Available online at www.sciencedirect.com

Nuclear Medicine and Biology 39 (2012) 560–569www.elsevier.com/locate/nucmedbio

68Ga-labeling and in vivo evaluation of a uPAR binding DOTA- andNODAGA-conjugated peptide for PET imaging of invasive cancers

Morten Perssona,b,c, Jacob Madsenb, Søren Østergaardd, Michael Plouga,e, Andreas Kjaera,b, c,⁎aThe Danish-Chinese Center for Proteases and Cancer, Denmark

bDepartment of Clinical Physiology, Nuclear Medicine and PET, Center for Diagnostic Investigations, Rigshospitalet, 2100-Copenhagen, DenmarkcCluster for Molecular Imaging, Faculty of Health Sciences, University of Copenhagen, 2100-Copenhagen, Denmark

dNovo Nordisk A/S, Diabetes Protein and Peptide Chemistry, 2760-Måløv, DenmarkeFinsen Laboratory, Rigshospitalet, 2100-Copenhagen, Denmark

Received 26 August 2011; received in revised form 3 October 2011; accepted 10 October 2011

Abstract

Introduction: The urokinase-type plasminogen activator receptor (uPAR) is a well-established biomarker for tumor aggressiveness andmetastatic potential. DOTA-AE105 and DOTA-AE105-NH2 labeled with 64Cu have previously been demonstrated to be able tononinvasively monitor uPAR expression using positron emission tomography (PET) in human cancer xenograft mice models. Here weintroduce 68Ga-DOTA-AE105-NH2 and

68Ga-NODAGA-AE105-NH2 and evaluate their imaging properties using small-animal PET.Methods: Synthesis of DOTA-AE105-NH2 and NODAGA-AE105-NH2 was based on solid-phase peptide synthesis protocols using theFmoc strategy. 68GaCl3 was eluted from a 68Ge/68Ga generator. The eluate was either concentrated on a cation-exchange column orfractionated and used directly for labeling. For in vitro characterization of both tracers, partition coefficient, buffer and plasma stability,uPAR binding affinity and cell uptake were determined. To characterize the in vivo properties, dynamic microPET imaging was carried out innude mice bearing human glioma U87MG tumor xenograft.Results: In vitro experiments revealed uPAR binding affinities in the lower nM range for both conjugated peptides and identical to AE105.Labeling of DOTA-AE105-NH2 and NODAGA-AE105-NH2 with

68Ga was done at 95°C and room temperature, respectively. The highestradiochemical yield and purity were obtained using fractionated elution, whereas a negative effect of acetone on labeling efficiency forNODAGA-AE105-NH2 was observed. Good stability in phosphate-buffered saline and mouse plasma was observed. High cell uptake wasfound for both tracers in U87MG tumor cells. Dynamic microPET imaging demonstrated good tumor-to-background ratio for both tracers.Tumor uptake was 2.1% ID/g and 1.3% ID/g 30 min postinjection and 2.0% ID/g and 1.1% ID/g 60 min postinjection for 68Ga-NODAGA-AE105-NH2 and 68Ga-DOTA-AE105-NH2, respectively. A significantly higher tumor-to-muscle ratio (Pb.05) was found for 68Ga-NODAGA-AE105-NH2 60 min postinjection.Conclusions: The use of 68Ga-DOTA-AE105-NH2 and

68Ga-NODAGA-AE105-NH2 as the first gallium-68 labeled uPAR radiotracers fornoninvasive PET imaging is reported, which combine versatility with good imaging properties. These new tracers thus constitute aninteresting alternative to the 64Cu-labeled version (64Cu-DOTA-AE105 and 64Cu-DOTA-AE105-NH2) for detecting uPAR expression intumor tissue. In our hands, the fractionated elution approach was superior for labeling of peptides, and 68Ga-NODAGA-AE105-NH2 is thefavored tracer as it provides the highest tumor-to-background ratio.© 2012 Elsevier Inc. All rights reserved.

Keywords: uPAR; MicroPET; NODAGA; DOTA; AE105; Gallium-68; Cancer xenograft

⁎ Corresponding author. Department of Clinical Physiology, NuclearMedicine & PET, Copenhagen University Hospital-4012, 2100-Copenha-gen Ø, Denmark.

E-mail address: akjaer@sund.ku.dk (A. Kjaer).

0969-8051/$ – see front matter © 2012 Elsevier Inc. All rights reserved.doi:10.1016/j.nucmedbio.2011.10.011

1. Introduction

The urokinase-type plasminogen activator (uPA) and itsreceptor (uPAR) have been implicated in cancer as a markerfor poor prognosis in a variety of humanmalignancies such asbreast, colorectal and gastric cancer [1–4]. uPAR expressionis particularly abundant at the invasive front of tumors or in

561M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

the surrounding stroma cells. uPAR is therefore generallyrecognized as a molecular marker for tumor invasion andmetastatic disease and is therefore also considered animportant target in cancer research [3]. The ability tovisualize and quantify uPAR expression noninvasively invivo is thus attractive from a clinical perspective [5–8].

