European Journal of Nuclear Medicine and Molecular Imaging

, Volume 38, Issue 7, pp 1303–1312

[68Ga]NODAGA-RGD for imaging αvβ3 integrin expression


  • Peter A. Knetsch
    • Department of Nuclear MedicineInnsbruck Medical University
    • Universitätsklinik für NuklearmedizinMedizinische Universität Innsbruck
  • Milos Petrik
    • Department of Nuclear MedicineInnsbruck Medical University
  • Christoph M. Griessinger
    • Laboratory for Preclinical Imaging and Imaging Technology of the Werner Siemens-Foundation, Department of RadiologyUniversity of Tübingen
  • Christine Rangger
    • Department of Nuclear MedicineInnsbruck Medical University
  • Melpomeni Fani
    • Department of Nuclear MedicineUniversity of Freiburg
  • Christian Kesenheimer
    • Laboratory for Preclinical Imaging and Imaging Technology of the Werner Siemens-Foundation, Department of RadiologyUniversity of Tübingen
  • Elisabeth von Guggenberg
    • Department of Nuclear MedicineInnsbruck Medical University
  • Bernd J. Pichler
    • Laboratory for Preclinical Imaging and Imaging Technology of the Werner Siemens-Foundation, Department of RadiologyUniversity of Tübingen
  • Irene Virgolini
    • Department of Nuclear MedicineInnsbruck Medical University
  • Clemens Decristoforo
    • Department of Nuclear MedicineInnsbruck Medical University
    • Department of Nuclear MedicineInnsbruck Medical University
    • Universitätsklinik für NuklearmedizinMedizinische Universität Innsbruck
Original Article

DOI: 10.1007/s00259-011-1778-0

Cite this article as:
Knetsch, P.A., Petrik, M., Griessinger, C.M. et al. Eur J Nucl Med Mol Imaging (2011) 38: 1303. doi:10.1007/s00259-011-1778-0



A molecular target involved in the angiogenic process is the αvβ3 integrin. It has been demonstrated in preclinical as well as in clinical studies that radiolabelled RGD peptides and positron emission tomography (PET) allow noninvasive monitoring of αvβ3 expression. Here we introduce a 68Ga-labelled NOTA-conjugated RGD peptide ([68Ga]NODAGA-RGD) and compare its imaging properties with [68Ga]DOTA-RGD using small animal PET.


Synthesis of c(RGDfK(NODAGA)) was based on solid phase peptide synthesis protocols using the Fmoc strategy. The 68Ga labelling protocol was optimized concerning temperature, peptide concentration and reaction time. For in vitro characterization, partition coefficient, protein binding properties, serum stability, αvβ3 binding affinity and cell uptake were determined. To characterize the in vivo properties, biodistribution studies and microPET imaging were carried out. For both in vitro and in vivo evaluation, αvβ3-positive human melanoma M21 and αvβ3-negative M21-L cells were used.


[68Ga]NODAGA-RGD can be produced within 5 min at room temperature with high radiochemical yield and purity (> 96%). In vitro evaluation showed high αvβ3 binding affinity (IC50 = 4.7 ± 1.6 nM) and receptor-specific uptake. The radiotracer was stable in phosphate-buffered saline, pH 7.4, FeCl3 solution, and human serum. Protein-bound activity after 180 min incubation was found to be 12-fold lower than for [68Ga]DOTA-RGD. Biodistribution data 60 min post-injection confirmed receptor-specific tumour accumulation. The activity concentration of [68Ga]NODAGA-RGD was lower than [68Ga]DOTA-RGD in all organs and tissues investigated, leading to an improved tumour to blood ratio ([68Ga]NODAGA-RGD: 11, [68Ga]DOTA-RGD: 4). MicroPET imaging confirmed the improved imaging properties of [68Ga]NODAGA-RGD compared to [68Ga]DOTA-RGD.


The introduced [68Ga]NODAGA-RGD combines easy accessibility with high stability and good imaging properties making it an interesting alternative to the 18F-labelled RGD peptides currently used for imaging αvβ3 expression.


68GaNODAGARGD peptidesαvβ3Molecular imagingAngiogenesis


Angiogenesis plays an important role in many pathological processes, such as rheumatoid arthritis [1], psoriasis [2], cardiovascular diseases (e.g. atherosclerotic plagues) [3] as well as cancer [4]. Integrin αvβ3 is overexpressed on activated endothelial cells where it plays a critical role in the angiogenic process. Moreover, several studies have shown that αvβ3 is an important receptor affecting tumour growth, local invasiveness and metastatic potential [5]. Several drug candidates selectively binding to αvβ3 are currently under investigation in clinical trials. Thus, monitoring of αvβ3 expression, which coincides with the angiogenic process, holds great potential for drug development as well as clinical studies that provide information on tumour aggressiveness and therapy response [6].

