S680

p=0.011p=0.011

p=0.00038p=0.00038

Fig. 6. In vivo imaging and quantification of IntegriSensesignal in A673 tumor-bearing mice. a Representative volumerenderings taken at the same color gating from tumor-bearing mice injected with IntegriSense, coinjected withIntegriSense, and 200 nmoles of parent compound orinjected with free dye only. b The total amount of fluores-cence (pmol) was quantified in specific ROIs encompassingeach tumor. Coinjection with the parent compound results ina significant 67% decrease in IntegriSense signal (p=0.011).Mice injected with free dye have significantly lower tumorsignal as compared to those injected with IntegriSense(p=0.00038).


imaging time, and their use of ionizing radiation. Tocircumvent these limitations, researchers have developedNIRF molecular agents, including small integrin antagonistsfor optical imaging [41] and instrumentation enabling theirdetection. This approach has been used successfully to detectintegrins in vivo either alone [22–27, 42, 43] or incombination with other imaging modalities [28, 44] but hasfaced limitations in the inability to quantify and spectrallyresolve distinct fluorescent agents in vivo. Fluorescencetomographic imaging, a method for three-dimensional local-ization and quantification of fluorochromes in deep tissues,was developed using photon measurements collected atdifferent projections through tissue and appropriate tomo-graphic schemes that take into account the diffusive natureof photon propagation [29, 30]. Optical tomography,performed in the context of appropriate laser/filter combina-tions, allows excitation and fluorescence detection at multi-ple wavelengths, thus, making it possible to image multiple

agents simultaneously and quantify multiple biologicalevents.

The major drawbacks of all RGD peptide-based imagingmolecules are their nonselectivity among a subset of integrinfamily members. There are 24 integrins, eight of which bindto RGD [2, 45]. They also exhibit relatively low affinity toRGD as compared to nonpeptide αvβ3 antagonists [23–26,41]. Fluorescently labeled RGD imaging agents were shownto have IC50 values in the 20–70 nM range when usingechistatin or αvβ3 as tracers [23, 24], compared to 4.1 nM forIntegriSense towards αvβ3 in the present study. Efforts toimprove affinity by attaching several RGD-peptides to onereporting molecule have been problematic due to theincrease in probe size, which can potentially alter thebiodistribution and increase nonspecific accumulation [46].Furthermore, such approaches to improve RGD-peptideaffinity do not improve their lack of selectivity amongvarious integrin family members. In order to avoid these two

a

b c

0

10

20

30

40

50

60

70

80

IntegriSense ProSense

Tum

or f

luor

esce

nce

(pm

oles

)

VehicleAvastin

p=0.0008

p=0.036

Day 1 Day 1Day 8 Day 80

100

200

300

400

500

600

700

Vehicle Avastin

Tum

or v

olum

e (m

m3 )

p=0.022

p=0.717

Day 1 Day 1Day 8 Day 80

100

200

300

400

500

600

700

Vehicle Avastin

Tum

or v

olum

e (m

m3 )

p=0.022

p=0.717

VehicleAvastin

Inte

griS

ense

Pro

Sens

e

415

344

273

201

130

415

344

273

201

130200

157

114

71

28

200

157

114

71

28

nM

Fig. 7. Effect of antiangiogenic Avastin treatment on integrin and cathepsin B signals in A673 tumor-bearing mice. aRepresentative volume renderings of IntegriSense and ProSense. Note the differential localization of fluorescent signal withIntegriSense and ProSense in the same animals (IntegriSense in blue and ProSense in red). b Quantification of tumorfluorescence shows that IntegriSense and Prosense signal in tumors is decreased following Avastin treatment (p=0.0008 andp=0.013, respectively). cAvastin also significantly inhibits tumor growth as determined by tumor volume calipersmeasurements (by58.4%, p=0.022).


important concerns with RGD peptide-based probes, wedecided to engineer an imaging agent by modifying lowmolecular weight peptidomimetics that have a strong affinityfor αvβ3 integrin and have been shown to be highly selectiveantagonists (ligands) of active αvβ3 integrin [35, 41].Covalent attachment of an optimal fluorescence moietyprovided optical imaging probes with high affinity andselectivity for αvβ3 integrin (Fig. 3 and Table 1).