Based on an unbiased selection in a naive phage displaylibrary by cell lines expressing high levels of uPAR, a familyof linear peptide antagonists of the uPA·uPAR interactionwas developed after affinity maturation [9]. The resulting 9-mer lead peptide denoted AE105 [9] forms a tight 1:1complex with purified human uPAR displaying a KD of 0.4nM with a koff of 2×10−4 s−1 as measured by surfaceplasmon resonance. AE105 is a potent competitive inhibitorof the uPA·uPAR interaction, displaying an IC50 value of 11nM in a purified system [9].

We have recently explored the use of this peptide forpositron emission tomography (PET) imaging of uPARexpression [10, 11]. In both studies, DOTA was conjugatedto the N-terminal of the targeting peptides (DOTA-AE105and DOTA-AE105-NH2), which were subsequently labeledwith the long-lived PET isotope 64Cu (T1/2=12.7 hr,β+=17.8%) (64Cu-DOTA-AE105) and investigated in ahuman cancer xenograft mice model. A quantitativecorrelation between uPAR expression and the tumor uptakeof 64Cu-DOTA-AE105-NH2 in several different humanxenograft in mice was recently reported [10], thus illustratingthe ability to noninvasively detect uPAR expression in vivo.Because of the limited availability due to the cyclotron-dependent production of 64Cu, the use in medical centersworldwide is complicated logistically. With the increaseduse of the 68Ge/68Ga generator during the last decade, theadvancement of 68Ga-based PET imaging agents has begunoffering a very cost-effective alternative to the on-sitecyclotron [12, 13]. 68Ga has some promising physicalcharacteristics (T1/2=68 min, β+=89%) for imaging since thephysical half-life more resembles the half-life of peptides invivo and it has a higher positron abundance than 64Cu.

Here we introduce the first 68Ga-labeled peptides for PETimaging of uPAR. The amide form of the small linearpeptide AE105 was conjugated with the macrocyclicchelators DOTA (DOTA-AE105-NH2) and NODAGA [14](NODAGA-AE105-NH2) in the N-terminal. Both peptideswere labeled with 68GaCl3 eluate after either a cation-exchange column purification step or a fractionation of theeluate [15] in order to compare the two different approaches.Finally, the in vitro uPAR binding properties and stabilitywere investigated together with dynamic in vivo PETimaging in nude mice bearing tumor xenograft of theuPAR-positive human glioblastoma cell line U87MG [10].

2. Materials and methods

2.1. Chemical and biological reagents

All commercial chemicals were of analytical grade. Theywere all used without further purification. 2-(4,7,10-tris(2-

tert-butoxy-2-oxoethyl)-1,4,7,10-tetraazacyclo-dodecan-1-yl)-acetic acid(DOTA-tris(tBu)ester and 4-(4,7-bis(2-(tert-butoxy)-2-oxoethyl)-1,4,7-triazacyclononan-1-yl)-5-(tert-butoxy)-5-oxopentanoic acid [NODAGA-tris(tBu)ester]were purchased from CheMatech (Dijon, France). 68GaCl3was eluted in 0.1 M HCl obtained from a 68Ge/Ga68

generator (Eckert & Ziegler) at the Department of NuclearMedicine & PET, Copenhagen University Hospital. GaCl3was purchased from Sigma-Aldrich.

2.2. Peptide synthesis of DOTA-AE105-NH2 andNODAGA-AE105-NH2

The amide form of the lead peptide [AE105-NH2: Asp-Cha-Phe-(D)Ser-(D)Arg-Tyr-Leu-Trp-Ser-CONH2] wassynthesized on Tentagel S RAM resin (Rapp Polymere,Germany) using traditional Fmoc solid-phase peptidechemistry, and the macrocyclic chelators (DOTA orNODAGA) were conjugated to the N-terminal coupled asdescribed previously [10]. Incubation of 20 nmol conjugatedpeptides with a twofold molar excess of the stable Ga3+ at 5mM in water at room temperature led to complete complexformation of Ga-DOTA-AE105-NH2 and Ga-NODAGA-AE105-NH2 peptides, respectively, as revealed by analyticalreverse phase high-performance liquid chromatography(HPLC) and Matrix-assisted laser desorption/ionizationmass spectrometry (MALDI-MS) (data not shown).

2.3. In vitro affinity binding

The IC50 values of the synthetic peptides for inhibition ofthe uPA·uPAR interaction were measured by surfaceplasmon resonance using a Biacore3000 as recentlydescribed [10]. In brief, a high density of purifiedrecombinant human pro-uPA was immobilized on a CM5sensor chip by amine coupling (∼6–-7000 RU). A lowconcentration of purified uPAR (0.5 nM) was preincubatedwith serial threefold dilutions of the synthetic peptidescovering 0.1 nM to 300 μM.