Integrins are heterodimeric transmembrane glycoproteins, which mediate cell-matrix and cell-cell interactions. It was found that the amino acid sequence arginine-glycine-aspartic acid (RGD) is an important binding epitope of several extracellular matrix proteins such as vitronectin, fibrinogen and fibronectin [7]. The cyclic pentapeptide cyclo(-Arg-Gly-Asp-dPhe-Val-) shows high affinity and selectivity for the αvβ3 integrin. Thus, this was the initial lead structure for the development of radiolabelled compounds for the noninvasive determination of αvβ3 integrin expression using nuclear medicine tracer techniques [8].

Meanwhile a great variety of different radiolabelled RGD peptides have been introduced. This includes peptides labelled with halogens as well as radiometals. [18F]Galacto-RGD is the most intensively evaluated compound which is already in clinical studies [9]. This tracer shows good pharmacokinetics and receptor-specific uptake, allowing noninvasive imaging of αvβ3 integrin expression. However, the synthesis of this tracer is complex and time consuming, making an automated production process, which is mandatory for routine clinical use, extremely difficult [10].

One approach to overcome this problem is the use of alternative labelling strategies. One straightforward technique is radiometallation of peptides. Therefore peptides are conjugated with chelating systems, which allow binding of the corresponding radioactive metal. A common chelating system for a variety of metals including 111In, 90Y, 177Lu and 68Ga is 1,4,7,10-tetraazacyclododecane-N,N′,N′′,N′′′- tetraacetic acid (DOTA). Thus, we introduced DOTA-RGD enabling labelling via 68Ga [11]. The synthesis was straightforward and the tracer showed receptor-specific tumour uptake in a murine tumour model. However, also high amounts of protein-bound activity were found in corresponding assays resulting in high background activity especially in blood. It is known that the ion radius of Ga3+ is too small to fit optimally in the DOTA cage. This may be an explanation of the unfavourable biodistribution of [68Ga]DOTA-RGD, especially when having in mind that the corresponding [111In]DOTA-RGD does not show such effects [11].

More stable 68Ga-labelled peptides may be synthesized by using smaller triazacyclononane cages [e.g. 1,4,7-triazacyclononane-N,N′,N′′-triacetic acid (NOTA)], which possess better binding properties for radiometals with an ion radius like 68Ga [12]. Here we introduce a cyclic RGD peptide conjugated with NODAGA, a NOTA-derivatized chelating system, for labelling with 68Ga and compare the imaging properties with [68Ga]DOTA-RGD using small animal positron emission tomography (PET).

Materials and methods

All reagents were used as supplied with no further purification. 9-Fluorenylmethoxycarbonyl (Fmoc)-protected amino acids were purchased from Novabiochem (La Jolla, CA, USA). Trityl chloride polystyrene (TCP) resin was obtained from PepChem (Reutlingen, Germany). The coupling reagents 1-hydroxy-7-azabenzotriazole (HOAt) and O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU) were purchased from GenScript Corporation (Piscataway, NJ, USA). All other organic reagents were obtained from VWR International GmbH (Vienna, Austria) or Sigma-Aldrich Handels GmbH (Vienna, Austria). Human M21 and M21-L melanoma cells were a kind gift from D. A. Cheresh, Departments of Immunology and Vascular Biology, The Scripps Research Institute, La Jolla, CA, USA.

Liquid chromatography-mass spectrometry (LC-MS) analysis was carried out using a fritless nanospray column (100 µm ID, packed to 10 cm with 3 µm C18 material) constructed in house to analyze the samples by MS (LTQ ion trap instrument; Thermo Finnigan; San Jose, CA, USA) equipped with a nanospray source and an UltiMate 3000 HPLC pump (Dionex, Germering, Germany).

For reversed-phase high-performance liquid chromatography (RP-HPLC) analysis, a Dionex P680 HPLC pump with Dionex UVD 170 U UV/VIS detector (Dionex, Germering, Germany) and a Bioscan radiometric detector (Bioscan, Washington D.C., USA) were used. An ACE Nucleosil C18 3 μm, 150 × 3.0 mm column (ACE, Innsbruck, Austria), flow rates of 0.5 ml/min, and UV detection at 220 nm were employed with the following acetonitrile (ACN)/H2O/0.1% trifluoroacetic acid (TFA) gradients: 0–2.0 min 0% ACN, 2.0–16.0 min 0–60% ACN (gradient A) and 0–20 min 30–80% ACN (gradient B).