Since the active form of αvβ3 integrin is on the surface oftumor cells as well as on new vessel cells [47, 48], it wasimportant to measure the affinity for αvβ3 integrin in thepresence and absence of Mn2+ ions, which facilitate theactive conformation of the integrin. As demonstrated inFig. 3b the affinity as well as the amplitude of thefluorescence increase is identical in the presence as well asthe absence of Mn2+ ions. Both experimental isotherms arebest described using a single binding site model providing aKd of 4.2 nM. This data indicates that IntegriSense binds toHEK293-αvβ3 cells via what is characterized in the literatureas an active form of the integrin [47–50]. However, the datadoes not provide any experimental evidence on the affinityof IntegriSense for the resting form of the receptor. Based onprevious results with other peptidomimetics [50], it is alsopossible that IntegriSense binds with much lower affinity tothe resting form of the receptor and then stimulates aconformational transition of the receptor to the active formand consequently high affinity binding of the IntegriSense.Such information cannot be obtained from equilibriumbinding measurements presented in this paper. The valueof Kd measured in equilibrium binding assay agreesremarkably well with the IC50 of 4.1 nM measured usinginhibition of adhesion of the same HEK293-αvβ3 cells usedin the binding assay to the endogenous ligand vitronectin.Since the binding of cells to vitronectin requires the activeform of the integrin, this may be an indirect indication thatKd obtained in binding measurements represents affinity ofthe IntegriSense for the active form of the receptor.

The functional assay has been further used to determinethe selectivity of IntegriSense in binding to other integrins.As listed in Table 1, the selectivity of IntegriSense is similarto the original peptidomimetic with the exception of αIIbβ3integrin, which controls the binding of platelets to fibrinogenbut plays no role in angiogenesis. The difference in theaffinities of IntegriSense to αvβ3 and αvβ5 integrin couldexplain the difference in the detected specific tumorfluorescence in MDA-MB231 (αvβ3-expressing) and HT29(low αvβ3, high αvβ5) tumor xenograft studies (Fig. 4a) inwhich MDA-MB231-bearing mice showed higher and moreprolonged tumor fluorescence. Selectivity of IntegriSensefor different integrins involved in tumor growth provide aunique opportunity to evaluate the individual functions ofintegrins and potentially serve in the evaluation of integrin-based cancer therapeutics. Cancer metastasis has beenassociated in many cases with the expression of αvβ3integrin [51–53]. Thus, the selective reporter of αvβ3 integrinactivity allows for the detection of metastatic activity of

individual cancer cells and identification of metastases witha potential for transition to clinical applications in thedetection of colon cancer using endoscopy.

The rate, constant for the dissociation of IntegriSensefrom αvβ3 integrin (Fig. 3; koff=1.08×10

−4 s−1), agrees wellwith the dissociation rate constant of fluorescence reporter ofαIIbβ3 integrin on human platelets [50], indicating again thatIntegriSense binds to the activated form of αvβ3 integrin.Ligand binding to the activated form of the αvβ3 integrinresults in ligand internalization (endocytosis) and accumu-lation of the ligand in the cells [25]. Indeed, incubation ofIntegriSense with HEK293–αvβ3 cells at 4°C, when the cellmembrane is in a “frozen stage,” results in the binding ofIntegriSense on the surface of the cells, while whenincubated at 37°C the probe is also internalized (Fig. 3a).The internalization of IntegriSense introduces a practicalupper limit to the frequency of repeat imaging of this agentin longitudinal studies. The amount of internalized probe iscontrolled by the affinity of the probe for the integrin and therate of the probe clearance from the circulation. Generally,the higher the affinity of the probe the more probe would beexpected to be internalized. Furthermore, internalization ofthe probe is also likely to increase with slower clearancefrom the circulation. Significant internalization of Integri-Sense in the cancer cells (Fig. 4b) is likely responsible forthe slow clearance of the probe from the tumors (Fig. 4a).This is supported by our findings that the probe bound tosurface αvβ3 integrins is cleared within 6 h after the probeinjection and that competition with parent compoundaccelerates the clearance of IntegriSense. Integrin αvβ3 andαvβ5 are involved in tumor angiogenesis and/or tumor cellmigration and metastasis [4, 54]. We have found thatIntegriSense detects tumor integrin in subcutaneous xeno-graft tumors on the flanks of NU/NU mice using severalcancer cell lines (MDA-MB231, A673, HCT116, andHT29), hinting towards selective imaging of αvβ3- overαvβ5-expressing tumor lines. IntegriSense binding to MDA-MB231 and A673 tumors can be blocked by competition ofthe high-affinity integrin-binding parent compound (Figs. 4and 6; 24 h). Thus, imaging IntegriSense fluorescence intumor allows detecting and quantifying activated αvβ3integrin within the tumor. The tomographic measurementof fluorophore concentrations provides a new avenue forpreclinical and potentially clinical, evaluation of αvβ3integrin-based cancer therapeutics.