2.4. Radiolabeling of DOTA-AE105-NH2 andNODAGA-AE105-NH2

Radiolabeling of DOTA-AE105-NH2 and NODAGA-AE105-NH2 with 68Ga was performed by either a cation-exchange column approach or using fractionated elution [15](Fig. 1A, B). For the cation-exchange column approach, the68Ga solution was eluted from the generator with 6 ml of0.1 M HCl. The generator eluate was directly passed onto aStrata XC column, dried with air and eluted with 700 μl0.02 M HCl/acetone (2:98). A total of 100 μl of this eluatewas then added to 5 nmol conjugated peptide dissolved in300 μl H2O. For DOTA-AE105-NH2, 100 μl 0.1 M sodiumacetate (pH=3.5) was added to adjust the pH in the solution,whereas for NODAGA-AE105-NH2, 100 μl 0.1 M sodiumacetate (pH=5.0) was used. The complex formations wereperformed at 90°C for 10 min or at room temperature for 15min for 68Ga-DOTA-AE105-NH2 and 68Ga-NODAGA-

Fig. 1. Radiochemical synthesis scheme of 68Ga-DOTA-AE105-NH2 (A) and68Ga-NODAGA-AE105-NH2 (B). AE105-NH2: [NH2-Asp-Cha-Phe-(D)Ser-(D)

Arg-Tyr-Leu-Trp-Ser-CO-NH2].

562 M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

AE105-NH2, respectively. For the fractionated approach, sixvials containing 1 ml eluate each was collected andmeasured in a dose calibrator. A total of 100 μl of thevial containing the highest amount of radioactivity wasadded to a vial containing 5 nmol of either DOTA-AE105-NH2 or NODAGA-AE105-NH2 in 300 μl H2O. Samebuffer concentration, pH, temperature and reaction time asdescribe above were subsequently used again.

The labeled peptides were purified using a Sep-Pak LightC18 cartridge (Waters), which was first washed with water toelute any free 68GaCl3 and other impurities before beingeluted with 0.5 ml 96% ethanol. Finally, the labeled peptidewas diluted with 10 volume of 0.9% saline solution. Allproducts were analyzed using a Dionex HPLC system with aLuna C18 RP column (5 μm, 250×10 mm). The mobilephase changed from 92% solvent A (0.1% trifluoroacetic

acid in 90%water, 10% acetonitrile) and 8% solvent B (0.1%trifluoroacetic acid in 90% aqueous acetonitrile) to 60%solvent B at 12 min at a flow rate of 1 ml/min.

2.5. Cell binding assay

U87MG cells were seeded at a density of 0.15×106 in 24-well tissue culture plates and were allowed to attachovernight. The cells were washed three times withphosphate-buffered saline (PBS) and incubated with 68Ga-DOTA-AE105-NH2 or 68Ga-NODAGA-AE105-NH2 (100kBq/well, in culture medium) at 37°C. Specific binding wasdetermined after 15, 30, 60 and 120 min. At each time point,cells were washed three times with chilled PBS containing0.2% bovine serum albumin followed by detaching the cellsby treatment with 0.5 M NaOH. Specific binding was

563M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

estimated by dividing the activity on cells by the activity inthe collected wash volume using a gamma counter (PerkinElmer). Experiments were performed in triplicate wells.

2.6. Octanol–water partition coefficient; Log P

To 0.5 ml of the radiolabeled peptide in PBS, 0.5 ml ofoctanol in a microcentrifuge tube was added. The tube wasstirred vigorously over a period of 15 min. An aliquot of theaqueous and the octanol layer was collected and thencounted in the gamma counter (Perkin-Elmer). Log P valueswere calculated by taking the Log value of the activity in theoctanol phase divided with the activity in the water phase(mean of n=5).

2.7. Buffer and plasma stability

68Ga-labeled DOTA-AE105-NH2 and NODAGA-AE105-NH2 were incubated in either PBS or nude mouseserum for 60 min at 37°C. For PBS stability, the solution wasanalyzed by direct injection onto the HPLC. For plasmastability, plasma was first mixed with acetonitrile (1:1) toparticipate plasma proteins. After centrifugation, the super-natant was injected onto the HPLC column using the sameconditions as described above.

2.8. Cell line and animal model

The human glioma cancer cell line U87MG was used forall in vitro and in vivo experiments. The cell line wasobtained from the American Type Culture Collection(Manassas, VA, USA), and the culture medium was obtainedfrom Invitrogen Co. (Carlsbad, CA, USA). U87MGglioblastoma cells were grown in Dulbecco's modifiedEagle medium (low glucose) supplemented with 10% (v/v)fetal bovine serum and 1% penicillin/streptomycin at 37°Cand 5% CO2. All animal experiments were performed undera protocol approved by the Animal Research Committee ofthe Danish Ministry of Justice. Xenografts of U87MG wereestablished by injection of 200 μl cells (1×107 cells/ml)suspended in 100 μl Matrigel (BD Biosciences, San Jose,CA, USA) subcutaneously in the left and right flank offemale NMRI nude mice obtained from Taconic, underanesthesia by Hypnorm/Doricum. When the tumor volumereached approximately 100–300 mm3 (2–3 weeks afterinoculation), the mice were enrolled in biodistribution ormicroPET/ studies.