Isolation of the peptide via preparative RP-HPLC was performed using a Gilson 322 HPLC pump with Gilson UV/VIS-155 detector (Gilson International B.V., Limburg, Germany) and a MultoHigh 100 RP 18 5 μm, 250 × 10 mm column (CS-Chromatographie Service GmbH, Langerwehe, Germany). The flow rate was 5.0 ml/min. The ACN/H2O/0.1% TFA gradient used was as follows: 0–1.0 min 0% ACN, 1.0–21.0 min 0–50% ACN (gradient C).

The radioactivity of the samples was measured using a 2480 Automatic Gamma Counter Wizard2 3″ (PerkinElmer, Vienna, Austria).

The 68Ga generator was purchased from Eckert & Ziegler (Berlin, Germany) with nominal activities of 1,100 MBq and was eluted with 0.1 N HCl (Biochemical grade; Fluka, Buchs, Switzerland) using the fractionated elution approach.

Peptide synthesis and radiolabelling

1-(1-Carboxy-3-carbo-tert-butoxypropyl)-4,7-(carbo-tert-butoxymethyl)-1,4,7-triazacyclononane (NODAGA(tBu)3)

NODAGA(tBu)3 was synthesized in a four-step procedure including L-glutamic acid-5-benzyl ester derivatization, conjugation of 1-tert-butyl-5-benzyl-α-bromoglutarate with triazacyclononane, synthesis of 1-(1-carbobenzyloxy-3-carbo-tert-butoxypropyl)-4,7-bis(carbo-tert-butoxymethyl)-1,4,7-triazacyclononane and final removal of the benzyl protection group following protocols from the literature [13].


The synthesis of the linear peptide was carried out on a solid support using a TCP resin and Fmoc strategy, as described previously [14]. For side chain protection of the amino acids 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl (Pbf) for arginine, tert-butyl (tBu) for aspartic acid, and benzyloxycarbonyl (Z) for lysine were used. Subsequently the peptide was cyclized in dimethylformamide (DMF) in the presence of diphenylphosphoryl azide (DPPA) and sodium hydrogen carbonate. Lysine was deprotected by hydrogenation in the presence of an activated charcoal palladium catalyst. LC-MS: m/z [M + H]+ = 912.5 [C44H65N9O10S; exact mass 911.46 (calculated)]; RP-HPLC: tR = 9.6 min (gradient B).

NODAGA conjugation and final deprotection

To a solution consisting of 20 mg (22 μmol) cyclo(-Arg(Pbf)-Gly-Asp(OtBu)-dPhe-Lys-) and 12 mg (22 μmol) NODAGA(tBu)3 in 2 ml DMF, 8 mg (22 μmol) HATU, 3 mg (22 μmol) HOAt and 12 μl; N,N-diisopropylethylamine (Dipea; 66 μmol) were added. After adjusting the pH to approx. 8 using Dipea the reaction solution was stirred for 8 h at ambient temperature, followed by removal of the solvent in vacuo, precipitation and washing of the peptide with water. Final centrifugation resulted in 30 mg of the crude peptide.

Removal of the remaining protecting groups was carried out by treating the peptide with 4.5 ml of a solution of 95% TFA, 2.5% water and 2.5% triisobutylsilane for 10 h at room temperature. Subsequently, the reaction mixture was evaporated, triturated with diethyl ether and dried in vacuo. The crude peptide was purified by RP-HPLC (gradient C) resulting in the desired product cyclo(-Arg-Gly-Asp-dPhe-Lys(NODAGA)-) (NODAGA-RGD): LC-MS: m/z [M + H]+ = 961.5 [C42H64N12O14; exact mass 960.46 (calculated)]; RP-HPLC: tR = 11.0 min (gradient A).

Radiolabelling with 68Ga


The labelling procedure was optimized concerning amount of peptide, reaction time and reaction temperature. Details can be found in Table 1 and in the “Results” section.
Table 1

Optimization of the labelling conditions



Reaction timeb


Peptide amountc




5 min


5 μg




10 min


10 μg




15 min


20 μg




30 min


40 μg


For optimization of the labelling conditions reaction temperature, reaction time and peptide amount have been modified

a10 μg peptide, 15 min reaction time

b10 μg peptide, room temperature

c15 min reaction time, room temperature

Optimization led to the following standard labelling protocol: to 10 μl NODAGA-RGD (1 μg/μl in water) 60 μl 1.9 M sodium acetate solution and 300 μl 68GaCl3 of the main fraction of the generator eluate (approx. 100–200 MBq in 0.1 M HCl) were added. The solution was allowed to react for 15 min at room temperature. For all experiments, with the exception of the imaging study, the solution was used without further purification.