The power of fluorescence tomography with multiplexdetection (i.e., simultaneously, using probes of differentspectral characteristics detecting different biologies) hasbeen demonstrated using IntegriSense (excitation 670 nm)and ProSense (excitation 750 nm) for the detection of theeffects of the commercial cancer therapeutic Avastin onhuman A673 tumor xenograft growth (Fig. 7). Avastin, ahumanized anti-VEGF monoclonal antibody (bevacizumab)has been shown to inhibit tumor growth in A673 rhabdo-myosarcoma tumors implanted in nude mice at the dosesused in the present study [55]. As expected, tumor growth


was inhibited after Avastin treatment (Fig. 7). Inhibition oftumor growth following Avastin treatment was detected as asignificant drop in tumor-localized IntegriSense signal andProSense fluorescence (detecting predominantly tumor andtumor-associated macrophage cathepsin B activity). Asexpected, the distribution of fluorescence within the tumorsdiffered between the ProSense and IntegriSense probes(Fig. 7a), revealing differences in the localization of αvβ3integrin and cathepsin B within the tumor. Thus, this newimaging technology in combination with newly designedspecific molecular imaging agents opens new ways ofdetecting, quantifying, and understanding the importantbiological mechanisms that occur during tumor developmenttherapeutic intervention.

ConclusionWe have generated and characterized a high-affinity,selective integrin-binding agent (IntegriSense) that is usefulas a near-infrared imaging probe. This novel optical agent, inthe context of noninvasive fluorescence tomography, estab-lishes it as a powerful tool in quantifying integrin levelsassociated with tumor growth and the changes that occurupon antiangiogenic treatment. The activation of αvβ3integrin is not limited to tumor growth but is detected inmany other physiological and pathophysiological states [54].Indeed, recent studies have already highlighted the potentialof IntegriSense imaging in the detection of lesions intransgenic mouse models of atherosclerosis [56, 57], andIntegriSense may also be useful in the study of acute lunginflammation, acute and chronic neutrophilia, osteoclast-driven bone remodeling, arterial injury, and arthritis.

.

References1. Ruoslahti E (1996) RGD and other recognition sequences for integrins.

Annu Rev Cell Dev Biol 12:697–7152. Lu X, Lu D, Scully MF et al (2008) The role of integrin-mediated cell

adhesion in atherosclerosis: pathophysiology and clinical opportunities.Curr Pharm Des 14:2140–2158

3. Stupack DG (2007) The biology of integrins. Oncology 21:6–124. Hood JD, Cheresh DA (2002) Role of integrins in cell invasion and

migration. Nat Rev Cancer 2:91–1005. Friedlander M, Theesfeld CL, Sugita M et al (1996) Involvement of

integrins αvß3 and αv ß5 in ocular neovascular diseases. Proc Natl AcadSci USA 93:9764–9769

6. Brooks PC, Clark RA, Cheresh DA (1994) Requirement of vascularintegrin alphavbeta3 for angiogenesis. Science 264:569–571

7. Gasparini G, Brooks PC, Biganzoli E et al (1998) Vascular integrinαvß3: a new prognostic indicator in breast cancer. Clin Cancer Res4:2625–2634

8. Vonlaufen A, Wiedle G, Borisch B, Birrer S, Luder P, Imhof BA (2001)Integrin αvß3 expression in colon carcinoma correlates with survival.Mod Pathol 14:1126–1132

9. Gutheil JC, Campbell TN, Pierce PR et al (2000) Targeted antiangio-genic therapy for cancer using Vitaxin: a humanized monoclonalantibody to the integrin αvβ3. Clin. Cancer Res 6:3056–3061

10. Mulgrew K, Kinneer K, Yao XT et al (2006) Direct targeting of αvβ3integrin on tumor cells with a monoclonal antibody, Abegrin™. MolCancer Ther 5:3122–3129

11. Hariharan S, Gustafson D, Holden S et al (2007) Assessment of thebiological and pharmacological effects of the αvβ3 and αvβ5 integrin

receptor antagonist, cilengitide (EMD 121974), in patients withadvanced solid tumors. Ann Onc 18:1400–1407

12. Haubner R, Wester HJ (2004) Radiolabeled tracers for imaging oftumor angiogenesis and evaluation of anti-angiogenic therapies. CurrPharm Des 10:1439–1455

13. Dijkgraaf I, Beer AJ, Wester HJ (2009) Application of RGD-containingpeptides as imaging probes for alphavbeta3 expression. Front Biosci14:887–899

14. Chen X, Park R, Tohme M, Shahinian AH, Bading JR, Conti PS (2004)MicroPET and autoradiographic imaging of breast cancer αv-integrinexpression using 18F- and 64Cu-labeled RGD peptide. Bioconjug Chem15:41–49