2.9. MicroPET imaging

PET scans were acquired with a microPET Focus 120scanner (Siemens Medical Solutions, Malvern, PA, USA).The energy window for the emission scans was set to 350–605 keVwith a time resolution of 6 ns. The acquired emissiondata set was automatically stored in list mode. MicroPETstudies were performed following intravenous injection of 5–7 MBq of either 68Ga-DOTA-AE105-NH2 or 68Ga-NODAGA-AE105-NH2 using sevoflurane anesthesia. Allanimals were PET scanned for 1 h immediately after injection

of the PET ligand. The 1-h dynamic list mode data weresorted in six frames of 10 min and processed into128×128×32 sinograms using a 3-dimensional maximum apriori (3D-MAP) algorithm into 256×256×95 matrices with avoxel size of 0.43 mm3. The resolution of the PET scannerwas 1.5 mm at center field of view, and 1.8 mm at 38 mm off-center using 3D-MAP. All emission data were corrected forscatter, dead time and decay time, and the system wascalibrated to provide the quantification unity in becquerelsper cubic centimeter (Bq/cc) by a cross calibration betweenthe microPET scanner and the well counter used for themeasurements of the injected activity. For this purpose, a 55-ml cylinder phantom containing a known activity concentra-tion of 18F obtained by the well-counter was scanned by themicroPET system. The acquired data of the phantom werethen reconstructed using the same reconstruction protocol asour later animal scans. A volume of interest was then drawnon the reconstructed images of the phantom in order to obtaina scale factor between the known activity (Bq/ml) and themicroPET-acquired values in counts/cc. This scaled factorcould then be used as a calibration factor for our animal PETscans to obtain the absolute values of the activity concentra-tion in the organs of interest.

All results were analyzed using Inveon software (SiemensMedical Solutions, Malvern, PA, USA). A 3D region ofinterest (ROI) was drawn on the image intending to cover thetumor. And the percent of injected dose per gram tissue (%ID/g) in the individual ROI was calculated by dividing themean specific activity in the ROI with injected dose.

2.10. Statistical analysis

Quantitative data are expressed as means±standard error ofthe mean (S.E.M.), and means were compared using Student'st test. P valuesb.05 were considered statistically significant.

3. Results

3.1. Affinity of the DOTA- andNODAGA-conjugated peptides

The interactions of the peptide AE105-NH2, together withthe chelator-conjugated analogue NODAGA-AE105-NH2,with immobilized human uPAR in solution were thenmeasured in real time by surface plasmon resonance (Fig.2). No reduction in the efficacy to compete the uPA·uPARinteraction was found between AE105-NH2 and the conjugat-ed versions (AE105-NH2: IC50=7.6±2.0 nM, DOTA-AE105-NH2: IC50=6.7±0.9 nM [10] and NODAGA-AE105-NH2:IC50=3.4±0.4 nM), indicating that neither DOTA norNODAGA conjugation to the N-terminus of AE105-NH2

had any effect on the affinity towards human uPAR.

3.2. Radiolabeling

Both DOTA-AE105-NH2 and NODAGA-AE105-NH2

were radiolabeled with 68Ga from a 68Ge/68Ga generator

Fig. 2. Inhibitory efficacy of AE105-NH2 and NODAGA-AE105-NH2 onthe uPA·uPAR interaction was investigated using surface plasmonresonance. No effect on affinity towards uPAR due to NODAGAconjugation (IC50=3.4 nM) was found compared with the core peptideAE105-NH2 (IC50=7.6 nM). ⁎IC50=6.7 nM for DOTA-AE105-NH2 is takenfrom Ref. [10].

564 M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

using the cation-exchange chromatography approach aswell as the fractionated elution approach [15] (Fig. 1).Using cation-exchange approach, a small radioactiveimpurity before the main product was generally observedfor 68Ga-NODAGA-AE105-NH2 (Fig. 3A), whereas thiswas not the case for 68Ga-DOTA-AE105-NH2. Using thefractionated elution approach, 68Ga-DOTA-AE105-NH2

and 68Ga-NODAGA-AE105-NH2 could be isolated aspure products. Especially for NODAGA-AE105-NH2,acetone used to elute 68Ga from the cation exchangecolumn had a negative effect on the degree of incorpora-tion of 68Ga (Fig. 3B). Based on these initial findings, allsubsequent experiments were performed using fractionatedelution approach. Both radiolabeled peptides were obtainedin greater than 95% radiochemical purity after purificationusing a Sep-Pak Light column. A specific radioactivity inthe final product of approximately 20 GBq/μmol for bothradiolabeled peptides with a radioactive concentration of10 MBq/ml was achieved.