For small animal PET studies 800 μl 68GaCl3 of the main fraction of the generator eluate (approx. 450 MBq in 0.1 M HCl) was combined with 85 μl 1.9 M sodium acetate solution and 20 μl NODAGA-RGD (1 μg/μl in water). After 8 min at room temperature the solution was fixed on a Sep-Pak C18 cartridge (Waters Corporation, Vienna, Austria) and washed with 2 ml 0.9% saline. [68Ga]NODAGA-RGD was eluted with 0.5 ml 95% ethanol. The solvent was removed at 70°C under argon and reconstituted with 125 μl phosphate-buffered saline (PBS) before use.


For small animal PET studies DOTA-RGD was labelled following the protocols described in the literature [11]. Briefly, 800 μl 68GaCl3 of the main fraction of the generator eluate (approx. 450 MBq in 0.1 M HCl) was combined with 85 μl 1.9 M sodium acetate solution and 20 μl DOTA-RGD (1 μg/μl in water). After 8 min at 80°C the solution was fixed on a C18 cartridge and washed with 2 ml 0.9% saline. [68Ga]DOTA-RGD was eluted with 0.7 ml 95% ethanol. The solvent was removed at 80°C under argon and reconstituted with 125 μl PBS before use.

In vitro evaluation

Partition coefficient determination

[68Ga]NODAGA-RGD (approx. 5 KBq, corresponding to approx. 1 pmol peptide) in 0.5 ml PBS, pH 7.4, was added to 0.5 ml octanol and the mixture was vigorously vortexed for 15 min. Subsequently, aliquots of the aqueous and the octanol layer were collected, measured in the gamma counter, and logD values were calculated (n = 5).

Protein binding assay

The protein binding studies were carried out by incubating [68Ga]NODAGA-RGD (1 MBq/ml) in fresh human serum at 37°C for several time points (30, 60, 120, and 180 min). After incubation the solution was passed through a size exclusion spin column (MicroSpin™ G-50 columns, GE Healthcare, Buckinghamshire, UK). Protein binding was determined by measuring the activity remaining on the column and the activity in the eluate in a gamma counter.

Stability studies

Stability of [68Ga]NODAGA-RGD in PBS, pH 7.4, FeCl3 solution, and human serum was determined by incubating the radiolabelled compound (2 MBq) in 2 ml of the corresponding solution/medium at 37°C. The peptide was incubated for 15 and 105 min in PBS or FeCl3 and 30 and 90 min in human serum, respectively. At the preselected time points an aliquot of the PBS and FeCl3 solution was directly injected onto the HPLC whereas the serum was passed through a Sep-Pak C18 cartridge, washed with 500 μl buffer and eluted with 500 μl ACN containing 0.1% TFA. Subsequently, the solvent was removed in vacuo and the residue was dissolved in 500 μl PBS before analysed on HPLC. For HPLC analysis gradient A was used. Extraction efficiency was calculated by dividing the activity measured for the fraction eluted with ACN/TFA by the total activity used.

Isolated receptor binding assay

In vitro binding affinity of cyclo(-Arg-Gly-Asp-dPhe-Val-) and cyclo(-Arg-Gly-Asp-dPhe-Lys(NODAGA)-) was determined by using immobilized integrin αvβ3 (Millipore-Chemicon, Temecula, CA, USA), and 125I-echistatin (Amersham-Pharmacia Biotech, Vienna, Austria) as radioligand, as described previously [15]. Briefly, 96-well plates were coated with αvβ3 integrins and incubated with a mixture of 125I-echistatin and increasing concentrations of the corresponding peptide. Unbound radioligand was washed out and receptor-bound activity was removed from the plate with 2 M sodium hydroxide solution (NaOH). Three independent measurements were made, and the IC50 values were calculated by fitting the per cent inhibition values using Origin software (Northampton, MA, USA).

Internalization studies using αvβ3-positive and αvβ3-negative cells

M21 (αvβ3 positive) and M21-L cells (negative control) were grown in RPMI 1640 (Gibco, Invitrogen Corporation, Paisley, UK) containing 1% glutamine and 1% bovine serum albumin (BSA) to a concentration of 2 × 106 cells/ml and aliquots of 1 ml were transferred to Eppendorf tubes. After addition of [68Ga]NODAGA-RGD (>100,000 cpm, 1 nM), cells were incubated at 37°C for 90 min in triplicate with either PBS with 0.5% BSA (150 μl, total series) or with 10 μM c(RGDyV) in PBS/0.5% BSA (150 μl, nonspecific series). Incubation was stopped by centrifugation, removal of medium, and rapid rinsing with ice-cold TRIS-buffered saline twice. Subsequently, cells were incubated at ambient temperature in acid wash buffer (20 mM acetate buffer pH 4.5) for 15 min at 37°C. The supernatant was collected (membrane-bound radioligand fraction) and the cells were washed with acid wash buffer. Cells were lysed by treatment in 1 N NaOH and cell radioactivity collected (internalized radioligand fraction). Protein content in the NaOH fraction was determined using spectrophotometric determination with Bradford reagent (Sigma, Vienna, Austria). Internalized radioactivity was determined and expressed as percentage of total activity per milligram protein.