15. Chen X, Park R, Shahinian AH et al (2004) 18F-labeled RGD peptide:initial evaluation for imaging brain tumor angiogenesis. Nucl Med Biol31:179–189

16. Haubner R, Wester HJ, Weber WA et al (2001) Noninvasive imaging ofαvβ3 integrin expression using 18F-labeled RGD-containing glycopeptideand positron emission tomography. Cancer Res 61:1781–1785

17. Janssen ML, Oyen WJ, Dijkgraaf I et al (2002) Tumor targeting withradiolabeled αvβ3 integrin binding peptides in a nude mouse model.Cancer Res 62:6146–6151

18. Massoud TF, Gambhir SS (2003) Molecular imaging in living subjects:seeing fundamental biological processes in a new light. Genes & Dev17:545–580

19. Leong-Poi H, Christiansen J, Klibanov AL, Kaul S, Lindner JR (2003)Noninvasive assessment of angiogenesis by ultrasound and microbubblestargeted to αv integrins. Circulation 107:455–460

20. Ellegala DB, Leong-Poi H, Carpenter JE et al (2003) Imaging tumorangiogenesis with contrast ultrasound and microbubbles targeted toαvβ3. Circulation 108:336–341

21. Sipkins DA, Cheresh DA, Kazemi M, Nevin LM, Bednarski MD, Li KCP(1998) Detection of tumor angiogenesis in vivo by alphavbeta3-targetedmagnetic resonance imaging. Nat Med 4:623–626

22. Jin ZH, Josserand V, Foillard S et al (2007) In vivo optical imaging ofintegrin alphaV-beta3 in mice using multivalent or monovalent cRGDtargeting vectors. Mol Cancer 12:41–49

23. Wu Y, Cai W, Chen X (2006) Near-infrared fluorescence imaging oftumor integrin αvβ3 expression with Cy7-labeled RGD multimers. MolImaging Biol 8:226–236

24. Chen X, Conti PS, Moats RA (2004) In vivo near-infrared fluorescenceimaging of integrin alphavbeta3 in brain tumor xenografts. Cancer Res64:8009–8014

25. Cheng Z, Wu Y, Xiong Z, Gambhir SS, Chen X (2005) Near-infraredfluorescent RGD peptides for optical imaging of integrin alphavbeta3expression in living mice. Bioconjug Chem 16:1433–1441

26. Wang W, Ke S, Wu Q et al (2004) Near-infrared optical imaging ofintegrin alphavbeta3 in human tumor xenografts. Mol Imaging 3:343–351

27. Hsu AR, Hou LC, Veeravagu A, Greve JM, Vogel H, Tse V, Chen X(2006) In vivo near-infrared fluorescence imaging of integrin alphavbeta3in an orthotopic glioblastoma model. Mol Imaging Biol 8:315–323

28. Cai W, Chen K, Li ZB, Gambhir SS, Chen X (2007) Dual-functionprobe for PET and near-infrared fluorescence imaging of tumorvasculature. J Nucl Med 48:1862–1870

29. Ntziachristos V, Ripoll J, Wang LV, Weissleder R (2005) Looking andlistening to light: the evolution of whole-body photonic imaging. NatBiotechnol 23:313–320

30. Graves EE, Ripoll J, Weissleder R, Ntziachristos VA (2003)Submillimeter resolution fluorescence molecular imaging system forsmall animal imaging. Med Phys 30:901–911

31. Montet X, Ntziachristos V, Grimm J, Weissleder R (2005)Tomographic fluorescence mapping of tumor targets. Cancer Res65:6330–6336

32. Montet X, Figueiredo JL, Alencar H, Ntziachristos V, Mahmood U,Weissleder R (2007) Tomographic fluorescence of tumor vascularvolume in mice. Radiology 242:751–758

33. Nahrendorf M, Sosnovik DE, Waterman P et al (2007) Dual channeloptical tomographic imaging of leukocyte recruitment and proteaseactivity in the healing myocardial infarct. Circ Res 100:1218–1225

34. Kos J, Lah TT (1998) Cysteine proteinases and their endogenousinhibitors: target proteins for prognosis, diagnosis and therapy in cancer.Oncol Rep 5:1349–1361

35. Coleman PJ, Brashear KM, Askew BC et al (2004) Nonpeptide αvβ3Antagonists. Part11: discovery and preclinical evaluation of potent αvβ3


antagonists for the prevention and treatment of osteoporosis. J MedChem 47:4829–4837