3.3. Cell binding

The in vitro cell binding of 68Ga-DOTA-AE105-NH2 and68Ga-NODAGA-AE105-NH2 (Fig. 4) was evaluated inU87MG cells that express high levels of uPAR [10].68Ga-NODAGA-AE105-NH2 had a slightly higher uptakecompared to 68Ga-DOTA-AE105-NH2. Both 68Ga-NODAGA-AE105-NH2 and 68Ga-DOTA-AE105-NH2 up-take reached a plateau of 1.37±0.1 and 1.46±0.1 at 60 minafter incubation and remained at a similar level for up to 2h, thus indicating similar binding kinetics.

3.4. Stability and Log P

68Ga-NODAGA-AE105-NH2 was more hydrophobiccompared with 68Ga-DOTA-AE105-NH2, assessed by theoctanol–water partition assay. No significant degradation ofany of the 68Ga-labeled peptides was observed in either PBSor mice plasma, assessed using radio-HPLC after 1-hincubation at 37°C (Table 1).

3.5. MicroPET imaging

Finally, a 1-h dynamic PET scan was performed in agroup of U87MG tumor-bearing mice after injection of either68Ga-DOTA-AE105-NH2 or

68Ga-NODAGA-AE105-NH2.The uptake in tumor (Fig. 5A), liver (Fig. 5B), muscle (Fig.5C) and blood (Fig. 5D) was determined based on manualdrawing of ROI based on the PET images (Fig. 5E). A higheruptake in tumor tissue was observed for 68Ga-NODAGA-AE105-NH2 compared to 68Ga-DOTA-AE105-NH2, withthe same pattern observed in liver and blood. Identicaluptakes in muscle tissue were seen for 68Ga-NODAGA-AE105-NH2 and 68Ga-DOTA-AE105-NH2 during the 1-hdynamic scan period, thus resulting in a significantly highertumor-to-muscle (T/M) ratio of 68Ga-NODAGA-AE105-NH2 compared with 68Ga-DOTA-AE105-NH2 (5.37±0.7vs. 3.34±0.16, Pb.05) 1 h postinjection (Fig. 6B). After30 min, no significant difference in tumor ratios wasfound (Fig. 6A).

4. Discussion

In this study, the radiolabeling and feasibility of using68Ga-labeled peptides as PET tracers for imaging of uPARexpression are reported as an interesting alternative tocyclotron-dependent radionuclides. The use of 68Ga-labeledpeptides for cancer imaging has recently attracted consider-able interest since 68Ga can be easily obtained from a68Ge/68Ga generator. 68Ga also possess a medium half-lifeof 68 min, which is optimal compared to the pharmacoki-netics of many synthetic peptides.

The cation-exchange approach and fractionated elutionapproach were used for 68Ga labeling. A negative effect ofacetone on labeling efficiency for NODAGA-AE105-NH2

was observed using the cation-exchange approach (Fig.3B). In addition, a small impurity with almost sameretention time as the main product was observed whenlabeling NODAGA-AE105-NH2 (Fig. 3A). Reduced label-ing efficiency was not observed for labeling DOTA-AE105-NH2. No differences in the labeling efficiency orany impurities were observed for any of the conjugatedpeptides after using the fractionated approach (Fig. 3A, B).Based on these findings, the amount of acetone somehowseems to affect the labeling of the NODAGA-conjugatedpeptide via a currently unknown mechanism. To the best ofour knowledge, no study describing this phenomenon hasbeen published. One hypothesis could be that two complex

Fig. 3. (A) Representative radio-HPLC chromatograms of final products of 68Ga-DOTA-AE105-NH2 and 68Ga-NODAGA-AE105-NH2 after cation-exchange approach and fractionation approach. (B) A huge effect of acetone concentration in the reaction vial when labeling NODAGA-AE105-NH2 wasfound, whereas no effect was seen for DOTA-AE105-NH2. Constant reaction volume (500 μl) was used, whereas increasing amount of elute was addedinstead of water.

565M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

isomers are formed when the reaction is performed inacetone solvent (e.g., cation-exchange column approach),whereas this is not the case when the reaction is done inwater (e.g., fractionation approach). This hypothesis ispartly based on the published work, where NOTA wasdescribed to form two complex isomers with 64Cu,resulting in very similar radio-HPLC chromatograms aswe have reported [16].