In vivo evaluation

All animal experiments were conducted in compliance with the Austrian animal protection laws and with approval of the Austrian Ministry of Science (BMWF-66.011/0135-II/10b/2008). Animal studies were performed using Balb/c nu/nu mice (Charles River, Sulzfeld, Germany). For the induction of tumour xenografts, M21 and M21-L cells were subcutaneously injected at a concentration of 5 × 106 cells/mouse and allowed to grow until tumours of 0.3–0.6 cm3 were visible.

Biodistribution studies

The receptor-specific uptake was determined using nude mice bearing human melanoma M21 and as negative control M21-L. A group of ten mice was injected with [68Ga]NODAGA-RGD (approx. 1 MBq/animal, corresponding to approx. 0.3 μg peptide) intravenously into the tail vein. The animals were sacrificed by cervical dislocation 60 min after injection. Organs (heart, stomach, lung, spleen, liver, pancreas, kidneys, and intestine), tissues (blood, muscle) and tumours were removed and weighed. Activity concentration in the samples was measured in the gamma counter. Results were expressed as percentage of injected dose per gram of tissue (%ID/g).

Small animal PET studies

In order to compare the in vivo uptake characteristics in M21/M21-L-bearing nude mice microPET studies were carried out. Thus, a group of four nude mice was injected on the right flank with M21 cells and on the left flank with M21-L cells at a concentration of 5 × 106 cells/mouse and allowed to grow until tumours of 0.5–1 cm3 were visible. On 2 consecutive days each mouse was injected with approx. 10 MBq of [68Ga]NODAGA-RGD and [68Ga]DOTA-RGD. One injection of radiotracer into the lateral tail vein was carried out per day.

Animals were positioned prone head first in the Inveon PET scanner (Siemens Medical Solutions, Knoxville, TN, USA). Coincidence-based list-mode acquisition started a few seconds prior to bolus injection of the radiotracer. The total measurement duration was 90 min. The list-mode data were sorted in 34 frames with a duration of 15 s up to 600 s each and subsequently reconstructed with 2-D ordered subset expectation maximization (OSEM). The image data set consisted per frame of a 128 × 128 matrix.

The final two frames of the image data set, corresponding to 20 min data acquisition time, were used to define volumes of interest (VOI): over the tumour at the right flank (M21), the tumour on the left flank (negative control tumour), a soft tissue reference region (in the same transaxial planes, in the upper part of the animal) and in the heart. Time-activity curves (TAC) were derived and visualized for each VOI with Inveon Research Workplace (Siemens Medical Systems, Knoxville, TN, USA).


Precursor synthesis and radiolabelling

Synthesis of the partially deprotected cyclo(-Arg(Pbf)-Gly-Asp(OtBu)-dPhe-Lys-) using Fmoc protocols, cyclization under high dilution conditions, and deprotection of the lysine side chain under hydrogen atmosphere was carried out in good yields. Conjugation of NODAGA(tBu)3 and cyclo(-Arg(Pbf)-Gly-Asp(OtBu)-dPhe-Lys-), accomplished by in situ activation, and subsequent deprotection resulted in NODAGA-RGD. The purity of the conjugate, after HPLC separation, was >95%.

For the optimization of the labelling conditions increasing amounts of peptide (5–40 μg), different reaction times (5–30 min) and different reaction temperatures (25-80°C) were studied (Table 1). It was found that NODAGA-RGD could be labelled with 68Ga even at room temperature within 5 min by using 10 μg precursor without further purification with radiochemical purities greater than 96%. However, for all further evaluation standard reaction conditions were 10 μg NODAGA-RGD, room temperature and 15 min reaction time. Moreover, it was found that, in contrast to the labelling of DOTA-RGD, a pH of 5, achieved by adding the double amount of sodium acetate buffer, resulted in a faster reaction. Thus, all labelling optimization as well as subsequent evaluation was carried out using that pH. Radiochemical purity of [68Ga]NODAGA-RGD (for proposed structure see Fig. 1) was always > 96% and specific activity ranged from 10 to 20 TBq/mmol.
Fig. 1

Proposed structure of [68Ga]NODAGA-RGD

In vitro characterization

The partition coefficient (logD) of [68Ga]NODAGA-RGD was −3.6 indicating high hydrophilicity.