36. Wang J, BreslinMJ, Coleman PJ et al (2004) Non-peptide αvβ3 antagonists.Part 7: 3-Substituted tetrahydro-[1,8]naphthyridine derivatives. BioorgMed Chem Lett 14:1049–1052

37. Simon KO, Nutt EM, Abraham DG, Rodan GA, Duong LT (1997) Thealphavbeta3 integrin regulates alpha5beta1-mediated cell migrationtoward fibronectin. J Biol Chem 272:29380–29389

38. Landegren U (1984) Measurement of cell numbers by means of theendogenous enzyme hexosaminidase. Applications to detection of lym-phokines and cell surface. antigens. J. Immunol. Methods 67:379–388

39. Rust WL, Carper SW, Plopper GE (2002) The promise of integrins aseffective targets for anticancer agents. J Biomed Biotech 2:124–130

40. Reardon DA, Fink KL, Mikkelsen T et al (2008) Randomized phase IIstudy of cilengitide, an integrin-targeting arginine-glycine-aspartic acidpeptide, in recurrent glioblastoma multiforme. J Clin Onc 26:5610–5617

41. Burnett CA, Xie J, Quijano J et al (2005) Synthesis, in vitro, and invivo characterization of an integrin alpha(v)beta(3)-targeted molecularprobe for optical imaging of tumor. Bioorg Med Chem 13:3763–3771

42. Achilefu S, Bloch S, Markiewicz MA et al (2005) Synergistic effects oflight-emitting probes and peptides for targeting and monitoring integrinexpression. Proc Natl Acad Sci USA 102:7976–7981

43. Aina OH, Marik J, Gandour-Edwards R, Lam KS (2005) Near-infraredoptical imaging of ovarian cancer xenografts with novel alpha 3-integrinbinding peptide “OA02”. Mol Imaging 4:439–447

44. Li C, Wang W, Wu Q et al (2006) Dual optical and nuclear imaging inhuman melanoma xenografts using a single targeted imaging probe.Nucl Med Biol 33:349–358

45. Humphries JD, Byron A, Humphries MJ (2006) Integrin ligands at aglance. J Cell Sci 119:3901–3903

46. Montet X, Funovics M, Montet-Abou K et al (2006) Multivalent effects ofRGDpeptides obtained by nanoparticle display. JMedChem49:6087–6093

47. Takagi J, Petre BM, Walz T, Springer TA (2002) Global conformationalrearrangements in integrin extracellular domains in outside-in andinside-out signaling. Cell 110:599–611

48. Adair BD, Xiong J-P, Maddock C, Goodman SL, Arnaout MA, YeagerM (2005) Three-dimensional EM structure of the ectodomain of integrinalphaV-beta3 in a complex with fibronectin. J Cell Biol 168:1109–1118

49. Xiong JP, Stehle T, Zhang R et al (2002) Crystal structure of theextracellular segment of integrin αVβ3 in complex with an Arg-Gly-Aspligand. Science 296:151–155

50. Bednar B, Cunningham ME, McQueney PA et al (1997) Flowcytometric measurement of kinetic and equilibrium binding parametersof arginine-glycine-aspartic acid ligands in binding to glycoprotein IIb/IIIa on platelets. Cytometry 28:58–65

51. Sloan EK, Pouliot N, Stanley KL et al (2006) Tumor specific expressionof αvβ3 integrin promotes spontaneous metastasis of breast cancer tobone. Breast Cancer Res 8:1–14

52. Cooper CR, Chay CH, Pienta KJ (2002) The role of alpha(v)beta(3) inprostate cancer progression. Neoplasia 4:191–194

53. Bakewell SJ, Nestor P, Prasad S et al (2003) Platelet and osteoclastbeta3 integrins are critical for bone metastasis. Proc Natl Acad Sci USA100:14205–14210

54. Carmeliet P, Jain RK (2000) Angiogenesis in cancer and other diseases.Nature 407:249–257

55. Presta LG, Chen H, O’Connor SJ et al (1997) Humanization of ananti-vascular endothelial growth factor monoclonal antibody for thetherapy of solid tumors and other disorders. Cancer Res 57:4593–4599

56. Sadeghi MM, Krassilnikova S, Zhang J et al (2004) Detection of injury-induced vascular remodeling by targeting activated avb3 integrin invivo. Circulation 110:84–90

57. Bednar B, Lin SH, Gaspar R et al (2008) In vivo imaging detection ofcathepsin activity and angiogenesis in mouse models of atherosclerosis.Abstract 2051 World Molecular Imaging Congress, Nice, France,September 10–14


Dual In Vivo Quantification of Integrin-targeted and Protease-activated Agents in Cancer Using Fluorescence Molecular Tomography (FMT)

Documents