Several receptor-based targets such as integrins, gastrin-releasing peptide receptor, somatostatin, HER-2, MMP-9and prostate-specific membrane antigen constitute promisingtargets for 68Ga-labeled PET ligands [17–22]. The majorityof these studies use DOTA as metal chelator for 68Ga.However, concerns regarding the in vivo stability of the68Ga-DOTA complex have been raised in the literature,suggesting that Ga3+ ion radius is too small to fit optimally in

Fig. 4. Cellular binding of 68Ga-DOTA-AE105-NH2 and68Ga-NODAGA-

AE105-NH2 using human glioblastoma U87MG (n=3, mean±S.E.M.).

able 1og P and stability

igand Partitioncoefficient

1-h stabilityin PBS

1-h stabilityin plasma

(Log P) (% intact) (% intact)8Ga-DOTA-AE105-NH2 −1.75±0.1 N98 N978Ga-NODAGA-AE105-NH2 −1.15±0.1 N98 N97

566 M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

the DOTA cage, which may decrease the inherent stabilityof these complexes [23]. However, good clinical resultsachieved using 68Ga-DOTA-conjugated peptides for im-aging of somatostatin receptors in neuroendocrine tumorpatients have been reported [24]. New macrocyclicchelators with improved properties based on a smallertricyclononane structure have been developed and used ina number of studies with some improvement in the tumor-to-background ratio compared to the correspondingDOTA-conjugated analogue [14]. Recently, the applicationof the NOTA-derived chelator (NODAGA) for PETimaging of αVβ3 integrin was described [14, 25].However, a lower tumor uptake in absolute values for68Ga-NODAGA-RGD compared to 68Ga-DOTA-RGDwas observed. A better tumor-to-background ratio due toimproved pharmacokinetics of the NODAGA-based 68Gatracer was obtained. In line with this, we also found asignificant difference (Pb.05) in the T/M ratios between68Ga-DOTA-AE105-NH2 and 68Ga-NODAGA-AE105-NH2 in our study, with 68Ga-NODAGA-AE105-NH2

yielding the more favorable ratio (5.37±0.72 vs. 3.34±0.17)(Fig. 6B).

A relatively low absolute tumor uptake was found 1 hpostinjection for both 68Ga-DOTA-AE105-NH2 (1.1% ID/g)and 68Ga-NODAGA-AE105-NH2 (2.0% ID/g) in this studycompared to the previously reported tumor uptake using64Cu-DOTA-AE105-NH2 (4.8% ID/g) in the same humancancer U87MG glioblastoma xenograft model [10]. Thisimplies that the use of 68Ga instead of 64Cu causes areduction in tumor uptake despite the fact that the targetingpeptide AE105-NH2 displays unaltered affinity for uPARirrespective of which macrocyclic chelator is used (Fig. 2A).Based on the substantial knowledge of the AE105-uPARinteraction [9], this observation seems contraintuitive, but it

may nonetheless be reasoned in the well-established in vivoinstability of the 68Ga complexes. Transmetallation ofgallium to the abundant plasma protein transferrin (concen-tration=0.25 g/L in blood) is thus possible due to the similarvan der Walls radius of Fe3+ (65 pm) and Ga3+(62 pm) [15].Elimination of 68Ga-DOTA-AE105-NH2 and 68Ga-NODAGA-AE105-NH2 was indeed very fast and occurredpredominantly via the kidney/bladder route, as seen in Fig. 5,which is distinct from the elimination via the liver/intestineroute observed for 64Cu-DOTA-AE105-NH2 [10]. Thisemphasizes the faster transmetallation for the radionuclidefrom the 68Ga complex into, e.g., transferring as the onepossible reason for the faster plasma elimination and lowertumor uptake.

Another source for the lower tumor uptake could beattributed to the specific activity (SA) of the tracers. Using68Ga, we could achieve an SA of 20 GBq/μmol, whereasfor 64Cu-based ligands, an SA above 25 GBq/μmol wasreported [10]. The SA activity has been found to be veryimportant for receptor-based imaging especially in mousemodels, where the volume of distribution and the numberof human receptors (tumor xenograft) are small. Therefore,a 20% reduction in the SA can indeed explain theobserved reduction in tumor uptake due to saturation ofreceptors by “cold” ligand [26]. However, this does notreduce the translational potential of 68Ga-based uPAR PETligands since an SA of 20 GBq/μmol would be fullyacceptable for human PET imaging due to the much lowerdose per kg body weight used in patients compared toanimal models.

Another explanation can be the known instability of the64Cu-DOTA complex, resulting in a relative high amountof free 64Cu accumulating in especially liver and tumortissue [27]. We have in our laboratory performed somepilot studies where 64CuCl2 was injected into nude micebearing human tumor xenograft, showing a tumor uptake ofapproximately 2%–3% ID/g (unpublished work). Theseresults indicate that any free 64Cu due to instability of64Cu-DOTA complex will result in a nonspecific accumu-lation in tumor tissue and therefore an artificially hightumor uptake value.