Incubation of [68Ga]NODAGA-RGD in human serum showed only low amounts of protein-bound activity (Fig. 2). The values are between 7-fold and 12-fold lower than for [68Ga]DOTA-RGD at 30 and 180 min incubation time, respectively, and comparable with the values found for [111In]DOTA-RGD [11].
Fig. 2

Comparison of protein binding properties. The amount of protein-bound activity, found by size exclusion chromatography, after different time points of incubation of the corresponding tracers in human serum showed for [68Ga]NODAGA-RGD comparable protein-bound activity as for [111In]DOTA-RGD, and for both significantly lower amounts than for [68Ga]DOTA-RGD. Data for [111In]DOTA-RGD and for [68Ga]DOTA-RGD are extracted from the literature [11]

[68Ga]NODAGA-RGD was stable in PBS, FeCl3 solution and human serum, regardless of whether the tracer was incubated for 15 and 105 min in buffer or FeCl3 solution or for 30 and 90 min in human serum.

Subsequent HPLC analysis showed that more than 97% of the labelled peptide was still intact. While aliquots of buffer and FeCl3 solution were analysed directly, for human serum analysis separation of the protein-bound fraction was carried out using Sep-Pak cartridges before HPLC analysis. Extraction efficiency for the latter was always greater than 97% (Fig. 3).
Fig. 3

Stability studies. HPLC analysis showed high stability of [68Ga]NODAGA-RGD up to approx. 1.5 h independently of whether it was incubated in PBS, pH 7.4 (a), 0.1 M FeCl3 solution (b) or human serum (c), at 37°C

The in vitro binding assays using increasing amounts of NODAGA-RGD or c(RGDyV) showed that the inhibitory peptides were able to fully suppress the binding of 125I-echistatin to the isolated immobilized αvβ3 integrin and that the binding kinetics follow a classic sigmoid path. The IC50 value found for NODAGA-RGD and for c(RGDyV) was 4.7 ± 1.6 and 2.8 ± 1.0 nM, respectively (Fig. 4).
Fig. 4

Binding affinity and cell uptake. A The in vitro binding assay using immobilized integrin αvβ3 and 125I-echistatin as radioligand demonstrated comparable binding characteristics for c(RGDfK(NODAGA)) (red) and the lead structure c(RGDyV) (black). B Cell uptake studies using human melanoma M21 (αvβ3-positive) and M21-L (αvβ3-negative) cells showed low but receptor-specific internalization for [68Ga]NODAGA-RGD (peptide concentration  = 20 nM). Blocking studies using 10 μM c(RGDyV) confirmed receptor-specific uptake

Determination of the cell uptake using M21 and M21-L human melanoma cells demonstrated receptor-specific internalization of [68Ga]NODAGA-RGD. This is confirmed by corresponding blocking experiments using 10 μM c(RGDyV). In this assay internalization can only be blocked for the αvβ3-positive cells, whereas for the negative control cells internalized activity was one third of the amount for the receptor-positive cells and very similar under blocked as well as under unblocked conditions (Fig. 4).

In vivo characterization

Biodistribution studies were carried out 60 min after [68Ga]NODAGA-RGD injection using nude mice bearing both, αvβ3-positive as well as αvβ3-negative human melanoma. Data are presented in Fig. 5. The experiment showed for [68Ga]NODAGA-RGD fourfold higher activity accumulation in the receptor-positive than the receptor-negative tumour confirming the receptor-specific uptake found in vitro. The absolute uptake of [68Ga]NODAGA-RGD in the tumour was approximately one half of the uptake found for [68Ga]DOTA-RGD. However, the activity accumulation found in blood was sixfold higher for [68Ga]DOTA-RGD than for [68Ga]NODAGA-RGD, resulting in higher activity concentration in almost all organs studied for [68Ga]DOTA-RGD.
Fig. 5

Comparison of biodistribution data. For biodistribution studies nude mice bearing the αvβ3-positive human melanoma M21 on the right flank and the negative control tumour M21-L on the left flank were used. Data collected 60 min after the injection of the corresponding tracer showed for all tracers receptor-specific tumour accumulation. Moreover, data demonstrate a clearly reduced activity concentration in blood for [68Ga]NODAGA-RGD compared with [68Ga]DOTA-RGD (from [11]). Moreover, almost all organs and tissues studied showed comparable values as found in the corresponding study using [18F]Galacto-RGD (data from [19])