However, when developing a new PET tracer, it is not theabsolute uptake but rather the contrast between target-to-nontarget tissue that defines the utility of the tracer. Both68Ga-DOTA-AE105-NH2 and 68Ga-NODAGA-AE105-NH2 could visualize tumor tissue (Fig. 4), but in a directcomparison with the previously published 64Cu-labeled

TL

L

6

6

Fig. 5. (A) Quantitative results based on manually drawn ROI analysis for the organ/tissue indicated during the 1-h dynamic PET scan. Higher uptake for 68Ga-NODAGA-AE105-NH2 (solid) compared with 68Ga-DOTA-AE105-NH2 (dashed) was observed, except for muscle tissue. Results are shown as mean for threeanimals. (B) Representative 10-min PET images during 1-h dynamic PET recording. Tumor could clearly be visualized for both labeled peptides (white arrowsindicate tumors), with a clear washout of the tracer over time. Each mouse received 5-7 MBq ligand in 0.2 ml (approx. 0.5 nmol peptide), and sevoflurane wasused as anesthesia during the entire 1-h dynamic PET scan.

567M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

peptide (64Cu-DOTA-AE105-NH2), the T/M ratio 1 hpostinjection for 68Ga-DOTA-AE105-NH2 was 7.4, with64Cu-DOTA-AE105-NH2 having a T/M ratio of 15.9 [10].The same pattern is found when comparing blood, liver andkidney ratios, thus illustrating the somewhat reducedcontrast for 68Ga-labeled PET tracers for imaging of uPARcompared with the 64Cu-labeled counterpart.

In conclusion, the first 68Ga-labeled PET tracers foruPAR imaging have been developed. A significantly higherT/M ratio was observed for 68Ga-NODAGA-AE105-NH2

compared with 68Ga-DOTA-AE105-NH2 1 h postinjection.Dynamic PET/CT imaging revealed that both radiolabeledpeptides were able to clearly visualize tumor tissue withuPAR expression in a human glioblastoma cancer xenograft

Fig. 6. (A) Tumor ratios for 68Ga-DOTA-AE105-NH2 and68Ga-NODAGA-

AE105-NH2 after 30 min (Top) and 60 min (Bottom) based on ROI analysison 1-h dynamic data. A significantly higher T/M uptake for 68Ga-NODAGA-AE105-NH2 60 min postinjection compared with 68Ga-DOTA-AE105-NH2 was found. No other difference in tumor ratios were observedin this study. Results are shown as Mean±S.E.M. (n=6). ⁎Pb.05.

568 M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

mouse model. Future directions in developing 68Ga-radiolabeled peptides for uPAR imaging may focus onimproving tumor uptake and tumor-to-background ratios.

Acknowledgments

This work was supported by The Danish NationalResearch Foundation (Centre for Proteases and Cancer),Danish Medical Research Council, the Danish NationalAdvanced Technology Foundation, the Novo NordiskFoundation, the Lundbeck Foundation, Svend AndersenFoundation, Research Foundation of Rigshospitalet, and theA.P. Moeller Foundation.

The authors declare that they have no conflict of interest.

References

[1] Dano K, Behrendt N, Hoyer-Hansen G, Johnsen M, Lund LR, PlougM, et al. Plasminogen activation and cancer. Thromb Haemost2005;93:676–81.

[2] Jacobsen B, Ploug M. The urokinase receptor and its structuralhomologue C4.4A in human cancer: expression, prognosis andpharmacological inhibition. Curr Med Chem 2008;15:2559–73.

[3] Romer J, Nielsen BS, Ploug M. The urokinase receptor as a potentialtarget in cancer therapy. Curr Pharm Des 2004;10:2359–76.

[4] Rasch MG, Lund IK, Almasi CE, Hoyer-Hansen G. Intact and cleaveduPAR forms: diagnostic and prognostic value in cancer. Front Biosci2008;13:6752–62.

[5] Jankun J, Merrick HW, Goldblatt PJ. Expression and localization ofelements of the plasminogen activation system in benign breast diseaseand breast cancers. J Cell Biochem 1993;53:135–44.

[6] Costantini V, Sidoni A, Deveglia R, Cazzato OA, Bellezza G, FerriI, et al. Combined overexpression of urokinase, urokinase receptor,and plasminogen activator inhibitor-1 is associated with breastcancer progression: an immunohistochemical comparison of normal,benign, and malignant breast tissues. Cancer 1996;77:1079–88.

[7] Dublin E, Hanby A, Patel NK, Liebman R, Barnes D. Immunohisto-chemical expression of uPA, uPAR, and PAI-1 in breast carcinoma.Fibroblastic expression has strong associations with tumor pathology.Am J Pathol 2000;157:1219–27.

[8] Fisher JL, Field CL, Zhou H, Harris TL, Henderson MA, Choong PF.Urokinase plasminogen activator system gene expression is increasedin human breast carcinoma and its bone metastases–a comparison ofnormal breast tissue, non-invasive and invasive carcinoma and osseousmetastases. Breast Cancer Res Treat 2000;61:1–12.

[9] Ploug M, Ostergaard S, Gardsvoll H, Kovalski K, Holst-Hansen C,Holm A, et al. Peptide-derived antagonists of the urokinase receptor.affinity maturation by combinatorial chemistry, identification offunctional epitopes, and inhibitory effect on cancer cell intravasation.Biochemistry 2001;40:12157–68.