Initial microPET studies combined with magnetic resonance (MR) imaging, showing the negative control tumour on the left flank and the αvβ3-positive tumour on the right flank, demonstrate receptor-specific uptake and rapid predominantly renal elimination of [68Ga]NODAGA-RGD (Fig. 6). This is also confirmed by analysing the TACs, revealing for [68Ga]NODAGA-RGD a comparable elimination from the muscle and the M21-L tumour (Fig. 6). Transaxial planes from the last two image frames, containing parts of the M21 and M21-L tumour of dynamic microPET studies over 90 min post-injection (p.i.), where the same mouse is imaged on 2 consecutive days with either [68Ga]DOTA-RGD or [68Ga]NODAGA-RGD, showed tracer uptake only in the M21 but not the M21-L region for both tracers, confirming the target selectivity of the tracers (Fig. 7). In contrast to the comparable receptor-specific uptake, differences in background activity between [68Ga]DOTA-RGD and [68Ga]NODAGA-RGD have been found. Basically, retention of tracer activity in the tumour was comparable and unspecific tracer uptake in the contralateral negative control tumour was low for both tracers. Analysis of the tumour to blood ratios [determined by defining regions of interest (ROI) in the tumour and in the heart] resulted in an approx. twofold higher ratio for [68Ga]NODAGA-RGD as found for [68Ga]DOTA-RGD. Additionally increased tumour to background ratios compared to [68Ga]DOTA-RGD are found for the negative control tumour and muscle. The ratios for liver and kidneys are approx. 1 for both tracers (Fig. 7).
Fig. 6

MicroPET and MR imaging. MR images confirm comparable size of the receptor-positive and negative control tumour. Studies demonstrated selective uptake in the receptor-positive tumour and rapid elimination from the body. PET image was smoothed with a 2-mm Gaussian filter. Hot spots in the PET and MR images outside the animals are the fiducial markers for PET and MR image fusions
Fig. 7

Comparison of dynamic microPET data. A Transaxial images 90 min p.i. acquired on 2 consecutive days after either injecting [68Ga]NODAGA-RGD or [68Ga]DOTA-RGD showed receptor-specific tumour uptake for both tracers. In contrast, background activity is found to be higher for [68Ga]DOTA-RGD than for [68Ga]NODAGA-RGD. B Quantitative ROI analysis confirmed improved tumour to background ratios for [68Ga]NODAGA-RGD


One of the most prominent tracers in imaging αvβ3 integrin expression is [18F]Galacto-RGD (for review see [9]). However, the synthesis of this PET tracer is very complex and time consuming making its use in routine clinical practice extremely difficult. Recently, we introduced [68Ga]DOTA-RGD [11] and demonstrated that this tracer allows receptor-selective monitoring of αvβ3 integrin expression. More importantly, the straightforward synthesis allows production of the tracer in a remote controlled system within 15 min including Sep-Pak purification. However, in subsequent in vitro and in vivo studies it was found that [68Ga]DOTA-RGD showed much higher protein-bound activity compared with the same peptide labelled with 111In. This effect resulted in higher activity concentration in blood leading to inferior imaging properties compared to [18F]Galacto-RGD.

Here we introduced [68Ga]NODAGA-RGD and showed that the replacement of DOTA by NODAGA resulted in a tracer with clearly reduced protein-bound activity which is in the same low range as found for [111In]DOTA-RGD [11]. It is not completely understood yet why increased protein binding properties are found for [68Ga]DOTA-RGD. Serum stability studies indicated that it does not seem to be due to a release of the metal from the chelator and transfer to proteins like transferrin, which is known to have also high binding affinity for Ga ions [16]. Anyway, it is known that the Ga3+ ion with an ion radius of 62 pm (e.g. the In3+ ion radius is 92 pm) [17] does not fit perfectly into the tetraazadodecane ring system of DOTA resulting in a hexadentate coordination sphere leaving one carboxylate function uncoordinated. A higher formation constant for Ga3+ ions is found for the triazanonane ring system (e.g. NOTA) which involves all functionalities for the complexation of the metal [18]. Indeed we found that the [68Ga]NODAGA-RGD, where NODAGA is a derivative of the NOTA system, showed high stability in human serum as well as when challenged using a FeCl3 solution and more importantly, an approx. 12-fold lower protein-bound activity after 180 min incubation in human serum compared to [68Ga]DOTA-RGD.