[10] PerssonM, Madsen J, Østergaard S, JensenMM, Jørgensen JT, Juhl K,Lehmann C, et al. P.E.T. Quantitative Imaging of human uPAR with64Cu-DOTA-AE105: implications for visualizing cancer invasion.Journal of Nuclear Medicine 2011 [Accepted].

[11] Li ZB, Niu G, Wang H, He L, Yang L, Ploug M, et al. Imaging ofurokinase-type plasminogen activator receptor expression using a64Cu-labeled linear peptide antagonist by microPET. Clin Cancer Res2008;14:4758–66.

[12] Maecke HR, Hofmann M, Haberkorn U. (68)Ga-labeled peptides intumor imaging. J Nucl Med 2005;46(Suppl 1):172S–8S.

[13] Fani M, Andre JP, Maecke HR. 68Ga-PET: a powerful generator-based alternative to cyclotron-based PET radiopharmaceuticals.Contrast Media Mol Imaging 2008;3:67–77.

[14] Eisenwiener KP, Prata MI, Buschmann I, Zhang HW, Santos AC,Wenger S, et al. NODAGATOC, a new chelator-coupled somatostatinanalogue labeled with [67/68Ga] and [111In] for SPECT, PET, andtargeted therapeutic applications of somatostatin receptor (hsst2)expressing tumors. Bioconjug Chem 2002;13:530–41.

[15] Roesch F, Riss PJ. The renaissance of the Ge/Ga radionuclidegenerator initiates new developments in Ga radiopharmaceuticalchemistry. Curr Top Med Chem 2010;10:1633–68.

[16] Schlesinger J, Rajander J, Ihalainen JA, Ramesh D, Eklund P,Fagerholm V, et al. Isomerism of [64Cu-NOTA-Bn]-labeled radio-tracers: separation of two complex isomers and determination of theirinterconversion energy barrier using ion pair chromatography. InorgChem 2011;50:4260–71.

[17] Li ZB, Chen K, Chen X. (68)Ga-labeled multimeric RGD peptides formicroPET imaging of integrin alpha(v)beta (3) expression. Eur J NuclMed Mol Imaging 2008;35:1100–8.

[18] Liu Z, Niu G, Wang F, Chen X. (68)Ga-labeled NOTA-RGD-BBNpeptide for dual integrin and GRPR-targeted tumor imaging. Eur JNucl Med Mol Imaging 2009;36:1483–94.

[19] Ugur O, Kothari PJ, Finn RD, Zanzonico P, Ruan S, Guenther I, et al.Ga-66 labeled somatostatin analogue DOTA-DPhe1-Tyr3-octreotideas a potential agent for positron emission tomography imaging andreceptor mediated internal radiotherapy of somatostatin receptorpositive tumors. Nucl Med Biol 2002;29:147–57.

569M. Persson et al. / Nuclear Medicine and Biology 39 (2012) 560–569

[20] Ren G, Zhang R, Liu Z, Webster JM, Miao Z, Gambhir SS, et al. A 2-helix small protein labeled with 68Ga for PET imaging of HER2expression. J Nucl Med 2009;50:1492–9.

[21] Ujula T, Huttunen M, Luoto P, Perakyla H, Simpura I, Wilson I, et al.Matrix metalloproteinase 9 targeting peptides: syntheses, 68Ga-labeling, and preliminary evaluation in a rat melanoma xenograftmodel. Bioconjug Chem 2010;21:1612–21.

[22] Banerjee SR, Pullambhatla M, Byun Y, Nimmagadda S, Green G,Fox JJ, et al. 68Ga-labeled inhibitors of prostate-specific membraneantigen (PSMA) for imaging prostate cancer. J Med Chem 2010;53:5333–41.

[23] Wadas TJ, Wong EH, Weisman GR, Anderson CJ. Coordinatingradiometals of copper, gallium, indium, yttrium, and zirconium forPET and SPECT imaging of disease. Chem Rev 2010;110:2858–902.

[24] Ambrosini V, Tomassetti P, Franchi R, Fanti S. Imaging of NETs withPET radiopharmaceuticals. Q J Nucl Med Mol Imaging 2010;54:16–23.

[25] Knetsch PA, Petrik M, Griessinger CM, Rangger C, Fani M,Kesenheimer C, et al. [(68)Ga]NODAGA-RGD for imaging alpha(v)beta (3) integrin expression. Eur J Nucl Med Mol Imaging2011;38:1303–12.

[26] Velikyan I, Beyer GJ, Bergstrom-Pettermann E, Johansen P,Bergstrom M, Langstrom B. The importance of high specificradioactivity in the performance of 68Ga-labeled peptide. Nucl MedBiol 2008;35:529–36.

[27] Shokeen M, Anderson CJ. Molecular imaging of cancer with copper-64 radiopharmaceuticals and positron emission tomography (PET).Acc Chem Res 2009;42:832–41.