Further studies demonstrated that the replacement of the DOTA chelating system by the NODAGA system has no influence on binding affinity and receptor-specific binding. The determined IC50 values are comparable with the lead structure cyclo(-Arg-Gly-Asp-dTyr-Val-) as well as with DOTA-RGD [11]. The cell uptake studies showed an approx. threefold higher uptake in the receptor-positive cells than in the receptor-negative human melanoma cells, which can only be blocked for the αvβ3-positive cells. This is also confirmed by the biodistribution studies using xenograft-bearing nude mice. Uptake in the receptor-positive tumour was fourfold higher than for the negative control tumour. This experiment demonstrated also the positive effect of the reduced protein binding properties on activity concentration in blood. The biodistribution data of [68Ga]NODAGA-RGD, [68Ga]DOTA-RGD and [18F]Galacto-RGD (Fig. 5) showed a clear reduction of the blood pool activity followed by a reduction in the activity concentration in all investigated organs and tissues including the tumours for [68Ga]NODAGA-RGD compared to [68Ga]DOTA-RGD. On the other hand, the reduction in blood pool activity for [68Ga]NODAGA-RGD results in a distribution pattern which is, with the exception of the spleen, comparable with the activity distribution found for [18F]Galacto-RGD [19]. As mentioned, due to the higher blood pool activity found for [68Ga]DOTA-RGD also uptake in the tumour is found to be higher. However, determination of the tumour to background ratios indicate better tumour to blood ratio for [68Ga]NODAGA-RGD than for [68Ga]DOTA-RGD. Moreover, most tumour to background ratios of [68Ga]NODAGA-RGD are in the same range as found for [18F]Galacto-RGD. This is confirmed by corresponding small animal PET images using nude mice bearing both the receptor-positive as well as the receptor-negative tumour (Fig. 7). The images taken on 2 consecutive days from the same mouse demonstrated receptor-specific uptake for both compounds, but also lower background activity for [68Ga]NODAGA-RGD compared to [68Ga]DOTA-RGD, leading to more contrasting images for the first. Because no direct comparison of imaging data between [68Ga]NODAGA-RGD and [18F]Galacto-RGD could be carried out, a final assessment of the tracer properties is difficult. But the qualitative comparison of the published data [11], where [68Ga]DOTA-RGD is directly compared with [18F]Galacto-RGD, indicates similar imaging properties for [68Ga]NODAGA-RGD and [18F]Galacto-RGD. Moreover, the imaging data are in accordance with the comparison of the data of the different biodistribution studies.

In addition to the improved imaging properties, optimization of the synthesis condition demonstrated that [68Ga]NODAGA-RGD can be synthesized straightforward at room temperature within 5 min with high radiochemical yield and purity making a fully automated processing of this synthesis easily achievable. Compared with the labelling of [68Ga]DOTA-RGD no increased temperature is necessary to form the corresponding complex in a very short reaction time. Thus, an overall production time of less than 20 min is possible, which is much faster than the production of any known 18F-labelled RGD peptide, including [18F]Galacto-RGD [10]. From other groups a monomeric [20] as well as some multimeric 68Ga-labelled RGD peptides [21, 22] are described. In contrast to [68Ga]NODAGA-RGD all these compounds use SCN-Bz-NOTA for the conjugation with the corresponding RGD peptides. Thus, the main difference of the chelating systems is found in the linker design (aliphatic vs aromatic linker moiety) and the connection of the chelator with the peptide (amide bond vs thiourea formation). However, production times are comparable for all the 68Ga-labelled derivatives. Chen and coworkers [21, 22] described that reactions were carried out at approx. 40°C, but it is not indicated if labelling did not work at room temperature with their compounds. Maybe one advantage of the NODAGA system lies in the use of an aliphatic linker, which may have lower effects on the hydrophilicity of the tracer than an aromatic system. However, a partition coefficient, which is −3.6 for [68Ga]NODAGA-RGD and thus even slightly better than for [18F]Galacto-RGD [10], is not found for any of the NOTA-Bz-SCN-conjugated RGD peptides for direct comparison. Due to the different tumour models used a direct comparison of the biodistribution data is not easy. However, tumour to blood ratios range between 7 and 11 with the highest ratio found for [68Ga]NODAGA-RGD.


In summary, [68Ga]NODAGA-RGD can be produced straightforward with high yields and radiochemical purity making it suitable for synthesis in an automated system. It shows high affinity for the αvβ3 integrin, receptor-selective tumour uptake, high metabolic stability, low protein-bound activity, rapid predominantly renal elimination and good tumour to background ratios indicating that [68Ga]NODAGA-RGD may be an attractive, easily available alternative to 18F-labelled RGD peptides, like [18F]Galacto-RGD, for imaging αvβ3 integrin expression.


Bettina Sarg and Sabine Hofer, Protein Micro Analysis Facility, Biocenter, Innsbruck Medical University are acknowledged for carrying out the LC-MS analysis. We thank Alexander Staaf, Department of Pharmaceutical Chemistry, University of Innsbruck for providing analytical data concerning synthesis of the chelator. David A. Cheresh, The Scripps Research Institute, La Jolla, CA is acknowledged for providing the human melanoma M21 and M21-L cells. Parts of the studies were financially supported by the BMBF-MoBiMed grant. This work was part of COST Action BM0607 “Targeted Radionuclide Therapy”.

Conflicts of interest


Copyright information

© Springer-Verlag 2011